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-SFE0003444
Implicit affect and alcohol outcome expectancies
h [electronic resource] /
by John Ray.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (M.A.)--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: Expectancy theory provides a useful framework within which to examine the link between cognitive representations of anticipated alcohol related outcomes and affective processes that ought to shape behavior at the level of implicit, or automatic, processing. The role of affect in alcohol expectancies is an important one as it reflects the approach-avoid contingency associated with reward learning presumed to underlie addictive processes. This study examined the relationship between affect and expectancy operation by using suboptimally presented alcohol related cues to prime affectively congruent evaluations of otherwise unrelated targets. Hypotheses predicted that drinkers who reported higher positive and arousing expectancies for alcohol outcomes would make affective evaluations (but not semantic categorizations) more accurately when target stimuli were preceded with an alcohol picture or word prime. Analysis of drinking and expectancy variables revealed positive relationships between drinking frequency and social/physical pleasure expectancies, as well as tension reduction expectancies. No relationships were found between drinking quantity and expectancies. Evaluation response accuracy was not related to alcohol expectancies. Discussion centers on potential reasons for lack of findings, including experimenter error and design limitations.
Advisor: Mark S. Goldman, Ph.D.
t USF Electronic Theses and Dissertations.
Implicit Affect and Alcohol Outcome Expectancies b y John M. Ray A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Psychology College of Arts and Sciences University of South Florida Major Professor: Mark S. Goldman Ph.D. Jon Rottenberg, Ph.D. Kristen Salomon, Ph.D. Date of Approval: March 30, 2010 Keywords: drinking, college, cognition, priming, cue reactivity Copyright 2010 John M. Ray
Table of Contents List of Tables i ii List of Figures v Abstract vi Introduction 1 Affect in Expectancies 2 Affective Priming 4 Suboptimal Affective Priming 8 Affective Priming Cues 9 Preliminary Findings 11 Specific Aims 13 Hypotheses 13 Method 14 Participants 14 Sample Characteristics 15 Procedure 16 Affective Priming W ith Words 17 Affective Priming W ith Pictures 18 Semantic Priming W ith Words and with Pictures 18 Measures 18 Response Window Procedure 18 Picture Stimuli 20 Word Stimuli 21 Questionn aires and Written Assessments 21 Post trial Measures 23 Debriefing 23 Results 23 Descriptive Statist ics for Independent Variables 23 Drinking Behavior 23 Alcohol Expectancies 24 Behavioral Inhibit ion/Activation 25 Within Se ssion Affect 27 i
Relationships Between Alcohol Expectancies and Drinking Variables 28 Descriptive Stati stics for Dependent Variables 30 Response Data 30 Relationships Between Alcohol Expectancies and Dependent Variables 31 Discussion 35 References 41 ii
List of Tables Table 1 Descriptive Statistics for Drinking Variables 24 Table 2 Descriptive Statistics for Alcohol Expectancy Questionnaire Scales 25 Table 3 Descriptive Statistics for Behavioral Inhibition and Activation Scales 26 Table 4 Correlations Between BIS/BAS and Drinking Variables 26 Table 5 Correlations Between BIS/BAS and AEQ Scales 26 Table 6 Correlations Between BIS/BAS and AEMax Scales 27 Table 7 Descripti ve Statistics for Pos itive and Negative Affect Scales 28 Table 8 Correlations Between AEMax Scales and Drinking Variables 28 Table 9 Quantity Correlations By Investigator 29 Table 10 Corr elations Between AEQ Scales and Drinking Variables 30 Table 11 Descriptiv e Statistics for Affective Task Response Accuracy 31 Table 12 Descripti ve Statistics for Semantic Task Response Accuracy 31 Table 13 Correlations Between AEMax Scales and Affective Task Response Accuracy 33 Table 14 Correlations Between AEMax Scales and Semantic Task Response Accuracy 33 iii
Table 15 Correl ations Between AEMax Scales and Affective Task Response Accuracy by Gender 34 Table 16 Correl ations Between AEMax Scales and Semantic Task Response Accuracy by Gender 34 i v
List of Figure s Figure 1 Trial Level Schematic of The Response Window 17 v
Implicit Affect and Alcohol Outcome Expectancies John M. Ray Abstract Expectancy theory provides a useful framework within which to examine the link between cognitive representations of anticipated alcohol related outcomes and affective processes that ought to shape behavior at the level of implicit, or automatic, processing The role of affect in alcohol expectancies is an important one as it reflects the approach avoid contingency associated with reward learning presumed to underlie addictive processes. This study examined the relationship between affect and expectancy operation by using suboptimally presented alcohol related cues to prime affectively congruent evaluations of otherwise unrelated targets. Hypotheses predicted that drinkers who reported higher positive and arousing expectancies for alcohol outc omes would make affective evaluations (but not semantic categorizations) more accurately when target stimuli were preceded with an alcohol picture or word prime. Analysis of drinking and expectancy variables revealed positive relationships between drinking frequency and social/physical pleasure expectancies, as well as tension vi
reduction expectancies. No relationships were found between drinking quantity and expectancies. Evaluation response accuracy was not related to alcohol expectancies. Discussion centers on potential reasons for lack of findings, including experimenter error and design limitations. v ii
1 Introduction The ability of animals to store information for later recall to aid in the interpretation of, and selection of behavioral response to, future events is the premise upon which Tolman (1932) emphasized the organizational aspect of learning. Memories of response outcome relationships enable an organism to predict outcomes from simil ar contingencies, often automatically. These memories and their associative linkages constitute expectancy templates, which guide behavior in response to familiarity derived from ongoing life events as they unfold (Goldman, 2002; Maddux, 1999). As reflecte d in expectancy operation, affect, which operates interactively with cognitive systems to guide decision making in the presence of multiple choices (Goldman, 2002; Goldman, Darkes, & Del Boca, 1999). Alcohol outcome expectancies comprise those templates representing direct or vicarious experiences with alcohol and anticipated effects of future use (Goldman, Brown, & Christiansen, 1987; Goldman, 2002). Expectancy theory provi des a useful framework within which to examine the link between cognitive representations of anticipated alcohol related outcomes and affective processes that ought to shape behavior at the level of implicit, or automatic, processing. Research has highligh ted several antecedent factors related to the onset and maintenance of problem drinking, including affect regulation, level of response (sensitivity) to alcohol, and tendency to engage in deviant behavior in
