|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Moderating the Effectiveness of Messages to Promote Physical Activity in Type 2 Diabetes by Rachel E. Myers A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy College of Nursing Univers ity of South Florida Major Professor: Jason W. Beckstead, Ph.D. Mary E. Evans, Ph.D., RN FAAN Frances M. Rankin, Ph.D., ARNP, FAANP William P. Sacco, Ph.D. Date of Approval: June 29, 2010 Key w ords: B ehavior C hange, H ealth B ehaviors, H ealth C ommuni cation, M essage F raming, M essage T ailoring, N ursing Copyright 2010, Rachel E. Myers
Dedication To my husband Alan . for always being there to encourage, support, and believe in me . for motivating me to persevere towards my goals and to ne ver give up . for being unselfish and making sacrifices so that I could fulfill my dreams . for loving me unconditionally and accepting me for who I am . for simply being you a lifetime partner, best friend and soul mate. To my parents and other family members . whose love support and encouragement helped carry me along this journey. To my cats Oreo and Goober . who made me laugh when I wanted to cry and who helped remind me to relax and enjoy the simple things in life.
Acknowledgements I would like to recognize and thank my major professor, Dr. Jason Beckstead, and my other committee members : Dr. Mary Evans, Dr. Frances Rankin, and Dr. William Sacco. I am sincerely grateful for and appreciative of their guidance support, encouragement, patience, and availability Without their expertise and mentorship, this work would not have been possible. In particular, I want to express my deepest gratitude to Dr. Beckstead for spending countless hours mentoring me in resear ch methodology and statistical analysis reviewing numerous dissertation drafts, and mindfully guiding me down the long and arduous yet very enlightening and rewarding doctoral path. He was a continuous source of strength, inspiration and motivation.
i Table of Contents List of Tables ..................................................................................................................... iv List of Figures ......................................................................................................................v Abstract .............................................................................................................................. vi Chapter 1: Introduction ........................................................................................................1 Identification of the Problem ...................................................................................1 Health Communication ............................................................................................3 Message tailoring .........................................................................................3 Message framing ..........................................................................................4 Moderators of framing effects .........................................................5 Consideration of future consequences .................................5 Motivational orientation .......................................................6 Physical Activity ......................................................................................................7 Call for Research ......................................................................................................8 The Present Study ....................................................................................................9 Purpose, specific aims, and hypotheses .....................................................10 Targeted population ...................................................................................11 Chapter Summary ..................................................................................................12 Chapter 2: Review of Literature ........................................................................................13 Message Fra ming Definitions and Typologies ......................................................13 Origins of Message Framing ..................................................................................14 Other Theories and Conceptual Models of Message Framing in Health Research .........................................................................................................18 Empirical Studies Involving Message Framing and Health Behaviors .................24 Detection behaviors ...................................................................................26 Prevention beha viors ..................................................................................27 Mixed function behaviors ..........................................................................29 Moderators of message framing .................................................................30 Motivational orientation .................................................................31 Temporal context ...........................................................................35 Consideration of future consequences ...........................................36 Discordant findings ....................................................................................39 Metaa nalytic Reviews o n Message Framing Research ........................................40 State of the Message Framing Literature ...............................................................45
ii Chapter Summary ..................................................................................................47 Chapter 3: Method .............................................................................................................48 Research Design .....................................................................................................48 Participants .............................................................................................................49 Materials ................................................................................................................50 Health message development .....................................................................50 Health message pilot study ........................................................................52 Consideration of future consequences .......................................................54 Motivational orie ntation .............................................................................54 BIS/BAS ........................................................................................54 RFQ ................................................................................................55 Intentions ....................................................................................................55 Evaluation of the health message ...............................................................56 Believability ...................................................................................56 Manipulation check ........................................................................56 Demographic and other variables ..............................................................56 Booklet assembly .......................................................................................56 Procedures ..............................................................................................................56 Institutional Review Board ........................................................................56 Recruitment ................................................................................................57 Chapter Summary ..................................................................................................57 Chapter 4: Results ..............................................................................................................58 Preliminary Analyses .............................................................................................58 Mail survey results .....................................................................................58 Missing data ...............................................................................................61 Description of sample ................................................................................61 Evaluation of the health message ...............................................................62 Hypotheses Testing ................................................................................................64 Calculation of congruence scores ..............................................................64 Descriptive stat istics ..................................................................................65 Scale scores ....................................................................................67 Congruence scores .........................................................................67 Intentions ........................................................................................68 Hypotheses .................................................................................................68 Supplemental Analyses ..........................................................................................71 Chapter Summary ..................................................................................................75 Chapter 5: Discussion ........................................................................................................76 Discussion of Findings ...........................................................................................76 Hypotheses testing .....................................................................................76 Mail survey ................................................................................................89 Supplemental analyses ...............................................................................90 Di rections for Future Research ..............................................................................96 Implications for Nursing Research, Education, and Practice ................................99
iii Study Limitations .................................................................................................106 List of References ............................................................................................................109 Appendices .......................................................................................................................121 Appendix A: Gain Framed/Distal Framed Health Message ................................122 Appendix B: Gain Framed/Immediate Frame d Health Message ........................123 Appendix C: Loss Framed/Distal Framed Health Message ................................124 Appendix D: Loss Framed/Immediate Framed Health Message ........................125 Appendix E: Pilot Study Questionnaire ...............................................................126 Appendix F: Consideration of Future Consequences Scale .................................127 Appendix G: BIS/BAS Scale ...............................................................................128 Appendix H: Regulatory Focus Questionnaire Scale ..........................................129 Appendix I: Intentions and Evaluation of the Health Message ...........................130 Appendix J: Demographic and Other Variables ..................................................131 Appendix K: Pre notice Letter (1st Mailing) ........................................................132 Appendix L: Cover Le tter for Initial Booklet (2nd Mailing) ................................133 Appendix M: Reminder/Thank You Postcard (3rd Mailing) ................................134 Appendix N: Cover Letter for Replacement Booklet (4th Mailing) .....................135 About the Author ................................................................................................... End Page
iv List of Tables Table 1 : Details from Four Health Message Versions to Illustrate Temporal Proximity and Valence Framing Manipulation .................................................51 Table 2 : Summary of Pilot Study Data for Evaluation of Health Message Believability and Framing Manipulation of Temporal Proximity and Valence ..............................................................................................................53 Table 3: Mail Survey Results Using Dillmans Tailored Design Method ......................60 Table 4: Characteristics of the Final Sample ...................................................................62 Table 5: Summary of Mail Survey Data for Evaluation of Health Message Believability and Framing Manipulation of Temporal Proximity and Valence ..............................................................................................................64 Table 6: Pea rson Correlations and Descriptive Statistics for Key Variables ..................66 Table 7: Means and (Standard Deviations) for Intentions to Increase Physical Activity as a Function of Framing Manipulation Conditions ............................68 Table 8: Summary of Multiple Regression Analyses P redicting Intentions to Increase Physical Activity from Valence Congruence and Temporal Congruence ........................................................................................................70 Table 9: Pearson Correlations for Supplemental Analysis Individual Differences Variables ............................................................................................................72 Table 10: Summary of Multiple Regression Analysis Predicting Intentions to Increase Physical Activity from Supplemental Analysis Variables ..................74
v List of Figures Figure 1: Logic Model of the Present Study .....................................................................10
vi Moderating the Effectiveness of Messages to Promote Physical Activity in Type 2 Diabetes Rachel E. Myers A bstract The world is experiencing a rapid rise in chronic health problems, which places an enormous burden on health care services. Modifiable health behaviors such as physical inactivity are largely responsible for this high prevalence and inc idence of chronic diseases. Message tailoring is a well established approach for constructing health communication and has been shown to increase the persuasiveness of messages i n the promotion of healthy behaviors. Message framing is an effective strategy that has been well studied in psychology over the past 20plus years across a breadth of health related behaviors but has received little attention in the nursing research literature Based on prospect theory, temporal construal theory, and motivational orientation theories, the present study examine d how two individual differences factors consideration of future consequences (CFC) and motivational orientation combine to moderate temporal proximity and valence framing effects on intentions to increase physical activity. A mail survey was conducted using Dillmans Tailored Design Method. Two hundred and eighteen adults with type 2 diabetes were randomly assigned to receive one of four versions of a health message aimed to increase regular physical activi ty. Messages were
vii framed us ing a 2 (immediate vs. distal framed) x 2 (gainvs. loss fram ed) design. After reading the message, participants rated their intention to increase physical activity. They also completed a measure of CFC and two measures of moti vational orientation. Participants who read a message with a temporal proximity or valence frame congruent with their CFC or motivational orientation respectively, did not show greater intentions to increase physical activity when compared to those who read a health message that was incongruent with these individual differences. Plausible expl anations f or these negative results are considered. S everal interesting findings e merged from supplemental analyses. For instance, participants who perceived the heal th message as more believable tended to have greater intention s to increase physical activity. Suggestions for future research applying message congruence to promote complex health behaviors in at risk populations are given. Implications of message framing and other message tailoring strategies for nursing research, education, and practice are discussed
1 Chapter 1: Introduction Identification of the Problem The world is experiencing a rapid rise in chronic health problems, which places an enormous burden on health care services (World Health Organization [WHO], 2005a). In 2005, an estimated 60% (35 million) of all global deaths were due to chronic diseases, primarily diabetes mellitus (DM) and cardiovascular diseases (32%), cancers (13%), and chronic respiratory diseases (7%) (Abegunde, Mathers, Adam, Ortegon, & Strong, 2007). Chronic diseases also place a grave economic burden on nations (Centers for Disease Control and Prevention [CDC], 2009 a ; WHO, 2005b). The WHO (2005a) calls for the health care workforce to transition from a traditional provider centered approach to a contemporary patient centered approach in order to lessen the occurrence and detrimental impact of these worldwide burdens. For example, it is estimated that at least 80% of all ty pe 2 DM and cardiovascular disease and over 40% of cancer can be prevented through changes in behavior (WHO, 2005b). Physical inactivity, u nhealthy diet, and tobacco use are three examples of modifiable behaviors that contribute to the prevalence of chroni c diseases. A patient centered approach, where care is coordinated across time and centered around patients needs, values, and preferences, strengthens patients role in managing
2 their health problems by empowering them to become active decision makers ra ther than just passive recipients of care (WHO, 2005a). Health education and self management/self care training, with an emphasis on promoting healthy behaviors to prevent future problems, are vital components of a patient centered approach. This educatio n and training may be informal (e.g., unplanned and unstructured during a clinic visit) or formal (e.g., structured group diabetes class) and range from simple (e.g., distribut ion of written materials) to complex (e.g., teach ing patients self blood glucose monitoring skills ). Nurses play an essential part in delivering such education and training, both independently and alongside other health care providers. Nurses are well placed and have extensive opportunities to deliver patient education and training in a variety of settings (Coster & Norman, 2009). Studies have revealed that nurses are perceived as credible sources of health information. For example, Jones, Sinclair, and Courneya (2003) conducted a pilot study and found that Registered Nurses (RNs) were not only perceived as credible sources but that this credibility did not differ from that of physicians. Research has also shown that patients find nurses easier to approach for health information than physicians. For example, Collins (2005) explored both nurse and physicianpatient communications and found that overall, patients more openly and freely communicated with nurses. Despite the fact that patient education and training are wellestablished key features of nursing and that nurses recognize t hese as important functions of their role, nurses often report difficulty providing education and training (see Coster & Norman, 2009; Kim, Heerey, & Kols, 2008). L ack of time is a common barrier to effective nursepatient communication. Nurses may only ha ve a few minutes to deliver an important
3 health education message. How should nurses go about getting t heir message across to their patients to maximally promote healthy behaviors? What well established, empirically tested, effective strategies can nurses employ to help them successfully achieve th is goal? The field of health psychology offers a theoretical and conceptual framework from which nursing can draw to answer these questions (Myers, 2010) Health psychology emphasizes health promotion and disease prevention and focuses on the development of theoretical constructs and empirically derived principles of behavior change (Matarazzo, 1980, 1982). Health psychology is also devoted to understanding psychological influences on how people stay healthy, why they become ill, and how they respond when they do get ill (Taylor, 2003, p. 17). Myers and Beckstead ( 2009) present an overview of the field of health psychology and highlight health psychologys utility for nursing research, education, and practice by providing examples of applications in nursing. Health Communication Message tailoring. Health behaviors and habits are complex, are determined by the interplay of multiple factors, and are resistant to change (see Rodin & Salovey, 1989; Taylor, 2003). Ef fective health communication uses theoretical based behavior modification principles to inform and influence individual and community decisions that enhance health (U.S. Department of Health and Human Services [DHHS], 2000b). Message tailoring is a health communication strategy that involves the customization of information i nterventions to best fit the characteristics and needs of specific target populations or individuals (Kreuter & Wray, 2003; Salovey, 2005). There is empirical evidence that tailored hea lth messages, compared to general, non tailored health
4 messages, are more persuasive and effective in promoting behavior change through various mechanisms such as enhanced salience and stimulation of greater cognitive activity or elaboration (e.g., Kreuter, Bull, Clark, & Oswald, 1999; Kreuter & Wray, 2003; Latimer, Katulak, Mowad, & Salovey, 2005). Nurses have long recognized the value of enhancing nurse patient communication and of utilizing approaches such as message tailoring (although not always labele d as such) in nursing interventions (e.g., Coster & Norman, 2009; Kim et al., 08; Shin, Hur, Pender, Jang, & Kim, 2006). Message f raming : Message framing is a widely studied method of message tailoring; over 150 studies have examined message framing effects on the promotion of self care behaviors (see Khberger, 1998; OKeefe & Jensen, 2006, 2007) Rothman and Salovey (1997) draw on Kahneman and Tverskys (1979) prospect theory and describe message framing as a strategy that involves manipulating how infor mation is presented t o affect peoples decisions and promote a specific behavior. Content of messages may be presented or framed in various ways, usually by valence emphasis on benefits (gains) or costs (losses) associated with health behaviors. Gain f ramed messages present benefits achieved by adopting a target behavior whereas lossframed messages convey costs of not adopting the target behavior (Salovey, 2005) Nearly all health related information can be framed in terms of gains and/or losses. In a ddition to valence a messages temporal proximity distance between performance of health behaviors and attainment of expected outcomes (immediate/short term or distal/longterm ) can be manipulated Predictions of valence framing and temporal proximity framing effects on health behaviors are often based on tenets from prospect theory ( Kahneman & Tversky, 1979)
5 and temporal construal theory (TCT) ( Liberman & Trope, 1998; Trope & Liberman, 2000), respectively These theories will be explained in Chapter 2. Early studies tested only main effects of framing manipulations on health behaviors, but this approach was overly simplistic and contributed to discordant findings (Myers, 2010) More recently research has focused on testing potential moderating variab les that influence message framing effects. These moderators can typically be categorized as situational or dispositional (individual differences) Historically nurses have recognized i ndividual differences in patient s (e.g., temperament or intelligence) and the impact these characteristics have on patients self care behaviors. Moderators of framing effects Consideration of future consequences (CFC) and motivational orientation are two moderators of framing effects that have been examined in the literat ure. These are described below. Consideration of f uture c onsequences CFC, a cognitive mindset refers to the extent to which people consider distant outcomes of their current behaviors and the extent to which people are influenced by these considerations (Strathman, Gleicher, Boninger, & Edwards, 1994) Tenets from TCT (Liberman & Trope, 1998; Trope & Liberman, 2000) help explain CFC effects on health behavior decisions. Strathman et al. (1994) hypothesized people low in CFC will focus more on immediate needs and concerns, acting to satisfy these and devaluing distant outcomes. Conversely, people high in CFC will consider future implications of behavior and act in accordance with distant goals. CFC has been found to moderate mes sage framing effects such as temporal proximity. For example, people high in CFC reported greater intentions to use sunscreen when positive outcomes were presented as distal and negative outcomes were presented as
6 immediate; the opposite was true for people low in CFC (Orbell & Kyria kaki, 2008). Nurse researchers recognize that many, if not most, behaviors result in short and longterm health consequences and have examined constructs similar to CFC such as subjective time experience (e.g., Sanders, 1986; Strumpf, 1987), future time pe rspective (e.g., Rew, Fouladi, & Yockey, 2002), and health temporal orientation (e.g., Russell, Champion, & Perkins, 2003; Russell, Perkins, Zollinger, & Champion, 2006). Motivational orientation. M otivational orientation e volved from and varies across several motivational theories. For example, s ome theories posit that approach orientation and avoidance orientation are two distinct motivational styles that influence decisions and behaviors; approachers respond more to rewards or incentives, whereas avoiders respond more to punishment or threat ( see Carver, Sutton, & Scheier, 2000, for a review ) Other theories ( e.g., Higgins, 1997, 1998, 1999) posit that goal orientation guides decision making and behaviors; people with a promotionfocus (promoter s) pursue goal s in a manner that ensures presence of positive outcome s whereas people with a prevention focus (preventers) pursue goal s in a manner that ensure s absence of negative outcome s. T hese two ideas appear related: an approach strategy is usual ly taken for promotion, and an avoidance strategy is usually taken for prevention (Higgins, 1997; Higgins et al., 2001 ) Similar behavioral predictions about message framing effects have been made for approachoriented and promotionfocused people, as well as for avoidance oriented and prevention focused people ( e.g., Lee & Aaker, 2004; Mann, Sherman, & Updegraff 2004; Rothman, Wlaschin, Bartels, Latimer, & Salovey, 2008). Although these two ideas have been examined in separate studies no health behavior study has drawn from an at risk community population to simultaneously
7 examine both constructs in the same sample. M otivational orientation has been found to moderate message framing effects such as valence. For example, avoiders (Mann et al., 2004) and preventers (Uskul, Sherman, & Fitzgibbon, 2009) were more persuaded by loss framed messages related to dental flossing, whereas approachers (Mann et al., 2004) and promoters (Uskul et al., 2009) were more persuaded by gainframed messages. Physical A ctivity According to the U.S. DHHS (2008), physical activity is defined as any bodily movement created by skeletal muscle contraction that increases energy expenditure above a basal level but it is often generally referred to as health enhancing activity. People who engage in only baseline activity ( sedentary or lightintensity activities of daily life such as sitting, standing, slowly walking, or lifting lightweight objects) are considered physically inactive Health enhancing physical activity (e.g., bri sk walking, dancing, weight lifting) is activity that, when added to baseline activity, produces health benefits Physical a ctivity can also yield other benefits such as providing opportunit ies to have fun, to be with friends, and to improve physical appea rance. Exercise is a form of physical activity but is narrower in scope. Exercise is planned, structure d, repetitive, and purposive with an overall goal to improv e or maintain physical performance, fitness, or health. All exercise is physical activity, but not all physical activity is exercise. The U.S. DHHS (2008) Physical Activity Guidelines for Americans suggest adults aged 18 years and older should perform at least 150 minutes per week of moderate intensity (e.g., brisk walking) or 75 minutes per week of vigorous intensity (e.g., running or jogging) aerobic physical activity or an equivalent combination of the
8 two. The guidelines also recommend adults should engage in muscle strengthening activities that work a ll major muscle groups and involve moderat e to high levels of intensity two or more days per week (e.g., resistance training or weight lifting). Adults aged 65 years and older or with disabilities are encouraged to follow the adult guidelines if possible or to be as physically active as the ir abilities and conditions allow Regular physical activity is a well established essential component of an overall healthy lifestyle that contributes to the promotion of health (e.g., increased energy, improved sleep, enhanced cognitive function) and the preve ntion of disease (e.g., type 2 DM, cardiovascular disease high blood pressure). Being physically active is one of the most important steps Americans of all ages can take to improve their health and fitness (U.S. DHHS, 2008). Despite this common knowledge many people remain inactive. According to the 2009 Behavioral Risk Factor Surveillance System (BRFSS) annual survey data (CDC, 2010), 49% of Americans did not meet the minimum recommendations for moderate or vigorous physical activity, and approximately 24% reported doing no physical activity in the preceding month. Call for Research Although effective health communication is a w idely r ecognized he alth behavior modification approach, the U.S. DHHS (1996, 2000b, 2000c) reported existing heal th communi cation efforts to increase healthy behaviors such as physical activity have fallen short of achieving their intended goals. Health communication, DM, and physical activity are 3 of the 28 focus areas in the U.S. Healthy 2010 initiative (U.S. DHHS, 2000a, 2000b, 2000c) The initiative calls for an increase in health communication evaluation
9 and research aimed at enhancing health care providers communication skills so that providers may design and deliver more effective messages to promote behavior changes. The Present Study The p resent study responded to the U.S. Healthy 2010 initiatives call for research (U.S. DHHS, 2000a, 2000b, 2000c) by examining two moderating factors that may help optimize message tailoring effects (specifically framing) on the promotion of physical activity in adults with type 2 DM. The present study attempted to replicate previous findings and extend these findings by exploring interactions not previously tested in message framing research. A summary of the key concepts and their hypothesized interrelationships is shown in Figure 1. The study was also novel in that it examined these relationships in a n at risk c ommunity population not previously targeted.
10 Inputs Outputs Moderating Variables (Individual Differences) Figure 1. Logic Model of the Present Study. CFC = Consideration of Future Consequences; BIS/BAS = Behavioral Inhibition System/Behavioral Activ ation System; RFQ = Regulatory Focus Questionnaire. Purpose, specific a ims and h ypotheses. The overall purpose of the present study was to employ a multi theoretical, integrated approach to message framing and to examine its effects on health behaviors. More specifically, th e aim was to examine how two individual differences factors CFC and motivational orientation may combine to moderate framing effects (temporal proximity and valence ) on intentions to increase physical activity. Based on the logic model in Figure 1 and the literature that will be reviewed in Chapter 2, the following hypotheses were posited: H1. When valence of the message is congruent with individuals motivational orientation (i.e., gain framed :: approachoriented /promotionfocu sed and loss framed :: avoidance oriented/preventionfocused ) intentions to increase physical Message Framing Manipulation Temporal proximity Valence Intentions to Increase Physical Activity Intentions index score Consideration of Future Consequences CFC scale score Motivational Orientation BIS/BAS scale score RFQ scale score
11 activity will be greater than when the message is incongruent with individuals motivational orientation H2. When temporal proximity of the message is congrue nt with individuals' status on CFC (i.e., immediate framed :: low CFC, and distal framed :: high CFC) intentions to increase physical activity will be greater than when the message is incongruent with individuals CFC. H3. When both valence and temporal p roximity of the message are congruent with individuals motivational orientation and CFC standing, intentions to increase physical activity will be at their highest. Targeted p opulation The targeted population in the present study was adults with type 2 D M. In the United States alone, an estimated 23.6 million people 7.8% of the population have DM; type 2 diabetes accounts for 90 95% of all diagnosed cases in adults; and approximately $174 billion was spent on diabetes costs in 2007 (National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK], 2008). Physical inactivity is one of several modifiable health behaviors that contribute to the prevalence of type 2 DM (WHO, 2005b). Based on the U.S. DHHS (2008) Physical Activity Guidelines for Am ericans and empirical studies, the American Diabetes Association (ADA, 2010) recommends people with diabetes perform at least 150 minutes per week of moderate intensity aerobic physical activity In addition, people with type 2 DM (in the absence of contra indications) are encouraged to engage in resistance training three times per week (ADA, 2010) It is wellestablished that regular physical activity is an effective diabetes self care strategy that will contribute to the attainment of several individual he alth benefits (e.g., improved blood glucose control, weight loss), the optimization of
12 overall management of the disease, and a reduction in the national economic burden of diabetes costs. Chapter Summary Modifiable health behaviors are largely responsi ble for high prevalence and incidence of chronic diseases such as type 2 DM (WHO, 2005b) Physical activity yields multiple health benefits and is an essential self care component of DM management (ADA, 2010 ; U.S. DHHS, 1996, 2008) However, many people wi th DM remain physically inactive. The present study draws from prospect theory, TCT, and motivation theories as a conceptual framework to test how two individual differences factors CFC and motivational orientation may combine to moderate framing effec ts (temporal proximity and valence ) on intentions to increase physical activity in adults with type 2 DM A multitheoretical, integrated approach was used to provide a richer, fuller understanding of framing effects on health behaviors and to guide nurses and other health care pro viders in designing and delivering effectively tailored health messages to promote healthy self care behaviors.
