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-SFE0004545
Examining the relationship between work-to-family conflict and parenting behavior
h [electronic resource] /
by Eunae Cho.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (MA)--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: Although work-family conflict (WFC) has been of particular interest to work-family researchers, little attention has been paid to the consequences of WFC that reside in the family domain. Research on WFC and child outcomes is especially scant. The current study addresses the gap in the literature by investigating the relationship between work-interfere-with-family (WIF) and three forms of parent-child interaction behavior (PB): physical and recreational PB (PRPB), cognitive and academic-oriented PB (CAPB), and passive and maintenance-oriented PB (PMPB). The mechanism by which WIF relates to PB was further investigated by examining negative emotion as a mediator and trait guilt as a moderator of the relationship. Employed parents of early school-aged children (n = 201) participated in the survey. Results indicated that both time- and strain-based WIF were negatively related to two types of active PB, PRPB and CAPB. However, negative emotion did not mediate the relationship between WIF and PB. With regard to the moderating role of trait guilt, support was found for PRPB. Theoretical and practical implications, as well as future directions, are discussed.
Advisor: Tammy D. Allen, Ph. D.
Parent-child interaction behavior
t USF Electronic Theses and Dissertations.
Examining the Relationship between Work to Family Conflict and Parenting Behavior by Eunae Cho A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sc iences University of South Florida Major Professor: Tammy D. Allen, Ph.D. Vicky Phares, Ph.D. Russell E. Johnson Ph.D. Date of Approval: June 08 20 1 0 Keywords: work nonwork conflict, employed parent, parent child interaction behavior, negati ve emotions, trait guilt Copyright 2010, Eunae Cho
Acknowledgements I would like to acknowledge the members of my thesis committee, for their time, guidance and support I would like to specially acknowledge my major advisor, Dr. Tammy Allen, for her patience, encouragement, and support that led me to learn and develop throughout this project. Finally, I appreciate time and willingness to help of all the peop le who participated in my study and those who assisted me in the participant recruitmen t
i Table of Contents List of Tables i i i List of Figures iv Abstract v Chapter One Introduction 1 Work Family Conflict 3 Parent al Time and Parent C hild Interaction Behavior 4 WIF and Parent Child Interaction Behavior 5 Negative Emotions as a Me diator 6 Trait Guilt as a Moderator 11 Chapter Two Method 14 Participants and Procedure 14 Measures 16 Work to family conflict 16 Negative emotion 16 Trait guilt 16 Parent c hild i nteraction behavior 17 Control variables 18 Chapter Three Results 19 Pilot Study 19 Scale development 19 Participants and Procedure 20 Measures 21 Results and discussion 22 Factor Analysis 25 Descriptive statistics 27 Control variables 27 Preliminary Analysis 29 Hypotheses Testing 29 Mediator hypo thesis 36 Moderator hypothesis 37 Exploratory Analysis 42
ii Chapter Four Discussion 47 Main Findings 49 WIF and parent child interaction behavior 49 Negative emotion 50 Trait guilt 51 Theoretical Implications 52 Practical Implications 54 Limit ations 55 Future Directions 56 Conclusion 57 References 58 Appendices 66 Appendix A: WIF Scale Items 67 Appendix B: Negative Emotion Scale Items 68 Appendix C: Trait Negative Affectivity Scale Items 69 Appendix D: Trait Guilt Scale Items 70 Appen dix E: Time with Children 71 About the Author End Page
iii List of Tables Table 1 Descriptive statistics of study variables 1 5 Table 2. Parent child interaction behavior measure used in pilot study 20 Table 3. Descriptive statistics in pilot study 2 2 Table 4. Interc orrelations among variables in pilot study 24 Table 5. Standardized factor loadings from exploratory factor analysis 26 Table 6. Results of confirmatory factor analysis 26 Table 7. Interc orrelations among study variables 28 Table 8. Regressi on of time based WIF on parent child interaction behavior 31 Table 9. Regression of strain based WIF on parent child interaction behavior 32 Table 10. Regression of time based WIF on negative emotion 33 Table 11. Regression of strain based WIF on negative emotion 34 Table 12. Regression of WIF ts on parent child interaction behavior 35 Table 13. Regression of WIF ts on negative emotion 36 Table 14. Regression results for mediation with time based WIF as predictor 38 Table 15. Regression results for mediation with strain based WIF as predictor 38 Table 16. Regression results for mediation with WIF ts as predictor 39 Table 17. Moderated regression results of trait guilt and negative emotion on PB 40 Table 18 Moderated regression results of trait guilt and WIFts on PB 43 Table 19. Analysis of simple effects in moderated m ed iation m od el 46
iv List of Figures Figure 1. Hypothesized relationships 1 3 Figure 2. Interaction of negative emotion on PRPB as a function of trait guilt 41 Figure 3. Interaction of WIFts on PRPB as a function of trait guilt 4 2 Figure 4. Interaction of WIFts on CA PB as a function of trait guilt 44 Figure 5. Mediated m oderation m odels for l ow and h igh t rait g uilt 45
v Examining the Relationship between Work to Family Conflict and Parenting B ehavior Eunae Cho Abstract Although work family conflict (WFC) has been of particular interest to work family researchers, little attention has been paid to the consequences of WFC that reside in the family domain. Research on WFC and child outcomes is esp ecially scant. The current study addresses the gap in the literature by investigating the relationship between work interfere with family (WIF) and three forms of parent child interaction behavior (PB): physical and recreational PB (PRPB), cognitive and ac ademic oriented PB (CAPB), and passive and maintenance oriented PB (PMPB). The mechanism by which WIF relates to PB was further investigated by examining negative emotion as a mediator and trait guilt as a moderator of the relationship. Employed parents of early school aged children ( n = 201) participated in the survey. Results indicated that both time and strain based WIF were negatively related to two types of active PB, PRPB and CAPB. However, negative emotion did not mediate the relationship between WI F and PB. With regard to the moderating role of trait guilt, support was found for PRPB. Theoretical and practical implications, as well as future directions, are discussed.
1 Chapter One Introduction One prevail ing belief about employed parents is that it is difficult, if not impossible, for them to be as attentive parents as those who are not employed. Parental responsibility increases overall workload for individuals and creates more opportunities for work family conflict (WFC), which occurs when deman ds from the work and the family domains are incompatible ( Greenhaus & Beutell, 1985 ). Due to its pervasiveness and the negative impact on individuals, WFC has been of particular interest to work family researchers. Given the considerable amount of research conducted on WFC it is surprising that only minimal attention has been paid to the consequences of WFC that reside in the family domain ( Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005). Such research is vital in the sense that family is a basic buildin g block of any given community. Furthermore, little research has focused on the relationship between WFC and child outcomes. This is important in that it has been suggested that children are hidden stakeholders within the workplace ( Major, Allard, & Carden as, 2004). From both theoretical and practical perspectives, research examining how via its effects on parents world views, oppor tunities and constraints, and daily experiences (Crouter & McHale, 2005). In
2 and behavior have been associated with child outcomes (Sallinen, Ronka, Kinnunen, & Kok ka 2007 ; Nock & Kingston, 1988). Research regarding children has practical merit as well. Major et al. (2004) suggested child health affect s employees work lives as well as the organization because child illnesses result in increased employee absenteeis m and insurance claims being will benefit not only as theoretical and practical benefits highlight the necessity of investigating th e impact of parents spend their time with children within the family domain is needed. As a finite resource, time has been considered a main source of WFC (Greenhaus & Beutell, 1985). Also, time is a valuable resource in the family domain, as parental time is considered as a major form of investment in children (Gauthier, Smeeding, & Furstenberg, 2004). However, it is not just the raw a mount of time that need s consideration because t he type of parental behaviors and the quality of parental child interaction are suggested to have strong influences on child outcomes ( Grossman, Pollack, & Golding, 1988; Booth, Alison Clarke Stewart, Vandell McCartney, & Owen, 2002 ).With this in mind, the first objective of the present study is to examine the relationship between WFC and parent child interaction behavior. In doing so, the study attempts to answer to the call for further research on family ou tcomes and actual behaviors related to work family conflict (Eby et al., 2005; Kossek & Ozeki, 1998 ).
3 The second objective of the present study is to investigate the process by which WFC and parental behaviors relate. Specifically, the present study exami nes emotions as a linking mechanism. Given that previous research suggests that emotions play an important role in human decision making and subsequent behavior (Frijda, 1988), it is likely that emotions generated from WFC may explain variability in parent behavior at home. As the lack of knowledge about affective experience has repeatedly been pointed out as a critical gap in the work family literature (MacDermid, Seery, & Weiss, 2002), the current study address es this missing link in the literature by e xamining the role of relevant emotions in the context of WFC Finally, the present study examines an individual difference variable that may alter the strength of the work emotion parental behavior link. Specifically, the moderating role of trait guilt is examined. This fills an important gap in the literature, as the importance of guilt is often discussed theoretically in work family contexts ( e.g., Judge, Ilies, & Scott, 2006; Allen, in press ), but has rarely been a topic of empirical study. In following sections, the relationship between key concepts of the study (i.e. work family conflict, negative emotions, and parent child interaction behaviors) will be discussed, followed by a discussion of trait guilt and its application to the current context. Wor k F amily Conflict Work family conflict (WFC) occurs when an individual has simultaneous requirements in multiple roles that are mutually incompatible (Greenhaus & Beutell, 1985). WFC is bidirectional: work can interfere with family (WIF) and family can in terfere with work (FIW). As Frone (2003) noted, the antecedents of the conflict tend to be found in the role in which the conflict originated, whereas the consequences tend to
4 reside in the role that is hindered. Given that parent child interaction behavio r is the outcome of current interest, WIF will be addressed in the present study. There are three forms of WFC (Greenhaus & Beutell, 1985). Time based conflict occurs when time devoted to one role hampers the fulfillment of responsibilities in another role In strain based conflict stress generated in one role inhibits effective performance in another role. Behavior based conflict occurs when an effective behavior in one role reduces effectiveness in another role. Among the three, time and strain based co nflicts have drawn the most extensive attention (Allen, in press). The emphasis placed on these two forms aligns with the scarcity hypothesis (Marks, 1977), which states that it is inevitable to experience resource shortage as a result of participation in multiple roles because time and energy are limited resources. As WIF is defined as a goal conflict due to finite resource, the focus of the current study will be on time and strain based WIF. Parent al Time and Parent C hild Interaction Behavior Parental ti me refers to the time that parents devote to their children, and it is thought to be a significant indicator of investment in children (Nock & Kingston, 1988). Previous research has established the positive influence of parental time on child development. For example, children in families that spend more time together and participate in more family activities have greater academic achievement and fewer behavior problems (Hofferth & Sandberg, 2001 ). Duncan, Duncan, and Strycker ( 2000 ) also reported a negativ e relationship between time adolescents spend with their families and adolescent problem behavior Beyond time parents spen d with children previous research has also found beneficial effect s of various activities that parents and children share including eating meals, talking, and leisure activities ( Gauthier et al. 2004;
5 Cooksey & Fondell, 1996 ) As the activities may di ffer in terms of the quality and importance distinguishing the types of parental behaviors and activities seems pivotal in order to un derstand their differential impacts on the child outcome. With this in mind, the present study focuses on a specific type of parenting behavior : parent child interaction behavior. Parent child interaction behavior (PB) is defined as an activity that requi res a parent to engage in either a cognitive process or a physical interaction with direct attention to the child. It includes a variety of activities such as those that are academic (e.g., h elping with homework ) to those that are primarily recreational (e .g., playing games together). PB is known to be crucial in child development: academic activities parental encouragement, activities, and interest at home posi tively influence achievement (Epstein, 198 5); r ecreational activities and parent child play serve emotional, communicative, social and cultural functions (Tamis LeMonda, Uzgiris, & Bornstein, 2002). WIF and Parent Child Interaction Behavior WIF describes a situation in which the demands in the wor k domain interfere with successful performance in the family domain (Greenhaus, Allen, & Spector, 2006). Therefore, the degree of WIF should be negatively associated with the PB. T ime is a major resource for fulfilling goals and a common source of WF C (Gre enhaus & Beutell, 1985). To the extent that parents jobs require them to devote time to their work, there may be less frequent interaction between parents and children. T he level of t ime based WIF may prescribe the kind of activities in which parents are able to and willing to participate. For instance, there may be less frequent family outings if parents have to