2 general, with each of these areas differing on d imensions of genetic contribution, environmental influence, and personality variability (Sher, Grekin, & Williams, 2005). Two decades of research have yielded abundant evidence that expectancies mediate the relationship between antecedent risk factors for drinking and actual drinking behavior (Brown, 1985b; Brown, Goldman, & Christiansen, 1985; Roehrich & Goldman, 1995), and that expectancies predict drinking behavior (Christiansen, Smith, Roehling, & Goldman, 1989). To demonstrate the mediating role of exp ectancies, Darkes and Goldman (1993; see also Dunn, Lau, & Cruz, 2000) used an expectancy challenge, which resulted in reduced drinking after six weeks among a sample of college students. rinker class delineations to alcohol expectancy dimensions: light drinkers tend to endorse the negative and sedating effects of alcohol, while heavy drinkers report more positive and arousing effects (Goldman, Darkes, & Del Boca, 1999). Affect in Expectancies The affective quality of alcohol expectancies has been demonstrated in generated by nearly 10,000 college aged drinkers (Rather, Goldman, Roehrich, & Brannick, 1992; Goldman & Rather, 1993; Rather & Goldman, 1994; Goldman words being affective in nature, (e.g, happy, horny, so cial) The role of affect in alcohol expectancies is an important one as it reflects the approach avoid
3 contingency associated with reward learning presumed to underlie addictive processes (Holland & Gallagher, 2004; e.g., Winkielman, Knutson, Paulus, & Tr ujillo, 2007). Robinson and Berridge (1993) discussed this relationship in terms of incentive sensitization. According to this hypothesis, drugs create real changes in the neural substrates of reward response, resulting in hypersensitization of the neural pathways associated with reward learning, so that drug related cues acquire salient properties previously associated with the drug itself (Berridge & Robinson, 2003). Essentially, anticipation of reward effects creates a state within the organism in which the cue activates behavior as effectively as if the reward relevant is the hypothesized role of these seeking behavior. Incentive sensitization theory posits that reward representation resulting in unplanned, unconscious stimulation of drug seeki ng behavior. Presumably, drinkers with stronger or more abundant associations between alcohol and positive outcome related representations would be particularly sensitive to such manipulation. Goldman (2002; Rather & Goldman, 1994) suggests that for heavi er of a drinking related concept is more likely to lead to activation of related representations and hence a greater range of positive and arousing expectations
4 for alcohol. Activation of expectancy network associations is not necessarily volitional, but is often automatic given the presence of a priming cue, e.g., environmental (bottle of beer) or internal (memory of drinking event). Several experiments have demonstrated the automaticity of expectancy activation through implicit priming (expectancy word priming; Stein, Goldman, & Del Boca, 2001; modified Stroop task; Kramer & Goldman, 2003; false me mory; Reich, Goldman, & Noll, 2004). As mentioned above, expectancies serve the basic purpose of guiding role of affect in this process is to afford the organism the ability to d iscriminate between an event that is to be approached (life preserving) and one that is to be avoided (life threatening). Obviously, the notion of subjective emotion at the t deleterious situations would bear little resemblance to what modern humans of mood), but is mo Affective Priming The study of automatic evaluations has increased significantly over the past few decades as soci al and cognitive theories of automatic processing have driven much research on human interactions, e.g., stereotyping and appraisals (Klauer & Musch, 2003). Work in this area has helped to generate a series of
5 procedures, such as the affective priming para digm, designed to study evaluative associations in memory. Affective priming, as first demonstrated by Fazio, Sanbonmatsu, Powell, and Kardes (1986) is defined as the facilitation of evaluative judgment of a stimulus following an affectively congruent prim ing stimulus. Theoretical discussion of the affective priming effect has involved some vigorous debate, centering mostly on mechanism. Spreading activation (Bower, 1991; Fazio et al., 1986; Neely, 1991), response competition (i.e., Stroop like mechanism; e .g., Klinger et al., 2000), and an affective matching mechanism (e.g., Klauer & Stern, 1992), are three models that have been proposed to explain the facilitation of affectively congruent prime target pairs. Fazio (2001) has argued that it is not likely th at one theory of mechanism explains affective priming. Rather, each likely contributes to the effect differentially, providing prime target relationship facilitates the spreading of like associations, or their congruency speeds processing as a result of their associative compatability, the significant outcome is the activation of a readiness state. Readiness confers upon an organism the ability to anticipate some outcome by calling on previously established contingencies. This is, of course, the fundamental premise of expectancy theory. Early affective priming research appeared to suggest that affective associations may be fragile and difficult to measure with more than minimal effe cts. The paradigm was scrutinized as a replicable measure when some researchers failed to replicate aspects of the findings of Fazio et al (1986). For
6 example, De Houwer and Eleen (1998) obtained associative, but not affective priming, and De Houwer, Herma ns, Rothermund, and Wentura (2002) failed to obtain affective priming in semantic tasks. Others failed to obtain effects using a pronunciation task (e.g., Glaser & Banaji, 1999; Glaser, 2003; Klauer & Musch, 2001). These apparent shortcomings may have been reflective of the complexities inherent in psychological phenomena rather than methodological or theoretical flaws (Wittenbrink & Schwarz, 2007). Further testing of the phenomenon has demonstrated an interactive effect of context such that the association set activated by a priming mechanism (as indexed by the presence of priming effects) depends on the instruction set provided to the participant (Klauer & Musch, 2003). Refined designs have revealed that activation of a set of associations in memory, such as that which facilitates evaluation of an associated stimulus, depends on how the participant has focused his attention; that is, what his operational goal is. For example, instructing the participant to focus on a non evaluative dimension of a stimulus ( e.g., whether it is a living or non living thing) typically shows no effect of affective congruency between the prime and target, while a focus on the affective dimension of the same pairing results in robust priming effects (e.g., De Houwer et al., 2002; Klinger, Burton, & Pitts, 2000; Klauer & Musch, 2002). The task dependence of the affective priming effect supports the notion that association sets do not operate independently in terms of the processes activating them, but according to the functional dem and being made of them. Associations can be activated in accordance with goal state, not merely as a