13 Chapter 2: Review of Literature This chapter presents the definitions typologies, and origins of message framing, theoretical and conceptual frameworks of message framing research, examples of e mpirical studies from the message framing literature pertinent to the present study, a summary of three metaanalytic reviews on message framing and a synopsis of the curre nt state of the message framing literature pertaining to health related behaviors Message Framing Definitions and Typologies Message framing involves manipulating the context in which i nformation is considered when presented. The ultimate goal of message framing is usually to promote a particular behavior (Rothman & Salovey, 1997). Message framing effects are complex, and the empirical literature lacks consistency, as later described. These discrepant findings may partially be attributed to the absence of a universal operational definition of message framing. Instead, a variety of definitions exist (Wilson, Purdon, & Wallston, 1988), ranging from loose to strict interpretations (Khberger, 1998). Researchers have developed typologies/classification sch emes of message framing in an attempt to operationally define it and demonstrate various ways to frame messages (e.g., Fagley, 1993; Levin, Schneider, & Gaeth, 1998; Rothman & Salovey, 1997; Rothman, Salovey, Antone, Keough, & Martin, 1993; Tversky & Kahne man, 1981; Wilson et al., 1988).
14 These various definitions and typologies reveal that messages can be framed in more than one way, but most commonly they are framed by valence in terms of gains (benefits) or in terms of losses (costs). Gain framed messag es typically present benefits achieved by adopting a target behavior whereas lossframed messages usually convey costs of not adopting the target behavior (Rothman & Salovey, 1997; Salovey, 2005). Nearly all health related information can be framed in ter ms of gains and/or losses. In addition, messages can be framed by temporal proximity, where the expected outcomes of the health behavior (benefits and/or losses) are presented as occurring immediately (short term) or distally (long term). The present study and corresponding review of literature primarily focused on the valence and temporal proximity definitions of message framing. Origins of Message Framing Decision making under risk involves a choice between prospects or gambles. Historically, expected uti lity theory has dominated the analysis of decision making under risk (Kahneman & Tversky, 1979) and involves assigning expected values to final assets of prospects (choice options). The utilities of outcomes are weighted by their probabilities of occurrenc e and are used to determine the overall utility of each choice option (see Keeney & Raifa, 1976, and von Neumann & Morgenstern, 1944, for more details about expected utility theory). Kahneman and Tversky (1979) demonstrated several phenomena that systemati cally violate the basic tenets of expected utility theory and thus concluded that the expected utility theory was an inadequate descriptive model of decision making under risk. As a result, they proposed an alternative model of risky choice prospect theo ry to better understand preference and decision making under conditions of uncertainty.
15 In comparison to expected utility theory, prospect theory examines the subjective values and subjective probabilities of choice options rather than the objective out comes of wealth and welfare. Kahneman and Tversky (1979) describe the risky choice process as consisting of two phases: editing and evaluation. The editing phase involves a preliminary analysis of all possible choices where options are organized and reform ulated and outcomes are coded in terms of gains or losses, relative to a neutral reference point or asset position. The subsequent evaluation phase involves evaluating each edited option for overall value and choosing the option of highest value. Prospect theory proposes that when potential losses of a situation are made salient and behavioral choices involve risk or uncertainty, people are generally riskseeking and will more likely assume these risks. Conversely, when potential gains of a situation are ma de salient and behavioral choices pose minimal risk or minimal uncertainty, people are generally risk averse and will more likely act to avoid the risks ( Kahneman & Tversky 1979; Tversky & Kahneman 1981) Tversky and Kahneman (1981) introduced and test ed how prospect theory could be applied to the framing of decisions by presenting a series of hypothetical decision problems to college students Tversky and Kahneman (1981) describe a decision problem as one that can be defined by the options or acts pe ople must choose from, the possible consequences or outcomes of these acts, and the conditional probabilities (contingencies) of outcomes occurring given a particular act. They use the term decision frame to refer to ones conception of the acts, outcome s, and contingencies related to a specific choice. They propose several factors that influence which frame a decision maker will adopt, including ones cultural norms, habits, and personal characteristics and the formulation of the problem. T heir overall f indings supported basic tenets of prospect
16 theory. People tended to be sensitive to whether the valence of alternatives w as framed in terms of associated costs or benefits when the situations were objectively equivalent. The prototypical example of risky choice framing effects comes from Tversky and Kahnemans (1981) Asian disease problem, in which a hypothetical disease was expected to kill 600 people. Different pairs of effectively identical alternative options for responding to the outbreak were prese nted to research participants, expressed as outcomes of either the number of lives saved (gain framed) or lost (loss framed) In addition, within each framing condition, the pair of options differed in terms of the probability (certain vs. uncertain) and m agnitude (number or proportion of the lives) of the outcome. Consistent with prospect theory, the majority of respondents chose the riskaverse option when gains were certain (i.e., lives saved) and the risk seeking option when losses were certain (i.e., l ives lost). To illustrate, in the gain framed condition, the prospect of saving 200 lives with certainty (riskaverse) was more appealing than a one in three chance of saving 600 lives and a twoin three chance of saving no lives (riskseeki ng). In the los s framed condition, the one in three chance that nobody would die and the two in three chance that 600 people would die (risksee king) was more acceptable than the certain death of 400 people (riskaverse) P rospect theory and these early empirical finding s laid the foundation for the future of message framing research. Temporal construal theory (TCT) (Liberman & Trope, 1998; Trope & Liberman, 2000) has also guided message framing research especially related to the manipulation of a health messages tempo ral proximity the distance between performance of health behaviors and attainment of expected outcomes. TCT posits that temporal proximity systematically changes the way people represent (construe) certain actions (e.g., health
17 behaviors) and events (e.g ., outcomes) which in turn alters how people evaluate judge, and choose among these actions and events According to TCT, people tend to use higher level ( more schematic, abstract ) construal s to represent information about distant future situations and l ower level (less schematic, concrete ) construal s to represent information about immediate future situations (Liberman & Trope, 1998; Trope & Liberman, 2000) To illustrate, a high level construal may represent engaging in regular physical activity as im proving overall health and well being, whereas a low level construal may represent the same behavior as walking two miles a day on the treadmill at the gym . High and low level construal s also differ in their emphasis on feasibility versus de sirability and probability versus prize value. Feasibility refers to the ease or difficulty of reaching the outcome and represents a low level construal. Conversely, desirability refers to the value of an action s outcome (endstate) and r epresents a high level construal (Liberman & Trope, 1998; Trope & Liberman, 2000) Sagristano, Trope, and Liberman (2002) applied the distinction between feasibility and desirability to gambles and found that people rated safe bets (high probability of winning a small priz e value) as more appealing in the near future and risky bets (low probability of winning but large prize value) as more appealing in the distant future. Thus, they proposed that probability represents a low level construal and prize value represents a high level construal. With regard to decisionmaking, TCT suggests that decisions about the immediate future are largely influenced by low level aspects (e.g., feasibility and probability ) of the involved actions and events, whereas decisions about the distant future are largely influenced by highlevel aspects (e.g., desirability and prize value ) (Liberman & Trope, 1998; Sagristano et al., 2002; Trope & Liberman, 2000). This suggestion may
18 help explain why people choose to remain physically inactive today despite the well known benefits of regular physical activity tomorrow People may perceive regular physical activity as having high desirability but low feasibility. With immediate future decisions, feasibility often trumps desirability. People may be w illing to sacrifice future benefits of being physically active such as more energy a nd a healthier heart (large prize value but abstract and less certain) in order to gain immediate benefits of not being physically active such as more time to watch televis ion and play computer games (small prize value but concrete and more certain). Other Theories and Conceptual Models of Message Framing in Health Research Historically, prospect theory has been a primary framework for understanding preference and decision making under conditions of uncertainty and the dominant underlying theoretical perspective for message framing (Kahneman & Tversky, 1979, 1984; Tversky & Kahneman, 1981) as previously described Empirical evidence suggests, however, that prospect theory does not solely explain the effects of message framing on all behavior s under all circumstances, specifically those regarding health related behaviors (Levin et al., 1998; Rothman & Salovey, 1997; Wilson et al., 1988). The theory inadequately addresses the mechanisms and conditions under which message framing alters peoples attitudes, beliefs, and behaviors. Some researchers have found that prospect theory alone is insufficient and have offered alternative theoretical perspectives to help explain why peopl e vary in their responses when presented with subjectively different but objectively equivalent descriptions of the same decision problem. A few of these alternative perspectives are described next.
19 Nearly all health related information can be framed in terms of either benefits (gains) or costs (losses). However, the literature is replete with inconsistent findings, as later described. Rothman and Salovey (1997) offer potential explanations of how and why prospect theory may contribute to these inconsiste nt predictions. One plausible possibility is the unique differences between healthrelated decision problems and decision problems initially tested in prospect theory. First, health decision problems are often more real istic dealing with personal issues r ather than hypothetical public health issues. Second, health decision options are sometimes non discrete and compound (i.e., may consist of more than two options). Third, perceived risk for health behavior decisions is more subjective because formal probabilities of outcomes occurring as a result of these decisions are often unknown. Fourth, in health behavior message framing research, experimenters often have less control over the situations in which framed messages are predicted to exert influence than in itial researchers working in laboratory settings This factor could undermine any systematic test of prospect theorys predictions and result in inconsistent patterns of findings in health behavior research. Rothman and Salovey (1997) conclude that despite limitations of prospect theory, its basic assumptions can be operationalized and tested in health behavior research if careful attention is paid to the context in which a health message is delivered Rothman and Salovey (1997) propose three stages in the decision making process during which the relative influence of gainand loss framed messages may be examined. The likelihood that people respond to message framing in a manner consistent with prospect theory varies over these three stages. First, the amount of attention or cognitive processing people direct to the message can influence the degree to which they integrate
20 the message into a mental representation of the health issue. Second, peoples receptivity to, or acceptance of, a particular frame th at a message advocates can also affect framing effects, and peoples past and current experiences influence this receptivity. Third, peoples perceived function of the advocated health behavior (i.e., prevention, detection, or recuperative) can influence m essage framing effects. Rothman et al. (1993) and Rothman and Salovey (1997) describe health behaviors as serving one of three functions: to prevent onset of a health problem (e.g., regular physical activity will help prevent high blood glucose), to detect the development of a health problem (e.g., self monitoring of blood glucose will detect abnormal blood glucose levels), or to cure or treat an ongoing health problem (e.g., insulin administration will help keep blood glucose levels within the desi red rang e). The effectiveness of message framing is partly based on whether taking action is perceived to involve risk or uncertainty. For example, people tend to perceive performance of a detection behavior as risky (e.g., it may reveal an unpleasant find ing) and performance of a prevention behavior as relatively safe (e.g., it maintains ones health status). Based on theoretical principles and the empirical literature, Rothman and Salovey (1997) draw the following conclusions: loss framed messages are predicted to be most effective in promoting detection behaviors and gainframed messages are predicted to be most effective in promoting prevention behaviors. Many behaviors are typically construed as having just one function (i.e., prevention, detection, or curing), but some behaviors may be perceived as serving multiple functions. For example, some women may perceive undergoing a Pap anicolaou (Pap) test as serving a detection function because it detects the presence or absence of cervical problems. However, some women may also perceive undergoing a
21 Pap test as serving a prevention function because if it reveals mild cervical abnormalities, early interventions can be done to try and prevent further, more severe abnormalities (e.g., cervical cancer). Wilson et al. (1988) emphasize consequences as a central feature in theoretical frameworks of healthbehavior change and identify two recurring dimensions associated with behavior change research that involve consequences: perceived value and perceived threat. Besi des prospect theory, numerous theories emphasize perceptions of the value or threat of an outcome contained in a recommendation, such as the health belief model (Hochbaum, 1958; Rosenstock, 1960), health promotion model (Pender, 1982), theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975), theory of planned behavior (Ajzen, 1985, 1988, 1991), and protection motivation theory (Rogers, 1975). Several researchers have integrated tenets from prospect theory with one or more of these other t heories to study message framing effects on health behaviors (e.g., Jones, Sinclair, Rhodes, & Courneya, 2004; McCall & Martin Ginis, 2004). Studies are also emerging that integrate persuasion theories with prospect theory. One example is Petty and Caciop pos (1986) elaboration likelihood model (ELM) of persuasion effects. The basic premise of ELM is that a messages persuasive ability to influence a persons change in attitude about a particular issue or argument depends on how likely the person will elaborate upon (i.e., think about) this issue or argument. Several variables drawn from ELM have been shown to produce moderating effects on message framing. For example, Jones et al. (2003) found that source credibility moderates the effect of message framing on exercise intentions, exercise behaviors, and cognitive response/elaboration measures. People who received a gain framed message
22 from a credible source elaborated the message the most and reported the greatest amount of exercise intentions and behaviors Studies are also emerging that integrate motivation theories with prospect theory For example, a ccording to several theories of motivation (see Carver et al. 2000, for a review), behavior is regulated by two distinct brain systems that guide responses to stimuli of reward and punishment : the behavioral activation system (BAS) and the behavioral inhibition system (BIS) (often referred to as the approach and a voidance system s, respectively) Both systems (presumably orthogonal) represent chronic di spositional motivation styles, where one system ( BAS ) regulates appetitive behavior toward actual or potential rewards and the other system ( BIS ) regulates aversive behavior away from potential threats or punishments (Carver & White, 1994; Gray, 1982, 1990). Therefore, people with a predominant approachorientation (high BAS) should respond more to cues of reward or incentive, whereas people with a predominant avoidance orientation (high BIS) should respond more to cues of punishment or threat (Carver e t al., 2000). Another example of motivation theory is the r egulatory focus theory (Higgins, 1997, 1998, 1999) which predicts that goal orientation (or regulatory focus) is the dominant motivation system that guides decision making and behavior. This theo ry distinguishes between two types (presumably orthogonal) of goal orientation: promotionfocus and prevention focus. People with a promotionfocus (promoters) are motivated by advancement and accomplishment and eagerly pursue goal s in a manner that ensu re s presence of positive outcomes (e.g., they regularly exercise to achieve optimal blood glucose control). People with a preventionfocus (preventers) are motivated by security
23 needs and vigilantly pursue goal s in a manner that ensure s absence of negati ve outcome s (e.g., they regularly exercise to avoid high blood glucose). The regulatory focus motivation system reflects specific types of desired end states (final outcomes) rather than broad dispositions as emphasized in the approach/avoidance system ( Hi ggins, 1997) Although regulatory focus is posited to be a stable dispositional characteristic, situational variables may also temporarily induce a preventionor promotionfocused mindset (Higgins, 1997; Higgins et al., 2001). Rothman, Bartels, Wlaschin and Salovey (2006) and Rothman et al. (2008) provide a new conceptualization of framing effects on health behaviors. They acknowledge two dominant perspectives that to date have guided researchers in understanding the conditions under which gainand los s framed messages should be maximally persuasive: situational factors (i.e., differences in health behavior s function) and dispositional factors (i.e., individual differences in sensitivity to favorable or unfavorable outcomes). Both factor types have bee n tested separately as potential moderators in regulating persuasiveness of gainand loss framed messages in the promotion of health behaviors. Rothman et al. (2006) and Rothman et al. (2008) suggest that both sets of moderating factors may rest on a sing le set of underlying cognitive and affective processes, based on tenets of the regulatory focus theory, and thus propose the two factors should be measured simultaneously when testing message framing effects. They hypothesize that health behaviors can evok e either a promotion or preventionfocus mindset, and this effect is influenced by both features of the behavioral domain and characteristics of the individual. More specifically, they predict when people consider performing a behavior intended to promote health (e.g., physical activity), people will
24 experience thoughts and feelings consistent with a promotionfocus mindset, whereas when people consider performing a behavior intended to detect the presence of a health problem (e.g., exercise stress test), people will experience thoughts and feelings consistent with a preventionfocus mindset. The variability in how people interpret a given behavior (as either promotion or detection) will moderate the relative effectiveness of the framed message. Considerati on of this integrated approach of situational and dispositional factors has only recently appeared in the literature In summary, many theories and conceptual frameworks have been proposed to explain message framing effects on health behaviors. Early st udies of message framing typically involved just one theoretical approach, namely prospect theory. More recently, studies have involved the integration of two or more theories. Strong evidence supports that no one theory or model can solely explain all the message framing effects found in the literature Multi theoretical integrated approaches are necessary to better understand the complexity of these effects. Empirical Studies Involving Message Framing and Health Behaviors Prospect theory was initially tested in laboratory settings using discrete choice decisions involving monetary outcomes (e.g., gambling and purchasing) and hypothetical situations. Since Tversky and Kahnemans (1981) original Asian disease problem, many studies have used message frami ng to test the preference reversal prediction of prospect theory across a broader range of decision problems (Rothman et al., 1993). In particular, there has been a recent increase in the study of message framing effects on health behaviors.
25 Myers (2010) conducted a review of the literature to assess the current state of message framing research in health behaviors. She used three methods to locate relevant empirical studies: computerized database searches of CINAHL, PsycINFO, and PubMed, using message fr aming as a search term; examination of previous reviews and textbooks; and inspection of reference lists in previously located studies. She found that over the past 20 years, 25 plus health related behaviors have been studied in message framing research Examples of these behaviors include smoking (e.g., Moorman & van den Putte, 2008), drinking (e.g., Gerend & Cullen, 2008), exercise/physical activity (e.g., Jones et al., 2003), eating habits/behaviors (e.g., Tykocinski, Higgins, & Chaiken, 1994), dental hygiene (e.g., Sherman, Mann, & Updegraff 2006), health information seeking (OConnor, Warttig, Conner, & Lawton, 2009), sexual health promotion (e.g., OConnor, Ferguson, & OConnor, 2005), HIV testing (e.g., Apanovitch, McCarthy, & Salovey, 2003), breast feeding (e.g., Wolf, 2007), prostate exams (e.g., Cherubini, Rumiati, Rossi, Nigro, & Calabro, 2005), testicle self exams (e.g., Steffen, Sternberg, Teegarden, & Shepherd, 1994), breast self exams (e.g., Meyerowitz & Chaiken, 1987), mammograms (e.g., Banks et al., 1995 ), Pap tests (e.g., Lauver & Rubin, 1990), colonoscopy screenings (e.g., Canada & Turner, 2007), skin cancer prevention (e.g., Rothman et al., 1993), vaccinations (e.g., Gerend & Shepherd, 2007), and hand hygiene (e.g., Jenner, Jones, Fletcher Miller, & Scott, 2005). Many of these studies examined message framing effects on i ntention s to perform behaviors rather than on actual behavior performance. According to the theory of reasoned action (Ajzen & Fishbein, 1980; Fishbein & Ajzen, 1975), intentions to perform (or not perform) a given behavior are the most proximal (immediate) determinant of
26 behavior. A recent metaanalysis of metaanalyses revealed a strong correlation ( r = .53) between intention and behavior, indicating that on average, int entions explain 28% of the variance in future behavior (Sheeran, 2002). A few examples of message framing research in health behaviors are presented next. First, studies are presented that found main effects of message framing in detection, prevention, and mixed function (detection and prevention) behaviors Next, studies are presented that found moderating effects on message framing in a variety of health behaviors. Finally, disc ordant findings in the message framing empirical literature are briefly dis cussed along with possible explanations Detection behaviors. Several studies revealed findings consistent with R othman and Saloveys (1997) prediction that loss framed appeals are more effective in promoting illness detection (screening) behaviors than gainframed appeals. For example, Meyerowitz and Chaiken (1987) were the first to examine the relative influence of gain and loss framed information on health behavior T hey hypothesized that a pamphlet emphasizing the negative consequences of not pe rforming breast self examination (BSE) (loss framed) would be more persuasive than a pamphlet emphasizing positive consequences of BSE ( gainframed). Seventynine undergraduate female college students were randomly assigned to read a gain framed message, a loss framed message, a no arguments message, or no message at all about BSE. I ntentions to perform BSE w ere assessed immediately after the intervention and 4 months later during a phone interview As predicted, women who read the lossframed pamphlet repo rted greater intentions and more frequent BSE behavior 4months after the intervention than women in the other three conditions.
27 Loss framed messages have also been found to be more effective to promote mammography screenings. Banks et al. (1995) compared the effectiveness of gain versus loss framed messages to persuad e women to obtain mammography screening. The y recruited 133 women 40 years and older not adhering to current guidelines for obtaining mammography screening. These women were randomly assigned to view either gain framed or loss framed factually equivalent educational video presentations on breast cancer and mammography. Mammography utilization was assessed at 6 and 12month intervals via phone interviews. As hypothesized, women in the loss fra med condition were more likely to have obtained a mammogram within 12 months of the intervention. M essage framing effects on dental detection behaviors have also been examined. Rothman Martino, Bedell, Detweiler, and Salovey (1999, Experiment 2) recruited undergraduate college students to test the relative effectiveness of gain and loss framed messages to promote use of a disclosing mouth rinse to detect plaque. Immediately after reading the pamphlet on dental health, behavioral intentions were assessed. As predicted, students who read the lossframed message reported stronger intentions to buy and use the disclosing mouth rinse within the next week than those who read the gain framed message. Furthermore, a significantly greater percentage of students i n the loss framed condition requested a sample of the product than did those in the gain framed condition. Prevention behaviors. Several studies have also revealed findings consistent with Rothman and Saloveys (1997) prediction that gain framed appeals have a more effective impact on health affirming (prevention) behaviors than loss framed appeals. For example, i n the same experiment just described, R othman et al. (1999, Experiment 2) recruited undergraduate college students to also test the relative ef fectiveness of gain and
28 loss framed messages in promot ing use of a mouth rinse to prevent (rather than detect) dental health problems. Immediately after reading the pamphlet on dental health, behavioral intentions were assessed. As predicted, students who read the gainframed message reported stronger intentions to buy and use the mouth rinse within the next week than those who read the loss framed message. A significantly greater percentage of students in the gainframed condition requested a sample of the product than did those in the loss framed condition. Gain framed messages have also been found to be more effective in promot ing sunscreen use to prevent skin cancer. Detweiler, Bedell, Salovey, Pronin, and Rothman (1999) compared the effectiveness of four different framed messages to persuade 217 adult beachgoers to obtain and use sunscreen. There were two gain framed conditions ( a gain or a nonloss) a nd two loss framed conditions ( a loss or a nongain ). Beach goers were approached while at the beac h and were assigned to read one of four brochures about skin cancer and the use of sunscreen. Intentions were assessed before and immediately after the intervention. There was also a behavioral measure, where participants were given a coupon upon completing the questionnaire that was redeemable later that day for a free sunscreen sample. As predicted, people who read either gain framed brochure reported higher intent to repeatedly apply sunscreen while at the beach and to use sunscreen with a sun protection factor of 15 or higher and they r equest ed sunscreen more often (i.e., redeem ed the coupon) than those who read either loss framed brochure. The gainframed advantage was strongest among beach goers who had not planned to use sunscreen that day prior to t he intervention.