6 work during the weekends. Experiencing t ime based WIF may also determine the degree of engagement of a parent in the interaction behaviors such t hat a parent may only be able to supervise her child playing rather than actively interact with the child while she is working on some work assignments to be done by the next day. Therefore, negative relationship between time based WIF and PB is likely. St rain has also been recognized as major source of WFC (Greenhaus & Beutell, 1985). Studies have emphas ized the unique contribution of strain based conflict to individuals behavior at home. I l ies, Schw ind, Johnson, Wagner, DeRue, and Ilgen ( 2007 ) reported t hat individuals who experience d high levels of WIF tended to reduce social interaction with their families, even when the amount of time spent at h ome was controlled As active participation in activities with children requires more energy, parents may wit hdraw from interaction behaviors with children to the extent that the work has drained their energy. For example, a parent may prefer taking some rest at home over playing soccer with his child. Therefore, strain based WIF is likely to constrain nteraction with their child. Hypothesis 1 : Time based WIF is negatively associated with PB. Hypothesis 2 : Strain based WIF is negatively associated with PB. Negative Emotions as a Mediator The relationship between WIF and PB may be further explained thro ugh negative emotions. Viewing WIF as a form of goal conflict, t he goal pursuit literature suggests that individuals are likely to feel negative emotions when they experience WIF (Emmons, 1986; Van Hook & Higgins, 1988)
7 Individuals hold representations o (Markus & Nurius, 1986). These possible selves are associated with goals that the more than one goal at a time. For inst ance, a person may strive to be a caring parent, a loving spouse, and striving behaviors as they function as standards to evaluate the current self and frameworks to select future beha part of their possible selves will be motivated to arrange prime time for family. cannot be attentive to all of their possibl e selves simultaneously P ossible selves compete over limited resources (e.g. time, energy, or cognitive resource s ) because each ideal self requires a substantive amount of resources for people to obtain it (Kahneman, 1973; Norman & Bobrow, 1975) Due to finite resources, the degree that people can spend t ime and energy on a particular goal is bounded. That is, the more resources that are allocated to one goal, the less resources that remain for other goals. This suggests that multiple goals are likely to be perceived as incompatible, and this incompatibility may lead individuals to commit to goal oriented behaviors toward the most salient possible self at a time (George & Brief, 1996) Various environmental factor s ( e.g. organizational context ) or persona l preference ( e.g. goal importance to the self ) may determine the salience of possible sel ves, which may result in an unequal resource allocation. As a lower priority goal becomes difficult to obtain because of resource shortage an individual w o uld fail to attain the nonfocal goal and experience negative affect due to the failure. This idea can be applied to W I F, where work takes a priority over family goals For instance, in the case
8 of a successful CEO who spends the majority of his time working, the wo rk may hinder him from participating in important events for his children. P erceived discrepancy between the ideal and the actual self as a parent may render him to experience negative emotion for not being a devoted father. To the extent that being a cari ng parent is an equally important goal for the individual, the goal actual discrepancy will induce negative emotion Empirical research supports the link between WFC and negative affect, with researchers examining two types of affective states: discrete em otions and moods. Although the two concepts share a common feature as a transient episode of feeling or affect, unique characteristics of each necessitate distinction between the two. Emotions are the result of specific events and represent intense affecti ve experiences of relatively short duration (Weiss, 2002). Negative emotions such as anger or guilt have been positively associated with both directions of WFC (Greenhaus et al., 2006; Gilbert, Holahan, & Manning, 1981; Livingston & Judge, 2008; Allen, in press). On the other hand, moods refer to more diffuse and less intensive states of a longer duration. Unlike emotions, moods may not have a distinct cause. Examining the relationship between mood and WFC I l ies et al. ( 2007 ) found that workload was positi vely associated with negative affect at work, which in turn related to WIF and to negative affect at home. As the current study proposes affective experiences as a bridge between WIF and PB investigating emotion seem s appropriate First, emotion fits well in that WIF is noted as a clear antecedent of negative affect. Frijda ( 1994 ) suggests that whether an affective phenomenon involves a particular event or object differentiates moods from emotions. Unlike moods, emotions refer to object related affective s tates of mind. Also, the concept
9 of emotion is better in explaining the link between affective experience and a particular behavior 1988). Furthermore, consideration of distinct ive emotions may provide incremental value given their unique properties such as distinctive antecedents, subjective experience, and behavioral consequences (Berenbaum, Fujita, & Pfennig, 1995; Lazarus & Cohen Charash, 2001). Among a variety of discrete em otions, three emotions seem particularly relevant to W I F : distress, depression, and hostility. Individuals are distressed when they suffer from daily hassles, physical pain, and psychological exhaustion (Zarski, 1984). Feeling distressed is plausible when one experiences W I F which has been conceptualized as a stressor (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). While s triving for goals in the work and the family domain, it may be not unusual for parents to be exhausted and feel distressed. Previous res earch supported this idea in that W I F has been associated with life distres s and psychological s train (Greenhaus et al., 2006) The next emotion of interest is depress ion A consistent finding is that the experience of WFC in both directions positively rel ates to higher levels of depression (Frone, Russell, & Barnes, 1996; Hammer, Cullen, Neal, Sinclair, & Shafiro, 2006). People tend to be depressed when they perceive lack of control over stressful events ( Peterson, Maier, & Se ligman, 1995). WIF may render people depressed in that it often appears to be something beyond one s control. For instance parent s who find themselves coming back home, too tired to play with children may think that no matter how hard they try they cannot meet expectations in both do mains.
10 Finally, h ostility reflects the experience of anger (Lazarus & Cohen Charash, 2001), which has been linked to obstructed efforts in goal pursuit (Frijda, 1988) The experience of hostility is likely to be prevalent when W I F is perceived as an obsta cle in achieving personal goals (Judge et al., 2006). For example, parents who strive to be available for their children may feel hostile if work prevents interacting with children by leaving an insufficient amount of time and energy at home. Or the very e xperience of WIF could be perceived as a goal failure if an individual desires a balanced state of work and family. Previous research has recognized the general relationship between affect and social activities in which people engage. As a form of affectiv e experience, emotion is behavior (Frijda, 1988). A consistent finding from past research is that positive affect motivates individuals to engage in social activities where as negative affect is linked to the tendency to withdraw from social activities (Watson, 1988; Cunningham, 1988; Kahneman, Krueger, Schkade, Schwarz, & Stone, 2004). For instance, Ilies et al. (2007) reported that the extent that an employee engaged in soc ial behavior at home (e.g., eating mood emotions generates changes in parenting b ehavior (Crouter & McHale, 2005; Dix, 1991). Building on the discussion from the two streams of research, it is hypothesized that WIF induces negative emotions, which in turn may relate to PB. Hypothesis 3 : Time based WIF is positively associated with nega tive emotions Hypothesis 4 : Strain based WIF is positively associated with negative emotions
11 Hypothesis 5 : T he relationship between time based WIF and PB is mediated by negative emotions Hypothesis 6 : T he relationship between strain based WIF and PB is mediated by negative emotions. Trait Guilt as a Moderator Trait Guilt (i.e., guilt proneness) is a dispositional tendency to experience guilt in As a stable personality trait, indi viduals are known to be different in terms of the capability to experience guilt. Guilt refers to an unpleasant emotion that people experience when they acknowledge responsibility for perceive d fail ure to meet norms or fulfill personal goals (Tangney & Dea ring). Trait and state guilt are closely related such that trait guilt has found to result in actual guilt feelings (Leith & Baumeister, 1998). Researchers have suggested that guilt is highly applicable to WFC ( Judge, Ilies, & Scott, 2006; Allen, in press ) When demands from work and family are incompatible, individuals have to make decisions to prioritize duties such as choosing work over family or vice versa In the case of work and family t he decision is challenging and complicated because both roles ar e of importance. When an individual perceives the goal actual discrepancy in the less emphasized domain, t he individual may feel guilty especially if he or she assumes responsibility for the choice between work and family Considering guilt is especially p ertinent to the context of WIF given t he interpersonal characteristic of guilt (Baumeister, Stillwell, & Heatherton, 1994) Experiencing guilt has been con sider ed as an interpersonal phenomenon that arises from reflections of a misdeed to a relationship pa rtner T he intensity of guilt is thought to be stronger in close relationships than in weak
12 relationships. In this sense, guilt may exert a powerful influence in the family domain, where people have very intimate relationships. With regard to the role of g uilt, an adaptive function of guilt implies that guilt may moderate the relationship between negative emotions induced from WIF and PB. Specifically, the adverse impact of negative emotions on PB may be attenuated with the existence of guilt. Although guil t is an aversive emotional state based on the responsibility for perceived failure, it has been suggested to facilitate improvement by prompting individuals to monitor their own behavio r to shift their motivational focus to the underperforming domain, and to engage appropriate action (George & Brief, 1996). Tangney and Dearing (2002) explained that guilt allows people to keep positive regard for the self and to be able to repair what they have done wrong because guilt is g enerated from s pecific, controllab le, and less stable attributions. In an attempt to reduce the level of guilt, individuals may employ different strategies. I n the context of interpersonal relationship s, people apologize, put forth more effort, or spend more time with the partner to restor e the relationship ( Baumeister et al ., 1994 ). Due to the nature of time as definite and valued resource, people can demonstrate the value that they place on the relationship via allocating more time on it. To engage in corrective behaviors may also re d uce guilt as it is an attempt to make situations closer to the ideal. The core intention that lies in the behaviors is that individuals try to show their consideration about others and minimize the goal actual discrepancy. Applying this idea to parent child re lationship, p arents would try to increase the time with their child ren or to engage in certain behaviors n order to compensate for perceived discrepancy and eventually to reduce feeling of guilt
13 Insofar as individua ls differ in terms of the degree that they experience guilt, investigating trait guilt seems essential to understand parents behavior in the work and family interface P ast research showed that people who are predisposed to experience guilt tend ed to cont emplate their role in the failure, feel responsibility, and get motivated to take corrective actions ( Tangney 1990). Also, Leith and Baumeister (1998) reported that guilt prone people demonstrat ed the tendency to take the other person s perspective which is associated with beneficial relationship outcomes This suggests that parents who are higher on trait guilt may engage in PB as compensatory behaviors to remedy the relationship with children more so than parents who are lower on trait guilt. Perceiving poor performance as a parent (i.e., WIF) and feeling guilty about it, they will actively interact with children even if they experience negative emotions Figure 1 demonstrates hypothesized model of the relationship between WIF and PB including proposed m ediator and moderator. Hypothesis 7 : T he relationship between the negative emotions and PB is moderated by t rait guilt such that the relationship is weaker for those with higher trait guilt than for those with lower trait guilt Figure 1 Hypothesized relationships WIF Negative Emotion PB Trait Guilt
14 Chapter Two Method Participants and Procedure The sample consisted of 201 employed parents who were recruited from 11 after school programs located in a large metropolitan area within the southeastern region of the U. S. Among them, 12 couples ( n = 24) were included. To be eligible for the study participants had to wo rk at least 20 hours a week and live with at least one child between seven and nine years old The age of the child was restricted becau se parent ing activities change a long with the age of the child ( Walters & Stinnett, 1971 ; Martin, 1975). Early school age is a critical developmental phase in the study of parent child interaction because parents tend to give greater autonomy to children and spend less time with their children as the child moves from childhood to adolescence ( Maccoby, 1980 ). Participation was entirely voluntary and no compensation was granted. Participants were requested to complete a hard cop y survey at the after school but were al lowed to complete it at home and bring it back to school on the next day. Of the 201 participants, 35.3% were male and 64.7% were female. The average age of the sample was 3 8 07 years ( SD = 6.72 ). The majority was White/ Caucasian ( 54.2 %), followed by Blac k/ African American ( 24.4% ), Hispanic (7. 5 %), Asian /Pacific Islander ( 10.9% ), and Other ( 3.0 %). In regards to marital status, 68.7% were married, 14.4% were living with a partner and 16.9% were single In terms of education, 10.4%
15 had a high school degree, 23.9% had attended some college, 45.8% had a college degree, 2.5% had a ttended some graduate school, and 17.4% had a graduate degree. On average, participants worked 41.04 hours per week ( SD = 9.31 ). Descriptive statistics for all demographic variables ar e listed in Table 1. Table 1 Descriptive Statistics of Study Variables Variable n # of items M SD Obs. Min. Obs. Max. Variables PRPB 201 5 .92 3.51 1.49 1 7 CAPB 201 4 .77 4.98 1.37 1.75 7 PMPB 201 4 .86 3.91 1.37 1.40 7 InteractionT a 197 1 -14.66 7.82 3 53 WIFts 201 6 .91 5.77 2.16 2 10 WIFtime 201 3 .89 3.01 1.21 1 5 WIFstrain 201 3 .91 2.76 1.17 1 5 NE 107 18 .89 1.84 .59 1 3.83 Trait guilt 201 6 .85 .91 .72 0 3 NA 201 5 .74 1.49 .53 1 4.2 Demographics Gender 201 1 -.6 5 .48 0 1 Age 197 1 -38.07 6.72 21 65 Ethnicity 201 1 -1.87 1.23 1 6 Marital status 201 1 -1.48 .77 1 3 Work hour 201 1 -41.04 9.31 20 87.5 Tenure b 182 1 -69.19 53.81 1 242 Salary c 137 1 -53471.53 40523.49 6000 400000 Note. PRPB = Physic al and recreational parent child interaction behavior ; CAPB = Cognitive and academic oriented parent child interaction behavior ; PMPB = Passive and maintenance oriented parent child interaction behavior ; InteractionT = Number of hours of direct interaction with child per week; WIFts = A composite variable of WIFtime and WIFstrain; WIFtime = Time based work interfere with family; WIFstrain = Strain based work interfere with family; NE = Negative emotion; NA = Negative affectivity; Work hour = Number of hours of working per week. a Interaction time was reported in hours. b Tenure was reported in months. c Salary was reported in dollars.