7 function of pure associative strength. This is a similar relationship to that suggested by the multidimensional space created by the mapping of alcohol out come expectancies, which are thought to reflect multiple, dynamic, and interactive affective and cognitive processes. Wittenbrink (2007) notes that this reflects a strength of the affective priming paradigm as an index of implicit processing in that it is not dependent on high accessibility of targeted concepts. regardless of relative strength between nodes. Wittenbrink ( ibid .) underscores this possibility in addressing the ap parent attentional conditionality (e.g., task dependence) of the paradigm, which suggests that it is not necessary that one holds a goal orientation toward a specific concept, but merely that an organism have a general attentional focus activated, for exam ple, to assess the goodness or badness of its surroundings. This is relevant to the instance of specific primes, such as alcohol cues, which are often quite complex and include interaction of internal and external, as well as personal and social goals. The automaticity of priming effects is supported by a number of studies that examined the interval between the onset of the prime and the onset of the target, the Stimulus Onset Asynchrony (SOA). In these studies, SOA was manipulated between 300 ms and 1000 m s (De Houwer, Hermans, & Eelen, 1998; Fazio et al., 1986, Experiment 2; Hermans et al., 1994, Experiment 1; Hermans, Spruyt, De Houwer,& Eelen, 2003). In each of these studies, priming was observed at the 300 ms SOA, but not at 1000 ms. Other research exam ining the effects of SOA variation found that priming effects are strongest between 0
8 ms and 300 ms, after which they begin to dissipate quickly (Hermans, De Houwer, & Eleen, 2001). Because conscious processes are presumed to be more time consuming, primin g effects observed at the shorter SOA, but not at the longer SOA provides strong, though indirect, evidence for automatic processing of the prime target relationship. Suboptimal Affective Priming Several studies have demonstrated the reliability of subopti mal priming (i.e, stimuli presented in such a way that conscious recognition is improbable) in eliciting basic affective reactions (Dimberg & Thunberg, 1998; Murphy & Zajonc, 1993; Niedenthal, 1990; Winkielman & Cacioppo, 2001). Rotteveel, DeGroot, Geutske ns, and Phaf (2001) found a stronger effect of suboptimal than optimal priming, as observed in both facial electromyography (EMG) values and subjective ratings of ideographs. Whether a priming stimulus can be considered subliminal has been subject to some (Murphy & Zajonc, 1993; Rotteveel et al., 2001), noting that a lack of awareness in tasks measuring conscious effects does not guarantee that all conscious processes have been circumvented. Winkielman, Zajonc, and Sch warz (1997), studied suboptimal affective priming using masked facial expressions. The authors found that the priming effect remained robust even when subjects were told what to expect to feel in response to suboptimal stimuli, suggesting that such priming Wilbarger (2005) conducted two experiments in which subjects increased both
9 consumption behavior (i.e., juice consumed by thirsty subjects) and willingn ess to pay for the juice immediately following suboptimal exposure to positive facial expressions, while subjects reported no change in subjective mood. Because mood ratings were obtained immediately after priming trials, the methodology utilized in these report any change in feeling was due to errors of attention or memory. While the salience of facial expressions makes evolutionary sense in terms of threat detection, Winkielman et al. (2005) suggest that, for modern humans, the influence of suboptimal facial expressions on approach avoidance behavior may involve more general changes in positive and negative affect. If this is the case, several classes of salient stimuli should evoke sim ilar behaviors even when presented suboptimally. Examples of potential stimuli include survival related pictures such as snakes or potential mates, social stimuli such as money or other such status symbols, and stimuli related to social behavior of specifi c groups. This study proposes to test this hypothesis by using alcohol pictures (in addition to words) to prime affective evaluations in drinkers whose alcohol expectancies presumably predispose them to attach positive, approach oriented meaning to represe ntations associated with alcohol related concepts. Affective Priming Cues Affective priming has been reliably demonstrated using words, drawings (Giner Sorolla, Garcia, & Bargh, 1999), photos of angry and happy faces (Murphy & Zajonc, 1993), familiar v. strange faces and names (Banse, 1999), and even odors (Hermans, Baeyens & Eleen, 1998), and several dependent
10 variables have been utilized in the affective priming paradigm, ranging from simple liking ratings (Murphy & Zajonc, 1 993) to consumption behavior (Winkielman et al., 2005). Much of this research has focused on the phenomenon itself, limiting experimental manipulations to those with the most robust effect sizes. This strategy has resulted in a rich literature supporting t he affective priming effect, but has left unexplored the role of other potentially influential cue types. Under normal viewing conditions, words are perceptually unambiguous. On the other hand, pictures are relativeley complex and thus potentially ambiguou s, especially given extremely brief exposures, such as those used in suboptimal priming. Much of the work demonstrating affective priming effects with pictures has utilized real facial expressions which are inherently salient stimuli (e.g., Winkielman et a l., 2005). Additionally, non face picture primes have typically involved simple line drawings, rather than life like depictions (e.g., Giner Sorolla et al., 1999). Of course, human environments are not limited to words, faces, and simple drawings. Rather, the stimuli these cues are theorized to represent are complex and often ambiguously perceived in most situations, given the sheer number of cues available at any moment in a given environment. Recently, affective priming has been shown using more varied pi ctures, for example, scenes of people and animals engaged in a variety of activities (Avero & Calvo, 2006), but the vast majority of studies have been limited to word pairings. The inclusion of alcohol pictures as affective primes in this study is apparent ly unique.
11 Preliminary Findings This study is a continuation of previous work examining the role of affect in activating alcohol outcome expectancies using pictures as primes (Ray, Darkes, & Goldman, 2007). In that recent study, participants grouped by exp ectancy endorsement (high v. low positive/arousing) viewed pictures of neutral objects and rated them on dimensions of valence and arousal. As hypothesized, affective priming with facial expressions (highly salient, potentially universal cues) was replicat ed; this was reflected in higher subjective ratings of face primed neutral pictures. Alcohol primed pictures were also rated higher, but not significantly, and no main effect of expectancy group was realized, though this also trended in the hypothesized di rection. The absence of hypothesized alcohol prime effects in the Ray et al. (2007) experiment was likely due to a combination of demand effects related to the subjective ratings instrument and a power deficit related to the diffusion of power across sourc es of variability within the trial presentation procedure. Participants were instructed to make their ratings as accurately as possible, but were given 2000 ms exposure time, plus 4000 ms intertrial interval time with which to make judgments. It is likely that even this apparently brief window allows for an unacceptable amount of deliberation, and variability, in ratings that are supposed to capture automatic processes. This study sought to minimize the influence of rating latency as a source of variability to increase the power needed to detect an effect of alcohol prime.