29 M essage framing effects on p hysical activity have also been examined For example, Latimer, Rench, et al. (2008) recruited 322 sedentary, healthy callers to the U.S. National Cancer Institutes Cancer Information Service to compare the ef fectiveness of gain, loss and mixedframed messages on promoting moderate to vigorous physical activity. Participants randomly received one of three framed messages on three separate occasions ( via telephone at baseline and via print at Week s 1 and 5). Intentions and self reported physical activity were assessed at baseline, Week 2 and Week 9 via phone interview At Week 2, gainand mixedframed messages resulted in stronger intentions to engage in physical activity than loss framed messages. In additi on, as predicted at Week 9, people who read gain framed messages self reported greater physical activity participation than those who read loss or mixedframed messages. Van t Riet, Ruiter, Werrij, and De Vries (2010) also found a gainframed advantage for physical activity. They conducted a webbased study with 787 adults living in the Netherlands, of which 299 completed all measures. Participants were randomly assigned to read either a gain or loss framed message about physical activity Intentions to be physically active were assessed prior and immediately after reading the message. Physical activity levels were also assessed at a 3 month follow up via email. As predicted, people who read gain framed messages had stronger intentions and marginally gre ater levels of physical activity than those who read loss framed messages. Mixed function behaviors. Researchers have also examined message framing effects on health behaviors that may b e perceived as having mixed functions (e.g., detection and pre vention ). For example, Rivers, Salovey, Pizarro, Pizarro, and Schneider (2005) conducted a n experiment with 441 women who attend ed an urban community
30 health clinic to test the hypothesis that loss and gainframed messages differentially influence Pap test utilization behaviors depending on the risk involved in performing the behavior. Participants were randomly assigned to view one of four video presentations about the importance of obtaining an annual Pap test. Self reported Pap test utilization was assessed at 6 and 12 months via telephone following the intervention Consistent with Rothman and Saloveys (1987) predictions, when Pap tests were presented as having a detection function, loss framed messages that emphasized costs of failing to detect cervical cancer early (a risky behavior) were more persuasive in motivating women to obtain a Pap test than were gain framed messages. However, when Pap tests were presented as having a prevention function, gainframed messages that emphasized the benefits of pre venting cervical cancer (a less risky behavior) were more persuasive than loss framed messages. Other researchers have found similar results with different mixed function health behaviors (e.g., Hsiao, 2003; Rothman et al., 1999). Moderators of message f raming. Whereas several early studies tested only main effects of framing on health behaviors, more recent research on message framing has focused on identifying and examining variables that moderate framing effects In Myers (2010) review of the literatu re, she found over 20 variables that have been examined as moderators of message framing effects. Some of these variables are situational such as health behavior type or function (e.g., Hsiao, 2003), framing method (e.g., Ferguson & Gallagher, 2007), temporal context (e.g., Gerend & Cullen, 2008), and type of value appeal (health vs. self esteem e.g., Robberson & Rogers, 1988). Other variables are dispositional and represent individual differences such as cognitive processing style (e.g., Meyers Levy & Mah eswaran, 2004), issue involvement (e.g., Rothman et al., 1993),
31 personal relevance (e.g., McElroy & Seta, 2003), stages of readiness to change (e.g., Hsiao, 2003), need for cognition (e.g., Rothman et al., 1999), perceptions of benefits, risks/threats, susceptibility, and/or disease severity (e.g., Toll et al., 2008), behavioral norms (e.g., Blanton, Stuart, & VandenEijnden, 2001), perceived source credibility (e.g., Arora, Stoner, & Arora, 2006; Jones et al., 2003), motivational orientation (e.g., Gerend & Shepherd, 2007; Mann et al., 2004 ; Uskul et al., 2009) and consideration of future consequences (e.g., OConnor et al., 2009) A few studies involving three of these variables motivational orientation, temporal context and CFC are presented next. Motivational orientation. Motivational orientation has been found to moderate message valence framing effects on health behaviors. For example, in relation to the approach/avoidance system previously described, Mann et al. (2004) developed a congruency hypothesis which predicts that health messages framed to be aligned with peoples approach/avoidance motivations will be the most effective in promoting health behaviors. For example, a gain framed message should be more persuasive with a person who is pred ominantly approachoriented, and a loss framed message should be more persuasive with a person who is predominantly avoidance oriented. S everal studies lend support for this hypothesis Carver and Whites (1994) BIS/BAS scale (which consists of a BIS scal e and a BAS scale) has been used to assess motivational orientation. Mann et al. (2004) tested the congruency hypothesis with 63 undergraduate college students who did not floss regularly Participants first completed the BIS/BAS scale and were then random ly assigned to read either a gain or loss framed article on dental flossing. They were also given samples of floss and were instructed to use them. Self reported flossing behavior
32 was measured via written questionnaire one week after the intervention. As predicted avoiders were more persuaded by lossframed messages, whereas approachers were more persuaded by gainframed messages. In a follow up study, Sherman et al. (2006) used a similar procedure with 67 undergraduate students who did not floss re gularly, but measures of message perception, self efficacy, and intentions were added in an effort to examine t he psychological pathways through which the interaction of dispositional motivations and message framing leads to health behavior change. A s pred icted, people who read a congruently framed message intended to floss more and used more dental flosses than those who read an incongruent message. There was also evidence that self efficacy and intentions mediated the congruency effect. In another follow up study, Updegraff, Sherman, Luyster, and Mann (2007) also explored if argument strength moderated the congruency effect One hundred and thirty six undergraduate college students who did not regularly floss read either a strong or weak message about den tal flossing with a frame (gain vs. loss) that either matched or mismatched their motivational orientation (approach vs. avoidance). Results showed participants were sensitive to argument strength in the matched but not in the mismatched conditions. A rgume nt strength moderated the congruency effect on self reported flossing behaviors. When arguments were strong, matching the message to motivations yielded favorable effects on flossing behavior; however, when arguments were weak, matching the message to moti vations led to noticeably worse effects on these outcomes. These findings suggest strong arguments may be necessary when attempting to optimize the impact of message frame and motivational orientation
33 Gerend and Shepherd (2007) tested the congruency hypothesis with 121 female undergraduate college students. Participants were randomly assigned to read a gain or loss framed booklet on the promotion of a vaccine against human papillomavirus (HPV). After reading the message, participants completed a post ma nipulation survey, including the BIS/BAS scale and a measure of intentions. As predicted, a lossframe advantage was observed for people high in avoidance motivation. Avoiders who read the loss framed message reported greater HPV vaccination intentions t han those who read the gain framed message. However, a gain frame advantage was not observed for people high in approach motivation. S imilar to the congruency hypothesis in the approach/avoidance system (Mann et al., 2004) Higgins (2000) propose d the value from fit hypothesis in the promotion/prevention system. The value from fit hypothesis posits when people pursue goals in a manner consistent with their regulatory focus, they experience a sense of fit that increases the value of the health behavior and thereby increases the behaviors likelihood of occurrence Several studies support this hypothesis and have found that a message framed to fit peoples regulatory focus is more persuasive. For example, a gain framed message should be more persuasiv e with a person who is predominantly promotionfocused, and a loss framed message should be more persuasive with a person who is predominantly preventionfocused. Scales such as the regulatory f ocus s cale (Lockwood, Jordan, & Kunda, 2002) and the R egulato ry Focus Questionnaire (RF Q) (Higgins et al., 2001) have been used to assess dispositional regulatory focus. Both scales consist of a promotion focus and a prevention focus scale. Latimer Rivers, et al. (2008) examined the effectiveness of
34 regulatory fit messages for increasing physical activity among inactive people. Two hundred and six sedentary adult participants recruited from the National Cancer Institutes Cancer Information Service were randomly assigned to receive either promotion (benefits) focuse d or prevention (costs) focused messages encouraging physical activity. Two weeks later, participants completed a follow up phone interview that assessed regulatory focus (using the RFQ) and amount of physical activity over the previous 7 days. As predicted, tailored messages that fit peoples regulatory focus led to greater physical activity participation than nonfit messages, particularly in the promotionfocused condition. In the preventionfocused condition, patterns of behavior were as predicted but w ere not significantly different. Uskul et al. ( 2009) examined the value from fit hypothesis with 100 undergraduate students from two cultural groups (White British and EastAsians) likely to differ in their chronic dominant regulatory focus Participants completed the regulatory focus scale and then read a randomly assigned gain or loss framed message about dental flossing. They then answered items related to attitudes towards flossing and intentions to floss over the following week. Attitude and intention scores were highly correlated and were therefore combined to form one index of persuasion. As predicted, the mediated moderation analysis revealed three two way interactions: message frame x cultural background, regulatory focus x cultural background, a nd message frame x regulatory focus (the latter which supports the value from fit hypothesis) In addition, the message frame x regulatory focus interaction mediated the message frame x cultural background interaction White British participants (who had an overall stronger promotion focus) assigned to read the gain framed message were more persuaded to floss than those who
35 read the loss framed message, whereas East Asian participants (who had an overall stronger prevention focus) assigned to read the loss framed message were more persuaded to floss than those who read the gainframed message. Latimer, Williams Piehota, et al. (2008) recruited 518 adults from the National Cancer Institutes Cancer Information Service, who were randomly assigned to receive either prevention or promotion oriented messages encouraging fruit and vegetable intake. Messages were mailed at 1 week, 2 months, and 3 months after a baseline interview. The RFQ was used to assess regulatory focus. Follow up interviews were conducted v ia telephone at 1 and 4 months to assess participants actual fruit and vegetable intake and to determine if they were meeting or failing to meet the A Day guideline. The pattern of findings was consistent with the value from fit hypothesis but the interactions were only marginally significant. At the 4 month follow up, promoters who read the promotionoriented message were somewhat more likely to meet the A Day guideline than preventers, whereas preventers who read the prevention oriented message w ere somewhat more likely to meet the guideline than promoters. No differences in behavior were noted between groups at the 1 month follow up. Temporal context. Temporal context is another variable that has been shown to moderate message valence framing effects. Gerend and Cullen (2008) evaluated the interactive effects of valence framing and temporal context on college student alcohol use. Two hundred and twentyeight participants were randomly assigned to read an alcohol prevention message tha t varied by message valence frame (gains vs. losses) and temporal context (short vs. longterm consequences). Alcohol drinking behavior was assessed one month post intervention. An interaction was found as predicted. Students
36 who read the gainframed message reported lower alcohol use as compared to those who read the loss framed message, but only when consequences were presented as short term. Valence frame had no effect when consequences were presented as long term. TCT predictions suggest that the leve l of construal (high vs. low) with which valence is associated, rather than valence itself, determines the effect of temporal distance on decision making. In a series of five nonhealth behavior studies with college students, Trope and Liberman (2000) found that for decisions involving the distant future, people chose options that ha d positive high level construals but negative low level construals and reject ed options that ha d negative highlevel construals but positive low level construals. The reverse wa s true for decisions involving the immediate future. Consideration of future consequences. CFC is a cognitive mindset that refers to the extent to which people consider distant outcomes of their current behaviors and the extent to which people are influe nced by these considerations (Strathman et al., 1994) CFC has been found to moderate message temporal framing effects. Strathman et al. (1994) hypothesized that people low in CFC focus more on their immediate needs and concerns and will therefore act to satisfy these immediate needs. Conversely, people high in CFC consider the future implications of their behavior and act in accordance with their distant goals. Seve ral studies lend support to this hypothesis. Strathman et al.s (1994) CFC scale is often u sed to assess CFC. Orbell, Perugini, and Rakow (2004) were the first to extend Strathman et al.s (1994) CFC hypothesis in the health domain. They used a 2 (time frame ) x 2 (order of positive and negative consequences) x 2 (CFC) design to construct four me ssages about bowel (colorectal) cancer screening Each message contained t wo positive and two negative consequences
37 where the order of these consequences was counterbalanced Two hundred and twenty 50 69 year old people from the community read one versi on of the message and then completed post manipulation measures, includ ing t he CFC scale. As predicted, a CFC main effect and CFC x time frame interaction were obtained. For the main effect, people high in CFC favored screening more than people l ow in CFC. For the interaction, people low in CFC (compared to those high in CFC) produced more positive thoughts and they were more likely to endorse colorectal screening when positive consequences were short term and negative consequences were long term People high in CFC (compared to those low in CFC) produced more positive thoughts, and they were more likely to endorse screening when positive consequences were long term and negative consequences were short term. Orbell and Hagger (2006) recruited 210 adult s from the community and assigned them to read one version of a message about a proposed type 2 diabetes screening program (developed using the same 2 x 2 x 2 design as Orbell et al., 2004) After reading the message, participants completed post manipulati on measures, including the CFC scale. High CFC individuals held more positive thoughts and showed greater intentions toward diabetes screening than low CFC individuals. There was also an expected CFC x time frame interaction. People low in CFC (compared to those high in CFC) produced more positive thoughts and showed greater intention to partake in diabetes screening when positive consequences were short term and negative consequences were long term. The opposite was true for people high in CFC (compared to those low in CFC) Furthermore, the CFC x time frame interaction effect on intentions was mediated by the net number of positive thoughts.
38 Orbell and Kyriakaki (2008) were the first to examine the CFC hypothesis in a preventi on (v s. detection) behavior They conducted two experiments (both using a similar 2 x 2 x 2 design as the previous two studies described above ) with staff, students, and visitors around a university campus ( N = 121 and N = 279) and assigned participants to read a message about sk in cancer and sunscreen use. Participants in both experiments completed the CFC scale after reading the message. Experiment one assessed i ntention to use sunscreen as the outcome measure, whereas experiment two assessed behavior (redemption of a voucher for free sunscreen) as the outcome measure. A main effect was observed in the first experiment where high CFC individuals showed greater intention to use sunscreen than low CFC individuals. The CFC hypothesis was supported in both experiments. In experiment one, people high in CFC (compared to those low in CFC) reported greater intentions to use sunscreen when positive outcomes were presented as distal and negative outcomes were presented as immediate; the opposite was true for people low in CFC (compared to those high in CFC). In experiment two, people low in CFC (compared to those high in CFC) were more likely to redeem vouchers for free sunscreen when positive outcomes were presented as immediate and negative outcomes were presented as future. Although the opposite pattern was seen for people high in CFC, the interaction was not significant. The CFC x time frame interaction effect on intentions (experiment one) and behavior (experiment two) was mediated by the net number of positive thoughts. CFC has also been found to moderate message valence and regulatory focus framing effects. OConnor et al. (2009) found that people high in CFC were more responsive to loss framed messages related to health information seeking behaviors, and
39 people low in CFC were more responsive to gain framed messages. Kees (2007) found people high in CFC reported greater levels of persuasion than people low in CFC when a message related to consuming unhealthy foods was preventionframed; however, when the message was promotion fr amed, people low and high in CFC responded similarly. Discordant findings. Despite many studies showing hypothesized effects, the overall pattern of results for message framings influence on health behavior decisions is inconsistent For example, a lthough several studies revealed findings consistent with Rothman and Saloveys (1997) predictions for detection and prevention behaviors (as previously presented) several studies failed to find an advantage for either frame (e.g., Lalor & Hailey, 1990; Lauver & Rubin, 1990) or found the message framing effect to be limited to a specific subset of individuals (e.g., Apanovitch et al. 2003; Finney & Iannot ti, 2002). P ossible explanations for these and other discordant results have been offered that include th e limitations of prospect theory (e.g., Rothman & Salovey, 1997), ambiguous theoretical terms such as risk (OKeefe & Jensen, 2006, 2007), inconsistent applications of prospect theory (e.g., OKeefe & Jensen, 2006, 2007), the influence of other theories/ conceptual models (e.g., Wilson et al., 1988), the lack of a universal operational definition of message framing (e.g., Levin et al., 1998), the variability in taxonomies/classification schemes of message framing (e.g., Levin et al., 1998), the diversity o f behaviors studied (e.g., Wilson et al., 1988), the variability in perceptions of health behavior functions (prevention vs. detection e.g., Rothman et al ., 2009), and the existence of omitted moderating variables (e.g., Rothman et al., 1997) A summary o f
40 three metaanalytic reviews on message framing research is presented next that further addresses discrepant findings in the literature. Meta analytic Reviews on Message Framing Research Three metaanalysis papers on message framing research have been p ublished in an attempt to systematically organize, analyze, and summarize the effects of message framing on health and other types of behaviors. The earliest metaanalysis was by Khberger (1998). He examined 136 empirical papers that reported framing expe riments. Based on studies described within these papers, he calculated 230 single effect sizes. His analysis encompassed a broad scope of domains (health, business, gambling, and social), study designs, and framing types. He defined and coded study charact eristics in the following areas and examined each as a potential moderating variable: risk characteristics (including risk manipulation, quality of risk, and number of risky events); task characteristics (including framing manipulation, response mode, comparison, unit of analysis, and problem domain); participant characteristics (including whether the sample was students or the target audience); and year of publication. His findings revealed that overall message framing produced small to moderate effects. I n addition, he found that taken as a whole, most of the study characteristics were significant moderators except participant characteristics (student vs. nonstudent population) and unit of analysis (individual vs. group). More recently, OKeefe and Jensen (2006) conducted a meta analytic review of the relative persuasiveness of gain and loss framed messages based on 165 studies. They classified the studies into six distinct broad categories: disease detection behaviors, disease prevention behaviors, oth er health related behaviors, sociopolitical subjects,
41 advertising of consumer products and services, and other (otherwise unclassified). They also coded each message as containing one of four types of kernel state phrasing ( i.e., l inguistic representations of the consequence/outcome): exclusively desirable, exclusively undesirable, combination of desirable and undesirable, or indeterminate (related to unavailability of sufficient message detail). They examined both message topic and kernel state phrasing as potential moderating variables. For each distinguishable message pair, they calculated an effect size to summarize the comparison between a gain framed message and its lossframed counterpart. Across all 165 studies, they did not find a significant persuasive advantage for one framing form over the other. However, they did find that message topic had a significant moderating effect. Of the five substantive behavior categories examined, only disease prevention showed a significant difference in persuasiveness. For messages advocating disease prevention behaviors, there was a significant persuasive advantage of gain framed messages over loss framed messages (consistent with theoretical predictions). Contrary to expectation, however, for messages advocating di sease detection behaviors, gain and loss framed messages did not significantly differ. In addition, they did not find any significant moderating effects of kernel state phrasing. OKeefe and Jensen (2006) offer several possible explanations of their findi ngs. The most plausible explanation is presented next. As previously discussed, health prevention and detection behaviors are commonly described in terms of their risk, with prevention behaviors typically perceived as less risky than detection behaviors. It is this distinguishing characteristic that has guided several theoretical predictions of message framing effects on these behaviors. OKeefe and Jensen (2006) suggest that the word risk and its variants (e.g., risky) are
42 ambiguous and have various interpretations. In the original prospect theory (Kahneman & Tversky, 1979, 1984; Tversky & Kahneman, 1981), risk refers to the association between action and outcome; an action is perceived as risky if its outcomes are perceived as probabilistic or not certain. In the more familiar application of prospect theory to gainloss message variation (e.g., Rothman & Salovey, 1997), risk refers to the perceived desirability or dangerousness of an outcome; a behavior is perceived as risky if its outcome is undesirable or dangerous. These different interpretations of prospect theory make some theoretical assumptions problematic. For example, some people may perceive exercise (a prevention behavior) as not risky because it is typically safe; others may perce ive exercise as risky because the outcomes are not certain. If people use level of uncertainty to classify a behaviors level of risk, this may result in no difference of perceived risk between disease detection and disease prevention behaviors and thus no effect of message framing. In an extension of this meta analysis, OKeefe and Jensen (2007) published another review, with a specific focus on only disease prevention behaviors. They analyzed 93 studies that examined gainframed and loss framed messages in advocating disease prevention behaviors. Unlike their previous analysis, they classified these prevention behaviors into eight health related categories: diet/nutrition behaviors, safer sex behaviors, skin cancer prevention behaviors, dental hygiene behaviors, exercise behaviors, smoking cessation or non initiation, inoculation (vaccination), and other (or multiple different) prevention behaviors. They also coded each message as containing one of four types of kernel state phrasing. They examined both specific behavior type and kernel state phrasing as design variables potentially moderating framing effects. For each
43 distinguishable message pair, they calculated an effect size to summarize the comparison between a gain framed message and its lossframed counterpart. They found a significant advantage of gainframed appeals, but the effect size was extremely small and limited to one prevention behavior (dental hygiene). For messages advocating dental hygiene behaviors, the analysis revealed expected resul ts: gain framed appeals were more persuasive than loss framed appeals. However, the analysis found no differences in persuasiveness between framed messages concerning any of the other seven prevention behavior categories. For kernel state phrasing, they found message framing effects do not dependably vary as a consequence of kernel state phrasing in gain framed appeals but that they do dependably vary in loss framed appeals. OKeefe and Jensen (2007) go on to discuss possible explanations for their findings, including unique characteristics of dental hygiene behaviors. The most plausible explanation they offer relates to the varying ambiguous interpretations of risk, as previously described in their 2006 meta analysis. They suggest that it cannot be assu med that all health prevention and health detection behaviors are considered riskaverse and risk seeking, respectively. Consequently, it cannot be predicted that matching gain and loss framed messages with prevention and detection behaviors, respectively will always be more effective. They also suggest the perceived protective outcomes of performing dental hygiene behaviors may be more certain (and thus less risky) than the protective outcomes of performing the other seven prevention behaviors, which would potentially make the dental hygiene behaviors more sensitive to the expected effects of gain framed messages.
44 A s with any meta analytic study, each of the three reviews had limitations to consider, such as the exclusion of potentially relevant studies, in sufficient message details to examine variables of interest such as dose of framing manipulation, the methods for calculating certain effect sizes (as noted in OKeefe & Jensen, 2007), and insufficient power to detect differences in all conditions (e. g., as noted in OKeefe & Jensen, 2007, for both exercise and skin cancer prevention behaviors). In addition, unlike Khbergers (1998) review, OKeefe and Jensen (2006, 2007) did not code for relevant study characteristics as potential moderat ors in eithe r of their reviews Furthermore, none of the reviews coded for individual difference s variable s as potential moderators of message framing effects. In light of the complexities surrounding framing effects, m etaanalytic reviews that narrowly examine the l i terature by only compar ing main effects have limited utility Latimer, Salovey, and Rothman (2007) suggest the effectiveness of framed messages hinges on how individuals thin k and feel about the behavior and not just the function or nature of the behavior per se. Failure to consider the i mpact of individual differences on message framing effects may suppress true framing effects and underestimate utility of gainand loss framed appeals. In summary, all three metaanalysis papers on message framing research fall short of providing an adequate, comprehensive review For example, these papers in consistently report and compare study characteristics such as research design, research setting, sample description, nature of the health problem (hypothetical or real ), dependent variables (e.g., attitudes, beliefs, intentions, and behavior changes), moderating variables, and measures used. I nclusion of these and other factors such as underlying theoretical/conceptual frameworks would have provide d a richer understandi ng of the
45 literature and better equip ped researchers to propose meaningful recommendations based on their synthesis. State of the Message Framing Literature Conclusions regarding the current state of the message framing literature and suggestions for advancing th is literature are presented next Re gardless of whether examining main effects or moderating effects, the overall pattern of results for message framings influence on health behavior decisions is inconsistent. Although numerous moderators of framing effects have been studied to help explain these discordant findings a comprehensive analysis and synthesis of these effects has not been published. More s ystematic reviews are needed to thoroughly summarize these relationships in order to produce a fuller understanding of message framings boundaries. In addition, other types of message tailoring ( such as temporal proximity framing ) need to be examined simultaneously with valence framing to test for interactions that may yield even more persuasive health promotion messages than either manipulation alone. S tudies such as these have only recently begun t o evolve in the literature M essage framing as a health communication strategy is underutilized by nurses and is severely lacking in nursing research. Most studies reviewed were from the psychology literature with only a few from the nursing literature (e.g., Jenner et al., 2005; Lauver & Rubin, 1990). In addition, s everal studies involved hypothetical health related situations (e.g., Rothman et al., Ex periment 1, 1999) rather than actual health problems (e.g., Finney & Iannotti, 2002). Many studies also involved college students (often undergraduate psychology students) (e.g., Sherman et al., 2006) rather than representative
46 samples of persons with or at risk for developing various diseases such as skin cancer (e.g., Detweiler et al., 1999). Finally, although message framing studies have been conducted on a breadth of detection and prevention behaviors, fewer studies have examined prevention (Rothman et al., 2008). This is of concern given that modifiable prevention behaviors such as physical inactivity are largely responsible for the high prevalence and incidence of chronic diseases like type 2 DM (WHO, 2005b). Although a few studies that examined physi cal activity yielded some main effects of message framing consistent with Rothman and Saloveys (1997) predictions (e.g., Latimer Rench, et al., 2008; Vant Riet et al., 2010), several d id not (e.g., Arora et al., 2006; Jones et al., 2004; McGall & Ginis, 2004). More importantly, the majority of these studies revealed one or more variables that moderated message framing effects on physical activity such as source credibility (e.g., Arora et al., 2006; Jones et al., 2003; Jones et al., 2004), type of value appeal (e.g., Robberson & Rogers, 1988), stages of readiness to change (e.g., Hsiao, 2003), need for cognition (e.g., Hsiao, 2003), and gender (e.g., Hsiao, 2003; Kroll, 2005). However, two variables of interest that have not been widely tested as moderato rs of framing effects on physical activity behaviors are CFC and motivational orientation. These and other potential moderators should be examined in physical activity and other prevention behaviors with salient populations (e.g., people with type 2 DM) to advance the message framing literature and to better guide health care providers in developing and delivering effectively tailored health messages to promote healthy behaviors.
47 Chapter Summary A comprehensive review of the message framing literature r evealed numerous gaps in which future research is needed to help fill. The present study draws from prospect theory, TCT, and motivation theories to address several of these gaps and to advance both the message framing and nursing literature. The next chap ter describes the method used to conduct the present study.