16 M easures All measures are included in the appendi x. S cores on each scale were obtained by averaging the score on each item H igh er scores indicat e a greater prevalence of the construct. Work to family conflict (WIF) Two subscales from Carlson, Kacmar, and Williams s (2000) WIF scale were used to assess time and strain based WIF Each subscale consists of three items. T he response to each item w as rated on a 5 poi nt Likert type scale that ranged from 1 ( strongly disagree ) to 5 ( strongly agree ) S ample items include The time I must devote to my job keeps me from participating equally in household responsibilities and activities an d When I get home from work I am often too frazzled to participate in family activities/responsibilities. scales were also examined together as an overall index of WIF. Negative emotion Negative emotions w ere measured with adjective s from t he p ositive and n egative a ffect s chedule Expanded form (the PANAS X; Watson & Clark, 1994) and the circumplex model of affect (Russell, 1980) T he response to each item was rated on a 5 point Likert type scale that ranged from 1 ( very slightly or n ot at all ) to 5 ( extremely ) Items for distress distressed, afraid, frustrated and annoyed. Hostility w as measured with the words of angry, hostile, irritable, scornful, disgusted, and loathing Finally sad, blue, downhearted, alone, lonely, d epressed, gloomy, and miserable w ere averaged for depression. Only participants who reported that they had experienced WIF during the past four weeks were asked to respond to these items. Trait guilt. The P ersonal Feelings Q u estionnaire 2 ( the PFQ 2; Har der & Zalma, 1990 ) w as used to assess trait guilt. The scale showed construct validity and a two factor
17 structure that distinguishes shame and guilt. It assesses how common the emotions are for the rater and the response for each item is rated on a 5 point Likert type scale that rang es from 0 ( never experience the feeling ) to 4 ( experience the feeling continuously or almost continuously ). Only the six items that pertain to guilt (i.e., m ild guilt, worry about hurting or injuring someone intense guilt regr et, remorse and feeling you deserve criticism for what you did ) were included in the present study. Parent c hild i nteraction behavior (PB) PB was measured with a scale that was developed for the current study. Based on results from the pilot study (shown below), 16 items were used to measure both active and passive parenting behaviors. However, factor analysis (shown below) suggested that a three factor solution fit better for the data and that three items had either low or crossed factor loadings. Accord ingly, the hypotheses were tested separately for each of the three types of PB. The first PB, physical and recreational parent child interaction behavior (PRPB), was measured with five items. An cognitive and academic oriented parent child interaction behavior (CAPB), was measured with four items. A help my child with his/her homework The final PB is passive and maintenance oriented parent child interaction behavior (PMPB) an d was measured with My child and I The scale assesse d the frequency of each behavior during the past four weeks using a 7 point Likert type scale (1 = never 2 = 1 5 times 3 = 6 10 times 4 = 11 15 tim es 5 = 16 20 times 6 = 21 25 times and 7 = 25 times or more ). Additionally, a single item was used to ask the number of hours per week that parents directly interact with child. While responding to the items participants were asked to cho o se one child based on his/her age and focus on
18 th at child. When there are multiple children whose age fall between seven and nine, the older child w as chosen C ontrol variables Due to their potential relationships with the criterion variables, g ender ethnicity age, marital status, number of children living at home, education, organizational tenure, salary and age of the focal child were considered as control variables. The highest level of education was asked (1 = some high school, 2 = high school diploma, 3= some c ollege, 4 = 2 year college degree, 5 = 4 year college degree, 6 = some graduate school, or 7 = graduate degree). With regard to ethnicity, participants identified themselves as White/Caucasian, Black/African American, Hispanic, Asian/Pacific Islander, Nati ve American, or Other. Marital status was reported as married, living with a partner/significant other, or single. In analyses, gender, ethnicity, and marital status were dummy coded (male = 0, female = 1; White/Caucacian = 0, all others = 1; single = 0, m arried or living with a partner/significant other = 1, respectively). Additionally, trait negative affect ivity was assessed using five items from the PANAS X ( Watson & Clark, 1994)
19 Chapter Three Results Pilot Study As no measure for parent child intera ction behavior exists, a scale was created for the present study. Hinkin (1995) suggested that a new scale must demonstrate the basic and essential psychometric properties of reliability and validity. The purpose of the pilot study was to assess internal r eliability of the parent child interaction behavior measure and to find evidence for construct validity. The internal reliability of the scale was measured with coefficient alpha (Cronbach, 1951). Construct validity of the scale was measured by testing hyp otheses that involve relationships between the focal construct and other theoretically related variables (Cronbach & Meehl, 1955). In the current study, it was hypothesized that parent child interaction behaviors are negatively associated with time and st rain based WIF. Scale development. A new scale for parent child interaction behavior was intended to capture a wide range of activities that parents do with a preschooler or an early school aged child (e.g., the ages between seven and nine) After reviewin g past research on parental time and behavior ( Nock & Kingston 1988 ; Gauthier et al. 2004 ; Sayer, Bianchi, & Robinson 2004 ), a list of 18 items that illustrate various interaction behaviors was constructed. All items are listed in Table 2 The behaviors varied in terms of the level of cognitive or physical involvement required by parents. Activities that are
20 presumed to require a gross engagement were considered as active behaviors whereas activities that can be handled with less involvement were regarde d passive. Table 2. Parent child Interaction Behavior Measure Used in Pilot Study No. Items 1. I help my child with his/her homework 2. I read to my child 3. My child and I have discussions about my child s achievement s or co ncern 4. My child and I play together ( e.g., bike riding, playing sports ). 5. I play outside with my child 6. My child and I exercise together 7. I go on outings with my child (e.g., museum, zoo, sporting event ). 8. I play indoor games with my chil d (e.g., board games, video games ). 9. My child and I discuss TV shows/movies 10. My child and I talk while we are driving together 11. I do creative activities with my child such as dancing, singing, and making c rafts 12. My child and I have meals together.* 13. s and extended fami ly 14. My child and I do housework together. 15. My child and I go shopping together. 16. My child and I watch TV together. 17. My child and I do grocery shopping together. 18. My child and I do not interact with each other when we are together.* Note. indicates removed item based on the pilot study results. Items 1 11 address active parent child interaction behavior; Items 12 17 address passive parent child interaction behav ior; Item 18 indicates no interaction behavior. Participants and p rocedure The sample consisted of 33 employed parents who liv e with at least one child whose age is between seven and nine. The parents were required to work at least 20 hours a week in ord er to participate. Participant recruitment
21 took place through multiple routes. First, individuals who enrolled in a large southern university participated in the study in order to fulfill course requirement. Also, an advertisement for recruitment was poste d on several websites that parents frequently visit. Finally, snowball sampling was used such that individuals who had already participated in the study were asked to introduce the study to others eligible and willing to participate. The average age of par ticipants was 38.59 years old ( SD = 7.17). Of the 33 participants, 57.6% were female and 42.4% were male. In regards to the ethnicity, 48.5% were White/Caucasian, 36.4% were Asian/Pacific islander, 9.1% were Hispanic, and 6.1% were Black/African American. Approximately 85% participants were either married or living with a partner while 15% reported self as single. In terms of education, 3.0% had a high school degree, 18.2% had attended some college, 48.5% had a college degree, 6.1% had a ttended some graduat e school, and 24.2% had a graduate degree. On average, they worked 40.85 hours per week ( SD = 10.66). Descriptive statistics for all demographic variables are listed in Table 3. Measures. With the exception of parent child interaction behavior, all measure s included in the pilot study were identical to those used in the primary study. Parent child interaction behavior was measured with an 18 item scale that was developed for the study. Among the 18 items, there were 11 active behaviors, six passive behavior s, and one item that indicated no interaction between parent and child. The response to each item was rated on a 7 point scale with frequency anchor s (1 = never 2 = rarely 3 = 1 2 times a month 4 = once a week 5 = 2 3 times a week 6 = 4 5 times a week and 7 = most days of the week ) While responding to the items participants were asked to choose one child
22 based on his/her age and focus on the child. When there are multiple children whose age fall s between seven and nine, the older one w as chosen Ta ble 3 Descriptive Statistics in Pilot Study Variable n # of items M SD Variables PB a 33 18 9.02 1.27 .72 Active PB 33 11 4.54 .7 0 .73 Passive PB 33 7 4.48 .76 .67 W I Ftime 32 3 3.18 .80 .74 W I Fstrain 32 3 2.63 .72 .65 NE 13 18 1.88 .53 .86 Trait g uilt 31 6 1.27 .68 .79 NA 31 5 1.66 .68 .83 Demographics Gender 33 1 .58 .50 -Age 32 1 38.59 7.17 -Ethnicity 33 1 .52 .51 -Marital status 33 1 .85 .36 -Work hour 33 1 40.85 10.68 -Tenure b 33 1 73.56 93.20 -Salary c 24 1 6130 8.33 36044.77 -Note. PB = Parent child interaction behavior; Active PB = Active parent child interaction behavior; Passive PB = Passive parent child interaction behavior; WIFtime = Time based work interfere with family ; WIFstrain = Strain based work int erfere with family ; NE = Negative emotions; NA = Negative affectivity; Work hour = Number of hours of working per week. a PB = A composite variable of Active PB and Passive PB. b Tenure was reported in months. c Salary was reported in dollars.