12 T he priming procedure found to achieve the strongest effects by minimizing power diffusion via speed/accuracy trade off is the response window technique developed by Greenwald, Draine, and Abrams (1996). The response window technique has been refined and used extensively by several investigators in subsequent years (e.g., Klauer & Musch, 2002; Klinger et al., 2000; Musch & Klauer, 2001). This procedure allows participants a very brief time to indicate whether they find the target to be positive or negative, with percentage correct being the dependent variable. Percentage correct, or accuracy, is defined in this paradigm by the number of evaluations which accurately reflect actual target vale nce. By restricting all respondents to similar latencies, speed is controlled and accuracy is used to index the priming effect. This procedure is outlined in detail in the Method section. This study utilize d a dissociation design in which identical affectivel y congruent prime target pairs we re presented in separate conditions that differ ed only on instruction set. Specifically, participants evaluate d the target affectively in one condition and categorize d it on a non affective dimension in another condition. It was anticipated that priming effects in the affect condition together with the absence of priming effects in the simple categorization condition would demonstrate the presence of an affective component activation in the evaluative trials. It sh ould be noted that this design wa s not intended to demonstrate independence of affect, but rather to illucidate the activation of affect beyond semantic activation alone.
13 Specific Aims Given the potentially impor tant role that implicit processing may play in alcohol expectanc ies, the purpose of this study wa s to further explore the relationship between alcohol outcome expectancies and the affective processes that influence behavior at a level beneath conscious awa reness. Much of the research on alcohol expectancies to date has focused on the cognitive activation of expectancies, while affective priming has been less well studied. By exploiting the automatic nature of expectancy activation, both cognitiv e and affect ive, this study aimed to elucidate this relationship using an affective priming paradigm. First, this study aimed to demonstrate the affective component of expectancy operation by using suboptimally presented alcohol related words to automatically activate (prime) affectively congruent evaluations of otherwise unrelated targets. Second, b ecause real world environments involve complex visual cues that cannot be adequately approximated by words, the inclusion of pictures as primes aimed to extend the research supporting alcohol expectancy theory by showing empirically that they are not limited to language based associations. Hypotheses 1. Drinkers who report greater positive and arousing expectancies will accurately evaluate a greater percentage of alcohol primed /positive target word presentations. a. There will be no expectancy related difference of accuracy for non affective categorization
14 2. Drinkers who report greater positive and arousing expectancies will accurately evaluate a greater percentage of alcohol primed /positive target picture presentations. a. There will be no expectancy related difference of accuracy for non affective categorization. Method Participants A sample of young adult drinkers was recruited from the University of South Florida campus via the rese arch participation pool during the Summer and Fall 2008 semesters. Age range was limited to 18 24 years, as this reflects the period of most frequent drinking among young adults (NIH, 2004) To examine potential differences in alcohol expectancies between genders, an effort was made to include equal numbers of males and females in the study. Previous studies have shown at least minimal gender differences within alcohol expectancies (e.g. Des Rosiers, Noll, & Goldman, 2002; Weinberger, Darkes, Del Boca, & Go ldman, 2003). Expectancy research suggests that males and females endorse alcohol expectancies similarly, but that variability in semantic meaning behind expectancy words may explain differences between genders. Standard drinking quantity and frequency que stions were included as part served as a screening instrument. Respondents were eligible if on the mass testing measure they reported being a drinker and right handed For the purpos es of this study, a drinker was defined as one who reported consuming
15 alcohol at least once per month Non drinkers were excluded via the mass testing screening procedure The Alcohol Expectancy Questionnaire (AEQ; Brown, Christiansen, & Goldman, 1987 ) was also administered as part of the participant pool mass testing battery. Only those potential participants who responded to the mass testing protocol and met eligibility requirements were able to view and sign up for this experiment via the participant poo l experiment system. Sample Characteristics The original sample included 101 participants, two of whom were excluded, one due to inability to successfully complete the practice session, the other because he fell asleep during the experiment Data from the remaining 99 participants were included for analyses. The final sample included ninety nine college aged students, with a mean age of 20.04 years (SD = 1.69). All participants were currently enrolled at the University of South Florida as full time, undergr aduate, college students. The sample was reflective of Tampa Bay Area demographics: 75.8 % Caucasian, 5.1% African American, 9.1% Hispanic, 4.0% Asian, and 6.1% other. Fifty three males and forty six females were enrolled in the study, and gender groups di d not differ in age [t(97) = .61, p > .05], or race Procedure Participants who completed the screening instrument and met minimum criteria were eligible to register for an ostensibly unrelated study in the Student Research Institute (SRI) lab (USF, PCD 2101). The true nature of the study (i.e.,
16 that it involves alcohol related stimuli) was concealed until debriefing to avoid potential contamination and demand effects related to alcohol use. Eligible participants attended a one time, fifty minute laboratory session All participants read and signed the IRB approved Informed Consent Document and were reminded of their volunteer status and given the opportunity to withdraw from the study They were then briefed on procedure, wh ich was described as part of a study of the effects of state affect on the ability to rapidly categorize briefly presented words and pictures; specifically, that participants would view words (condition 1) and pictures (condition 2) on a computer screen af ter each of which they would use a standard keyboard or keypad to make a categorization. Following the intake procedure, participants completed the trait version of the Positive And Negative Affect Scale (PANAS trait version; see Questionnaires and Written Assessments ) A state version of the PANAS was completed following each condition. Affective priming with words Each participant completed at least three forty trial practice blocks of irrelevant prime target pairings to establish baseline response tende ncies and to calibrate the response window (the Response Window Procedure is outlined in detail below). Four sixteen trial test blocks of word presentations followed, in which affective words were paired with either affectively congruent alcohol prime word s, affectively incongruent alcohol prime words, affectively congruent non alcohol prime words, or affectively incongruent non alcohol prime words For example, the prime word BEER might be paired