48 Chapter 3: Method This chapter describes the method used to explore how two individual differences factors CFC and motivational orientation may combine to moderate temporal proximity and valence framing effects on intentions to increase physical activity. The research design, participants and study m aterials are first described followed by a discussion of research procedures. Research Design The present study was an experimental design using random assignment. V alence (gains vs. losses) w as crossed with temporal proximity (immediate vs. distal) to form four versions (frames) of the health message. Two individual differences constructs ( CFC and motivational orientation ) w ere measured usi ng established instruments, each hypothesized to moderate the influence of the framing manipulation on intentions to increase physical activity D ata collection consisted of a mail survey that was conducted using Dillmans Tailored Design Method (TDM) (Dil lman, Smyth, & Christian, 2009). This method is based on social exchange theory and over 30 years of empirical research aimed at maximizing survey response rates. The method entails up to f ive different contacts: 1) an initial prenotice letter to respondents describing the importance of the forthcoming survey
49 booklet; 2) the booklet itself which arrives a few days later with a personalized cover letter, token incentive ($1 $5 is typically recommended ), and stamped return envelope; 3) a reminder/thank you postcard sent to all recipients a few days to a week following the booklet; 4) a replacement booklet and different cover letter, with a slightly more insistent tone, sent to those who have still not responded by 2 4 weeks ; and 5) a final contact made by a different mode of delivery 2 4 weeks after the previous mailing (e.g., special mail delivery, telephone follow up call ). Th e first four contacts were used in th e present study; no telephone calls were made. Although th e TDM approa ch is more expensive a nd labor intensive than a one shot bulk mailing approach that typically yields only 2535% response rate, Dillmans method routinely produces response rates over 50%. ( S ee Dillman et al., 2009, for examples of empirical studies that used this method). Participants Eligible p articipants were recruited through the University of South Florida (USF) Medical Clinic database. In collaboration with Anthony Morrison, MD and Nancy Grove, ARNP (who specialize in diabetes management) a list was extracted from t his database and provided to the principal investigator which included n a mes and mailing addresses of a dult patients (18 years and older) with type 2 DM (based on an ICD9 code of 250.00 or 250.02) who were seen at the USF Medical Clinic by Dr. Morrison or Ms. Grove between January 1 and December 22, 2009. No protected health information was sought or obtained. The original list of patients contained 695 names Duplicate names were removed from the list, resulting in 649 unique names. A power analysis was conducted in order to determine minimum sample size for the present study. Sherman et al. (2006) reported an interaction of motivational
50 orientation with (valence) framing on promotion of flossing. These estimates of congruence effects ( R2) ranged from .07 to .13. Assuming alpha = .05 and an R2 of .07 for each main effect in the present study (valence congruence and temporal proximity congruence), it was determined a total of N = 187 would provide power all three hypotheses using multiple regression. Materials Health m essage development Four versions of a health message aimed to increase physical activity of people with diabetes were constructed based on published diabetes materials from organizations such as the ADA, NIDDK, and the Ame rican Association of Diabetes Educators (AAD E). Key principles from the National Cancer Institute (NCI 1994) were also used as a guide when designing the message content and layout. The first paragr aph in the body of text was identical across all four ve rsions and contain ed information such as examples of physical activit ies and recommendations for how often to engage in these activities. The second paragraph contain ed the same factual content but differ ed among the four versions only in valence and temporal proximity frames as follows: gain framed/distal, gain framed/immediate, lossframed/distal, and loss framed/immediate. (Refer to Table 1 for details of the framing manipulation). The two gainframed conditions listed eight statements regarding the bene fits people may gain by doing regular physical activity whereas the two loss framed conditions list ed these same eight statements but framed as benefits people may lose by not being physically active The two distal framed conditions present ed these gains or losses as outcomes that may occur in the future whereas the two immediate framed conditions present ed them as outcomes that may occur immediately
51 Table 1 Details from Four Health Message Versions to Illustrate Temporal Proximity and Valence Framing Manipulation Valence Temporal Gain Loss Distal By doing regular physical activity, you may gain these benefits in the years to come: Brighter mood and more positive attitude Greater self esteem, pride, and confidence More energy, better sleep, b etter sex life, and less stress Increased burning of extra calories to help keep a healthy weight Improved blood pressure, blood cholesterol, and circulation Less need for diabetes medication due to better blood sugars Stronger bones, more strength, and in creased flexibility and balance Reduced arthritis pain By not doing regular physical activity, you may lose benefits in the years to come by: Missing out on a brighter mood and more positive attitude Missing out on greater self esteem, pride, and confide nce Missing out on more energy, better sleep, better sex life, and less stress Missing out on increased burning of extra calories to help keep a healthy weight Missing out on improved blood pressure, blood cholesterol, and circulation Missing out on less n eed for diabetes medication due to better blood sugars Missing out on stronger bones, more strength, and increased flexibility and balance Missing out on reduced arthritis pain Immediate By doing regular physical activity, you may gain these benefits imm ediately (within 1 2 weeks) : Brighter mood and more positive attitude Greater self esteem, pride, and confidence More energy, better sleep, better sex life, and less stress Increased burning of extra calories to help keep a healthy weight Improved blood pr essure, blood cholesterol, and circulation Less need for diabetes medication due to better blood sugars Stronger bones, more strength, and increased flexibility and balance Reduced arthritis pain By not doing regular physical activity, you may lose benefi ts immediately (within 1 2 weeks) by: Missing out on a brighter mood and more positive attitude Missing out on greater self esteem, pride, and confidence Missing out on more energy, better sleep, better sex life, and less stress Missing out on increased bu rning of extra calories to help keep a healthy weight Missing out on improved blood pressure, blood cholesterol, and circulation Missing out on less need for diabetes medication due to better blood sugars Missing out on stronger bones, more strength, and i ncreased flexibility and balance Missing out on reduced arthritis pain
52 According to the NCI (1994), use of headers, footers, and graphics on print materials can enhance readability and effectiveness of the message. Three concise statements ( boldfaced, underlined, and/ or italicized) were positioned above and below the body of text to not only reiterate the messages main emphasis but to also strengthen the framing manipulation. Among the four message versions, these statements differed only in temporal p roximity and valence frames (e.g., Physical Activity Can Improve Your Health Immediately vs. Lack of Physical Activity Can Jeopardize Your Health in the Years to Come ) Furthermore, simple graphics were used that were culturally sensitive and relevant to the text, illustrating five e xamples of physical activit ies (i.e., weight lifting, walking, biking, golfing, and gardening). These graphics were i dentical across all four versions (See Appendices A D f or the four health message versions). Heal th message p ilot study. The four versions of the health message were pilot tested with 20 nurse practitioners (NPs) familiar with the diabetes population ( n = 5 per version using random assignment ) The purpose of this pilot was to obtain feedback regardin g the proposed framing manipulation and format in order to determine the need for revisions, if any, prior to conducting the present study. After reading the message, each NP completed a 6 item questionnaire. (See Appendix E). One item assessed perceived believability of the message. Two items assessed the temporal proximity framing manipulation, where NPs rated the extent to which the message emphasized physical activity outcomes that can happen in the future versus immediately Two more items assessed the valence framing manipulation, where NPs rated the extent to which the message emphasized benefits of being physically active versus risks of being physically in active. Responses to the five items (listed in Table 2) were provided on a 7 point scale
53 ranging from 3 (strongly disagree) to +3 (strongly agree), with the midpoint labeled neither agree nor disagree The sixth item was an openended question asking NP s their thoughts about the graphics included on the message. Table 2 Summary of Pilot Study D ata for Evaluat ion of Health Message Believability and Framing Manipulation of Temporal Proximity and Valence Condition ( n = 5 each) Questionnaire Item G D GI L D LI 1. The message was very believable. 2.60 1.20 2.20 2.40 2. The message emphasiz ed things that can happen in the future. 1.80 1.40 2.80 1.60 3. The message emphasized things that can happen immediately. 2.20 1.20 1.60 2.40 4. The message emphasized the benefits of being physically active. 2.80 3.00 2.60 2.00 5. The message e mphasized the risks of not being physically active. 0.40 0.60 2.60 1.60 Note. N = 20. Each of the five items was rated on a 7 point scale ranging from 3 (strongly disagree) to +3 (strongly agree). Cell means a re presented for each condition. G D = g ain framed /distal ; GI = gain framed /immediate; L D = lossframed / distal; LI = loss framed /immediate R esults of the pilot are shown in Table 2. Overall, the pilot group perceived t he message as believable ( M = 2.10, SD = 1.12). Marginal means were examined in order to assess framing manipulation. The pattern of means was as expected for three of the four items: item 2 [(distal M = 2.3) > (immediate M = 1.5)], item 4 [(gain M = 2.9) > (loss M = 2.3)], and item 5 [(loss M = 2.1) > (gain M = 0.1)]. For item 3 the marginal means were
54 nearly identical, but the pattern was not as expected [(distal M = 1.9) > (immediate M = 1.8)] Finally, overall invited comments were positive, and the graphics were reported to be a ppropriate and not distracting. In summary, pilot results suggested the health messages were perceived as intended and graphics were appropriate; therefore, no changes were made to the original health message versions prior to conducting the present study Consideration of f uture c onsequences. Strathman et al.s (1994) 12item CFC scale w as used to measure participants CFC. Strathman et al. (1994) reported reliability across four independent college samples: = .80, .82, .86, and .81, respectively. Participants in the present study provided r atings for the items on a 5point scale ranging from 1 (extremely uncharacteristic) to 5 (extremely characteristic) with the midpoint labeled uncertain Items 3, 4, 5, 9, 10, 11, and 12 were reverse scored. Responses to the 12 items were averaged into a single index of CFC s uch that higher scores indicate d a greater focus on distal outcomes. Reliability in the current study was = 83. (See Appendix F for the full CFC scale). Motivational orientation M otivational orientation was operationalized us ing two measures: Carver and Whites (1994) BIS/BAS scale and Higgins et al. s ( 2001) R FQ scale. BIS/BAS. The 20item BIS/BAS scale consists of the 7item B IS scale (see App endix G items 1 7) and the 13 item BAS scale (see Appendix G items 8 20) with reported reliabilities of = 80 and = 84, respectively (Sherman et al., 2006). Participants in the present study provided r esponses to the items on a 4point scale ranging from 1 (strongly agree) to 4 (strongly disagree) with items 5 and 7 reverse
55 scored Consistent with othe r studies (e.g., Mann et al., 2004), scores on the two scales were combined into a single index of motivational orientation by subtracting the BAS mean from the BIS mean A positive score indicate d a person was predominantly approach oriented, whereas a ne gative score indicated a person was predominantly a voidance oriented. In the present study, the reliabilities of the BIS and BAS s cales were = .79 and .87, respectively. RFQ. The 11 item RFQ scale consists of the 6item promotion scale (see Appendix H items 1, 3, 7, 9, 10, and 11) and the 5item prevention scale (see Appendix H items 2, 4, 5, 6, and 8) with reported reliabilities of = 73 and .80, respectively (Higgins et al., 2001) Participants in the present study provided r esponses to the items on a 5 point scale ranging from 1 (never or seldom) to 5 (very often), with the midpoint labeled sometimes Items 1, 2, 4, 6, 8, 9, and 11 were reverse scored Consistent with other studies (e.g., Higgins et al., 2001), scores on the two scales were combined into a single index of motivational orientation b y subtracting the prevention mean from the promotion mean. A positive score indicated a person was predominantly promotionfocused, whereas a negative score indicated a person was predominantly preventionfocused I n the present study, the reliabilities of the promotion and prevention scales were = 69 and .80, respectively. Intentions. Eight items were developed to assess participants intentions to increase physical activity. Using a 7 point scale ranging from 3 (strongly disagree) to +3 (strongly agree), with the midpoint labeled neither agree nor disagree, participants indicated ho w likely they were to increase physical activity over the next 4 weeks after reading one of the four health message versions (See Appendix I, items 1 8) Responses
56 to the eight items were averaged into a single index of intentions with items 1, 4, 6, a nd 8 reversescored ( = .94). Evaluation of the health message. Believability. The item used in the pilot study was used in the present study to assess perceived believability of the message. Participants provided a response to this item on a 7point sca le ranging from 3 (strongly disagree) to +3 (strongly agree), with the midpoint labeled neither agree nor disagree. (See Appendix I, item 9). Manipulation check. The four items used in the pilot study were used in the present study as a manipulation check to assess the extent to which the four health message versions were perceived as intended. Participants provided responses to the items on a 7point scale ranging from 3 (strongly disagree) to +3 (strongly agree), with the midpoint labeled neither agree nor disagree (See Appendix I, items 10 13). Demographic and o ther variables Several items were developed for participants to self report a ge, gender, race/ ethnicity, years diagnosed with diabetes, current physical activity height, and weight ( S ee Appendix J ) Booklet a ssembly M easures w ere compiled into a booklet in the following order: 1) demographics and other variables; 2) one version of the health message (chosen at random) ; 3) intentions and evaluation of the health message (combined on the same page) ; and 4 ) the individual difference s scales (CFC, BIS/BAS, and RFQ respectively ) Procedures Institutional Review Board The present study was reviewed and approved by the USF Institutional Review Board before any pilot or actual data were collected
57 Recruitment Based on results of empirical studies that used Dillman s T DM ( Dillman et al., 2009) to conduct mail surveys, a 53% return rate was estimated for the present study. I n order to obtain the minimum sample size of 187 (as calculated by the power analysis), it was determined the first sample would consist of 352 eligible participants. A second sample would be selected if mail survey results from the first sample did not yield at least 187 completed booklets. The original list of 649 unique patient names obtained from the USF Medical Clinic database was not sorted in any known order. However, the list was s plit into two subsets (odd vs. even) to help ensure homogeneity, should a second sample be needed. The initial prenotice letter (see Appendix K ) was mailed to all 325 patients from the first subset and the first 27 patients listed in the second subset About one week later, the second mailing was sent, including a personal cover letter (see Appendix L ), the booklet of measures, a $1 to ken incentive, and a stamped return envelope. Participants were advised study involvement was voluntary and anonymous and to not put any personal identifiers on the booklet. Within approximately another week, a reminder/thank you postcard was mailed out (s ee Appendix M ). Finally, about 1 month after mailing the first booklet, a replacement booklet with different cover letter (see Appendix N ) and a stamped return envelope was sent to eligible participants who had not yet responded. Chapter Summary This c hapter described the method used in the present study, including research design, participants, study materials, and research procedures. The next chapter describes results of the study, including preliminary analyses, hypotheses testing, and supplemental analyses and findings.
58 Chapter 4: Results This chapter describes results of the present study. Preliminary analyses are presented first, including mail survey results, identification of missing data, description of the final sample, and respondents perceptions of the health message. Results of hypotheses testing are presented next, followed by supplemental analyses and findings. U nless otherwise noted, significance for all analyses was set at the .05 level (2 tailed). Preliminary Analyses Mail s urvey r esults. Table 3 shows the mail survey results for each of the four step s in Dillmans T DM ( Dillman et al., 2009). Survey response rates were calculated based on the total number of deliverable booklets. T he first sample initially yielded a response rate of 48% upon completion of step two with 163 participants who returned booklets with responses. The subs equent mailing of postcards and replacement booklets (steps three and four ) resulted in an additional 28 participants who returned booklets with respons es, increasi ng the response rate to 56% for the first sample. However, at least 20 of these booklets had a substantial amount of missing data ; t herefore, a second sample was selected by choosing the next 100 eligible participants o n the second subset from the USF Medical Clinic list. The f our step mail survey procedure w as repeated with the second sample of 100 as with the first sample. The second sample initially yielded a
59 response rate of 40% upon completion of step two with 40 participants who returned booklets with responses. The subsequent mailing of postcards and replacement booklets resulted in an additional 8 participants who returned booklets with responses, increasing the response rate to 49% for the second sample. Across both samples, the overall response rate for the mail survey was 54%, with 239 participants who returned booklets with responses.
60 Table 3 Mail Survey Results Using Dillmans Tailored Design Method N o. Returned Dillmans TDM Steps No. Sent No. Not Returned Not De liver able Blank/ Declined With Responses First Sample 1. Prenotice letter 352 2. Booklet 352 15 9 9 2 1 16 3 3. Postcard a 341 4. Replacement Booklet 159 11 7 0 14 2 8 Total 352b 117c 9 35 191 Second Sample 1. Prenotice letter 100 2. Booklet 100 51 1 8 40 3. Postcard a 98 4. Replacement Booklet 51 38 1 4 8 Total 100b 38c 2 12 48 Grand Total 452 155 11 47 239 Note TDM = Tailored Design Method. aPostcards were not sent if previous mailings were not delive rable or person had already declined participation. bTotal number sent refers to the total number of booklets initially mailed out in each sample. cTotal number not returned ref ers to the replacement booklets only, not the initial booklets.
61 Missing d ata The 239 returned booklets with responses were screened for missing data. Participants needed to have valid responses for at least 80% of the items on all of the scale s ( intentions, CFC scale, the two BIS/BAS scales, and the two RFQ s cales ) If parti cipants did not meet this criterion, they were excluded from analyses. T his criterion yielded a final sample size of 218 upon which hypotheses were tested. (Participants were not excluded if they had missing data on only demographic or message evaluation v ariables; therefore, some preliminary and supplemental analyses have an N < 218). Several of the excluded participants did not complete any demographic items ; therefore, a comparison could not be made between this group and the final sample. Description of s ample Characteristics of the final sample of 218 participants are summarized in Table 4. Part icipants ranged from 24 92 years old ( M = 58.63, SD = 11.96). Gender was fairly equal, with 104 males and 114 females. The majority (70%) was White/ Caucasian, 12% were Hispanic/Chicano/Latino, and 11% were Black/African American Participants reported having diabetes an average of 14.19 years. Total weekly physical activity ranged from 0 to 1400 minutes ( M = 175.04, SD = 227.40). However, the frequenc y most commonly reported was 0 minutes (38, or 17%), with a median of 120 minutes. This wide variation may partially be attributed to differences in perception as to what constitutes physical activity. For example, some people may think simply being on the ir feet is physical activity, regardless of movement. Others may think physical activity involves actions that elevate the heart rate above baseline. Still others may think only planned, structured exercise constitutes physical activity. Finally, participa nts on average were obese ( M = 33.73, SD = 7.88) which is defined as a body mass index ( BMI ) of 30 or above (CDC, 2009b) With the exception of age, there were
62 no significant differences in demographic variables among the four conditions. P articipants who read the immediateframed information were significantly older than those who read the future framed information, F (1, 216) = 4.70, p = .031. Table 4 Characteristics of the Final Sample Characteristic s n (%) Range Mean ( SD ) Age (years) 24 92 58.63 (11.96) Gender Male 104 (4 8 ) Female 114 (52) Race/Ethnicity White/Caucasian 153 (70) Black/African American 23 (1 1 ) Hispanic/Chicano/Latino 25 (1 2 ) Asian or Pacific Islander/Asian American 10 ( 5 ) American Indian or Alaskan Native 3 ( 1) Other 2 ( 1 ) Years with Diabetes 0 50 14.19 (9.25) Total Weekly Activity (min utes ) a 0 1400 175.04 (227.40) Body Mass Index (BMI)b 18.759.6 33.73 (7.88) Note N = 218 for all characteristics except race ( N = 216), years with diabetes ( N = 216), and total weekly exercise ( N = 213) related to missing data a Total weekly physical activity was calculated by multiplying number of usual days of physical activity per week by number of usual minutes of physical activity on each of these days. b BMI was calculated as follows: ( weight in pounds x 703) / ( total height in inches )2. Evaluation of the health message. Results of the health message evaluation are shown in Table 5. An ANOVA revealed the four conditions did not produce significant differences on ratings of message believability (item 1). Overall, participants perceived t he message as believable ( M = 2.44, SD = 0.82, scale range 3 to +3). Marginal means
63 were examined in order to assess framing manipulation s The pattern of means was as expected for all four items: item 2 (distal M = 2.36) > (immediate M = 2.15) ; item 3 (immediate M = 1.99) > (distal M = 1.48), F (1, 214) = 6.06, p = .015; item 4 (gain M = 2.63) > (loss M = 2.61) ; and item 5 (loss M = 2.29 ) > (gain M = 1.26) F (1, 214) = 17.78, p < 001. Participants who read the immediateframed information rated the message to be more immediateoriented th an those who read the future framed information. P articipants who read t he lossframed information rated the message to have an emphasis on risks more so than those who read the gain framed information. In summary, results of the se checks suggest the health messages were perceived as intended.
64 Table 5 Summary of Mail Surv ey Data for Evaluation of Health Message Believability and Framing Manipulation of Temporal Proximity and Valence Condition Questionnaire Item G D ( n = 59 ) GI ( n = 52 ) L D ( n = 55 ) LI ( n = 50 ) 1. The message was very believable. 2.51 2.48 2.46 2.32 2. The message emphasized things that can happen in the future. 2.22 2.25 2.51 2.05 3. The message emphasized things that can happen immediately. 1.22 2.00 1.76 1.98 4. The message emphasized the benefits of being physically active. 2.54 2.73 2.69 2.52 5. The message emphasized the risks of not being physically active. 1.24 1.29 2.47 2.08 Note. N = 216. Each of the five items was rated on a 7 point scale ranging from 3 (strongly disagree) to +3 (strongly agree). Cell means are presen ted for each condition. G D = gain framed/distal; GI = gain framed/immediate; L D = lossframed/ distal; LI = loss framed/immediate. Hypotheses Testing Calculation of c ongruenc e s cores. F raming manipulation and individual differences measures w ere used to construct variables representing temporal proximity c ongruence and valence congruence To illustrate, conditional transformation on the CFC score produced the temporal proximity congruence score. For participants in the distal framed condition, their CFC sco re represented the degree of temporal proximity congruence with the message ( i.e., higher CFC score mean t greater congruence), but for participants in the immediate framed condition, their CFC score represented the degree of
65 temporal proximity in congruence and was therefore reverse scored Similar conditional transformations were performed for both motivational orientation scores (i.e., BIS/BAS and RFQ) to produce two separate valence congruence scores. For participants in the gainframed condition, both th eir RFQ and BIS/BAS scores represented the degree of valence congruence with the message (i.e., higher RFQ or BIS/BAS scores meant greater congruence ). For participants in the loss framed condition, both their RFQ and BIS/BAS scores represented the degree of valence in congruence and were therefore reverse scored In all three transformations, a constant was added to transformed congruence scores to produce a range of positive values comparable to untransformed scores. Descriptive statistics. Table 6 provides descriptive statistics for key variables. ANOVAs were conducted to test for differences across the four conditions on each variable.
66 Table 6 Pearson Correlations and D escriptive Statistics for Key Variable s Variable 1 2 3 4 5 6 7 1. Int entions 2. CFC .169* 3. BIS/BAS .014 .147* 4. RFQ .069 .090 .441* 5. Temporal proximity congruence (CFC) .013 .026 .048 .016 6. Valence congruence (BIS/BAS) .033 .071 .071 .048 .078 7. Valence congruence (RFQ) .127 .022 .059 .028 .074 .450* M 1.24 3.43 0.10 0.06 3.02 3.93 5.02 SD 1.42 0.71 0.65 0.99 0.83 0.66 0.99 Scale range -3 to 3 1 to 5 -3 to 3 -4 to 4 1 to 5 1 to 7 1 to 9 N ote Listwise N =218. For intentions, higher score represents greater intention to increase physical activity within 4 weeks after reading the health message. For the CFC scale, higher score represents more future oriented. For the BIS/BAS scale, higher sco re represents more approach oriented. For the RFQ scale, higher score represents more promotionfocused. For temporal congruence, higher score represents greater congruence of message with CFC score. For valence congruence, higher score represents greater congruence of message with BIS/BAS or RFQ score. CFC = Consideration of Future Consequences score; BIS/BAS = Behavioral Inhibition System/Behavioral Activation System score; RFQ = Regulatory Focus Questionnaire score. *p < .05.