23 Results and di scussion. Descriptive statistics for all study variables are listed in Table 3. Intercorrelations among the variables are demonstrated in Table 4 Internal consistency reliability of the parent child interaction behavior measure was assessed with coefficie nt alpha (Cronbach, 1951). Alpha of .70 or higher indicates that the scale is reliable (Nunnally, 1973). The scale demonstrated an adequate level of reliability ( = .72). With regard to construct validity, correlations between parent child interaction beh avior and WIF were examined. Given low statistical power due to small sample size ( n = 33), results were interpreted in terms of the trend, rather than significance of the relationship. Active PB was not correlated with time based WIF ( r = .00 p = 99) Contrary to hypothesis, time based WIF was positively correlated with passive PB ( r = .38 p < 05). This suggests that p arents do more passive PB when they experience time based WIF. This unexpected finding could be explained by the nature of passive PB. Most activities considered as passive PB were maintenance behaviors that are essential for everyday life. Experiencing time based WIF, parents might have substituted passive PB for active PB because passive PB must be carried out in any circumstances. On t he other hand, relationships between strain based WIF and both active and passive PB were negative as hypothesized ( r = .22 p = 23; r = .03 p = 86, respectively). It is important to note that the effect size was larger for active PB. This suggests th at active PB may be more adversely impacted by strain based WIF than was passive PB because active PB requires more energy. Taken together, correlations among WIF and PB variables provided preliminary evidence for construct validity of the new measure. Two items were removed based on qualitative feedback from participants and the analyses results. Specifically, participants commented on items that were difficult to
24 Table 4. Intercorrelations among Variables in Pilot Study Note Results based on n = 33. Gender: Male = 0, Female = 1; Ethnicity: White/Caucasian = 0, All others = 1; Marital status: Single = 0, Married or Living with a partner/significant other = 1. a n = 13 p < .05, ** p < .01. Variable 1. 2 3 4 5 6. 7 8. 9. 10. 11. 12. 1 3. 14. 15. 1. PB -2. Active PB .86** -3. Passive PB .88** .50** -4. W I Ftime .23 .00 .38* -5. W I Fstrain .14 .22 .03 .16 -6. NE a .59* .43 .60* .22 .57 -7. Trait Guilt .01 .03 .05 .28 .47** .41 -8. NA .13 .10 .31 .26 .46* .55 .52** -9. Gender .40* .26 .42* .07 .10 .22 .07 .05 -10. Age .29 .19 .31 .15 .27 .28 .01 .02 .34 -11. Ethnicity .19 .20 .13 .07 .18 .09 .04 .16 .10 .03 -12. Marital status .01 .04 .05 .20 .07 .01 .03 .07 .19 .24 .10 -13. Work hour .39* .27 .40* .03 .13 .22 .37* .18 .47** .30 .14 .02 -14. Tenure .32 .16 .38* .23 .31 .60 .02 .27 .08 .46** .15 .29 .01 -15. Salary .53** .37 .55** 16 .35 .23 .03 .05 .60** .80** .16 .43* .50* .40 -
25 understand. In addition, participants pointed out that some behaviors on the list were everyday activities whereas others occur less frequently. Accordingly, frequency anchors were changed so that all behaviors are rated on a m onthly time frame. In the primary study, the final measure that consists of 16 items was used with the new anchors. Factor Analysis Primary Study Factor analyses for PB were conducted to examine the structure of the new scale. Exploratory factor analysis (EFA) was conducted with principal axis factoring method and oblimin rotation. Although the scale was designed to address active and passive PB results suggested a three factor model that exp lain s approximately 65% of the variance. D ue to low factor load ings, three items were eliminated (i.e., My child and I discuss TV shows/movies I do creative activities with my child such as dancing, singing, and making crafts ) Table 5 shows stan dardized factor loadings for each item. Next, confirmatory factor analysis (CFA) was conducted on the shortened, 13 item measure. Specifically, 2 factor model and 3 factor models were compared. Fit indices for the two models are found in Table 6. Although the 2 was significant for both models ( 2 (64 ) = 266.96, p < .01 and 2 (62 ) = 176.96, p <. 01 for 2 factor and 3 factor model, respectively), the 3 factor model fit the data better than 2 factor model Further examination revealed that two factors emerge d from nine items that were originally considered active PB. The first factor was named oriented parent child interaction behavior (CA oriented parent child interaction behavior
26 Table 5. Standardized Factor Loadings from Exploratory Factor Analysis No. Items Factors 1 2 3 5. I play outside with my child .996 .069 .063 4 My child and I play together ( e.g., bike riding, playing sports) .910 .162 .050 6. My child and I exercise together .814 .027 .008 7. I go on outings with my child (e.g., museum, zoo, sporting event) .679 .229 .019 8. I play indoor games with my c hild (e.g., board games, video games) .644 .136 .101 2. I read to my child .056 .788 .096 1. I help my child with his/her homework .036 .678 .020 3. achievement s or concern .102 .655 .048 10. My ch ild and I talk while we are driving together .003 .547 .077 14. My child and I go shopping together .024 .047 .881 16. My child and I do grocery shopping together .089 .048 .861 13. My child and I do housework together. .099 .030 .676 15. My child and I watch TV together. .040 .071 .643 9. My child and I discuss TV shows/movies.* .215 .239 .269 11. I do creative activities with my child such as dancing, singing, and making crafts.* .333 .262 .184 12. d extended family.* .176 .139 .393 Eigen value 6.10 1.93 1.11 Percent of variance explained 46.91 14.84 8.55 Note. indicates removed item based on the factor analysis results. Bold indicates items that were used for confirmatory factor analysis. Ite ms for Factor 1 and Factor 2 Table 6. Results of Confirmatory Factor Analysis Model 2 df CF I GFI NFI RMSEA SRMR 2 Factor 267.36 64 .87 .81 .83 .13 .09 3 Factor 176.29 62 .93 .87 .89 .10 .07 Note. CFI = C omparative fit index ; GFI = The goodness of fit index; NFI = Normed fit index; RMSEA = Root Mean Square Error of Approximation; SRMR = Standa rdized Root mean square residual.
27 Descriptive Statistics Descriptive statistics for all study variables are listed in Table 1 Internal reliability analyses for study variables indicated that all measures, including the new parent child interaction behavi or measure, had an acceptable level of coefficient alpha (Nunnally, 1973). Intercorrelations among study variables are reported in Table 7 With few exceptions, WIF and PB variables were all significantly correlated. Sample size for negative emotion is 107 because only participants who had experienced WIF during past four weeks were asked to respond to the items of negative emotion. Control Variables G ender eth nicity age, marital status, number of children living at home, education, organizational tenure, salary negative affectivity (NA), and age of the focal child were measured as potential control variables. Correlation analysis was conducted to examine relationships a mong the variables and criterion variables. Results showed that g ender and age were si gnificantly associated with CAPB ( r = .31 p < 01; r = .17 p < 05, respectively). Gender, age, education, and salary were all significantly associated with PMPB ( r = .16 p < 05; r = .26 p < 01; r = .19 p < 01; r = .2 3, p < 01, respectively). Finally, NA was significantly correlated with PRPB ( r = .15 p < 05). The rest of demographic variables (ethnicity, marital status, number of children living at home, organizational tenure, and age of the focal child ) were not significantly related to an y PB. In order to preserve power, only variables that were significant ly associated with the criterion variable in question were included as controls in analyses.
28 Table 7. Interc orrelations among Study Variables Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1 1. 12. 13. 14. 15. 16. 17. 18 1. PRPB -2. CAPB .62 ** -3. PMPB .47 ** .46 ** -4. Inter T .28 ** .31 ** .32 ** -5. WIFts a .43 ** .19 ** .01 .27 ** -6. WIFtime .36 ** .17 .03 .21 ** .91 ** -7. WIFstrain .43 ** .17 .00 .28 ** .91 ** .66 ** -8. NE b .11 .06 .18 .02 .44 ** .43 ** .35 ** -9. Trait guilt .13 .04 .14 .22 ** .11 .08 .13 .20 -10. NA .15 .01 .06 .16 .10 .03 .15 .33 ** .52 ** -11. Gender .04 .31 ** .16 .21 ** .06 .11 .00 .05 .08 .01 -12. Age .13 .17 .26 ** .20 ** .08 .11 .04 .21 .01 .05 .32 ** -13. Ethnicity .00 .01 .09 .15 .07 .22 .11 .13 .16 .03 .09 .25 ** -14. Education .10 .06 .19 .18 .08 .0 4 .10 .01 .24 ** .18 ** .01 .27 ** .02 -15. Marital .04 .07 .01 .16 .13 .19 ** .05 .08 .05 .09 .08 .10 .07 .02 -16. Wor k h r .18 .23 ** .09 .15 .29 ** .32 ** .21 ** .01 .04 .04 .19 ** .02 .02 .04 .06 -17. Tenure .05 .05 .05 .07 .01 .03 00 .03 .04 .05 .18 .19 .14 .08 .05 .10 -18. Salary .01 .12 .23 ** .19 .01 .07 .06 .09 .07 .04 .31 ** .30 ** .12 .23 ** .17 .30 ** .22 -Note. Results based on n = 201. PRPB = Physical and recreational paren t child interaction behavior; CAPB = Cognitive and academic oriented parent chi ld interaction behavior; PM PB = Passive and maintenance oriented parent child i nteraction behavior; Inter T = Number of hours of direct interaction with child per week; W I Ftime = Time based work interfere with fam ily; W I Fstrain = strain based work interfere with family; NE = Negative emotions; NA = Negative affectivity; Work h r = Numb er of hours of working per week a WIFts = A composite variable of WIFtime and WIFstrain. b n = 107. p < .05, ** p < .01.
29 Prelimin ary Analysis Prior to hypotheses testing, data were inspected to identify outliers and violations of assumptions. Data points that fall three standard deviations above or below the mean were considered as potential outliers. Initially one outlier was ident ified for negative emotion but it was not removed as further examination revealed that its value was still plausible for the negative emotion measure. Assumption of n ormality was examin ed with the values of skewness and kurtosis Values of two standard err ors of skewness/kurtosis or more were considered to signal nonnormality. Negative emotions and trait guilt were positively skewed, which suggested that majority participants reported low level of the constructs. Also, distribution of negative emotions appe ared to be a platykurtic However, given that the product moment correlation is robust with respect to the normality assumption (Cohen, 1988), analyses were conducted without transformation of the data. Additionally, assumptions of regression were examined Given the design of the study, independence of observation was assumed to be met. Normality of residual was examined with q q plots. The plots for all variables indicated normality of residuals. Next, scatterplots of the study variable were inspected to test the linearity and homoscedasticity. The plots appeared linear. Also, the variance of the predictor variables appeared to be constan t across levels of the criterion variable, providing evidence for the assumption of homoscedasticity. Hypotheses Testing An alpha level of .05 was used for all analyses Separate analyses were conducted for three types of PB All hypotheses were tested using a composite variable of time and strain based WIF (WIFts) as well as each WIF variable.