17 with the target word HAPPY a presumably affectively congruent pairing for drinkers with high positive/arousing expectancies. Each trial consisted of four components (see Figure 1) presented in order as follows: fixation icon (1000 ms cross), forward mask (400 ms), suboptimal prime (32 ms alcohol or non al cohol word), backward mask (32 ms), 1000 ms affectively polarized word (SOA = 64 ms). Participants were instructed to evaluate each target word as positive or negative (i.e., for valence) within the 133 ms response window, after which they would prepare f or presentation of the subsequent trial. Valence ratings/ categorization were made with designated key strokes on a standard computer keyboard. Figure 1 Trial Level Schematic of The Response Window
18 A ffective priming with pictures Picture priming sessio ns were identical to picture priming sessions with the exception that pictures were used instead of words, both as primes and as targets. For example, a prime picture of a BEER might be paired with a target picture of a PUPPY an affectively congruent pair ing for drinkers with high positive/arousing expectancies. Semantic priming with words and with pictures The semantic priming manipulation was identical to the affective priming manipulation, including identical primes, targets, and prime target pairings, except that participants were instructed to categorize targets according to a non affective dimension [i.e., single v. multiple syllables (words) or subjects (pictures)]. Following the completion of the experiment, participants completed the remainder of the written assessments, including the BIS/BAS, and the SAQ ( see Questionnaires and Written Assessments for a description of each ). Measures Response window procedure Greenwald, Draine, and Abrams (1996) designed the Response Window Technique, in which participants are given a very brief time to indicate whether they find a target to be positive or negative, with percentage correct being the DV. Percentage correct, or accuracy, is defined in this paradigm by the number of evaluations which accurately ref lect actual target valence or category. For example, a positive evaluation of a positive prime positive target pairing within the response window would be scored as correct response, because the target is positive Conversely, a negative evaluation of the same pairing would be scored as an incorrect response A
19 positive response to a positive prime negative target pairing would be scored as an incorrect response, because the target is negative Any response made outside the response window would be scored a s an incorrect response. By restricting all respondents to similar latencies, speed is controlled and accuracy is used to index the priming effect. All congruent pairings are hypothesized to facilitate responding, so that accuracy should be higher for thes e pairings For example, alcohol related prime positive target pairings should lead to more accurate responding for participants with higher positive/arousing expectancies for alcohol. In this procedure, participants are allowed a window of 133 ms within w hich to evaluate the target item. The response window is set with its center at 400 ms after target onset, so that the participant is to respond between 333 ms and 467 ms following presentation of the target. In order to minimize potential floor and ceilin g effects resulting from restriction of latency ranges, Musch and Klauer (2001), following Draine and Greenwald (1998), modified the window procedure to adapt to changes in individual performance. This adaptive response technique, initially centered at 400 ms following target onset, increases or decreases the window center by 33 ms at the end of each block according to performance in that block. The window center is decreased when the error percentage is less than or equal to 20% and mean response latency for that block does not exceed the current window by more than 100 ms. The window center is increased when the error percentage is greater than or equal to 45% and the mean response latency exceeds the current window by more than 100ms. If neither of these sets of conditions is met, the
20 window center is not changed. Only trials in which the participant responds in the interval between 100ms and 1000 ms after target onset are included in determining percentage correct scores. Participants ar e instructed that their goal should be to respond as accurately as possible and that all responses falling outside of the response window are considered incorrect. Opening of the window is marked by a change in the target from grayscale to color. Early res ponses result in no change in target properties, while an on time response allows the target to change to color 333ms after target onset, marking the beginning of the 133ms response window. When the response occurs during the window, the target is overlaid with a trial is initiated after an additional 400 ms have passed. When the participant fails to respond during the window, the target changes to back grayscale for 300ms afte r the end of the response window. The screen is then cleared, and the next trial is initiated after an additional 1000ms interval. Participants perform a minimum of three practice blocks of 40 irrelevant trials. Practice continues until there is no longer any adjustment of the window center. Participants then perform four 48 trial blocks, per the priming paradigm described. Picture s timuli Thirty two alcohol related pictures to be used as primes were selected from advertisements and the internet. Ninety six neutral pictures (thirty two primes and sixty four targets) were selected from the International Affective Picture Set (IAPS; Lang, hma n, & Vaitl, 1988).
21 Word s timuli Thirty two alcohol words to be used as primes were selected from The University of South Florida Word Association Norms (Nelson, McEvoy & Schreiber, 1998). Ninety six neutral words (thirty two primes and sixty four targets ) were selected from the Affective Norms for English Words (ANEW; Bradley & Lang, 1999). Questionnaires and written assessments Alcohol Expectancy Multiaxial Assessment Short Version (AEMax Short ; Goldman & Darkes, 2004). This measure includes 24 expectancy words which complete the phrase, Participants indicate how frequently they believe the newly constructed statement is true (7 point Likert: never to always). The AEMax has been shown to be both reliable and valid, dire ctly predicting later alcoho l use (Goldman & Darkes, 2004). Alcohol Expectancy Questionnaire (AEQ; Brown, Christiansen & Goldman, 1987; Goldman, Greenbaum & Darkes, 1997). The AEQ consists of 68 true/false statements to which the participant responds regarding the effects of alcohol. Items correlate with alcohol consumption and related behavior, as well as alcohol abuse, with a mean reliability of 0.84. This measure is comprised of six factors: global positive changes, sexual enhancement, physical and social pleasure, increased social assertiveness, relaxation and tension reduction, and arousal and aggression. The AEQ was administered as part of the participant pool mass testing battery. BIS/BAS Scale (Carver & White, 1994) This 20 item instrument is d esigned to assess sensitivity to the behavioral inhibition and activation systems