67 Scale scores. T he four c onditions did not differ on the three individual difference s scale scores ( CFC, BIS/BAS or RFQ). On average, participants w ere somewhat future oriented as evidenced by a mean score above the range midpoint on the CFC scale ( M = 3.43, SD = 0.71, scale ran ge 1 to 5). For motivational orientation, participants as a whole were approachoriented and promotionfocused as evidenced by positive mean scores on the BIS/BAS ( M = 0.10, SD = 0.65) and RFQ ( M = 0.06, SD = 0.99), respectively. Furthermore, BIS/BAS and R FQ scores were positively correlated ( r = .441 p < .001) As theory would suggest (Higgins, 1997) the more approachoriented participants tended to be more promotionfocused and the more avoidance oriented participants tended to be more preventionfoc us ed CFC scores were also positively correlated with B IS/BAS scores ( r = .147, p = .031) but not with RFQ scores ( r = .090, p = .188) The more future oriented participants tended to be more approach oriented Furthermore, CFC score had a significant positive correlation with intentions ( r = .169, p = .012). The more future oriented participants tended to have greater intentions to increase physical activity. Congruence scores. Th e four conditions did not differ on the degree of valence congruence (RFQ) wit h the health message. However, F tests revealed significant differences in marginal means for temporal proximity congruence and valence congruence (BIS/BAS). P articipants in the distalframed condition had a greater degree of temporal proximity congruence with health message ( M = 3.43) than participants in the immediate framed condition ( M = 2.57), F (1, 21 6) = 79.25, p < .001. P articipants in the gainframed condition had a greater degree of valence congruence (BIS/BAS) with health
68 message ( M = 4.02) than participants in the loss framed condition ( M = 3.82), F (1, 21 6) = 5.17, p = .024. A s expected, the two valence congruence scores had a significant positive correlation ( r = .450, p < .001) Intentions. T he four conditions did not differ on the intention s index score, as shown in Table 7. Overall, participants in each condition intended to increase physical activity over the next 4 weeks after reading the health message as evidenced by a positive mean score ( M = 1.24, SD = 1.42). Table 7 Means and (Standard Deviations) for Intentions to Increase Physical Activity as a Function of Framing Manipulation Conditions Valence Temporal Gain Loss Marginal Means Distal 1.03 (1.35) n = 59 1.24 (1.52) n = 55 1.13 Immediate 1.40 (1.44) n = 53 1.30 (1.39) n = 51 1.35 Marginal Means 1.20 1.27 Note. N = 218. The scale score for intentions to increase physical activity over the next 4 weeks after reading the health message was based on a 7 point scale ranging from 3 to +3. Hypotheses Correlations and multiple regression were used to test the three hypotheses: H1. When valence of the message is congruent with individuals motivational orientation ( i.e., gain framed :: approachorient ed/ promotionfocused, and loss framed :: avoidance orient ed/ prev entionfocused) intentions to increase physical
69 activity will be greater than when the message is incongruent with individuals motivational orientation H2. When temporal proximity of the message is congruent with individual s status on CFC (i.e., immed iate framed :: low CFC, and distal framed :: high CFC) intentions to increase physical activity will be greater than when the message is incongruent with individuals CFC. H3. When both valence and temporal proximity of the message are congruent with indi vidual s motivational orientation and CFC standing, intentions to increase physical activity will be at their highest. The zero order correlation of intentions to increase physical activity with valence congruence and temporal proximity congruence was exam ined to test h ypotheses one and two respectively Support for these two hypotheses was based on tests of significance of the correlations. Hypothesis three was tested using two regression models with two predictors (the valence and temporal proximity congruence variables) due to the correlation between the two valence congruence measures ( r = .450). Support for hypothesis three was based on tests of significance applied to the regression coefficients.
70 Table 8 Summary of Multiple Regression Analyses P redicting Intentions to Increase Physical Activity from Valence Congruence and Temporal Congruence Variable B SE Intercept 1.55 8 0.66 5 Valence c ongruence (BIS/BAS) 0.069 0.14 8 .032 Temporal c ongruence (CFC) 0.017 0.11 7 .010 R 2 = .001 Intercept 2.16 2 0.5 88 Valence c ongruenc e ( RFQ ) 0.1 81 0.09 7 .12 6 Temporal c ongruence (CFC) 0.00 6 0.11 6 .00 3 R2 = .016 Note. N =218. BIS/BAS = Behavioral Inhibition System/Behavioral Activation System score; CFC = Consideration of Future Conse quences score; RFQ = Regulatory Focus Questionnaire score. *p < .05. Neither hypothesis one nor two was supported as evidenced by nonsignificant zero order correlations of intentions to increase physical activity with v alence congruence (BIS/BAS) (r = .033, p = .632), valence congruence (RFQ) ( r = .127, p = .062), and temporal proximity congruence ( r = .013, p = .854). A s shown in Table 8, hypothes is three was also not supported. In the first regression equation, main effects were not significant for valence congruence (BIS/BAS) ( = .032, p = .642) or temporal proximity congruence ( = .010, p = .883), and the overall model was not significant R2 = .001, F ( 2, 215) = 125, p = 882. Similarly, in t he second equation, main effects were not significant for valence congruence (RFQ) ( = .12 6, p = .064) or temporal proximity congruence ( = .003, p = .962), and the overall model was not significant, R2 = .016, F ( 2, 215) = 1.753, p = 176.
71 Supplemental Analyses Despite the disappointing results of the hypotheses testing, s everal in teresting correlations among other individual differ ences variables emerged from the present study, as shown in Table 9 .
72 Table 9 Pearson Correlations for Supplemental Analysis Individual Differences Variables Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Intent ions 2. Age -.089 3. Years .123 .236* 4. BMI .039 .217* .073 5. Activity -.111 .064 .025 -.216* 6. BIS -.151* .260* .104 -.198* .145 7. BAS .213* .076 .066 .118 .095 .203* 8. Promotion .155* .066 .082 .206* .157* .359* .189* 9. Prevention .021 .023 .074 -.044 -.002 .030 .266* .129 10. Believe .229* .137* .000 .0 54 .003 .036 .024 .178* .109 11. Future .019 .003 .021 .111 .046 .045 .032 .056 .098 .484* 12. Immediate .101 -.027 -.070 .137* .073 -.041 -.110 .059 -.045 .379* .201* 13. Benefits .138* .139* .054 014 .000 .005 .103 .204* .082 .669* .375* .309* 14. Risks .019 .124 .030 .051 .001 .004 .030 .026 .019 .293* .348* .258* .258* Note. Listwise N = 209. Intentions = intentions to increase physical activity; Years = y ears wi th diabetes; BMI = Body Mass Index; Activity = baseline physical activity; BIS = Behavioral Inhibition System scale score; BAS = Behavioral Activation System scale score; Promotion and Prevention = scale scores from the Regulatory Focus Questionnaire; Beli eve = perceived message believability; Future = perceived message emphasis on future; Immediate = perceived message emphasis on immediate ; Benefits = perceived message emphasis on benefits ; Risks = perceived message emphasis on risks. *p < .05.
73 Contrary t o Carver and White s (1994) predicted independence between the BIS and BAS scales, the present study revealed a significant positive correlation between the se two scales ( r = .203, p = .003) People with a stronger avoidance orientation (i.e., lower BIS sc ore) tended to also have a stronger approachorientation (i.e., lower BAS score). For the RFQ, findings were somewhat consistent with Higgins et al. s (2001) predicted independence between the promotion and prevention scales, with a positive but only margi nally significant correlation ( r = .129, p = .063). People with a stronger promotionfocus (i.e., higher promotion score) tended to have a somewhat stronger preventionfocus (i.e., higher prevention score). B etween the BIS/BAS and RFQ scales, the BAS s cale was negatively correlated with the promotion scale as expected ( r = .189, p = .006) People with a stronger approach orientation (i.e., a lower BAS scale score) tended to also have a stronger promotionfocus (i.e., a higher promotion scale score). A simi lar negative correlation between the BIS and prevention scales was expected but not supported ( r = .030, p = .666). Furthermore, there were significant positive correlations between the BAS and prevention s cales ( r = .266, p < .001) and the BIS and promoti on scales ( r = .359, p < .001). People with a weaker approach orientation (i.e., a higher BAS scale score) tended to have a stronger p revention focus (i.e., a higher prevention scale score), whereas people with a weaker avoidance orientation (i.e., a highe r BIS scale score) tended to have a stronger promotion focus (i.e., a higher promotion scale score). Another interesting finding shown in Table 9 is that f ive variables had a significant correlation with intentions to increase physical activity: BIS sca le score ( r = .151, p = .029), BAS scale score ( r = .213, p = .002) promotion scale score ( r = .155,
74 p = .025), perceived message believability ( r = .229, p = .001) and perceived message emphasis on benefits of being physically active ( r = .138, p = .046). As previously reported, CFC score was also significantly correlated with intentions ( r = .169, p = .012). In addition, years with diabetes had a marginally significant correlation with intentions ( r = .123, p = .075). A regression model was used to examine the unique influence of these s even variables on intentions while controlling for the others. Age was also included in the model because of its significant correlation with ye ars ( r = .236, p = .001), BIS scale score ( r = .260, p < .001) percei ved message believability ( r = .137, p = .048) and perceived message emphasis on benefits ( r = .139, p = 045). Table 10 presents results of the regression analys is Table 10 Summary of Multiple Regression Analysis Predicting Intentions to Increase Physi cal Activity from Supplemental Analysis Variables Variable B SE Intercept 1. 440 0. 8 54 BIS scale score 0.374 0.191 .147 BAS scale score 0.444 0.205 .150* Promotion s cale score 0.206 0.168 .098 Perceived message believability 0.413 0.150 .242* Perceived message emphasis on benefits 0.132 0.183 .064 CFC score 0. 240 0.1 41 122 Years with diabetes 0.011 0.011 .070 Age 0.008 0.008 .067 R2 = 149 Note. N = 215. BIS = Behavioral Inhibition System; BAS = Behavioral Activation System; CFC = Consideration of Future Consequences *p < .05.
75 When the eight variables were simultaneously entered as predictor s the overall regression model significantly explained 14.9% of the total variance in intentions to increase physical act ivity, F (8, 206) = 4.525, p < .001. Perceived message believability had the strongest influence on predicting intentions ( = .242, p = .006) People who perceived the message as more believable tended to have greater intentions to increase physical activi ty. The only other variable with a significant main effect was the BAS scale score ( = .150, p < .032) People with a stronger approach orientation (i.e., a lower BAS scale score) tended to have greater intentions to increase physical activity. Chapter Summary D ata were analyzed to test the hypotheses that when message valence and temporal proximity are congruent with individuals motivational orientation and CFC, respectively, intentions to increase physical activity w ill be greater than when the message is incongruent with these individual differences Although a nalyses did not support the h ypotheses supplemental analyses revealed several interesting correlations among other variables. The next chapter discusses possible explanations for these findings, directions for future research, implications for nurs ing, and study limitations.
76 Chapter 5: Discussion The present study examined how two individual differences factors CFC and motivational orientation combine to moderate message framing effects ( temporal proximity and valence) on intentions to increase physical activity. An attempt was made to replicate previous findings (largely from laboratory settings with undergraduate college students) in an at risk, c ommunity population and to extend these findings by exploring relationships not previously tested in message framing health behavior research. The present study was the first to examine message framing effects in an at risk, community population composed of adults with type 2 DM It was also the first study to examine the additive effects of both valence and temporal proximity message congruence with individual differences in the same sample. This study was the first health behavior study to draw from an at risk community population to si multaneously examine the message framing moderator effects of both the approach/avoidance and promotion/prevention motivational orientation systems in the same sample Discussion of Findings Hypotheses testing. Unfortunately, data did not lend support to the hypotheses presented above Participants who read a health message with a valence frame congruent with their motivation al orientation did not show greater intention s to increase physical
77 activity when compared to those who read a health message that was incongruent with their motivational orientation. P articipants who read a health message with a temporal proximity frame congruent with their CFC did not show greater intention s to increase physical activity relative to those who read a health message that was incongruent with their CFC. Several plausible explanations for these discordant findings are explored next. First, the health behavior examined in the present study (physical activity) may be more complex than behaviors examined in previous studies. For example, most studies that found a moderating effect of motivational orientation or CFC on message framing involved simpler behaviors such as dental flossing (e.g., Mann et al., 2004), vaccination (e.g., Gerend & Sheperd, 2007), sunscreen use (e.g., O rbell & Kyriakaki, 2008) and colorectal cancer screening (e.g., Orbell et al., 2004). In comparison to physical activity, these simpler behaviors typically take less time to perform and may involve just a one time or as needed behavior whereas increasi ng physical activity represents a commitment to a lifesty le change with less certain outcomes Furthermore, the term physical activity is more susceptible to multiple interpretations than simpler behaviors such as flossing. As r eported above the wide va riability of total weekly baseline physical activity may be attributed in part, to differences in participants perception as to what constitutes physical activity, which may, in turn, influence intentions and behaviors. D ecisions to engage in physical activity on a day to day basis may be more sensitive to contextual factors (e.g., physical limitations, acute or chronic health problems, competing demands, family/social support accessibility, preconceived attitudes beliefs and perceptions ) than are decisions regarding simpler, less imposing behaviors such as flossing A few participants in the present study wrote comments on
78 their completed booklet suggesting their intentions to increase physical activity were thwarted by such factors. For example: Poo l down right now. I have muscle myopathy, degenerative disc disease, and plantar fasciitis, so I am in pain and it is difficult to move ; Because of a crushed left foot, I have to wear a brace. I can only do light walking around the house. Thus i t is pl ausible that message framing effects on complex behaviors such as physical activity are attenuated to a greater degree compared to simpler behaviors due to the influence of such contextual variables. V ery few studies have assessed the congruence effect of message frame with individual differences on physical activity, t hus limiting meaningful comparisons. For example, unlike the present study, Latimer Rivers, et al. (2008) found support for the value from fit hypothesis, particularly in the promotionfocused condition. However, the researchers tested this hypothesis in a low risk community sample (v s. an at risk, c ommunity population) i n which people who had a physical impairment or physicians recommendation contraindicating unsupervised physical acti vity participation were excluded from the study. A second plausible explanation for the present studys unexpected findings is that a dults with type 2 DM have several unique characteristics that distinguish them from the more widely s tudied college studen ts in controlled laboratory settings or from health y, low risk community samples. For example, people with DM often have comorbid conditions that may affect their decisions to engage in physical activity such as uncontrolled blood pressure, peripheral neuropathy with painful and/or numb lower extremities, uncontrolled blood glucose, and visual problems. In addition, the regimen for self managing diabetes is one of the most challenging compared to that of other chronic
79 illnesses (Schillinger et al., 2002). P ersons with diabetes may perceive this self management as invasive because it often requires making difficult lifestyle changes and perform ing numerous functions In addition to being physical ly active, people with diabetes may have to perform tasks such as monitoring blood glucose taking multiple medications, visiting several health care providers, performing daily foot care, and adhering to a special diet on a daily basis These and other distinguishable characteristics of the diabetes population pose a challenge to researchers who attempt to extend findings from message framing studies that examined younger, more homogenous healthier populations Strength of the present studys message framing manipulation was explored as a third plausible explanation for discordant findings. Although the overall manipulation check suggested the health messages were perceived as intended, results (mean differences on items in Table 5) also showed that these manipulations have room for improvement. One possible strategy that may have enhanced manipulation strength of the present studys messages would have been to add a survey item asking participants to list their thoughts about the message immediately after reading it. Thought listing procedures have been used in many message framing studies to elicit greater elaboration upon, and deeper processing of the message, which in turn c ould strengthen framing manipulation and persuasiveness of the health message. P etty and Cacioppos (1986) elaboration likelihood model (ELM) posits that a messages persuasive ability to influence behavior change depends on how likely readers will think about the issue or argument. Several variables drawn from ELM have been shown to yield mediating or moderating effects on message framing such a s thoughts source credibility, and argument strength For
80 example, Orbell and K yriakaki ( 2008) found that thoughts mediated the CFC x temporal proximity framing manipulation. Jones et al. (2003) found a source credibility x valence frame interaction such that people who read a gainframed message from a credible source elaborated more and reported more positive exercise intentions and behaviors than participants in the other three conditions (gain framed/noncredible, loss framed/credible, loss framed/noncredible). Simple graphics with a fun, light tone were included in the present study to enhance readability of the message, to more actively engage the reader, and to illustrate examples of physical activity (in order to promote consistency in interpretat ions). The style of graphics used was similar to styles commonly used in diabetes education materials published by credible health organizations. Although the pilot study with 20 NPs suggested these graphics were appropriate, it is possible the cartoonlik e nature of the graphics decreased the messages source credibility, which in turn may have weakened effectiveness of framing manipulation. Laypeople in the community may perceive messages with graphics that have a more neutral, professional quality (e.g. pictures of real objects or real people) as having higher source credibility than messages with graphics that have a less serious, more casual quality (e.g., drawings of objects or cartoon figures). Empirical studies are needed to further explore this id ea. Updegraff et al. (2007) found argument strength moderated the valence congruence effect, suggesting that strong arguments (vs. weak arguments ) in favor of a health behavior are needed for motivational orientation to exert an impact on health behavior Compared to Updegraff et al.s (2007) operational definitions of weak and strong arguments about dental flossing, the strength of the present studys arguments
81 about physical activity appear average This average argument strength may have contributed to the weak congruence effects A rguments in the present studys message could have possibly be en strengthened by including more definitive rather than tentative statements (e.g., will vs. may or can) or by adding empirical evidence for being/not being physically active. However, peoples perceptions of what makes an argument strong may vary across populations (especially between college students and at risk community populations) and should be considered when designing health messages. P erceptions may also vary based on the health behavior being studied. Consideration of how a messages outcome focus may moderate framing effects is another s trategy that could enhance framing manipulation. Yi and Baumgartner (2009) found that outcome focus (gain end st ate vs. loss end state) moderated valence framing effects. Gain framed messages were more persuasive with a gain end state (presence of gain) than with a loss endstate (absence of loss), whereas loss framed messages were more persuasive with a loss end st ate (presence of loss) than with a gain end state (absence of gain). All versions of the health message in the present study presented arguments with a gain endstate (i.e., presence or absence of gain). The framing manipulation may have possibly be en stre ngthened if actual losses versus missed gains would have been presented in the loss framed messages (e.g., less energy and worse sleep v s. miss out on more energy and better sleep). A nother strategy that may improve framing manipulation is to consider possible effects that peoples beliefs about health behaviors may have on their decisions to perform these behaviors. People who believe physical activity is painful, is not beneficial, and is limited to only rigorous exercise may be l ess likely to increa se physical activity
82 after reading a congruent framed message than people who believe physical activity is fun, easy, and yields multiple benefits. Had a pre manipulation measure t o assess beliefs about physical activity been included in the present study, beliefs could have been examined as a covariate, which in turn may have s tatistically strengthen ed the framing manipulation. Consideration should also be given to the effects various operational definitions and uses of message framing may have on manipul ation strength. As the literature indicates, there is wide variation among studies with regard to how message frames are manipulated and how these frames (e.g., gains and losses) are operationally defined. In studies that examined approach/avoidance motiva tional orientation systems Rothman and Saloveys (1997) traditional approach (based on prospect theory, Kahneman & Tversky, 1979) was used to define gains and losses in terms of valence frame (e.g., Mann et al., 2004; Sherman et al., 2006). With this appr oach, gain framed messages can be presented as presence of gains or absence of losses, and loss framed messages can be presented as presence of losses or absence of gains. An alternative operational definition of gains and losses has been used in studies t hat e xamined promot ion/prevention motivational orientation systems (e.g. Latimer, Rivers, et al., 2008). Based on Higgins ( 1997, 1998, 1999) regulatory focus theory, gains and losses refer to end states of a behavior. With such an approach, messages wi th a gain end state (positive outcome focus) depict a presence or absence of gains, whereas messages with a loss end state (negative outcome focus) depict a presence or absence of loss Study designs also vary as to whether messages present a single frame or a mixed frame. Mos t studies that involved only valence framing used just one valence frame per message. Single valence framed
83 messages (gain or loss) have been shown to be more persuasive than mixed valence framed messages (gain and loss) (e.g., Latimer Rench, et al., 2008). However, in studies that examined CFC, mixed valence and mixed temporal (immediate and future) framed messages have been shown to moderate effects of the CFC x temporal frame interaction on the promotion of health behaviors (e.g., O rbell & Kyriakaki, 2008). The present study used R othman and Saloveys (1997) t raditional approach to define gains and losses but did not use Higgins ( 1997, 1998, 1999) approach. H ad Higgins concept of outcome focus /endstates been considered when designing the four message versions, valence congruence of messages with peoples regulatory focus (promotion or prevention ) may have been strengthened. Several studies have tested the moderating effect of regulatory focus on outcome focus, but findings ar e mixed. Some studies lend support for the predicted moderating effect (e.g., Latimer, Williams Piehota, et al., 2008 ) while other studies d o not (e.g., Yi & Baumgartner, 2009). The present study presented only one valence and one temporal frame per message. Had this study presented mixed versus single frames, temporal proximity congruence of messages with peoples CFC may have been strengthened, but valence congruence with motivational orientation may have simultaneously been weakened. Furthermore, where this mixed frame approach may be more persuasive with college students and healthy community samples, it may be less persuasive or even counter productive with at risk populations When delivering a message to people with diabetes about the promotion of re gular physical activity, emphasizing losses they will experience as well as gains (regardless of temporal frame) may be less effective than a message that emphasizes only benefits of physical activity, and in fact may actually dissuade someone from doing p hysical activity
84 at all. This reasoning is consistent with Rothman and Saloveys (1997) hypothesis that gainframed messages are more persuasive than loss framed messages for prevention behaviors Dose of message fram e may also affect strength of the m anipulation. Dose refers to the frequency of delivering framed messages. Most studies have used a single dose approach, where participants read a framed message at a single point in time and outcome variables were assessed one time after reading the message. A few studies have used a multi dose approach, where participants read a framed message at multiple points in time and outcome variables were assess ed at different intervals. S tudies have shown that framing manipulation was more effective in promoti ng healthy behaviors when a multi dose approach was used compared to a singledose (e.g., Latimer, Rench, et al., 2008). Methodological differences among studies were explored as a f ourth possible explanation for the present studys discordant findings. Th is study was the first to r epresent temporal proximity and valence congruence with variables constructed from conditional transformations on the corresponding individual differences scale scores ( CFC, BIS/BAS and RFQ) and to examine the congruenc e e ffect by testing for significance of zero order correlations of the congruence variable s with the outcome variable This innovative intuitive approach differs from the usual approach of representing congruence with a message frame x individual differences interaction term and examining the congruenc e effect by testing for significance of the interaction. Researchers also vary in how they treat individual differences scale scores. Some researchers analyze the scores a s continuous variables and u se multiple regression to test
85 the congruenc e effect (e.g., Mann et al., 2004), while others treat the scores as dichotomous va riables (usually via median split) and use an ANOVA to test the congruence effect (e.g., Orbell et al., 2004; Sherman et al., 2006) Maxwell and Delaney (1993) caution researchers who use bivariate median splits that they may lose power to detect true predictor criterion relationships in some situations or they may dramatically increase the probability of Type I errors in other situations. To a void such potential problems, the present study treated scale scores as continuous. I n order to e valuate the present studys novel approach and to ensure this approach d id not obscure message framing effects, data were also analyzed using the two traditio nal approaches. Even when the data were treated the same as in earlier studies (tests of interaction), there was no support for any congruence effect. This suggests the present studys approach for operationalizing congruence was not responsible for the ne gative findings obtained. Another way researchers vary in how they treat individual differences scale scores relates particularly to scales that are made up of two or more subscales. For example, with the BIS/BAS and RFQ scales, some researchers analyze separately the two subscales to test hypotheses (e.g., Yi & Baumgartner, 2009) whereas other researchers (as in the present study) combine the two subscales into a single index and analyze the index to test hypotheses (e.g., Latimer, Williams Piehota et al., 2008; Mann et al., 2004). Variations also exist in how researchers combine these subscales ; s ome subtract while others add the two scores Still another difference among studies relates specifically to those that use the BIS/BAS scale. Carver and Whites (1994) original BAS scale consist ed of three separate scales: BAS Reward Responsiveness, BAS Drive, and BAS
86 Fun Seeking. Because these three scales appear strongly related, researchers often combine them and use one BAS scale to analyze data ( as in the present study) (e.g., Gerend & Shepherd, 2007; Sherman et al., 2006). Other researchers analyze one or more BAS subscales separately (e.g., Yi & Baumgartner, 2009). Another methodological difference among studies that should be consider ed is the use of various measures to assess the same individual differences constructs. In addition to Strathman et al.s (1994) CFC scale, the Zimbardo Time Perspective Inventory (Zimbardo & Boyd, 1999) is another commonly used measure of future orientation. Lockw ood et al.s (2002) regulatory focus scale and Higgins et al.s (2001) RFQ scale are the most widely used measures to assess regulatory focus. However, other measures have also been used such as Carver and Whites (1994) B IS/BAS scale ( e.g., Yi & Baumgartn er, 2009). Based on convergent and discriminant validation principles (Campbell & Fiske, 1959), one would expect these measures of regulatory focus to behave similarly ; however, inconsistencies have been found. While the ideas of approach/avoidance and pr omotion/prevention motivational orientation have distinct differences, they also appear related. An approach strategy is usually taken for promotion, and an avoidance strategy is usually taken for prevention (Higgins, 1997; Higgins et al., 2001). Similar behavioral predictions about message framing effects have been made for approach oriented and promotionfocused people, as well as for avoidance oriented and prevention focused people (e.g., Lee & Aaker, 2004; Mann et al., 2004; Rothman et al., 2008). Although these two ideas have been examined in separate studies, only one health behavior study has simultaneously examined both constructs in the same sample.