30 Hypotheses 1 and 2 were tes ted by examining zero order correlations between WIF and parenting behaviors. Furthermore hierarchical regression was used to test whether the relationships remain significant after taking control variables into account Hypotheses 3 and 4 were also exami ned using z ero order correlation s between WIF and negative emotion Hierarchical regression was conducted in order to further investigate robustness of the relationships after controlling for the effect of trait negative affectivity. For each regression, c ontrol variables were entered in step one, followed by predictor variabl es in step two. Hypothesis 1 stated that time based WIF would be negatively related to PB. The hypothesis was supported for PRPB ( r = 36 p < .01; = 36 p < .01). This hypothesis was also supported for CAPB using correlation ( r = .17 p < 01); however, the relationship was no longer significant when the effects of age and gender were controlled in regression ( = .13 p = .06). No support was found for the negative relationship between time based WIF and PMPB ( r = 03 p =. 71; = 05 p =. 54 ). Regression results are shown in Table 8. Hypothesis 2 predicted that strain based WIF would be negatively related to PB. Again, the prediction was supported for PRPB ( r = .43, p < 01; = 41 p < .01). For CAPB, support for the hypothesis was found from both correlation and regression analyses ( r = 17 p < 01; = 17 p < .05). However, the hypothesis was not supported for PMPB ( r = 00 p = 95; = 06 p = 46 ). Results are sho wn in Table 9. Hypothesis 3 addressed positive relationship between time based WIF and negative emotion. This hypothesis was supported using correlation analysis ( r = 43 p < .01). The relationship was still significant after the effect of negative affect ivity was controlled for ( = .43, p < 01). Regression results are shown in Table 10
31 Table 8. Regression of Time based WIF on Parent child Interaction Behavior Variable PRPB CAPB PMPB Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Control Variables Gender .30** .29** .07 .07 Age .07 .06 .12 .12 Education .09 .09 Salary .16 .16 NA .15* .14* Predictor WIFtime .36** .13 .05 F 4.49* 17.69** 11.85** 9.19** 3.00* 2.47* df 1, 199 2, 198 2, 194 3, 193 4, 130 5 129 Overall R 2 .02 .15 .11 .13 .09 .09 2 .13** .02 .00 Note. NA = Negative affectivity. *p < .05 **p < .01
32 Table 9. Regression of Strain based WIF on Parent child Interaction Behavior Variable PR PB CA PB PMPB Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Control Variables Age .30** .30** .07 .07 Gender .07 .06 .12 .12 Education .09 .09 Salary .16 .16 NA .15* .09 Predictor WIF strain .41** .17* .06 F 4.49* 22.96** 11.85** 10.28** 3.00* 2.50* df 1, 199 2, 198 2, 194 3, 193 4, 130 5, 129 Overall R 2 .02 .19 .11 .14 .09 .09 2 .17** .03* .00 Note. NA = Negative affectivity. *p < .05 **p < .01.
33 Table 1 0 Regression of Time based WIF on Negative Emotion Variable Negative emotion Step 1 ( ) Step 2 ( ) Control Variable Negative affectivity .33** .32** Predictor WIF time .43** F 12.46** 21.29** df 1, 105 2, 104 Overall R 2 .11 .29 2 .18** *p < .05 **p < .01 Hypothesis 4 p ropos ed that strain based WIF would be posi tively related to negative emotion This proposition was also supported by both corr elation and regression analyses ( r = 35 p < .01; = 30 p < 01 ). Regression results are shown in Table 11. In general, analyses using WIFts demonstrated identical results with the findings from the original analyses. In terms of relationships with par enting behaviors (Hypotheses 1 and 2), WIFts was negatively associated with PRPB ( r = 43 p < .01; = 42 p < .01). CAPB was also negatively related to WIFts ( r = .19 p < 01); the relationship remained significant after controlling age and gender ( = .16 p < .05). Consistent with the original finding, there was no significant relationship between WIF and PMPB ( r = 01 p = 86; = 00 p = 97 ). Next, relationship between WIFts and
34 negative emotion was investigated (Hypotheses 3 and 4). As hyp othesized, WIFts was positively related to negative emotion ( r = 44 p < .01). The relationship was still significant after the effect of negative affectivity was controlled ( = .41 p < 01). Regression results are shown in Table 12 and 13. Table 1 1 R egression of Strain based WIF on Negative Emotion Variable Negative emotion Step 1 ( ) Step 2 ( ) Control Variable Negative affectivity .33** .27** Predictor WIF strain .30** F 12.46** 12.48** df 1, 105 2, 104 Overall R 2 .11 .19 2 .08 ** *p < .05 **p < .01
35 Table 1 2 Regression of WIFts on Parent child Interaction Behavior Variable PR PB CA PB PMPB Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Step 1 ( ) Step 2 ( ) Control Variables Age .07 .06 .12 .12 Gender .30 ** .29** .07 .07 Education .09 .09 Salary .16 .16 NA .15* .11 Predictor WIF ts .42** .16* .00 F 4.49* 24.63** 11.85** 10.08** 3.00* 2.38* df 1, 199 2, 198 2, 194 3, 193 4, 130 5, 129 Overall R 2 .02 .20 .11 .14 .09 .09 2 .18** .03* .00 Note. NA = Negative affectivity. *p < .05 **p < .01.
36 Table 13 Regression of WIFts on Negative Emotion Variable Negative emotion Step 1 ( ) Step 2 ( ) Control Variable Negative affectivity .33** .28** Predictor WIF t s .41** F 12.46** 19.20** D f 1, 105 2, 104 Overall R 2 .11 .27 2 .16** *p < .05 **p < .01 Mediator Hypothesis In order t o test Hypothes e s 5 and 6, a series of regression analyses were conducted with mean centered predictors (Aiken & West, 199 1). Specifically, PB was regressed onto WIF. Then, negative emotion was regressed onto WIF. Lastly, PB was regressed onto both WIF and negative emotion To support the hypotheses, s ignificant rela tionship in the first two steps as well as a nonsignifi cant relationship between the WIF and PB in the last step is necessary (Baron & Kenny, 1986 ) In addition to the test of the simple mediation model, indirect effect of negative emotions was examined using the Sobel test (Sobel, 1982), as recommended by Preacher and Hayes (2004). Hypotheses 5 and 6 proposed mediation effect of negative emotions in the relationship between time based WIF and PB The results are summarized in Table 14
37 and 15 Although both WIF variables were significant predict ors of PRPB and CAPB as well as negative emotions, negative emotions, the mediator, did not significantly predict all three types of PB. That is, one of the criteria for the test of simple mediation model, significant relationship between the mediator and outcome variable, was not met (Baron & Kenny, 1986). Therefore, hypotheses could not be tested with simple mediation test method Results from the Sobel test suggested that the mediation effect of negative emotions is significant only in the relationship between time based WIF and PMPB. In sum, Hypotheses 5 and 6 were not supported. Hypotheses 5 and 6 were also tested using WIFts as a predictor. Again, the test of simple mediation model could not be conducted due to nonsignificant relationships between the proposed mediator and outcome variables. The Sobel test indicated that negative emotion does not have significant indirect effect in the relationship between WIFts and all types of PB. Results are summarized in Table 16. Moderator Hypothesis Hypothes i s 7 was tested using moder ated hierarchical regression (James & Brett, 1984) with the cross product of negative emotion and trait guilt as the interaction term. All predictor variables were mean centered and the interaction term was created based on the centered variables (Aiken & West, 1991) In order to test moderating effect, 2 ) was examined. That is, control variables and predictor s were entered in step one and two, respectively, followed by the inter action term in step three. The results are displayed in Table 17.
38 Table 14 Regression Results for Mediation with Time based WIF as Predictor PRPB C A PB P M PB Model B SE t p B SE t p B SE t p WIFtime NE .23 .05 .43 4.93 .00 .23 .05 .43 4.93 .00 .2 3 .05 .43 4.93 .00 NE PB .27 .24 .11 1.10 .27 .14 .23 .06 .60 .55 .41 .21 .18 1.92 .06 WIFtime PB .45 .08 .36 5.51 .00 .19 .08 .17 2.40 .02 .03 .08 .03 .37 .71 WIFtime PB adding NE a .45 .14 .34 3.19 .00 .16 .13 .13 1.25 .21 .26 .16 .2 2 1.64 .11 Model summary R 2 = .10 p = .01 R 2 = 12 p = .01 R 2 = .20 p = .02 Sobel test z = 1.09 p > .05 z = .60 p > .05 z = 1.80 p < .05 Note. WIFtime = Time based WIF; PB = Parent child interaction behavior; NE = Negative emotion; PRPB = Physica l and recreational paren t child interaction behavior; CAPB = Cognitive and academic oriented parent chi ld interaction behavior; PM PB = Passive and maintenance oriented parent child interaction behavior a Model includes control variables. Table 15. Regres sion Results for Mediation with Strain based WIF as Predictor PRPB C A PB P M PB Model B SE t p B SE t p B SE t p WIFstrain NE .17 .04 .35 3.82 .00 .17 .04 .35 3.82 .00 .17 .04 .35 3.82 .00 NE PB .27 .24 .11 1.10 .27 .14 .23 .06 .60 .55 .41 .21 .18 1.92 .06 WIFstrain PB .54 .08 .43 6.63 .00 .20 .08 .17 2.45 .02 .01 .08 .00 .06 95 WIFstrain PB adding NE a .58 .11 .49 5.23 .00 .21 .11 .20 1.94 .06 .02 .14 .02 .15 .88 Model summary R 2 = .22 p = .00 R 2 = .14 p = .00 R 2 = .17 p = .05 Sobel test z = 1.09 p > .05 z = .60 p > .05 z = 1.77 p > .05 Note. WIFstrain = Strai n based WIF; PB = Parent child interaction behavior; NE = Negative emotion; PRPB = Physical and recreational paren t child interaction behavior; CAPB = Cognitive and academic oriented parent chi ld interaction behavior; PM PB = Passive and maintenance oriente d parent child interaction behavior a Model includes control variables.