22 the BIS/BAS and its subscales ranging from .65 to .83 (Jorm, Christensen, Henderson, Jacomb, Korten, & Rodgers, 1999). High scores on the BAS subscales (Drive, Fun, and Reward) have been associated with higher levels of sensitivity to reward in reaction to alcohol related cues (Kambouropolous & Stager, 2001). The BIS/BAS was administered following the priming procedure. Positive And Negative Affect Scale (PANAS) (Watson, Clark, & Tellegen, 1988). The PANAS is a state and trait affect measure containing twenty negative inde x. The state and trait versions are differentiated by whether the instruction set refers to current, recent, or long term judgments of affect. The PANAS scale sensitive to chan ges over time and is considered one of the best measures of current mood (Crawford & Henry, 2004; Watson et al., 1988), as well as trait affect over time (Watson & Walker, 1996). The trait version of the PANAS was administered immediatley following the i ntake procedure. The state version was administered once before the testing session and once following each condition in order to capture change or stability of affect during the experimental protocol. Stimulus Awareness Questionnaire This measure was cre ated for this study and consists of a series of questions designed to assess the extent to which a participant was able to detect the presence of a priming stimulus. It is designed to be progressively specific, beginning with a general question of whether the participant noticed anything unusual at all, and building in the event
23 of affirmative responses to direct questions regarding the nature of the stimulus or stimuli. Post trial m easures Following test trials, the participant completed a post trial PANA S Scale to assess perceived state affect. By administering the affect self rating scale immediately post trial, any failure to report change in affect is not likely to be attributable to errors of memory, motivation, or attention. Subsequent behavior or ph ysiological indicators of emotion can be assumed to have occurred outside of conscious awareness (Berridge & Winkielman, 2003). The participant then completed the AE Max and BIS/BAS instruments. Debriefing Following the experiment protocol, participants were informed of the true nature of the study and completed the Stimulus Awareness Questionnaire to determine whether any of the subliminal stimuli were detected during presentation. No participants indicated detection of priming stimuli. Results Descripti ve Statistics for Independent Variables Drinking Behavior Drinking behavior was assessed using single item, multiple choice quantity and frequency measures (see Table 1). Drinking frequency ranged from one to seven days per week [M = 2.00(1.28)] and did n ot differ between males [M = 2.00(1.24)] and females [M = 2.02(1.36); t (97) = 0.08( p > .05)]. Drinking quantity ranged from one to eight or more drinks per occasion [M = 3.93(1.93)] Males [M = 4.53(2.05)] reported drinking more than females [M = 3.24(1.5 5); t (97) = 3.48 ( p < .05)]. Elevated skewness and kurtosis
24 values for the drinking behavior variables indicated non normal distribution. These variables were subjected to a natural log transformation, which were used in all subsequent analyses. Table 1 Descriptive Statistics for Drinking Variables N Min Max Mean SD Skewness Kurtosis Quantity Males 53 1 8 4.53 2.05 0.17 0.90 Females 46 1 8 3.24 1.55 1.04 1.09 Males(ln) 53 1 8 1.63 0.41 0.48 0.60 Females(ln) 46 1 8 1.38 0.35 0.15 0.29 Frequency Males 53 1 6 2.00 1.23 1.31 1.25 Females 46 1 7 2.02 1.36 1.80 3.47 Males(ln) 53 1 6 1.03 0.36 0.71 0.57 Females(ln) 46 1 7 1.03 0.38 0.95 0.21 Alcohol e xpectancies Alcohol outcome expectancies were assessed using the Alcohol Expectancy Questionnaire (AEQ) and the Alcohol Expectancy Multi Axial Assessment (AEMax; see Table 2). The Subscales of the AEQ included Global Positive, Sexual Enhancement, Social and Physical Pleasure, Social Assertion, Tension Reduction, and Aggression/ Arousal. The AEMax included three second order factors (Positive/Arousing, Negative, and Sedating) and eight first order factor subscales (Social, Woozy, Sick, Egotistical, Horny, Attractive, Sleepy, and Dangerous). Subscales reflected elevated social, pos itive and arousing subscale means across this sample, a pattern consistent with
25 college aged populations. Expectancy means did not differ between genders for any subscale. Table 2 Descriptive Statistics for Alcohol Expectancy Questionnaire Scales Behavioral inhibition/activation Behavioral inhibition and activation were assessed using the BIS/BAS questionnaire (Carver & White, 1994) BIS (inhibition), and BAS (activation) subscales (Drive, Fun, and Reward) were analyzed. Consistent with previous research ( Jorm et al., 2001; Leone et al., 1999 ), data indicated greater reported inhibition among females and greater sensitivity to reward among males. Femal es scored significantly higher than males (see Table 3 for means) on the BIS [ t (97) = 2.53, p < .05] whereas males scored significantl y higher on both the BAS Drive [ t (97) = 2.19, p < .05 ] and BAS Reward [ t (97) = 2.8, p < .01 ] scales than did females. Behavioral inhibition and activation were not related to reported drinking behavior or expectancy variables (see Tables 4 6). Min Max Mean SD Skewness Kurtosis Global Positive 0 20 8.68 4.92 0.33 0.77 Sexual Enhancement 0 7 2.71 2.12 0.36 1.13 Social & Physical Pleasure 4 9 7.47 1.45 0.74 0.34 Social Assertion 0 10 6.99 2.89 0.94 0.13 Tension Reduction 0 9 5.83 2.28 0.37 0.72 Aggression/Arousal 0 9 4.57 2.14 0.04 0.51
26 Table 3 Descriptive Statistics for Behavioral Inhibition and Activation Scales Range Mean(SD) Ske wness Kurtosis Male Female Male Female Male Female Male Female BIS 4 19 4 21 11.47(3.70) 13.57(4.54) 0.04 0.41 0.77 0.67 BAS Drive 2 12 1 12 7.58(2.28) 6.5(2.65) 0.19 0.23 0.14 0.09 BAS Fun 2 12 0 12 8.47(2.31) 7.87(3.18) 0.59 0.86 0.75 0.02 BAS Reward 1 15 0 15 12.72(2.21) 10.78(4.41) 2.89 1.25 14.39 0.17 Table 4 Correlations Between BIS/BAS and Drinking Variables BIS BAS Drive BAS Fun BAS Reward Frequency 0.08 0.02 0.13 0.06 Quantity 0.09 0.07 0.11 0.13 Table 5 Correlations Between BIS/BAS and AEQ Scales BIS BAS Drive BAS Fun BAS Reward Global Positive 0.02 0.04 0.00 0.02 Sexual Enchancement 0.20 0.04 0.03 0.13 Social/Physical Pleasure 0.11 0.01 0.03 0.09 Social Assertion 0.05 0.06 0.02 0.10 Tesnion Reduction 0.11 0.07 0.11 0.17 Aggression/Arousal 0.03 0.07 0.10 0.15
27 Table 6 Correlations Between BIS/BAS and AEMax Scales BIS BAS Drive BAS Fun BAS Reward Positive/Arousing 0.16 0.00 0.01 0.05 Horny 0.10 0.01 0.00 0.04 Social 0.16 0.00 0.02 0.08 Attractive 0.13 0.02 0.00 0.07 Sedating 0.00 0.05 0.15 0.00 Sick 0.05 0.02 0.06 0.04 Sleepy 0.06 0.09 0.16 0.01 Woozy 0.00 0.00 0.15 0.02 Negative 0.02 0.01 0.01 0.01 Dangerous 0.05 0.04 0.00 0.02 Egotistical 0.08 0.06 0.02 0.05 Within session affect Positive and negative affect were measured using the PANAS (see Table 7 for descriptive statistics). Trait scores, obtained prior to the experimental protocol, indicated no differences between genders. State scores, obtained following each task (i.e., Affective and Semantic), indicated that positive affect decreased significa ntly between the two tasks [ t (96) = 2.75, p < .01 ] while negative affect remained unchanged [ t (97) = .22, p > .05 ] State affect did not differ b etween genders at either point.