87 Yi and Baumgartner (2009) conducted a laboratory experiment with 120 undergraduate college students t o examine whether the moderator effects of regulatory focus on message framing effectiveness depended on the way regulatory focus was measured. They used a 2 (overall valence) x 2 (outcome focus) x 2 (type of benefit/harm emphasized) within subjects factor ial design to develop messages about regular fruit and vegetable intake. Participants read all versions of the health message in a randomly assigned order and rated perceived persuasiveness of each message using a oneitem measure. Chronic regulatory focus was then assessed in a questionnaire that included Higgins et al.s (2001) RFQ scale, Lockwood et al.s (2002) regulatory focus scale, and Carver and Whites (1994) BIS/BAS scale. Discrepancies were revealed among the preventionrelated scales but not the promotionrelated scales. Yi and Baumgartner (2009) found partial support for Mann et al.s (2004) congruency hypothesis and Higgins (2000) value from fit hypothesis. The valence main effect was moderated by regulatory focus when this individual difference was assessed using Lockwood et al.s (2002) prevention focus scale and Carver and Whites (1994) BIS scale The gainframed advantage was less pronounced among participants with a strong prevention focus than those with a weak prevention focus. However, this interaction was not observed when using Higgins et al.s (2001) prevention scale. Regardless of which promotion scale was used, the valence main effect was not moderated by regulatory focus. The gainframed advantage was not more pronounce d among participants with a strong promotion focus than those with a weak promotion focus.
88 The present study extend s Yi and Baumgartners (2009) research beyond a laboratory setting with college students to a community setting with an at risk population by administering both the RFQ and BIS/BAS scale to participants and comparing valence congruence effects. W hereas Yi and Baumgartner (2009) separately analyzed the two RFQ and BIS/BAS subscales to test for congruence effects, the present study analyzed co nditionally transformed s ingle index scores to test the congruence hypotheses Unlike Yi and Baumgartner (2009), t he present study found congruence effects were absent regardless of the measure used to assess regulatory focus. Neither valence congruence va riable yielded an effect as evidenced by non significant (and very small) zero order correlations with intentions. These inconsistent findings between studies could be related to several factors such as differen ces among participants, different study setti ngs and different approaches in using and analyzing the RFQ and BIS/BAS scales. A fifth and final possibility that was explored as a potential contribution to the present studys findings was whether the participants in th e sample varied on individual differences measures when compared to participants in other studies. Normative values for the CFC RFQ, and BIS/BAS scales have not been reported in the literature; therefore, mean scores from this study were compared to mean scores in other health behavior studies. It is important to note that some studies failed to report mean scores, so only a few comparisons could be made. S tudies in which comparisons were made involved university campus or healthy community samples rather than at risk populations; t herefore, caution should be taken when interpreting results.
89 For the CFC scale ( scale range 1 to 5), mean score ( SD ) in the present study was 3.43 ( 0.71) compared to the following studies: 3.20 (0.66) (Orbell et al., 2004); 3.25 (0.96) (Orbell & Hagger 2006); 3.30 (0.69) (Orbell & Kyriakaki, 2008, Experiment 1) ; 3.31 (0.54) (Orbell & Kyriakaki, 2008, Experiment 2) ; and 3.20 (0.70) (Adams & White, 2009). For the RFQ scale (scale range 1 to 5), mean scores ( SD ) in the present study were 3.53 (0.66) and 3.46 (0.83) for the promotion and prevention scales, respectively, compared to the following studies: 3.72 (0.59) and 3.20 (0.88) (Fuglestad, Rothman, & Jeffery, 2008, Study 1A ) and 3.69 (0.60) and 3.47 (0.87) (Fuglestad et al., 2008, Study 1B). Neither the m eans nor standard deviations on the CFC and RFQ scales in this studys sample appear to differ from means and standard deviations in other studies samples. These data suggest at risk populations are similar to healthier community and university campus populations in both CFC and regulatory focus B ecause response metrics widely vary a cross studies employing the BIS/BAS scale, it is difficult to make meaningful comparisons among the scale means. Mail survey. The present study also provided some information on the effectiveness of Dillmans Tailored Design Method (TDM) (Dillman et al., 2009) as a recruitment procedure (results in Table 3). This study was one of the first in the health behavior message framing literature to use the TDM. Consistent with other studies that have used th is approach, an overall response rate of 54% was obtained, well above the 2535% response rate that one shot bulk mailing approaches typically yield. R esults also indicate the multiple contact design of the TDM was effective. The initial prenotice letter and booklet of measures with cover letter yielded an initial response rate of 46%. T he subsequent mailing of reminder/ thank you postcards and replacement booklets (as
90 applicable) increased the response rate to 54% Although the TD M approach used in this study was more expensive and more labor intensive than a one time mail survey, it appeared to be more successful. Over the past decade, there has been a dramatic rise in use of the Internet and email to conduct surveys. While electronic surveys may offer potential benefits (e.g., cost savings, greater efficiency) over mail surveys in certain populations (e.g., professional associations, university students), this mode of delivery is not always a suitable replacement for the mail m ode. Although both Internet access and computer operation skills have substantially improved in recent years, significant proportions of the U.S. population remain without the technology or ability. Other challenges are the lack of standards on how to crea te email addresses (in order to develop sampling algorithms) and the lack of a systematic list of Internet users from which to draw a sample (Dillman et al., 2009). These complications and potential difficulties for the target population of adults with type 2 DM were weighed in the decision to use the mail mode over the electronic mode Supplemental analyses. Several interesting findings emerged from supplemental analyses that were conducted in the present study. Perceived believability of the health me ssage yielded the strongest influence on predicting intentions to increase physical activity while controlling for the influence of seven other variables using multiple regression (results in Table 10). Strategies on how to increase believability of messag e s need to be further explored. In the present study (results in Table 9), believability was positively correlated with age ( r = .137, p = .048). People who were older tended to perceive the health message as more believable. This pattern was not significantly
91 different across the four conditions. This finding raises important questions What features of a message design make it appear more believable to younger people versus older people? Wha t individual differences vary between younger and older people that may affect perceptions of message believability (e.g., thought processes, beliefs, attitudes, past events)? Do these differences exist in other at risk or healthier populations? These and other relevant questions need to be studied i n order to guide nur sing practice. Message believability was also positively correlated with the RFQs promotion scale score ( r = .178, p = .010). People who had a stronger promotion focus tended to perceive the health message as more believable. This suggests the possibi lity that if a persons promotion focus could be i ncreased prior to reading a message, believability of the message may be increased, which may in turn increase intentions to perform healthy behaviors. Although regulatory focus is thought to be a relative ly stable construct Higgins (1997) posits that momentary situations can temporarily induce either a promotion or prevention focus. A few h ealth behavior studies lend support for this hypothesis (e.g., Lee & Aaker, 2004) Message believability was also po sitively correlated with CFC scale score ( r = .225, p = .001). People who were more future oriented tended to perceive the health message as more believable. Although a persons CFC is less likely than regulatory focus to be affected by situational manipul ation, recent findings have emerged that suggest CFC may be a changeable construct over time (e.g., Toepoel, 2010) such as when people experience a significant event (e.g., develop a chronic disease) or a dramatic change in their life (e.g., change in soci oeconomic position). In the present study, CFC score was
92 negatively correlated with BMI ( r = .193, p = .005). People with a higher future orientation tended to have a lower BMI Adams and White (2009) found a similar correlation as well as a positive cor relation between socioeconomic position (SEP) and CFC score P eople who were less deprived tended to be more future oriented. In addition, they found that CFC score partially mediated the negative relationship between SEP and BMI. These findings suggest CF C may play an important role in promot ing healthy behaviors that will lead to desirable health outcomes. These and other situational, health, and lifestyle variables need to be further explored to identify factors that may increase a persons CFC over time and thus improve their attitude towards healthy behaviors. BAS scale score yielded the second strongest influence on predicting intentions to increase physical activity while controlling for all other variables in the supplemental analysis regression mod el BAS and BIS brain systems have been found to be stable constructs, and little is known as to whether they are changeable. As reported in Table 6, BIS/BAS and RFQ single index scores were positively correlated as expected ( r = .441, p < .001) The more approachoriented people tended to be more promotionfocused, and the more avoidance oriented people tended to be more preventionfocused. As an extension of Yi and Baumgartners (2009) research, correlations within and between the scales subscales were examined to further explore the relationship between these two motivational orientation measures (results in T able 9 ). Due to reverse coding between scales, comparisons are not intuitive however. A low score on the BIS or BAS scale represents a stronger avoidance or approach orientation, respectively, whereas a low score on the RFQs prevention or promotion scale represents a weaker prevention or promotion focus, respectively.
93 According to Carver and White (1994) t he B A S and B I S scales should be orthogonal where any combination of scores is possible (e.g., strong approach/strong avoidance, strong approach/weak avoidance, weak approach/weak avoidance, strong avoidance/weak approach). The present study contradicts this prediction. B A S and B I S scores had a significant positive correlation (r = .203, p = .003), suggesting they were not orthogonal. This finding differs from studies that have found nonsignificant correlations between B A S and B I S scales (e.g., Gerend & Shepherd, 2007; Mann et al., 2004; Shen & Bigsby, 2010; Shen & Dillard, 2009; Updegraff et al., 2007; Yi & Baumgartner, 2009). All of these studies were with college students, suggesting the concept of orthogonality between the B A S and B I S scales may not generalize from a controlled laboratory setting to a noncontrolled, community setting. Similar to the B AS and B IS scales, t he RFQs promotion and prevention scales are expected to be orthogonal (Higgins et al., 2001 ), where any combination of scores is possible (e.g., strong promotion/strong pre vention, strong promotion/weak prevention, weak promotion/weak prevention, strong prevention/weak promotion) Promotion and prevention scores had a marginally significant positive correlation ( r = .129, p = .063) in the present study. In comparison to othe r studies conducted in community settings, some have found similar significant positive correlations between the promotion and prevention scales (e.g., Fuglestad et al., 2008, Studies 1A & 1B) whereas others have not (e.g., Haaga, Friedman Wheeler, McIntos h, & Ahrens, 2008 Study 1). As with the B A S and B I S scales, these findings suggest the concept of orthogonality between the promotion and prevention scales may not consistently emerge in community settings as they do in laboratory settings with college st udents
94 To further explore the relationship between BIS/BAS and RFQ scales, intercorrelations between these scales subscales were examined Based on predictions that pro moters are similar to approachers and preventers are similar to avoiders and that BAS and BIS scales and promotion and prevention scales are orthogonal, the following correlations were expected (using the reverse coding interpretation as previously described and convergent and divergent validation principles [Campbell & Fiske, 1959]): BAS and promotion scales should be negatively correlated; BIS and prevention scales should be negatively correlated; BAS and prevention scales should be weakly correlated; and BIS and promotion scales should be weakly correlated. Only one of these four predictions was supported in the present study: BAS scores were negatively correlated with promotion scores (r = .189, p = .006). Contrary to expectations, BIS scores were not correlated with prevention scores ( r = .030, p = .666), BIS scores were co rrelated with promotion scores (r = .359, p < .001), and BAS scores were correlated with prevention scores ( r = .266, p < .001). These results were compared to Yi and Baumgartners (2009) findings, who split the BAS scale into two separate scales when maki ng comparisons rather than using a single BAS index score (as in the present study). Yi and Baumgartner (2009) also found the expected correlation between the promotion and BAS scales, but only with one of the BAS scales (BA S drive). Similar to the present study, no correlation was found between the BIS and prevention scales as expected. Furthermore, a similar unexpected significant correlation between the BIS and promotion scales was also found, where weaker avoiders tended to be stronger promoters. However, a correlation was not found between either of the BAS scales and
95 the prevention scale as it was with the present study (where weaker approachers tended to be stronger preventers). Summerville and Roese (2008) conducted a principal component s ana lysis (nonhealth behavior study) with college students to compare the BIS/BAS and RFQ scales (using subscale scores for comparison). As with both the present study and Yi and Baumgartners (2009) study, Summerville and Roese (2008) found the expected co rrelation between the BAS and promotion scales but found no correlation between the BIS and prevention scales. (They did not report correlations between the BAS and prevention and BIS and promotion scales). In summary, findings among these three studies s uggest correlations within and between the BIS/BAS and RFQ subscales are unclear and do not consistently behave across populations and settings as originally predicted. These findings may partially explain the discordant results in the present study. W hereas neither the BIS/BAS nor RFQ single motivational index score was correlated with intentions to increase physical activity (results in Table 6), three (BIS, BAS, and promotion) of the four subscales were correlated with intentions (results in Table 9). These findings suggest that using separate subscale scores versus single index scores may better represent peoples true approach/avoidance and promotion/ prevention motivational orientation systems, which may in turn strengthen congruence effects of a framed message with motivational orientation. Summerville and Roese (2008) suggest another possible explanation for inconsistent findings between the BIS/BAS and RFQ scales. Although both scales are common measures of the broad motivational orientation construct, unique differences exist between the more specific constructs that each scale measures. Therefore, caution
96 should be exercised when comparing results between different measures of motivational orientation. Directions for Future Research As the literature indicates, there is a paucity of studies that have evaluated the congruence effect of message framing with individual differences in at risk populations involving complex health behaviors. Findings from the present study need to be replicated a nd extended to other at risk populations and behaviors Several themes emerged from this study that can serve as a useful guide to future message framing and message tailoring in health care research First, contextual variables associated with at risk p opulations and complex health behaviors need to be identified and their effects controlled for, which may in turn strengthen message framing effects on behaviors. Second, various strategies need to be tested under multiple conditions to identify effective ways to strengthen manipulation of framed messages such as the following: 1) inclusion of a thought listing task post manipulation; 2) manipulation of graphics (no graphics vs. cartoonlike graphics vs. more neutral and real life graphics) ; 3) manipulation of argument strength (e.g., tentative vs. definitive arguments; empirical data vs. no empirical data); 4) consideration of a messages outcome focus (presence of loss vs. absence of gain; presence of gain vs. absence of loss); 5) manipulation of how mes sages are framed and designed in various combinations (traditional gain framed vs. loss framed; gain endstate vs. loss end state; single frame vs. mixed frame vs. no frame) ; and 6) manipulation of message dose (single dose vs. multiple doses).
97 Third, various strategies need to be tested to identify effective ways to strengthen perceived believability of health messages, which may in turn optimize persuasiveness of the se messages and ultimately lead to an increase in healthy behaviors. Consideration should be given to variables with which believability has been shown to correlate In the present study, age positively correlated with believability. This relationship needs to be more closely examined to identify how perceptions of specific message character istics differ among various age groups. This, in turn, will guide nurses and other health care providers to design tailored messages that match specific age groups. Fourth, the relationship s of individuals promotionfocus motivational orientation and CFC with their intentions to perform healthy behaviors need to be further explored. In the present study, these two individual differences were positively correlated with both intentions and message believability. Findings from studies that have shown a persons regulatory focus can be temporarily manipulated to enhance the valence congruence effect with a particular message need to be replicated in at risk populations. S ituational, health, and lifestyle variables that may be related to CFC should also be expl ored to i dentify nursing interventions that may contribute to increasing a persons CFC over time, which may in turn strengthen effectiveness of gainframed messages on promoting healthy prevention behaviors. Fifth, relationships within and between motivational orientation measures (e.g., RFQ and BIS/BAS scales) need to be further examined in at risk populations. Data need to be analyzed using both composite single index scores and separate subscale scores in the same sample to determine if one approach more truly represents peoples motivational disposition than the other. A more precise measurement may in turn
98 strengthen valence congruence effects. In the future, researchers could replicate the present studys design (incorporating the suggestions to s trengthen framing manipulation as discussed above) but also include Lockwood et al.s ( 2002 ) regulatory focus scale as a third measure of motivational orientation. Studies suggest the regulatory focus scale behaves similarly to Carver and Whites (1994) BI S/BAS scale in laboratory settings with college students (e.g., Summerville & R oese, 2008; Yi & Baumgartner 2009). Strong evidence has shown that wide variability exists a mong message framing health behavior studies. This variability poses a challenge when trying to compare previous findings, which in turn makes it difficult to draw firm conclusions and formulate definitive clinical practice recommendations. Although three broad meta analysis pa pers on message framing research were conducted in an atte mpt to synthesize the literature ( Khberger, 1998; OKeefe & Jensen, 2006, 2007) all three fell short of providing an adequate, comprehensive review (as presented in Chapter 2). S ystematic reviews are needed that include comparisons of study characteristi cs such as research design, underlying theoretical conceptual framework, research setting, sample description, nature of the health problem (hypothetical vs. real), dependent variables, moderating variables, and measures employed. Latimer, Brawley, and Ba ssett (2010) recently conducted a focused systematic review to evaluate the effectiveness of three approaches for constructing physical activity messages: message tailoring, message framing (in terms of gains vs. losses) and targeting messages to change self efficacy. They reviewed only s ix studies t hat examined the effects of valence framed messages on physical activity behavior and/or intentions Compared to the other three meta analyses, Latimer et al.s (2010) review included more
99 meaningful comparisons across studies to facilitate identification of trends and patterns. However, their review was underpowered, so definitive practice reco mmendations could not be made. S tudies that involved at risk clinical populations (e.g., people with diabetes) were excluded from their review, thus further limiting generalizability of findings. Future systematic reviews using similar methods employed by Latimer et al. (2010) need to be conducted with studies that involved at risk populations and complex health behaviors. T he present studys inconsistent findings w ith theor etical predictions about congruence effects of message framing with individual differences su ggest that underlying theories and conceptual frameworks may need refinement to more accurately predict congruence effects on complex health behaviors in at risk community populations Predictions as originally hypothesized may not generalize from controlled laboratory settings to reallife, volatile settings. Fu ture research is essential t o guide these theoretica l revisions and to continue advancing the message framing literature Implications for Nursing Research, Education, and Practice Findings from the present study offer important implications for nursing research, education, practice. There is an internatio nal call for an increase in health communication research as one strategy to address the global epidemic of chronic diseases (e.g., DHHS, 2000b). There is a paucity of published empirical studies that adequately examine health care provider educational int erventions and the effectiveness of these interventions in contributing to desired outcomes. Coster and Norman (2009) report on findings of a review of 30 Cochrane systematic reviews of educational interventions designed to improve patients knowledge and skills to manage chronic disease, with particular reference to nursing contribution and practice. The majority of reviews (60%) were
100 judged to provide inadequate evidence of the effectiveness of the interventions. This insufficient evidence, coupled with l acking quality of several trials that were evaluated, limited Cochrane reviewers ability to draw firm conclusions on the effectiveness of educational interventions and to make specific clinical practice recommendations. Future experimental studies are needed to help identify the ingredients of successful messages (e.g., gain vs. loss framed) that health care providers deliver to their patients, with an emphasis on how to tailor these ingredients based on individual differences among their learners M essage framing is a promising health communication strategy that has been well studied in the psychology literature over the past 20plus years across a breadth of health behaviors. However, most of these studies were conducted in a laboratory setting with college students (sometimes using hypothetical health problems) rather than in a n at risk community population with actual health problems. Nurse researcher s are well connected to various clinical settings and thus have an opportunity to help fill in this literature gap by addressing questions previous findings of message framing effects have raised in relation to health care provider educational interventions. For example, when delivering diabetes self management education to adults with type 2 DM is a g ain framed or loss framed message more effective in promoting regular physical activity ? What individual differences among these learners moderate or mediate message framing effects, such as future orientation, motivational orientation, age, and beliefs? W hat other types of message manipulation may relate to valence framing, such as temporal proximity, argument strength, source credibility, and use/appearance of graphics? How should providers customize their message s based on differences in dispositional and
101 situational factors in order to maximize persuasiveness of the message on increasing physical activity ? Studies such as this example which involve other highly prevalent chronic diseases (e.g., cardiovascular disease and cancer) and associated modifiable health behaviors (e.g., unhealthy diet and tobacco use) may provide empirical answers to these questions and strengthen researchers ability to ascertain what educational interventions will work for whom and in what situations. Such studies will also resp ond to the international call for increased health communication research. Nurses have a prime opportunity to be on the cutting edge of message framing research in clinical settings involving at risk populations that will help guide the practice of health care providers in various disciplines in delivering influential messages to their patients. Research alone, however, is insufficient to achieve this desired goal. The valuable role of education must also be considered. Nurses and other health care provide rs require adequate knowledge and skills to successfully apply and incorporate research findings into their communication and education practices. However, providers vary in their own communication abilities, and there is a scarce amount of suitable traini ng opportunities to enhance these skills (Astin & Closs, 2007; Kim et al., 2008; WHO, 2005a). The WHO (2005a) reports that training of the international health care workforce has generally not kept pace with the rapid escalation of chronic health problems and that the workforce demonstrates a lack of training, education, and skill set to effectively manage patients with chronic conditions. This is largely related to challenges encountered during the recent transformation from the traditional provider center ed approach (which emphasized treating acute, episodic illnesses) to the contemporary patient centered approach (which emphasizes promoting
102 health and preventing chronic conditions and associated complications). The WHO (2005a) presents a new and expanded training model that consists of five core competencies which augment rather than replace existing competencies and can be implemented in a variety of training contexts. This model is designed to help expand the skills of all health care providers to meet t he new complexities associated with the chronic disease epidemic. The first competency in this model patient centered care is particularly relevant to the health communication literature as it includes the following main components: interviewing and co mmunicating effectively, assisting changes in health related behaviors, supporting self management, and using a proactive approach. Astin and Closs (2007) comment on how the WHO report (2005a) can be specifically applied to nursing, particularly the patien t centered care competency. They suggest that little has been done to equip nurses with the adequate knowledge and skills required to deliver self management education and thus call for greater training opportunities. Health care providers in general need to be competent in delivering educational messages. Offering multiple training opportunities in various settings and contexts will enhance the ability of providers to effectively educate and motivate their patients to better care for themselves. For examp le, in nursing academic programs, therapeutic nurse patient communication and education is a learning objective that cuts across a broad spectrum of health behaviors associated with chronic disease management and prevention. Therefore, components of health communication research should be woven throughout the curricula of all relevant courses in the program. In clinical settings, onsite workshops might be offered or off site education could be made available to nursing staff. Interactive exercises may especially be useful when teaching communication skills,
103 such as audio /videotaping learners during role play scenarios and playing back these tapes for learners and peers to critique their own performance. Less formal educational opportunities such as independent learning (e.g., handbooks and online literature reviews) and real time, on the job mentoring may also be valuable and promote reinforcement and enhancement of these learned skills. In addition to health communication approaches, providers also need to be familiar with situational and dispositional factors that have been shown to impact persuasiveness of approaches like message framing. As with any training, initial and ongoing competence needs to be assessed, which is consistent with the WHOs (2005a) model. For example, ability to execute effective communication strategies might be evaluated via methods such as asking learners to develop tailored written or verbal health behavior messages (e.g., gainvs. loss framed; immediate vs. distal framed) an d then deliver these messages in various contexts such as simulation, role playing, or demonstration with the target audience. Similar activities could be done where learners practice increasing believability of their message by t ailor ing it to specific in dividual differences such as age specific characteristics. Opportunities for learners to practice strategies shown to temporarily alter patients regulatory focus to increase congruence and effectiveness of a framed message should also be provided, as well as strategies shown to increase CFC over time. Health care providers competence in skills such as these is necessary to help maximize the potential practice implications of wellestablished health communication strategies Researchers have found that he alth care provider educational interventions for patients with chronic conditions such as diabetes, asthma, and epilepsy can increase knowledge, promote adoption of healthy self management behaviors, and improve health
104 status (see Barlow, Wright, Sheasby, Turner, & Hainsworth, 2002; Coster & Norman, 2009). Most importantly, these benefits help patients experience better overall health and quality of life. T hese desired outcomes are also valuable to providers themselves. For example, reimbursement for servic es to clinicians such as nurse practitioners and physicians is often tied to how successful their patients are at achieving targeted clinical goals such as normal hemoglobin A1C, blood pressure, and blood lipid levels. Clinicians are well aware of these go als but struggle to obtain them, largely related to patient education issues. As part of a patient centered approach, a vast amount of teaching aimed at promoting healthy behaviors may be required for patients with chronic conditions and related co morbi dities. Wellestablished health communication strategies such as message framing and other types of message tailoring may accelerate the speed in which patients adopt these behaviors. Health care providers report several barriers to providing effective edu cational interventions such as lack of knowledge of well established strategies, lack of confidence in ability to implement these strategies, lack of time to personally deliver health education, and lack of communication aids to reinforce and support this education (see Coster & Norman, 2009; Kim et al., 2008). Message tailoring i s an approach that offers promise in its potential to help lessen these barriers. Health care providers often use printed materials to deliver patient education. It is usually pre ferable to use these materials in conjunction with personally delivered education to help reinforce and support the messages. In reality, however, printed materials are often used as the only means to educate patients, largely due to insufficient time for personal delivery of the messages. Thus, it is essential to have adequate materials
105 that either alone or coupled with personal education are highly persuasive in promoting healthy behaviors. Most printed materials utilized are from the health care industry (e.g., pharmaceutical and medical supply companies) and are not always readily available to providers. In addition, these materials often contain biased, branded, inadequate information, are written at unacceptable reading and readability levels, and are not modifiable to account for individual differences in learners. These limitations weaken the persuasive power of these materials and may even render them ineffective and meaningless to patients. As a result, providers may be required to design their own printed materials. Principles of message framing and other message tailoring strategies can help guide nurses and other providers in developing customized materials to enhance overall effectiveness of educational interventions. In addition to written mate rials, message framing has potential for delivering less formal, verbal health messages. As part of the initial assessment, where providers should routinely assess patient characteristics like readiness to learn and preferred learning styles prior to provi ding education, providers could also evaluate individual characteristics that have been shown to moderate message framing effects on health behaviors (e.g., motivational orientation and CFC). Many of these characteristics can quickly be assessed using shor t, wellestablished instruments. P roviders could then use these findings to customize the verbal message (e.g., gain vs. loss or immediate vs. distal) to be most congruent with their patients characteristics (e.g., approach vs. avoidance oriented or low vs. high CFC). This strategy can be used to deliver education in a variety of clinical settings such as informally at the bedside or formally during a structured teaching session.