39 Table 16. Regression Results for Mediation with WIFts as Predictor PRPB C A PB P M PB Model B SE t p B SE t p B SE t p WIFts NE .13 .03 .44 5.00 .00 .13 .03 .44 5.00 .0 0 .13 .03 .44 5.00 .00 NE PB .27 .24 .11 1.10 .27 .14 .23 .06 .60 .55 .41 .21 .18 1.92 .06 WIFts PB .30 .04 .43 6.78 .00 .12 .04 .19 2.67 .01 .01 .05 .01 .18 .86 WIFts PB adding NE a .35 .07 .49 5.01 .00 .13 .07 .20 1.88 .06 .08 .09 .1 3 .96 .3 4 Model summary R 2 = .21 p = .00 R 2 = 14 p = .00 R 2 = .18 p = .04 Sobel test z = 1.09 p > .05 z = .60 p > .05 z = 1.78 p > .05 Note. WIFts = A composite variable of time and strain based WIF; PB = Parent child interaction behavior; NE = Negative emotion; PRPB = Physical and recreational parent child interaction behavior; CAPB = Cognitive and academic oriented parent chi ld interaction behavior; PM PB = Passive and maintenance oriented parent child interaction behavior a Model includes con trol variables
40 Table 17. Moderated Regression Results of Trait Guilt and Negative Emotion on PB Variable PRPB CAPB PMPB Step 1 ( ) Step 2 ( ) Step 3 ( ) Step 1 ( ) Step 2 ( ) Step 3 ( ) Step 1 ( ) Step 2 ( ) Step 3 ( ) Control Variables Age .20* .20 .20 .16 .08 .07 Gender .20 .19 .19 .20 .18 .17 Education .11 .01 .02 Salary .25 .25* .24* NA .05 .01 .01 Predictors Trait guilt .02 .02 .03 .04 .33* .34* Negative emotion .10 .17 .00 .01 .28* .27* Interaction Term Trait guilt X NE .29** .06 .07 F .25 .41 2.56* 6.33** 3.13* 2.57* 2.62* 3.53** 3.06** df 1, 105 3, 103 4, 102 2, 104 4, 102 5, 101 4, 67 6, 65 7, 64 Overall R 2 .00 .01 .09 .11 .11 .11 .14 .25 .26 2 .01 .08** .00 .00 .11* .01 p < .05, ** p < .01
41 Hypothesis 7 predicted that trait guilt would moderate the relationship between negative emotion and PB. Support was found for PRPB ( R 2 = .08, p < .01) but not for CAPB ( R 2 = .00, p = .53) an d PMPB ( R 2 = .01, p = .53 ). As hypothesized, parents who are low on trait guilt reduced the amount of PRPB as the level of negative emotion increased. On the other hand, parents who are high on trait guilt maintained the level of PRPB despite negative em otion. Therefore, Hypothesis 7 was partially supported. This significant interaction is plotted in Figure 2. Figure 2. Interaction of Negative Emotion on PRPB as a Function of Trait Guilt
42 Exploratory Analyses In order to gain greater insight into the data, additional analyses were conducted for exploratory purposes. First, the moderating role of trait guilt was tested in the relationship between WIF and PBs T he cross product of WIFts and trait guilt was used as the interaction term in moderated hierar chical regression (James & Brett, 1984). As was in the main analysis, control variables and predictor variables were entered in step one and two, respectively, followed by the interaction term in step three. Results revealed that trait guilt moderate relat ionships of WIFts with PRPB ( R 2 = .06, p < .01) and CAPB ( R 2 = .03, p < .05 ). As expected, guilt prone parents did not decrease the level of participation in PRPB and CAPB whereas parents who are lower on trail guilt reduced the amount of participation in those PBs. The results of moderated regression are shown in Table 18, and significant interactions are plotted in Figure 3 and 4. Figure 3. Interaction of WIFts on PRPB as a Function of Trait Guilt
43 Table 18. Moderated Regression Results of Trait Guilt and WIFts on PB Vari able PRPB CAPB PMPB Step 1 ( ) Step 2 ( ) Step 3 ( ) Step 1 ( ) Step 2 ( ) Step 3 ( ) Step 1 ( ) Step 2 ( ) Step 3 ( ) Control Variables Age .07 .06 .04 .12 .12 .11 Gender .30** .30** .27** .07 .06 .06 Education .09 .03 .03 Salary .16 .16 16 NA .15* .09 .03 Predictors Trait guilt .03 .07 .05 .06 .21* .21* WIFts .42** .40** .16* .14* .02 .02 Interaction Term Trait guilt X WIFts .22** .17* .01 F 4.49* 16.42 ** 15.99 ** 11.85 ** 7.69** 7.52 * 3.00* 3.02* 2.58* df 1, 199 3, 197 4, 196 2, 194 4, 192 5, 191 4, 130 6, 128 7, 127 Overall R 2 .02 .20 .25 .11 .14 .17 .09 .12 .12 2 .18** .05** .03* .03* .03 .00 *p < .05 **p < .01
44 Figure 4 Interaction of WIFts on CAPB as a Function of Trait Guilt Next, the entire hypothesized model was tested using a general path analytic framework for combining moderation and mediat ion. This framework has recently been proposed to address drawbacks of conventional methodology (Edwards & Lambert, 2007). First, regression analyses with mean centered variables were conducted separately for the three types of PB. In the first regression, negative emotion (mediator) was regressed onto WIF (predictor) followed by the second analysis, which predicted PB from WIF, negative emotion, trait guilt, and the interaction term. The interaction term was the cross product of negative emotion and trait guilt. Next, regression coefficients from the two regression analyses were utilized for the bootstrap (Efron & Tibshirani, 1993). The bootstrap generates a sampling distribution by repeatedly estimating the coefficients with bootstrap samples, which are cr eated by random sampling with replacement from the
45 original sample. Analysis of simple effects for moderated mediation was conducted based on the coefficient estimates from the bootstrap method in order to test the model. The results are listed in Table 19 In support of the hypotheses and the findings from the main analyses, paths between WIF and PRPB and between WIF and negative emotion were significant. Furthermore, differences were found with regard to participation in PRPB between parents who were high er and lower on trait guilt. The opposite direction of path coefficients in Figure 8 displays sharp contrast between the two groups of parents A. Simple effects for low trait guilt B. Simple effects for high trait guilt Figure 5 Mediated Moderation Models for Low and High Trait Guilt Note. Coefficient in boldface shows significant difference ( p < .05) across levels of the moderator variable 0.41 0.13** WIF PRPB 0 3 2 ** Negative Emotion 0.65* 0.13** WIF Negative Emotion 0 3 2 ** PRPB
46 Table 19 Analysis of Simple Effects in Moderate d Mediation Model PRPB C A PB P M PB Moderator variable Stage Effect Stage Effect Stage Effect 1st 2nd Direct Indirect Total 1st 2nd Direct Indirect Total 1st 2nd Direct Indirect Total Trait guilt Low .13 ** .41 .32 ** .05 .37 ** .13 ** .15 .12 .02 .10 .13 ** .04 .11 .01 .12 High .13 ** .65 .32 ** .08 .24 ** .13 ** .40 .12 .05 .07 .13 ** .54 .11 .07 .18 Differences 0 1.06 0 .13 .13 0 .25 0 .03 .03 0 .50 0 .06 .06 Model summary R 2 = .25 p = .00 R 2 = .04 p = .40 R 2 = .12 p = 01 Note. Z = 0.72 and 0.72 for low and high trait guilt, respectively (i.e., one standard deviation above and below the mean of the centered trait guilt variable). Differences in simple effect were computed by subtracting the effects for high trait guilt from the effects for low trait guilt. Tests of differences for the indirect and total effect were based on confidence intervals derived from b ootstrap estimates. 1 st Stage: Path from WIF to negative emotion; 2 nd Stage: Path from negative emotion to PB. p < 05. **p < .01
47 Chapter Four Discussion The purpose of the current study was to examine the relationship between of WFC that reside in the family domain has b een considered as a gap in the work family literature (Eby et al., 2005). Investigating the relationship between WIF and different types of PB, this study adds meaningful contribution to the literature. Additionally, the mediating role of negative emotion in the relationship between WIF and the PBs as well as the moderating role of trait guilt were examined to further understand the mechanism Some general findings deserve attention before discussi ng results of hypotheses testing. First, there was a significant negative relationship between the number of work hours and the number of hours of direct interaction with the child. Thus, the longer parents worked the less time they spent to interact with the child. Interestingly, the impact of hours of work on PB seemed to be different depending on the type of behavior in that only the two types of active parenting behavior were negatively associated with the working hours. Even when parents have to devote more time for their work, they may still need to carry out PMPB. Results suggest that this may not be the case for PRPB and CAPB as parents do not necessarily go beyond PMPB.
48 Second, parents tended to engage in more CAPB than PRPB. Data suggested that par ents do CAPB 2 3 times a week on average while doing PRPB once or less than once a week. This highlights two key differences between PRPB and CAPB. Although the both require active involvement of parents, physical activity (e.g., to exercise together) migh t be perceived more demanding and taxing than cognitive activity (e.g., to have discussions about achievements or concern). Moreover, recreational activities (e.g., playing indoor games together) are more likely to be thought as optional whereas academic o riented activities (e.g., helping the child to do homework) may be regarded as essential. Lastly, gender was significantly associated with PB. Consistent with previous research on time usage in the parenting literature (e.g., Bianchi Milkie, Sayer, & Robi nson, 2000 ), women did more CAPB and PMPB than did men. Also, women tended to spend longer time in interacting with their children than men. Interestingly, there was no gender difference with regard to PRPB. No gender difference in PRPB may be due to diffe rentiated parental roles; women tend to be the primary caregiver whereas men primarily serve a playmate role (Lamb, 1996). Accordingly, relatively equal participation across men and women is expected in PRPB compared to other PBs. As hypothesized, both tim e and strain based WIF were negatively associated with active parenting behaviors (i.e., PRPB and CAPB). The relationship between WIF and CAPB, however, was no longer significant when control variables were taken into consideration. Contrary to expectatio n, PMPB was not significantly correlated with WIF. Support was also found for the hypotheses that maintain positive relationships between time and strain based WIF and negative emotion.