28 Table 7 Descriptive Statistics for Positive And Negative Affect Scales Range Mean(SD) Skewness Kurtosis Trait Positive 20 49 36.54(6.08) 0.44 0.08 State Positive ( Task1 ) 14 49 30.72(7.04) 0.09 0.29 State Positive ( Task 2 ) 11 48 27.65(8.09) 0.09 0.22 Trait Negative 10 45 20.55(6.19) 0.85 1.84 State Negative ( Task 1 ) 10 32 15.16(4.59) 1.07 0.99 State Negative ( Task 2 ) 10 34 15.25(4.59) 1.37 2.46 Relationships Between Alcohol Expectancies and Drinking Variables Based on prior alcohol expectancy research ( e.g., Goldman & Darkes, 2004 ), it was expected that AEMax and AEQ subscales indexing positive, arousing, and social expectancies would correlate positively with drinking variables. AEMax subscale scores did not correlate with either of the drinking variables assessed (see Table 8 ). The single item Social, of the AEMax approached significant correlation with Drinking Quantity ( r = .19, p = .06). Examination of these correlations per individual investigator (i.e., principal investigator and four research assistants) revealed a possible experimenter effect, as several of the expected correlations were present for participants run by the lead investigator and some research assistants, whereas none were present for those run by others (see Table 9 ). Due to ambiguity within the AEMax data, all further analyses wer e based on the AEQ, which participants completed online prior to the experimental protocol. AEQ subscales Social and Physical Pleasure ( r = .22) and Tension Reduction ( r = .27) were positively correlated with Drinking Frequency, but not with Drinking Quant ity (see Table 10 ).
29 Table 8 Correlations Between AEMax Scales and Drinking Variables Frequency Quantity Positive/Arousing 0.04 0.13 Negative 0.08 0.09 Sedating 0.07 0.04 Horny 0.02 0.07 Social 0.13 0.11 Attractive 0.06 0.13 Sick 0.07 0.02 Sleepy 0.08 0.00 Woozy 0.02 0.09 Dangerous 0.02 0.08 Egotistical 0.14 0.09 Table 9 Quantity Correlations By Investigator PI (N=18) RA 1 (N=6) RA 2 (N=14) RA 3 (N=22) RA 4 (N=39) Frequency 0.71** 0.90** 0.19 0.17 0.08 AEMax Positive/Arousing 0.20 0.35 0.28 0.29 0.02 AEMax Horny 0.17 0.15 0.01 0.26 0.05 AEMax Social 0.34 0.66 0.32 0.13 0.04 AEMax Attractive 0.02 0.14 0.21 0.31 0.12 p <.05, ** p <.01
30 Table 10 Correlations Between AEQ Scales and Drinking Variables Frequency Quantity Global Positive 0.01 0.09 Sexual Enhancement 0.14 0.02 Social and Physical Pleasure .22* 0.10 Social Assertion 0.10 0.11 Tension Reduction .27** 0.05 Aggression/Arousal 0.17 0.02 p <.05, ** p <.01 Descriptive Statistics for D ependent Variables Response d ata Task response data reflects the percentage of responses falling within the response window and accurately reflecting target valence Responses falling outside of the response window were counted as incorrect responses, so total percent correct was calculated as the number of correct responses divided by the total number of trials in that block Response Accuracy was computed for each block within each task domain and descriptive statistics are displayed in tables 11 & 12 Examination of non alcohol prime conditions revealed that the overall priming effect did not occur. Whereas affectively congruent prime target pairings should have facilitated response accuracy, these conditions did not di ffer significantly from their incongruent counterparts (although non alcohol incongruent pairings were more highly associated with accuracy than were non alcohol congruent pairings ; [ t (98) = 2.19, p < .05] ) suggesting that congruency did not affect respo nse accuracy differentially.
31 Table 11 Descriptive Statistics for Affective Task Response Accuracy Range Mean(SD) Skewness Kurtosis Alcohol Words Congruent 0 1.00 .61(.22 ) 0.36 0.58 Alcohol Words Incongruent 0 1.00 .67 (.21) 0.52 0.13 Non alcohol Words Congruent 0 1.00 .60 (.22) 0.50 0.32 Non alcohol Words Incongruent .13 1.00 .65(.22 ) 0.16 0.53 Alcohol Pictures Congruent 0 1.00 .57(.25 ) 0.66 0.02 Alcohol Pictures Incongruent 0 1.00 .62 (.22) 0.33 0.34 Non alcohol Pictures Congruent .13 1.00 .59 (.21) 0.28 0.73 Non alcohol Pictures Incongruent .13 1.00 .59(.20 ) 0.27 0.47 Table 12 Descriptive Statistics for Semantic Task Response Accuracy Range Mean(SD) Skewness Kurtosis Alcohol Words Congruent 0 1.00 .71(.20 ) 0.28 0.54 Alcohol Words Incongruent 0 1.00 .70(.22 ) 0.73 0.26 Non alcohol Words Congruent 0 1.00 .6 8 (.2 5 ) 0.83 0.74 Non alcohol Words Incongruent 0 1.00 .6 9 (. 24 ) 0.72 0.25 Alcohol Pictures Congruent .13 1.00 .61(.23 ) 0.70 0.30 Alcohol Pictures Incongruent 0 1.00 .67 (.2 1 ) 0.74 0.07 Non alcohol Pictures Congruent .25 1.00 66(.23 ) 0.44 0.73 Non alcohol Pictures Incongruent 0 1.00 .66(.23 ) 0.95 0.59 Relationships Between Alcohol Expectancies and Dependent Variables It was hypothesized that positive expectancies would be positively correlated with response accuracy for Affective task trials, but not Semantic task trials, in which alcohol primes were paired with positively valenced targets Bivariate correlations perfo rmed on these variables revealed a significant relationship between
32 affective task response accuracy for alcohol prime/positive target pairs and the AEQ Tension Reduction scale ( r = .38, p <.01), but only for males and only in the picture condition Semanti c task accuracy for alcohol prime/positive target pairs was negatively correlated with this scale ( r = .31, p <.05), but only in the word condition and, again, only for males Among all remaining variables, bivariate correlations revealed no significant re lationships That is, prime target congruency was not related to level of positive alcohol expectancy endorsement aside from the Tension Reduction scale of the AEQ (see Tables 13 through 16 ) which was not systematically related to accuracy across tasks. A univariate ANOVA performed on Drink Quantity and the Affective/Alcohol Positive Congruent block revealed significant group differences ( F = 2.67, p <.05). Bonferroni corrected post hoc tests indicated that participants who reported consuming five drinks p er occasion differed significantly in accuracy compared to those who reported drinking one and those who reported drinking more than five, suggesting a possible non linear relationship between Drink Quantity and accuracy for the Affective/Alcohol/Congruent block. Quadratic regressions revealed significant relationships for this block in both the picture ( = 1.99, p < .01) and word ( = 1.90, p < .01) conditions, but also for the Affective/Alcohol/Non Congruent block in the word condition ( = 1.97, p < .01 ).