106 All messages, regardless of the medium through which they are delivered need to be appropriately customized. Knowing the right combination of ingredients that are maximally congruent and most effective with specific populations under a variety of circumstances will assist providers and the health care industry to formulate optimally tailored messages to help maximize their persuasiveness in promoting behavior change and hastening attainment of targeted goals. Study Limitations Some limitations should be kept in mind when considering results of the present study. F irst, the sample was nonrandom and comp osed of adults with type 2 DM from a single metropolitan area in the Southeastern United States (Tampa Bay) who had been seen by one of two health care providers (a n endocrinologist or an ARNP ) specializing in diabet es management As a result of sampling bias participants in this study may not have be en representative of all adults with type 2 DM in the Tampa Bay area. Second, the mail survey showed a s election bias ; o f the 441 deliverable surveys 155 (35%) were not returned. No demographic information was obtained from the nonrespondents; therefore, comparisons could not be made to those who did respond. Because the study design compared groups of respondents that were created by random assignment, the selection bi as issue seems somewhat trivial and most likely had a minimal effect on the findings. T hird, the data collected were obtained from self reports. As with all self report information, these data may reflect bias in reporting certain thoughts, feelings, and a ctions. Fourth, the effects of omitted variables (e.g., beliefs about physical activity, physical limitations, and current health condition) are unpredictable. Their absence could
107 contribute to misleading conclusions with regard to message framing effects on intentions to increase physical activity. F ifth, all participants received a booklet of materials presented in the same exact order, with individual differences scales placed after the health message, manipulation check, and intention items. Whereas many other message framing studies used a similar order (e.g., Gerend & Shepherd, 2007; Orbell & Kyriakaki 2008; Yi & Baumgartner, 2009), a few presented individual differences measures before the health message (e.g., Mann et al., 2004; Uskul et al., 2009) W hile it is unlikely that reading a health message would affect responses on scales that measure relatively stable i ndividual differences having not counterbalanced the sequence of materials, the possibility of order effects cannot be ruled out. T he pre sent study contributes to the message framing literature in that it is only one of a few that examined framing effects on a complex health behavior in a widely understudied at risk population. This study also introduced a new approach for operationalizing valence and temporal proximity congruence as a continuous variable using conditional transformations. P lausible explanations for this studys negative findings were thoroughly explored, and supplemental analyses were conducted to help guide future research and theory refinements. Implications of message framing and other message tailoring strategies for nursing research, education, and practice were also discussed. Although health psychology has generated many interesting approaches for improving people's health through promoting behavioral change, much of the work in this field may be described as basic research (using college students) that is a imed at clarifying constructs and establishing principles. In contrast, much of nursing research
108 may be describ ed as applied research aimed at solving challenging problems using practical means in "real world" settings. Psychology theory coupled with nursing practice is an untapped partnership that has exciting possibilities. Collaboration between nurses and health psychologists will bring about a richer understanding of how to employ a patient centered approach to effectively get health messages across to at risk community populations. This enhanced understanding will better equip nurses and other health care provi ders to design and deliver appropriately tailored health messages in order to optimize promotion of healthy self management behaviors and ultimately contribute to a reduction in the burdensome impact of chronic diseases throughout the world.
109 List of References Abegunde, D. O., Mathers, C. D., Adam, T., Ortegon, M., & Strong, K. (2007). The burden and costs of chronic diseases in low income and middle income countries. Lancet, 370, 19291938. Adams, J., & White, M. (2009). Time perspective in soc ioeconomic inequalities in smoking and body mass index. Health Psychology, 28, 8390. Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp. 1139). Heid elberg: Springer. Ajzen, I. (1988). Attitudes, personality, and behavior. Chicago: Dorsey. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50, 179211. Ajzen, I., & Fishbein, M. (Eds.). (1980). Un derstanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice Hall. American Diabetes Association. ( 2010). Clinical practice recommendations 2010. Diabetes Care, 33 (Suppl 1), S1S100. Apanovitch, A. M., McCarthy, D., & Salovey, P. (2003). Using message framing to motivate HIV testing among low income, ethnic minority women. Health Psychology, 22, 6067. Arora, R., Stoner, C., & Arora, A. (2006). Using framing and credibility to incorporate exercise and fitness in individuals li festyle. Journal of Consumer Marketing, 23, 199207. Astin, F., & Closs, S. J. (2007). Guest editorial: Chronic disease management and self care support for people living with longterm complications: Is the nursing workforce prepared? Journal of Clinical Nursing, 16(7b), 105106. Banks, S. M., Salovey, P., Greener, S., Rothman, A. J., Moyer, A., Beauvais, J., . .Epel, E. (1995). The effects of message framing on mammography utilization. Health Psychology, 14, 178184.
110 Barlow, J., Wright, C., Sheasby, J ., Turner, A., & Hainsworth, J. (2002). Self management approaches for people with chronic conditions: A review. Patient Education and Counseling, 48, 177187. Blanton, H., Stuart, A. E., & VandenEijnden, R. J. J. M. (2001). An introduction to deviance r egulation theory: The effect of behavioral norms on message framing. Personality and Social Psychology Bulletin, 27, 848 858. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitraitmultimethod matrix. Psychologica l Bulletin, 56, 81 105. Canada, R. E., & Turner, B. J. (2007). Talking to patients about screening colonoscopy Where conversations fall short. Journal of Family Practice, 56(8), E1E9. Carver, C. S., Sutton, S. K., & Scheier, M. F. (2000). Action, emot ion, and personality: Emerging conceptual integration. Personality and Social Psychology Bulletin, 26, 741751. Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment : The BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319333. Centers for Disease Control and Prevention (2009a ). Chronic disease prevention and health promotion. Retrieved June 13, 2010 from http://www.cdc.gov/nccdphp/overview.htm Ce nters for Disease Control and Prevention (2009b). Overweight and obesity. Retrieved May 15, 2010 from http://www.cdc.gov/obesity/defining.html Centers for Disease Control and Prevention (2010). Behavioral risk factor surveillance system: Turning inform ation into health. Retrieved May 22, 2010 from http://www.cdc.gov/brfss/index.htm Cherubini, P., Rumiati, R., Rossi, D., Nigro, F., & Calabro, A. (2005). Improving attitudes toward prostate examinations by loss framed appeals. Journal of Applied Social Ps ychology, 35, 732744. Collins, S. (2005). Explanations in consultations: The combined effectiveness of doctors and nurses communication with patients. Medical Education, 39, 785796. Coster, S., & Norman, I. (2009). Cochrane reviews of educational and self management interventions to guide nursing practice: A review. International Journal of Nursing Studies, 46, 508 528.
111 Detweiler, J. B., Bedell, B. T., Salovey, P., Pronin, E., & Rothman, A. J. (1999). Message framing and sunscreen use: Gain framed me ssages motivate beach goers. Health Psychology, 18, 189196. Dillman, D. A., Smyth, J. D., & Christian, L. M. (2009). Internet, mail, and mixedmode surveys: The tailored design method (3rd ed.). Hoboken, NJ: John Wiley & Sons. Fagley, N. S. (1993). A note concerning reflection effects versus framing effects. Psychological Bulletin, 113, 451452. Ferguson, E., & Gallagher, L. (2007). Message framing with respect to decisions about vaccination: The roles of frame valence, frame method and perceived risk. British Journal of Psychology, 98, 667680. Finney, L. J., & Iannotti, R. J. (2002). Message framing and mammography screening: A theorydriven intervention. Behavioral Medicine, 28, 514. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: AddisonWesley. Fuglestad, P. T., Rothman, A. J., & Jeffery, R. W. (2008). Getting there and hanging on: The effect of regulatory focus on performance in smoking and weight loss interventio ns. Health Psychology, 27, S260S270. Gerend, M. A., & Cullen, M. (2008). Effects of message framing and temporal context on college student drinking behavior. Journal of Experimental Social Psychology, 44, 11671173. Gerend, M. A., & Shepherd, J. E. ( 2007). Using message framing to promote acceptance of the human papillomavirus vaccine. Health Psychology, 26, 745752. Gray, J. A. (1982). The neuropsychology of anxiety: An enquiry into the functions of the septo hippocampal system New York: Oxford Uni versity Press. Gray, J. A. (1990). Brain systems that mediate both emotion and cognition. Cognition and Emotion, 4, 269288. Haaga, D. A. F., FriedmanWheeler, D. G., McIntosh, E., & Ahrens, A. H. (2008). Assessment of individual differences in regulat ory focus among cigarette smokers. Journal of Psychopatholog y and Behavior al Assessment, 30, 220228. Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 12801300. Higgins, E. T. (1998). Promotion and prevention: Regulatory focus as a motivational principle. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 30, pp. 146). New York: Academic Press.
112 Higgins, E. T. (1999). Promotion and prevention as a motivation duality: Implications for evaluative processes. In S. Chaiken & Y. Trope (Eds.), Dual process theories in social psychology New York: Guilford Press. Higgins, E. T. (2000). Making a good decision: Value from fit. American Psychologist, 55, 12171230. Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduk, O. N., & Taylor, A. (2001). Achievement orientations from subjective histories of success: Promotion pride versus prevention pride. European Journal of Social Psychology, 31, 323. Hochbaum, G. M. (1958). Public participation in medical screening programs: A sociopsychological study (PHS Publication No. 572). Washington, DC: Government Printing Office. Hsiao, E. T. (2003). Using message framing to promote regular physical activity in collegeage women and men (Doctoral dissertation). R etrieved from ProQuest Dissertations and Theses database. (Publication No. AAT 3059265) Jenner, E. A., Jones, F., Fletcher, B. C., Miller, L., & Scott, G. M. (2005). Hand hygiene posters: Motivators or mixed messages? Journal of Hospital Infection, 60, 218225. Jones, L. W., Sinclair, R. C., & Courneya, K. S. (2003). The effects of source credibility and message framing on exercise intentions, behaviors, and attitudes: An integration of the elaboration likelihood model and prospect theory. Journal of App lied Social Psychology, 33, 179196. Jones, L. W., Sinclair, R. C., Rhodes, R. E., & Courneya, K. S. (2004). Promoting exercise behaviour: An integration of persuasion theories and the theory of planned behavior. British Journal of Health Psychology, 9, 5 05521. Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47, 263291. Kahneman, D., & Tversky, A. (1984). Choice, values, and frames. American Psychologist, 39, 341350. Keeney, R. L., & Raiffa, H. ( 1976). Decisions with multiple objectives: Preferences and value tradeoffs. New York: Wiley. Kees, J. (2007). Consideration of future consequences as a moderator of temporal framing and regulatory focus in a health domain (Doctoral dissertation). Retrieve d from ProQuest Dissertations and Theses database. (Publication No. AAT 3262980)
113 Kim, Y. M., Heerey, M., & Kols, A. (2008). Factors that enable nurse patient communication in a family planning context: A positive deviance study. International Journal of N ursing Studies, 45, 14111421. Kreuter, M. W., Bull, F. C., Clark, E. M., & Oswald, D. L. (1999). Understanding how people process health information: A comparison of tailored and nontailored weight loss materials. Health Psychology, 18, 487 494. Kreuter M. W., & Wray, R. J. (2003). Tailored and targeted health communication: Strategies for enhancing information relevance. American Journal of Health Behavior, 27(Suppl 3), S227S232. Kroll, E. (2005). The effects of message framing and gender on physica l exercise communications to high school students (Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses database. ( Publication N o. AAT 3161768) Khberger, A. (1998). The influence of framing on risky decisions: A meta analysis. Organiz ational Behavior and Human Decision Processes, 75, 2355. Lalor, K. M., & Hailey, B. J. (1990). The effects of message framing and feelings of susceptibility to breast cancer on reported frequency of breast self examination. International Quarterly of Community Health Education, 10, 183192. Latimer, A. E., Brawley, L. R., & Bassett, R. L. (2010). A systematic review of three approaches for constructing physical activity messages: What messages work and what improvements are needed? International Journal of Behavioral Nutrition and Physical Activity. Advance online publication. doi: 10.1186/14795868736 Latimer, A. E., Katulak, N. A., Mowad, L., & Salovey, P. (2005). Motivating cancer prevention and early detection behaviors using psychologically tailor ed messages. Journal of Health Communication, 10, 137155. Latimer, A. E., Rench, T. A., Rivers, S. E., Katulak, N. A., Materese, S. A., Cadmus, L., . Salovey, P. (2008). Promoting participation in physical activity using framed messages: An applicat ion of prospect theory. British Journal of Health Psychology, 13, 659681. Latimer, A. E., Rivers, S. E., Rench, T. A., Katulak, N. A., Hicks, A., Hodorowski, J. K., . Salovey, P. (2008). A field experiment testing the utility of regulatory fit messag es for promoting physical activity. Journal of Experimental Social Psychology, 44, 826832. Latimer, A. E., Salovey, P., & Rothman, A. J. (2007). The effectiveness of gainframed messages for encouraging disease prevention behavior: Is all hope lost? Jour nal of Health Communication, 12, 645649.
114 Latimer, A. E., Williams Piehota, P., Katulak, N. A., Cox, A., Mowad, L., Higgins, E. T., & Salovey, P. (2008). Promoting fruit and vegetable intake through messages tailored to individual differences in regulatory focus. Annals of Behavioral Medicine, 35, 363369. Lauver, D., & Rubin, M. (1990). Message framing, dispositional optimism, and follow up for abnormal Papanicolaou tests. Research in Nursing & Health, 13, 199207. Lee, A. Y., & Aaker, J. L. (2004). Bri nging the frame into focus: The influence of regulatory fit on processing fluency and persuasion. Journal of Personality and Social Psychology, 86, 205218. Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not created equal: A typology and critical analysis of framing effects. Organizational Behavior and Human Decision Processes, 76 (2), 149188. Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability considerations in near and distant future decisions: A test of te mporal construal theory. Journal of Personality and Social Psychology, 75, 518. Lockwood, P., Jordan, C. H., & Kunda, Z. (2002). Motivation by positive or negative role models: Regulatory focus determines who will best inspire us. Journal of Personality and Social Psychology, 83, 854864. Mann, T., Sherman, D., & Updegraff, J. (2004). Dispositional motivations and message framing: A test of the congruency hypothesis in college students. Health Psychology, 23, 330334. Matarazzo, J. D. (1980). Behaviora l health and behavioral medicine: Frontiers for a new health psychology. American Psychologist, 35, 807 817. Matarazzo, J. D. (1982). Behavioral healths challenge to academic, scientific, and professional psychology. American Psychologist, 37, 114. M axwell, S. E., & Delaney, H. D. (1993). Bivariate median splits and spurious statistical significance. Psychological Bulletin, 113, 181190. McCall, L. A., & Martin Ginis, K. A. (2004). The effects of message framing on exercise adherence and health belie fs among patients in a cardiac rehabilitation program. Journal of Applied Biobehavioral Research, 9, 122 135. McElroy, T., & Seta, J. J. (2003). Framing effects: An analytic holistic perspective. Journal of Experimental Social Psychology, 39, 610617.
115 M eyerowitz, B. E., & Chaiken, S. (1987). The effect of message framing on breast self examination attitudes, intentions, and behavior. Journal of Personality and Social Psychology, 52, 500510. Meyers Levy, J., & Maheswaran, D. (2004). Exploring message fr aming outcomes when systematic, heuristic, or both types of processing occur. Journal of Consumer Psychology, 14, 159167. Moorman, M., & van den Putte, B. (2008). The influence of message framing, intention to quit smoking, and nicotine dependence on the persuasiveness of smoking cessation messages. Addictive Behaviors, 33, 12671275. Myers, R. E. (2010). Promoting healthy behaviors: How do we get the message across? International Journal of Nursing Studies, 47, 500 512. Myers, R. E., & Beckstead, J. W. ( 2009). Health psychology and its applications to nursing research, education, and practice Unpublished manuscript, College of Nursing, University of South Florida, Tampa. National Cancer Institute. (1994). Clear and simple: Developing effective prin t materials for low literate readers (NIH Publication No. 953594). Rockville, MD: U.S. Department of Health and Human Services. National Institute of Diabetes and Digestive and Kidney Diseases. (2008). National diabetes statistics, 2007 fact sheet. Retri eved from http://diabetes.niddk.nih.gov/dm/pubs/statistics/index.htm OConnor, D. B., Ferguson, E., & OConnor, R. C. (2005). Intentions to use hormonal male contraception: The role of message framing, attitudes and stress appraisals. British Journal of Psychology, 96, 351 369. OConnor, D. B., Warttig, S., Conner, M., & Lawton, R. (2009). Raising awareness of hypertension risk through a webbased framing intervention: Does consideration of future consequences make a difference? Psychology, Health & Medi cine, 14, 213219. OKeefe, D. J., & Jensen, J. D. (2006). The advantages of compliance or the disadvantages of noncompliance? A meta analytic review of the relative persuasive effectiveness of gain framed and loss framed messages. Communication Yearbook, 30, 143. OKeefe, D. J., & Jensen, J. D. (2007). The relative persuasiveness of gain framed lossframed messages for encouraging disease prevention behaviors: A metaanalytic review. Journal of Health Communication, 12, 623 644.
116 Orbell, S., & Hagger, M (2006). Temporal framing and the decision to take part in type 2 diabetes screening: Effects of individual differences in consideration of future consequences on persuasion. Health Psychology, 25, 537548. Orbell, S., & Kyriakaki, M. (2008). Temporal fr aming and persuasion to adopt preventive health behavior: Moderating effects of individual differences in consideration of future consequences on sunscreen use. Health Psychology, 27, 770779. Orbell, S., Perugini, M., & Rakow, T. (2004). Individual diffe rences in sensitivity to health communications: Consideration of future consequences. Health Psychology, 23, 388396. Pender, N. J. (1982). Health promotion in nursing practice Norwalk, CT: Appleton CenturyCrofts. Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 19, pp. 123205). New York: Academic Press. Rew, L., Fouladi, R. T., & Yockey, R. D. (2002). Sexual health practices of homeless youth. Journal of Nursing Scholarship, 34, 139145. Rivers, S. E., Salovey, P., Pizarro, D. A., Pizarro, J., & Schneider, T. R. (2005). Message framing and Pap test utilization among women attending a community health clinic. Journal of Health Psychology, 10, 6577. Robberson, M., & Rogers, R. W. (1988). Beyond fear appeals: Negative and positive persuasive appeals to health and self esteem. Journal of Applied Social Psychology, 18, 277287. Rodin, J., & Salovey, P. (1989). Health psychology. Annual Review of Psychology, 40, 533579. Rogers, R. W. (1975). A protection motivation theory of fear appeals and attitude change. Journal of Psychology, 91, 93114. Rosenstock, I. M. (1960). What research in motivation suggests for public health. American Journal of Public Health, 50, 295302. Rothman, A. J., Bartels, R. D., Wlaschin, J., & Salovey, P. (2006). The strategic use of gainand loss framed messages to promote healthy behavior: How theory can inform practice. Journal of Communication, 56, S202S220.
117 Rothman, A. J., Martino, S. C., Bedell, B. T., Detweiler, J. B., & Salovey, P. (1999). The systematic influence of gain and loss framed messages on interest in and use of different types of health behavior. Personality and Social Psychology Bulletin, 25, 13551369. Rothman, A. J., & Salovey, P. (1997). Shaping perceptions to motivate healthy behavior: The role of message framing. Psychological Bulletin, 121, 319. Rothman, A. J., Salovey, P., Antone, C., Keough, K., & Martin, C. D. (1993). The influenc e of message framing on intentions to perform health behaviors. Journal of Experimental Social Psychology, 29, 408433. Rothman, A. J., Wlaschin, J. T., Bartels, R. D., Latimer, A., & Salovey, P. (2008). How persons and situations regulate message framing effects: The study of health behavior. In A. J. Elliot (Ed.), Handbook of approach and avoidance motivation (pp. 475486). New York: Psychology Press. Russell, K. M., Champion, V. L., & Perkins, S. M. (2003). Development of cultural belief scales for mammography screening. Oncology Nursing Forum, 30, 633640. Russell, K. M., Perkins, S. M., Zollinger, T. W., & Champion, V. L. (2006). Sociocultural context of mammography screening use. Oncology Nursing Forum, 33, 105112. Sagristano, M. D., Trope, Y., & Liberman, N. (2002). Time dependent gambling: Odds now, money later. Journal of Experimental Psychology: General, 131, 364 376. Salovey, P. (2005). Promoting prevention and detection: Psychologically tailoring and framing messages about health. In R. Bibace, J. D. Laird, K. L. Noller, & J. Valsiner (Eds.), Science and medicine in dialogue: Thinking through particulars and universals (pp. 1742). Westport, CT: Praeger. Sanders, S. A. (1986). Development of a tool to measure subjective time experience. Int ernational Journal of Nursing Studies, 35, 178 182. Schillinger, D., Grumbach, K., Piette, J., Wang, F., Osmond, D., Daher, C., . Bindman, A. B. (2002). Association of health literacy with diabetes outcomes. Journal of the American Medical Associatio n, 288, 475482. Sheeran, P. (2002). Intentionbehavior relations: A conceptual and empirical review. European Review of Social Psychology, 12, 136. Shen, L., & Bigsby, E. (2010). Behavioral activation/inhibition systems and emotions: A test of valence vs. action tendency hypotheses. Communication Monographs, 77, 126.