49 No evidence was found for the mediating role of negative emotion in t he relationship between WIF and PB, with an exception ; the Sobel test indicated that the indirect effect of negative emotion was significant in the relationship between time based WIF and PMPB. It is notable that nonsignificant relationships were found bet ween negative emotion and parenting behaviors. That is, negative emotions that parents experience did not seem to be determinants of PBs. Support for hypothesis that concerns the moderating role of trait guilt in the relationship between negative emotion a nd PB was found only for PRPB. Specifically, when negative emotion due to WIF was experienced, parents who are high on trait guilt engaged in more PRPB whereas parents who are low on trait guilt did less PRPB. Interestingly, exploratory analysis found that trait guilt moderates relationships between WIFts and both types of active PB. With regard to PRPB, all parents reported reduced amount of PRPB when experiencing high WIFts. However, the decrease was larger for parents who are low on trait guilt compared to those who are high on trait guilt. In terms of CAPB, only parents who are low on trait guilt engaged in less CAPB when they experience high WIFts; parents who are high on trait guilt reported approximately the same amount of CAPB regardless of the level of WIFts. Main Findings WIF and parent child interaction behavior. A negative relationship was hypothesized between time and strain based WIF and PB. As WIF describes a situation where work demands inhibits effective functioning in family domain, parents who report high level of WIF were expected to engage in less PB. Results indicate that work indeed hampered successful performance of the parent role by bringing strain and leaving no
50 sufficient time and energy for interaction with the child. This was par ticularly true for PB that requires gross involvement of parents. (Crouter & McHale, 2005). Findings from the current study shed light on the mechanism by which the work influence s child development; the work may place a barrier for parents in carrying out certain PB. When WIF occurs, activities that parents are able to or willing to engage in with their child seems to be limited. Especially, the results showed that such negative r elationship is strongest between PRPB and strain WIF. Fatigue and strain from work may preclude participation in physical and recreational activities for parents. Interestingly, PMPB showed no relationship with both time and strain based WIF. This indicat es that the impact of work is not necessarily identical across different types of PB. Therefore, future researchers should consider unique characteristics of each Negative emotion. It was hypothesized that WIF and negative emotion are positively correlated because WIF signals that individuals failed to meet the goal of being a good parent. The hypothesis was supported. Both time and strain based WIF demonstrated positive relationship with negative emotion. WIF predicted negative negative emotions, into consideration. Highlighting affective aspect of WIF, the finding echoes the call for more research on affective experience in the work family literature (MacDermid, Seery, & Weiss, 2002). Also hypothesized was the mediating role of negative emotion in the relationship between WIF and PB. Based on prior research on emotion and parenting, negative
51 emotion th at is elicited from WIF was expected to explain variability in PB. No evidence was found for this hypothesis. Nonsignificant relationship between negative emotion and PB seemed to be one of possible reasons for the null results. Contrary to the past resear ch that has established the link between emotion and behavior (Frijda, 1988), the current There are several reasons to explain the results. Emotion may affect PB under certain circumst ances only. For example, negative emotion might need to be caused by children to make parents act in a certain way towards them. D ix G ershoff M eunier and M iller (2004) argued that it is important to consider not only type of specific emotion but also re ason and motivation different PB even though they are experiencing the same emotion. In the case of the current study, negative emotion might have not exerted a strong influence on PB because it occurred due to work. As an alternative explanation, it is perhaps the target of the behavior that matters. Granted concerns of parents for their ch ildren, parents may not act solely based on their emotion when they interact with their children. In sum, the results underscore that the origin of emotion and the target of behavior should be considered in understanding the association between emotion and behavior. Trait guilt. The moderating role of trait guilt was hypothesized in the relationship between negative emotion and PB. In support of previous research (Tangney, 1990), the hypothesis was supported such that the relationship between negative emoti on and PRPB is weaker for parents who are higher on trait guilt than for those who are lower on trait guilt. Parents who are predisposed to experience guilt might have not decreased the level
52 of PB when experiencing negative emotions because they felt resp onsibility for WIF and had higher motivation to take corrective actions. The very choice of PRPB may be explained by the tendency of guilt Leith & Baumeister 1998) When parents who are predisposed to feel guilt experience negative emotion that elicited from WIF, they might tried to correct the situation by doing something that children like most, such as playing a fun game together or going to the zoo together. Considering demands and difficulty to engage i n PRPB, the results suggest that trait guilt is a powerful driver for PB. Theoretical Implications The current study expands the literature by examining PB as a consequence of icity hypothesis, which posits that factors associated with a given domain relate to conflict originating from that domain and that the consequences of the conflict tend to be found in the domain that is interfered (Frone, 2003). Supporting the hypothesis, a negative relationship was found between WIF and active PBs. That is, WIF that is originated from the work domain related to PB in the family domain. Identifying a relationship between WIF and child outcomes is especially meaningful in that it addresses an understudied topic in the work family literature (Eby et al., 2005). The current study also expanded the work family literature by investigating emotions associated with WIF. A number of scholars have called for more research on affective experience of the work family interface (MacDermid, Seery, & Weiss, 2002). The present study answers this call and underscores the affective nature of the work family interface by demonstrating a significant relationship between WIF and negative
53 emotion. This finding ha s an implication for the goal pursuit literature as well. WIF was expected to elicit negative emotion because people experience negative emotion when they fail to achieve their goals (Emmons, 1986). WIF represents a form of goal conflict, which occurs due to the fact that people pursue more than one goal at a time (Markus & Nurius, 1986) and that they need to spend substantive resources in terms of time and energy in order to obtain the goals (Kahneman, 1973). The positive relationship between WIF and negat ive emotion is consistent with this literature. The present study further provides important contributions to the literature by gaining insights into the process by which WIF relates to PB. Although limited support was found for both mediation and moderati on hypotheses, the results are still theoretically meaningful. Generally speaking, findings from the current study raise an important issue that specific types of PB need to be considered in the future research that investigates the determinants of PB. Wit h regard to the mediating hypothesis, this study provides preliminary evidence that negative emotion mediates the relationship between WIF and PMPB. Given that very little research has examined the underlying mechanisms linking the work domain and child ou tcomes, the current study expands the literature and shows a fruitful area for future research. In terms of the moderating hypothesis, results showed that trait guilt is a significant moderator in the relationship between negative emotion and PRPB. The exa mination of guilt contributes to the literature because little empirical research has been done despite the fact that several scholars have discussed theoretical relevance of the construct in the context of work and family (Allen, in press; Judge, Ilies, & Scott, 2006). Especially, the current study is theoretically meaningful as it expands the literature
54 by investigating guilt as a stable individual difference. Although empirical evidence has been found for the relationship between negative affect and beha vior at home (Ilies et al., 2007), dispositional variables that may attenuate the relationship have rarely been a topic of research. With support for the idea that trait guilt may serve a boundary condition in the relationship between emotion and behavior in the family domain, the present study calls for more research that takes individual differences into account. Practical Implications In addition to the abovementioned implications for the literature, the current study has practical implications as well. Results indicated that behavior in the family and strain based WIF, which was positively associated with the duration of work, appeared to be a significant predictor of active PB. Considering fundamental role of the active PB in child development (Epstein, 1985; Tamis LeMonda, Uzgiris, & Bornstein, 2002) and the work that deprives parents of the active interaction with their children is expected to bring negative consequences to the organization in the long run. Therefore, it is imperative for organizations to consider multiple roles that their employees serve outside the work. For instance, organizations may want to offer family friendly policies such as flextime that enable employees to participate in a variety of activities with their children. The results also highlighted the potential benefit of trait guilt. In current study, guilt pronen ess appeared to attenuate negative relationship between WIF and PRPB. This suggests that an adequate level of guilt may contribute to parent child relationship and child development as it ensures a necessary developmental interaction for children.
55 Limitati ons There are several limitations that should be mentioned. First, the research design used in the current study is cross sectional. Although it is theoretically sound to consider WIF as an antecedent of PB, the nature of the study precludes any inferences about causality. For instance, it is possible that parents who participated in less PB perceive more WIF. The cross sectional design is especially limiting because one time survey may not be an adequate method to study emotion that is a transient affectiv e state. Future researchers should use the experience sampling method (ESM) to capture more dynamic relationships among the study variables. Second, the present study collected data via self report only, which raises an issue of common method bias. However correlation analysis revealed that the hours of work had similar relationships with other study variables (PB, negative emotion, and trait guilt) as time and strain based WIF did. Taking the hours of work as an objective indicator of work, this suggests that WIF was a subjective report but reflected objective reality as well. With regard to emotion, self report was necessary as individual was assumed to be most knowledgeable about their own feeling. On the other hand, PB could be collected from multiple sources, including the focal child or the spouse, in order to minimize potential memory bias in the self report. Future study should also consider using objective data to gain an accurate and comprehensive picture of PB. For instance, d ifferent indicators of child health or school achievement can be incorporated into the future research to further examine outcomes.
56 A f inal l imitation involves the scale for parent child interaction behavior that was developed for the present study. Although effort had been put throughout the scale development to ensure validity of the measurement, further research endeavor is warranted to find further evidence for validity of the scale. Also, the current scale measures the freq uency of each behavior, with an assumption that certain behaviors are always active and others are passive. However, some activities could be either active or passive depending on the level of engagement that actually occurs while parents are participating For example, some parents may not actually interact with their children while they are exercising together. Or, some parents could have a good conversation with their children while they are doing housework together. Therefore, future research should try to examine qualitative differences in each PB. Future Directions The current study is one of few attempts to examine links between work and parenting behavior. Key findings of the study provide many ideas for future research. First, future study should fu Such information will enable us to pinpoint specific feature of each occupation that affects PB more adversely than others. By utilizing other data sources such as O*Net, future research may gain insight into whether there are particular groups of parents who are obstructed to do PB due to their work. work affects PB. The current study did not find evidence for th e mediating role of emotion between WIF and PB. Therefore, future research will benefit from delving into the reason for no significant relationship between negative emotion and PB. For instance,
57 future research may investigate if PB is not an emotion driv en behavior at all or if the majority of parents successfully employ coping strategies to regulate their emotion. Third, the impact of trait guilt on parents themselves is another fruitful area for the future research. Although trait guilt turned out to be beneficial for children in that guilt prone parents maintained the level of active PB despite WIF and negative emotion, constant contemplation over the failure and taking responsibility may not be healthy for parents themselves. Especially, if the fundame ntal cause for WIF is something that employees cannot easily change, engaging in PB is could take a toll on parents in the long run. Finally, the positive influence of work on the family domain should also be further studied. The work family literature has paid much more attention to negative interface of work and family (Eby et al., 2005). Along with recent movement in the work family literature that attempts to address facilitation between work and family (WFF), future study will contribute to the literat ure by investigating the relationship between WFF, positive emotion, and PB. Conclusion The present study addressed the consequence of WFC in family domain, an important but understudied topic in the work family literature. Specifically, the link between W work was significantly related to parenting behavior and that a dispositional tendency to experience guilt attenuates the negative relationship between the two. Reiterating findings suggests that further investi gation may be worthwhile for work family research.
58 Reference s Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting inte ractions. Thousand Oaks, CA: Sage. Allen, T. D. (in press). The work family interface. In S. Kozlowski (Ed). Oxford Handbook of Industrial and Organizational Psychology Oxford University Press. Baron, R. M., & Kenny, D. A. (1986). The moderator mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173 1182. Baumeister, R. F., Stillwell, A. M., & Heatherton, T. F. (1994). Guilt: An interperso nal approach. Psychological Bulletin, 115, 243 267. Berenbaum, H., Fujita, F., & Pfennig, J. (1995). Consistency, specificity, and correlates of negative emotions. Journal of Personality and Social Psychology, 68, 342 352. Bianchi, S. M., Milkie, M. A., S ayer, L. C., & Robinson, J. P. ( 2000 ). Is a nyone doing the housework? Trends in the gender division of h ousehold labor. Social Forces, 79, 191 228. Booth, C. L., Alison Clarke Stewart, K., Vandell, D. L., McCartney, K., & Owen, M. T. (2002). Child care usa ge and mother infant quality time. Journal of Marriage and the Family, 64, 16 26.
59 Carlson, D. S., Kacmar, K. M., & Williams, L. J. (2000). Construction and initial validation of a multidimensional measure of work family conflict. Journal of Vocational Be havior, 56 249 276. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). New York: Academic Press. Cooksey, E. C., & Fondell, M. M. (1996). Spending time with his kids: Effects of family lives. Journal of Marriage and the Family, 58, 693 707. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika 16 297 334. Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychologi cal Bulletin, 52, 281 302. Crouter, A. C., & McHale, S. M. (2005). The long arm of the job revisited: Parenting in dual earner families. In T. Luster., & L. Okagaki (Eds). Parenting: An ecological perspective (pp. 275 296 ). Mahwah, NJ : Lawrence Erlbaum As sociate s. Cunningham, M. R. (1988). What do you do when you re happy or blue? Mood, expectancies, and behavioral interest. Motivation and Emotion, 12, 309 331. Dix, T. (1991). The affective organization of parenting: Adaptive and maladaptive processes. Psy chological Bulletin, 110 3 25. D ix T., G ershoff E. T., M eunier L. N., and M iller, P. C. (2004). The affective structure of supportive parenting: Depressive s ymptoms, immediate emotions, and child oriented m otivation Developmental Psychology, 40, 1212 1227
60 Duncan, S. C., Duncan, T. E., & Strycker, L. A. (2000). Risk and protective factors influencing adolescent problem behavior: A multivariate latent growth curve analysis. Annals of Behavioral Medicine, 22, 103 109. Eby, L. T., Casper, W. J., Lockwood, A., Bordeaux, C., & Brinley, A. (2005). Work and family research in IO/OB: Content analysis and review of the literature (1980 2002). Journal of Vocational Behavior, 66 124 197. Edwards, J. R., & Lambert, L. S. (2007). Methods for integrating moderation and mediation: A g eneral analytical framework using moderated path a nalysis Psychological Methods, 12, 1 22. Efron, B., & Tibshirani, R. (1993). An introduction to the bootstrap. New York: Chapman & Hall. Emmons, R. A. (1986). Personal Strivings: An appr oach to personality and subjective well being. Journal of Personality and Social Psychology, 51, 1058 1068. Epstein, J. L. (1985). Home and school connections in schools of the future: Implications of research on parent involvement. Peabody Journal of Educ ation, 62 18 41. Frijda, N. H. (1988) The laws of emotion. American Psychologist, 43, 349 358. Frijda, N. H. (1994). Varieties of affect: Emotions and episodes moods, and sentiments. In P. Ekman & R. J. Davidson (Eds.), The nature of emotion: Fundamenta l questions (pp. 59 67). New York: Oxford University Press. Frone, M. R. (2003). Work family balance. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health psychology (pp. 143 162 ). Washington, DC: Ameri can Psychological Association.