118 Shen, L, & Dillard, J. P. (2009). Message frames interact with motivational systems to determine depth of message processing. Health Communication, 24, 504514. Sherman, D. K., Mann, T., & Updegraff, J. A. (2006). Approach/avoidance motivation, message framing, and health behavior: Understanding the congruency effect. Motivation and Emotion, 30, 165169. Shin, Y. H., Hur, H. K., Pender, N. J., Jang, H. J., & Kim, M S. (2006). Exercise se lf efficacy, exercise benefits and barriers, and commitment to a plan for exercise among Korean women with osteoporosis and osteoarthritis. International Journal of Nursing Studies, 43, 3 10. Steffen, V. J., Sternberg, L., Teegarden, L. A., & Shepherd, K (1994). Practice and persuasive frame: Effects on beliefs, intention, and performance of a cancer self examination. Journal of Applied Social Psychology, 24, 897925. Strathman, A., Gleicher, F., Boninger, D. S., & Edwards, C. S. (1994). The considerati on of future consequences: Weighing immediate and distant outcomes of behavior. Journal of Personality and Social Psychology, 66, 742752. Strumpf, N. E. (1987). Probing the temporal world of the elderly. International Journal of Nursing Studies, 24, 201 214. Summerville, A., & Roese, N. J. (2008). Self report measures of individual differences in regulatory focus: A cautionary note. Journal of Research in Personality, 42, 247254. Taylor, S. E. (2003). Health psychology (5th ed.). New York: McGraw Hill Toepoel, V. (2010). Is consideration of future consequences a changeable construct? Personality and Individual Differences, 48, 951956. Toll, B. A., Salovey, P., OMalley, S. S., Mazure, C. M., Latimer, A., & McKee, S. A. (2008). Message framing for s moking cessation: The interaction of risk perceptions and gender. Nicotine & Tobacco Research, 10, 195200. Trope, Y., & Liberman, N. (2000). Temporal construal and time dependent changes in preference. Journal of Personality and Social Psychology, 79, 876889. Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211, 453458. Tykocinski, O., Higgins, E. T., & Chaiken, S. (1994). Message framing, self discrepancies, and yielding to persuasive messages: The m otivational significance of psychological situations. Personality and Social Psychology Bulletin, 20, 107115.
119 Updegraff, J. A., Sherman, D. K., Luyster, F. S., & Mann, T. L. (2007). The effects of message quality and congruency on perceptions of tailored health communications. Journal of Experimental Social Psychology, 43, 249257. U.S. Department of Health and Human Services. (1996). Physical activity and health: A report of the Surgeon General Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention (CDC), National Center for Chronic Disease Prevention and Health Promotion. Retrieved from http://www.cdc.gov/nccdphp/sgr/sgr.htm U.S. Department of Health and Human Services. (2000a). Diabetes. In Healthy Peopl e 2010, Vol. 1. Washington, DC: U.S. Government Printing Office. Retrieved from http://www.healthypeople.gov/document/pdf/Volume1/05Diabetes.pdf U.S. Department of Health and Human Services. (2000b). Health communication. In Healthy People 2010, Vol. 1. W ashington, DC: U.S. Government Printing Office. Retrieved from http://www.healthypeople.gov/document/pdf/Volume1/11HealthCom.pdf U.S. Department of Health and Human Services. (2000c). Physical activity and fitness. In Healthy People 2010, Vol. 2. Washingt on, DC: U.S. Government Printing Office. Retrieved from http://www.healthypeople.gov/document/pdf/Volume2/22Physical.pdf U.S. Department of Health and Human Services. (2008). 2008 physical activity guidelines for Americans. Retrieved from http://www.health.gov/paguidelines Uskul, A. K., Sherman, D. K., & Fitzgibbon, J. (2009). The cultural congruency effect: Culture, regulatory focus, and the effectiveness of gainvs. loss framed health messages. Journal of Experimental Social Psychology, 45, 535 541. van't Riet, J., Ruiter, R. A. C., Werrij, M. Q., & de Vries, H. (2010). Health Education Research, 25, 343354. von Neumann, J., & Morgenstern, O. (1944). Theory of games and economic behavior. Princeton: Princeton University Press. Wilson, D. K., Purdon, S. E., & Wallston, K. A. (1988). Compliance to health recommendations: A theoretical overview of message framing. Health Education Research, 3 (2), 161171. Wolf, J. B. (2007). Is breast really best? Risk and total motherhood in the national breastfeeding awareness campaign. Journal of Health Politics, Policy and Law, 32, 595636.
120 World Health Organization. (2005a). Preparing a health care workforce for the 21st century: The challenge of chronic conditions. Geneva: Author. World Health Organization. (2005b). Preventing chronic diseases: A vital investment. Geneva: Author. Yi, S., & Baumgartner, H. (2009). Regulatory focus and message framing: A test of three accounts. Motivation and Emotion, 33, 435443. Zimbardo, P. G., & Boyd, J. N. (1999). Putting tim e in perspective: A valid, reliable individual differences metric. Journal of Personality and Social Psychology, 77, 12711288.
122 Appendix A: GainFramed/Distal Framed Health Message Physical Activity Can Improve Your Health in the Years to Come Regular physical activity can make it easier to achieve good health in the future !! Researchers have found that physical activities like golfing, walking, gardening, bicycling, dancing, swimming, and skiing are an important part of a healthy lifestyle for people with diabetes. For example, just 15 minutes of moderate walking can lower your blood sugar. Regular physical activity can also help you avoid expensive medical treatm ents by preventing problems such as heart disease, stroke, bone loss, and some cancers. Health care providers agree that most adults should do physical activity at least 30 minutes a day 5 or more days a week.* By doing regular physical activity, you m ay gain these benefits in the years to come: Brighter mood and more positive attitude Greater self esteem, pride, and confidence More energy, better sleep, better sex life, and less stress Increased burning of extra calories to help keep a healthy weight Improved blood pressure, blood cholesterol, and circulation Less need for diabetes medication due to better blood sugars Stronger bones, more strength, and increased flexibility and balance Reduced arthritis pain *Always check with your health care provide r before starting a physical activity to make sure it is safe for you. Want to achieve good health in the future? . Be active!
123 Appendix B : Gain Framed/Immediate Framed Health Message Physical Acti vity Can Improve Your Health Immediately Regular physical activity can make it easier to achieve good health now !! Researchers have found that physical activities like golfing, walking, gardening, bicycling, dancing, swimming, and skiing are an important part of a healthy lifestyle for people with diabetes. For example, just 15 minutes of moderate walking can lower your blood sugar. Regular physical activity can also help you avoid expensive medical treatments by preventing problems such as heart disease, stroke, bone loss, and some cancers. Health care providers agree that most adults should do physical activity at least 30 minutes a day 5 or more days a week.* By doing regular physical activity, you may gain these benefits immediately (within 1 2 we eks): Brighter mood and more positive attitude Greater self esteem, pride, and confidence More energy, better sleep, better sex life, and less stress Increased burning of extra calories to help keep a healthy weight Improved blood pressure, blood cholest erol, and circulation Less need for diabetes medication due to better blood sugars Stronger bones, more strength, and increased flexibility and balance Reduced arthritis pain *Always check with your health care provider before starting a physical activity to make sure it is safe for you. Want to achieve immediate good health? . Be active!
124 Appendix C: Loss Framed/ Distal Framed Health Message Lack of Physical Activity Can Jeopardize Your Health in the Years to Come Lack of r egular physical activity can make it harder to achieve good health in the future !! Researchers have found that physical activities like golfing, walking, gardening, bicycling, dancing, swimming, and skiing are an important part of a healt hy lifestyle for people with diabetes. For example, just 15 minutes of moderate walking can lower your blood sugar. Regular physical activity can also help you avoid expensive medical treatments by preventing problems such as heart disease, stroke, bone loss, and some cancers. Health care providers agree that most adults should do physical activity at least 30 minutes a day 5 or more days a week.* By not doing regular physical activity, you may lose benefits in the years to come by: Missing out on a brighter mood and more positive attitude Missing out on greater self esteem, pride, and confidence Missing out on more energy, better sleep, better sex life, and less stress Missing out on increased burning of extra calories to help keep a healthy weight M issing out on improved blood pressure, blood cholesterol, and circulation Missing out on less need for diabetes medication due to better blood sugars Missing out on stronger bones, more strength, and increased flexibility and balance Missing out on reduced arthritis pain *Always check with your health care provider before starting a physical activity to make sure it is safe for you. Want to miss out on good health in the future? . Dont be active!
125 Appendix D: Loss Framed/Immediate Framed Health Message Lack of Physical Activity Can Jeopardize Your Health Immediately Lack of r egular physical activity can make it harder to achieve good health now !! Researchers have found that physical activities like golfi ng, walking, gardening, bicycling, dancing, swimming, and skiing are an important part of a healthy lifestyle for people with diabetes. For example, just 15 minutes of moderate walking can lower your blood sugar. Regular physical activity can also help you avoid expensive medical treatments by preventing problems such as heart disease, stroke, bone loss, and some cancers. Health care providers agree that most adults should do physical activity at least 30 minutes a day 5 or more days a week.* By not doing regular physical activity, you may lose benefits immediately (within 1 2 weeks) by: Missing out on a brighter mood and more positive attitude Missing out on greater self esteem, pride, and confidence Missing out on more energy, better sleep, better sex life, and less stress Missing out on increased burning of extra calories to help keep a healthy weight Missing out on improved blood pressure, blood cholesterol, and circulation Missing out on less need for diabetes medication due to better blood sugar s Missing out on stronger bones, more strength, and increased flexibility and balance Missing out on reduced arthritis pain *Always check with your health care provider before starting a physical activity to make sure it is safe for you. Want to miss out on immediate good health? . Dont be active!
126 Appendix E: Pilot Study Questionnaire Please read each statement carefully and use the rating scale below to choose a numbered response that best describes your o pinion. For example, if you strongly agreed with a particular statement you would indicate so by choosing the number +3. If you slightly disagreed with a particular statement you would indicate so by choosing the number 1. If you neither agree, nor dis agree, with a particular statement you would indicate so by choosing the number 0. -3 -2 1 0 +1 +2 +3 strongly disagree slightly neither agree slightly agree strongly disagree disagree nor disagree agree agree Please write your responses in the blank spaces provided. Be sure to use positive numbers if you agree; use negative numbers if you di sagree. ____ 1. The message was very believable. ____ 2. The message emphasized things that can happen in the future. ____ 3. The message emphasized things that can happen immediately. ____ 4. The message emphasized the benefits of being physicall y active. ____ 5. The message emphasized the risks of not being physically active. For the following question, please write your answer on the blank lines provided. If you need more space, feel free to attach an additional page. What are your thought s about the graphics on the message you just read? (For example: Do they seem appropriate? Are they distracting? ). Do you have any suggestions for changing the graphics? If yes, please describe those changes. ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________ ________________________________________________________________________
127 Appendix F: Consideration of Future Consequences Scale For each of the statements below, please indicate whether or not the statement is characteristic of you. If the statement i s extremely un characteristic of you (not at all like you) please write a 1 to the left of the question; if the statement is extremely characteristic of you (very much like you) please write a 5 next to the question. And, of course, use the numbers in the middle if you fall between the extremes. Please keep the following scale in mind as you rate each of the statements below. 1 extremely uncharacteristic 2 somewhat uncharacteristic 3 uncertain 4 somewhat characteristic 5 extremely characteristic ____ 1. I consider how things might be in the future, and try to influence those things with my day to day behavior. ____2. Often I engage in a particular behavior in order to achieve outcomes that may not result for many years. ____3. I only act to satisfy immediate concerns, figuring the future will take care of itself. ____4. My behavior is only influenced by the immediate (i.e., a matter of days or weeks) outcomes of my actions. ____5. My convenience is a big factor in the decisions I make or the actions I take. ____6. I am willing to sacrifice my immediate happiness or well being in order to achieve future outcomes. ____7. I think it is important to take warnings about negative outcomes seriously even if the negative outcome will not occur for many ye ars. ____8. I think it is more important to perform a behavior with important distant consequences than a behavior with less important immediate consequences. ____9. I generally ignore warnings about possible future problems because I think the problems will be resolved before they reach crisis level. ____10. I think that sacrificing now is usually unnecessary since future outcomes can be dealt with at a later time. ____11. I only act to satisfy immediate concerns, figuring that I will take care of future problems that may occur at a later date. ____12. Since my day to day work has specific outcomes, it is more important to me than behavior that has distant outcomes.
128 Appendix G: BIS/BAS Scale For each of the statements below, please indicate whether or not you agree with the statement. If you strongly agree with the statement please write a 1 to the left of the question; if you strongly disagree with the statement please write a next to the question. And, of course, use the numbers in the middle if you fall between the extremes. Please keep the following scale in mind as you rate each of the statements below. 1 strongly agree 2 agree 3 disagree 4 strongly disagree ____1. If I think something unpleasant is going to happen I usually get pr etty worked up. ____2. I worry about making mistakes. ____3. Criticism or scolding hurts me quite a bit. ____4. I feel pretty worried or upset when I think or know somebody is angry at me. ____5. Even if something bad is about to happen to me, I rarely experience fear or nervousness. ____6. I feel worried when I think I have done poorly at something. ____7. I have very few fears compared to my friends. ____8. When I get something I want, I feel excited and energized. ____9. When Im doing well at something, I love to keep at it. ____10. When good things happen to me, it affects me strongly. ____11. It would excite me to win a contest. ____12. When I see an opportunity for something I like, I get excited right away. ____13. When I want something, I usually go all out to get it. ____14. I go out of my way to get things I want. ____15. If I see a chance to get something I want, I move on it right away. ____16. When I go after something I use a no holds barred approach. ____17. I will often do things for no other reason than that they might be fun. ____18. I crave excitement and new sensations. ____19. Im always willing to try something new if I think it will be fun. ____20. I often act on the spur of the moment.
129 Appendix H: Regul atory Focus Questionnaire Scale This set of questions asks you about specific events in your life. Please indicate your answer to each question by circling the appropriate number below it. 1. Compared to most people, are you typically unable to get what yo u want out of life? 1 never or seldom 2 3 sometimes 4 5 very often 2. Growing up, would you ever cross the line by doing things that your parents would not tolerate? 1 never or seldom 2 3 sometimes 4 5 very often 3. How often have you accomplished things t hat got you psyched to work even harder? 1 never or seldom 2 3 a few times 4 5 many times 4. Did you get on your parents nerves often when you were growing up? 1 never or seldom 2 3 sometimes 4 5 very often 5. How often did you obey rules and regulations that were established by your parents? 1 never or seldom 2 3 sometimes 4 5 always 6. Growing up, did you ever act in ways that your parents thought were objectionable? 1 never or seldom 2 3 sometimes 4 5 very often 7. Do you often do well at different things that you try? 1 never or seldom 2 3 sometimes 4 5 very often 8. Not being careful enough has gotten me into trouble at times. 1 never or seldom 2 3 sometimes 4 5 very often 9. When it comes to achieving things that are important to me, I find that I dont perform as well as I ideally would like to do. 1 never true 2 3 sometimes true 4 5 very often true 10. I feel like I have made progress toward being successful in my life. 1 certainly false 2 3 4 5 certainly true 11. I have found very few hobbies or activiti es in my life that capture my interest or motivate me to put effort into them. 1 certainly false 2 3 4 5 certainly true
130 Appendix I: Intentions and Evaluation of the Health Message Please read each statement carefully and use the rating scale below to choose a numbered response that best describes your opinion. For example, if you strongly agreed with a particular statement you would indicate so by choosing the number +3. If you slightly disagreed with a particular statement you would indicate so b y choosing the number -1. If you neither agree, nor disagree, with a particular statement you would indicate so by choosing the number 0. -3 -2 1 0 +1 +2 +3 strongly disagree slightly neither agree slightly agr ee strongly disagree disagree nor disagree agree agree Please write your responses in the blank spaces provided. Be sure to use positive numbers if you agree; use negative numbers if you disagree. ____ 1. After reading the message, I have no intention of increasing the amount of my physical activity over the next 4 weeks. ____ 2. I am more likely to do physical activity more often over the next 4 weeks t han I currently do based on the content in the message I just read. ____ 3. After reading the message, I am more likely to do more physical activity over the next 4 weeks than I currently do. ____ 4. I have no intention of increasing the amount of m y physical activity over the next 4 weeks based on the content in the message I just read. ____ 5. After reading the message, I plan to do more physical activity over the next 4 weeks than I currently do. ____ 6. I am less likely to do physical activity more often over the next 4 weeks than I currently do based on the content in the message I just read. ____ 7. I plan to do physical activity more often over the next 4 weeks than I currently do based on the content in the message I just read. ____ 8. After reading the message, I am less likely to do more physical activity over the next 4 weeks than I currently do. ____ 9. The message was very believable. ____10. The message emphasized things that can happen in the future. ____11. The message emphasized things that can happen immediately. ____12. The message emphasized the benefits of being physically active. ____13. The message emphasized the risks of not being physically active.
131 Appendix J: Demographic and Other Variables Below are so me questions which are asked for background and descriptive purposes only. Please write or mark your answers to the following questions in the spaces provided. 1. Your Age: ______ 2. Gender (check one) male____ female ____ 3. Race/ethn icity (check one) ____ White/Caucasian ____ Black/African American ____ Hispanic/Chicano/Latino ____ Asian or Pacific Islander/Asian American ____ American Indian or Alaskan Native ____ Other (specify) _________________ 4. Number of years you have had type 2 diabetes: _______ years 5. Current amount of physical activity: a. How many days a week do you usually do physical activity (walking, biking, golfing, gardening, etc.) ? _____ days b. How many minutes of physical activity do you usually do on each of these days? _____ min. 6. How tall are you? ______ feet, ______ inches 7. How much do you weigh? ______ pounds
132 Appendix K: Pre notice Letter (1st Mailing) (Date) Dear Researchers at the University of South Florida (USF) study many topic s. Among them, how best to design health care messages. T o do this, we need the help of people who agree to take part in research. A few days from now you will receive in the mail a packet that includes a health care message about physical activity and diabetes, along with a brief questionnaire. You are being asked to take part in this important study because you are an adult with type 2 diabetes who has visited Anthony Morrison, MD or Nancy Grove, ARNP at the USF Medical Clinic. This research is being conducted as part of my dissertation in the USF College of Nursing. Dr. Morrison and Ms. Grove are helping me with this study. I am writing in advance because we have found many people like to know ahead of time that they will be receiving a questionn aire in the mail. The study is an important one that will help health care providers design useful education messages to encourage patients to do healthy behaviors like physical activity. Thank you for your time and consideration. It is only with the generous help of people like you that our research can be successful. Sincerely, Rachel E. Myers, RN, MSN, CDE USF College of Nursing P.S. We will be enclosing a small token of appreciation with the questionnaire as a way of saying thank you in advanc e for your help.
133 Appendix L: Cover Letter for Initial Booklet (2nd Mailing) (Date) Dear A few days ago, you should have received a letter inviting you to participate in a University of South Florida (USF) research study about physical activity and diabetes health care messages. We are asking you to take part in this study because you are an adult with type 2 diabetes who has visited Anthony Morrison, MD or Nancy Grove, ARNP at the USF Medical Clinic. This study will help health care providers find out the best way to design useful education messages to encourage patients to do healthy behaviors like physical activity. The enclosed packet contains a health care message and a brief questionnaire. If you choose to participate in this study, plea se read the enclosed information, fill out the questionnaire, and mail it back to us in the self -addressed stamped envelope provided. It should take you about 20 minutes to complete the questionnaire, which contains three parts. PART 1 has a few question s about your background (your age, number of years youve had diabetes, etc.). In PART 2, you will read a message about physical activity and diabetes and then answer a few questions about the message. PART 3 contains questions about your thoughts, feeli ngs, and actions. Your responses are very important to us. Please answer all questions, even if you have to guess. When you are finished, please mail the questionnaire back to us in the stamped envelope provided. This will end your part in the study. There are no direct benefits to you for taking part in this study. There are no known risks to you should you choose to participate. Whether or not you participate in this study will in no way affect the care that the USF Medical Clinic provides to y ou. The information you provide will be combined with responses from about 200 other volunteers. Please DO NOT put your name anywhere on the questionnaire. Your participation in thi s study is strictly voluntary and anonymous. If for some reason you prefer not to take part, please let us know by returning the blank questionnaire in the stamped envelope provided. We have enclosed a small token of appreciation as a way of saying thank you in advance for your help. If you have any questions, concerns or complaints about this study, call Rachel Myers at __________ If you have questions about your rights as a participant in this study, general questions, or have complaints, concer ns or issues you want to discuss with someone outside the research, call the Division of Research Integrity and Compliance of the USF at (813) 974 -9343. Thank you very much for helping us with this important study. Sincerely, Rachel E. Myers, RN, MSN, CDE USF College of Nursing P.S. We ask that you return the questionnaire within 1 week so we can make sure your responses are included in our research.
134 Appendix M: Reminder/Thank You Postcard (3rd Mailing) (Date) A few days ago, you should have received an invitation from me to participate in a research study by reading a message about physical activity and completing a questionnaire. It was sent to you as part of a study to help health care providers design useful education messages to encourage pa tients to do healthy behaviors like physical activity. If you have already returned the questionnaire, please accept my sincere thanks. If you have not yet completed and returned the questionnaire, please do so within one week so we can make sure your responses are included in our research. Your participation is important to the success of our study. Thank you very much for volunteering to help us. Sincerely, Rachel E. Myers, RN, MSN, CDE University of South Florida, College of Nursing
135 Appendix N: Cover Letter for Replacement Booklet (4th Mailing) (Date) Dear About four weeks ago, I sent you information about participating in a research study. The information included a brief questionnaire. To the best of our knowledge, the questionnai re has not yet been returned. We think the results of this study are going to be very useful to health care providers who design education messages to encourage patients to do healthy behaviors like physical activity. We are writing again because of the importance that your responses have for helping us get accurate results. Although we sent questionnaires to a lot of people with type 2 diabetes, its only by hearing from nearly everyone in the sample that we can be sure the results are truly representative. We hope that you will volunteer to fill out and return the questionnaire soon, but if for any reason you prefer not to take part in this study, please let us know by returning a note or blank questionnaire in the enclosed stamped envelope. Sincerel y, Rachel E. Myers, RN, MSN, CDE USF College of Nursing P.S. If you have any questions, please feel free to contact me. The number where I can be reached is ____________.
About the Author Rachel E. Myers received a Bachelor of Science in Nursing degree from Florida State University, Tallahassee, Florida in 1990 and a Master of Science in Nursing degree from George Mason University, Fairfax, Virginia in 1999. She also received a Nursing Education post graduate certificate from George Mason University in 2000. She has been a Certified Diabetes Educator since 1993. She has worked as a patient and staff educator and a clinical staff nurse in medical surgical, diabetes, womens health, and mental health settings. In addition, she has held several leadership positions such as Director of Nursing and Director of Performance Improvement and Research. She has been a nursing consultant since 2001 in facilities for individuals with intellectual disabilities and mental illness. She was the recipi ent of a Tampa Bay Organization of Nurse Executives scholarship in 2008. After receiving her Ph.D., Dr. Myers continued her work as a consultant and obtained a nursing faculty position.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004619
Moderating the effectiveness of messages to promote physical activity in type 2 diabetes
h [electronic resource] /
by Rachel Myers.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Dissertation (PHD)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: The world is experiencing a rapid rise in chronic health problems, which places an enormous burden on health care services. Modifiable health behaviors such as physical inactivity are largely responsible for this high prevalence and incidence of chronic diseases. Message tailoring is a well-established approach for constructing health communication and has been shown to increase the persuasiveness of messages in the promotion of healthy behaviors. Message framing is an effective strategy that has been well-studied in psychology over the past 20-plus years across a breadth of health-related behaviors but has received little attention in the nursing research literature. Based on prospect theory, temporal construal theory, and motivational orientation theories, the present study examined how two individual differences factors consideration of future consequences (CFC) and motivational orientation combine to moderate temporal proximity and valence framing effects on intentions to increase physical activity. A mail survey was conducted using Dillman's Tailored Design Method. Two hundred and eighteen adults with type 2 diabetes were randomly assigned to receive one of four versions of a health message aimed to increase regular physical activity. Messages were framed using a 2 (immediate- vs. distal-framed) x 2 (gain- vs. loss-framed) design. After reading the message, participants rated their intention to increase physical activity. They also completed a measure of CFC and two measures of motivational orientation. Participants who read a message with a temporal proximity or valence frame congruent with their CFC or motivational orientation, respectively, did not show greater intentions to increase physical activity when compared to those who read a health message that was incongruent with these individual differences. Plausible explanations for these negative results are considered. Several interesting findings emerged from supplemental analyses. For instance, participants who perceived the health message as more believable tended to have greater intentions to increase physical activity. Suggestions for future research applying message congruence to promote complex health behaviors in at-risk populations are given. Implications of message framing and other message tailoring strategies for nursing research, education, and practice are discussed.
Advisor: Jason W. Beckstead, Ph.D.
t USF Electronic Theses and Dissertations.