61 Frone M. R., Russell, M., & Barnes, G. M. (1996). Work0family conflict, gender, and health related outcomes: A study of employed parents in two community samples. Journal of Occupational Health Psychology, 1, 57 69. Gauthier, A. H., Smeeding, T. M., & Furstenb erg, F. F. (2004). Are parents investing less time in children? Trends in selected industrialized countries. Population and Development Review, 30, 647 671. George, J. M., & Brief, A. P. (1996). Motivational agendas in the workplace: The effects of feeling s on focus of attention and work motivation. In B. M. Staw & E. E. Cummings (Eds). Research in organizational behavior (pp. 75 109). JAI Press. Gilbert, L. A., Holahan, C. K., & Manning, L. (1981). Coping with conflict between professional and maternal rol es. Family Relations, 30, 419 426. Greenhaus, J. H., Allen, T. D., & Spector, P. E. (2006). Health consequ ences of work family conflict: The dark side of the work family interface. In P. L. Perrewe & D. C. Ganster (Eds). Research in occupational stress and well being 5 (pp. 61 99). JAI Press/Elsevier. Greenhaus, J. H. & Beutell, N. J. (1985). Sources and conflict between work and family roles. Academy of Management Review 10 76 88. Grossman, F. K., Pollack, W. S., & Golding, E. (1988 ). Fathers and child ren: Predicting the quality and quantity of fathering. Developmental Psychology, 24, 82 91. Harder, D. W., & Zalma, A. (19 90 ). Two promising shame and guilt scales: A construct validity comparison. Journal of Personality Assessment, 55, 729 745. Hammer, L. B., Cullen, J. C., Neal, M. B., Sinclair, R. R., & Shafiro, M. (2006). The longitudinal effects of work family conflict and positive spillover on experiences
62 of depressive symptoms among dual earner couples. Journal of Occupational Health Psychology, 11, 249 265. Hinkin, T.R. (1995). A review of scale development practices in the study of organizations. Journal of Management 21 967 988. Hofferth, S. L., & Sandberg, J. F. (2001). How American children spend time. Journal of Marriage and Family, 63, 295 3 08. Ilies, R., Schwind, K., Wagner, D. T., Johnson, M., DeRue, D. S., & Ilgen, D. R. (2007). When can employees have a family life? The effects of daily workload and affect on work Family conflict and social activities at home. Journal of Applied Psycholog y, 92 1368 1379 James, L. R., & Brett, J. M. (1984). Mediators, moderators, and tests for mediation. Journal of Applied Psychology, 69, 307 321. Judge, T. A., Ilies, R., & Scott, B. A. (2006). Work family conflict and emotions: Effects at work and at hom e. Personnel Psychology, 59, 779 814. Kahn, R. L., Wolfe, D. M., Quinn, R., Snoek, J. D., & Rosenthal, R. A. (1964). Organizational stress. New York: Wiley. Kahneman, D. (1973). Attention and effort. Englewood Cliffs, JN: Prentice H all. Kahneman, D., Krueg er, A. B., Schkade, D. A., Schwarz, N., & Stone, A. A. (2004). A survey method for characterizing daily experience: The day reconstruction method. Science, 306, 1776 1780. Kossek, E., & Ozeki, C. (1998). Work family conflict, policies, and the job life sat isfaction relationship: A review and directions for organizational behavior human resources research. Journal of Applied Psychology, 83, 139 149.
63 Lamb, M. E. (1996). The role of the father in child development (3 rd ed .) New York: Wiley Lazarus, R. S. & C ohen Charash, Y. (2001). Discrete emotions in organizational life. In R. L. Payne & G. L. Cooper (Eds.), Emotions at work: Theory, research, and applications for management (pp. 45 81). Chichester, UK: Wiley. Leith, K. P., & Baumeister, R. F. (1998). Empat hy, shame, guilt, and narratives of interpersonal conflicts: Guilt prone people are better at perspective taking. Journal of Personality, 66, 1 37. Livingston, B. A., & Judge, T. A. (2008). Emotional responses to work family conflict: A n examination of gender role orientation working men and women. Journal of Applied Psychology, 93, 207 216. Maccoby, E. (1980). Social development Psychological growth and the parent child relationship. New York: Harcourt Brace Jovanovich. MacDermid, S. M., Seery, B. L., & Weiss, H. M. ( 2002 ). An emotional examination of the work family interface. I n R. G. Lord, R. J. Klimoski, & R. Kanfer (Eds.), Emotions in the workplace (pp.402 427). San Francisco: Jossey Bass. Major, D. A., Allard, C. B., & Cardenas, R. A. (2004). Child health: A legitimate business concern. Journal of Occupational Health Psychology, 9, 306 321. Marks, S. R. (1977). Multiple roles and role strain: So me notes on human energy, time, and commitment. American Sociological Review, 42, 178 19 2. Markus, H., & Nurius, P. (1986). Possible selves. American Psychologist, 41, 954 969. Martin, B. (1975). Parent child relations. In F. D. Horowitz (Ed.), Review of child development research (Vol. 4). Chicago: University of Chicago Press.
64 Nock, S. L., & time commitments. Social Forces, 67, 59 85. Norman, D. A., & Bobrow, D. B. (1975) On data limited and resource limited processes. Cognitive Psychology, 7 44 64. Nunnally, J. C. ( 197 3 ) Research strategies and measurement methods for investigating human development. In J. R. Nesselroade, & H. W. Reese (Eds.), Life span developmental psychology: Methodological issues Oxford, U.K. : Academic Press Peterson, C., Maier, S.F., & Se ligman M. E. P. (1995). Learned helplessness: A theory for the age of personal control. New York: Oxford University Press. Preacher, K. J. & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Resear ch Methods, Instruments, & Computers, 36, 268 277. Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39, 1161 1178. Sallinen, M., Ronka, A., Kin nunen, U., & Kokka, K. (2007). Trajectories of depressive mood in adolescents: Does parental work or parent adolescent relationship matter? A follow up study through junior high school in Finland. International Journal of Behavioral Development, 31 181 190. Sayer, L. C., Bianchi, S. M., & Robinson, J. P. (2004). Are parents investing less in American Journal of Sociology, 110, 1 43. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effect in structural equation models. Sociological Methodolo gy, 13, 290 312.
65 Tamis LeMonda, C. S., Uzgiris, I. C., & Bornstein, M. H. (2002). Play in parent child interactions. In M. H. Bornstein (Ed.), Handbook of Parenting (pp. 221 241). Mahwah NJ: LEA publishers. Tangney, J. P. (1990). Assessing individual diffe rences in proneness to shame and guilt: development of the self conscious affect and attribution inventory. Journal of Personality and Social Psychology, 59, 102 111. Tangney, J. P., & Dearing, R. L. (2002). Shame and guilt New York: Guilford Press. Van H ook, E., & Higgins, E. T. (1988). Self related problems beyond the self concept: Motivational consequences of discrepant self guides. Journal of Personality and Social Psychology, 55, 625 633. Walters, J., & Stinnett, N. (1971). Parent child relationships: A decade review of research. Journal of Marriage and the Family, 33, 70 90. Watson, D. (1988). Intraindividual and interindividual analyses of positive and negative affect: Their relation to health complaints, perceived stress, and daily activities. Journ al of Personality and Social Psychology, 54, 1020 1030. Watson, D., & Clark, L. A. (1994). The PANAS X: Manual for the positive and negative affect schedule Expanded form. Unpublished manuscript, University of Iowa. Weiss, H. M. (2002). Conceptual and empi rical foundations for the study of affect at work. In R. G. Lord, R. J. K limoski & R. Kanfer (Eds.), Emotions in the workplace: Understanding the structure and role of emotions in organizational behavior (pp. 20 63). San Francisco: CA: Jossey Bass. Zarski, J. J. (1984). Hassle and health: A replication. Health Psychology, 3, 243 251.
67 Appendix A WIF Scale Items Time based Work to f amily Conflict 1. My work keeps me from my family activities more than I would like. 2. The time I mus t devote to my job keeps me from participating equally in household responsibilities and activities. 3. I have to miss family activities due to the amount of time I must spend on work responsibilities. Strain based Work to f amily Conflict 1. When I get home fro m work I am often too frazzled to participate in family activities/responsibilities 2. I am often so emotionally drained when I get home from work that it prevents me from contributing to my family. 3. Due to all the pressures at work, sometimes when I come hom e I am too stressed to do the things I enjoy. Carlson, Kacmar, & Williams (2000).
68 Appendix B Negative Emotion Scale Items Please r ead each item and then write the appropriate answer in the space next to that word. Indicate to what extent you have felt this way wh en you r work interfer ed with family within the past 4 weeks Use the following scale to record your answers: 1 2 3 4 5 V ery slightly A little M oderately Q uite a bit E xtremely or not at all 1. S ad 2. D isgusted 3. A fraid 4. D ownhearted 5. Lo nely 6. D istressed 7. H ostile 8. S cornful 9. A lone 10. I rritable 11. L oathing 12. A ngry 13. B lue 14. D epressed 15. M iserable 16. Gl oomy 17. F rustrated 18. A nnoyed Watson & Clark (1994 ) ; Russell (1980)
69 App endix C Trait Negative Affectivity Scale Items This scale consists of words that describe different feelings and emotions. Read each item and then mark the appropriate answer in the space next to that word. Indicate to what extent you have felt this way in general that is, on the average Use the following scale to record your answers: 1 2 3 4 5 V ery slightly A little M oderately Q uite a bit E xtremely or not at all 1. S cared 2. N ervous 3. J ittery 4. U pset 5. A shamed Watson & Clark (1994 ).
70 Appendix D Trait Guilt Scale Items Read each item and then write appropriate answer in the space next to that word. Please indicate how common the feeling is for you in genera l 0 1 2 3 4 Never Rarely Some of the time Frequently but not Continuously or almost continuously continuously 1. M ild guilt 2. W orry a bout hurting or injuring someone 3. I ntense guilt 4. R egret 5. R emorse 6. F eeling you deserve criticism for what you did Harder & Zalma (1990).
71 Appendix E Time with Children We are interested in the types of activities that you and your child do together. While co mpleting this survey p lease think about THE PAST 4 WEEKS and focus on ONE CHILD whose age is between 7 and 9 If you have more than one child within the age range, focus on the older child. A list of child activities is provided below. Some activities m ay occur more often than others. Please indicate how many times you participated in each activity with this child during the past 4 weeks. U se the following scale to record your answers: 1 2 3 4 5 6 7 Never 1 5 times 6 10 times 11 15 times 16 20 t imes 21 25 times 25 times or more 1. I help my child with his/her homework 2. I read to my child 3. My child and I have discussions about my child s achievements or concern 4. My child and I play together (e.g., bike riding, playing sports ). 5. I play out side with my child 6. My child and I exercise together 7. I go on outings with my child (e.g., museum, zoo, sporting event) 8. I play indoor games with my child (e.g., board games, video games). 9. My child and I talk while we are driving together 10 My child and I do housework together 1 1 My child and I go shopping together 12 My child and I watch TV together 13 My child and I do grocery shopping together Direct interaction with the child: Average weekday time spent together: Average weekend day time spent together:
About the Author in 2007. She is a Ph.D. student at the University of South Florida since 2007. Her researc h interests are work and family issues with a focus on its health implication and emotion at the workplace. She has been coa uthored articles in top tier industrial and organizational psychology journals, including Personnel Psychology and Human Performance She has also presented at several professional conferences, including the Society for Industrial and Organizational Psychology and Academy of Management.