33 Table 13 Correlations Between AEMax Scales and Affective Task Response Accuracy Social/Physical Pleasure Tension Reduction Alcohol Words Congruent 0.04 0.15 Alcohol Words Incongruent 0.03 0.01 Non alcohol Words Congruent 0.03 0.01 Non alcohol Words Incongruent 0.09 0.08 Alcohol Pictures Congruent 0.02 0.06 Alcohol Pictures Incongruent 0.02 0.07 Non alcohol Pictures Congruent 0.03 0.03 Non alcohol Pictures Incongruent 0.04 0.10 Table 14 Correlations Between AEMax Scales and Semantic Task Response Accuracy Social/Physical Pleasure Tension Reduction Alcohol Words Congruent 0.02 0.09 Alcohol Words Incongruent 0.04 0.03 Non alcohol Words Congruent 0.02 0.04 Non alcohol Words Incongruent 0.09 0.07 Alcohol Pictures Congruent 0.02 0.08 Alcohol Pictures Incongruent 0.04 0.04 Non alcohol Pictures Congruent 0.06 0.09 Non alcohol Pictures Incongruent 0.03 0.01
34 Table 15 Correlations Between AEMax Scales and Affective Task Response Accuracy by Gender Social/Physical Pleasure Tension Reduction Males Females Males Females Alcohol Words Congruent 0.01 0.02 0.10 0.01 Alcohol Words Incongruent 0.12 0.03 0.05 0.10 Non alcohol Words Congruent 0.04 0.14 0.03 0.09 Non alcohol Words Incongruent 0.03 0.14 0.11 0.08 Alcohol Pictures Congruent 0.20 0.17 0.38 ** 0.06 Alcohol Pictures Incongruent 0.14 0.26 0.14 0.16 Non alcohol Pictures Congruent 0.00 0.09 0.02 0.01 Non alcohol Pictures Incongruent 0.03 0.18 0.00 0.14 Table 16 Correlations Between AEMax Scales and Semantic Task Response Accuracy by Gender Social/Physical Pleasure Tension Reduction Males Females Males Females Alcohol Words Congruent 0.03 0.04 0.31 0.19 Alcohol Words Incongruent 0.07 0.02 0.01 0.08 Non alcohol Words Congruent 0.12 0.02 0.19 0.00 Non alcohol Words Incongruent 0.09 0.01 0.09 0.06 Alcohol Pictures Congruent 0.05 0.02 0.01 0.16 Alcohol Pictures Incongruent 0.18 0.16 0.23 0.18 Non alcohol Pictures Congruent 0.07 0.01 0.10 0.17 Non alcohol Pictures Incongruent 0.19 0.02 0.03 0.16
35 Discussion Drinking behaviors, a lcohol expectancies, trait and state affect, behavioral inhibition and activation and response accuracy to primed affective and semantic evaluation tasks were measured in a sample of 18 24 year old college student drinkers The primary aim of this study was to demonstrate the affective component of expectancy operation by using a suboptimal priming paradigm in which alcohol related cues were hypothesized to automatically facilitate evaluations of affectively congruent targets The paradigm was based on previous research supporting automatic cognitive and affective priming with both words and pictures ( see Musch & Klauer 2003 ), whereas the use of alcohol related cues as affective primes in this study was novel Of particular interest was th e relationship between positive/arousing alcohol outcome expectancy endorsement and response accuracy in the affective task This sample reported drinking twice weekly at a moderately high level, just below NIAAA defined binge levels for both males [M = 4.53(2.05)] and females [M = 3.24(1.55) ] These levels are consistent with boundary conditions regarding the relationship between drinking and positive expectancies. Despite this, expected relationships were not borne out. Most notably and critically, these basic boundary conditions were not met for alcohol expect ancies endorsed via the AEMax, the expectancy measure (of the two utilized here ) most closely P revious research ( see Goldman & Darkes, 2004 ) has consistently shown a positive relationship between measures of current drinking and the Positive/Arousing subscales of the
36 AEMax, whereas in this study, drinking was not significantly related to any aspect of the AEMax C orrelations between these measures at the level of individual investigator (PI and fou five investigators but these did not reach statistical significance, likely because the resultant sample division sacrificed power. Although this pattern suggests experimenter error examination of th e raw data, together with interviews of each experimenter, did not indicate any systematic difference in the way the measures were delivered. Participants had also completed the AEQ at an earlier timepoint (sometimes as distally as 90 days or more) as par t of online mass testing through the Psychology Department. Both t he Social and Physical Pleasure scale and Tension Reduction scale were significantly correlated with reported drinking frequency, but not quantity, at least partial ly establishing boundary conditions necessary for further analysis T his dataset does not contain information necessary to determine the cause of the lack of correspondence between expectancies and drinking quantity, despite their significant correlation s with frequency, but it is reasonable to assume that the gap in time between expectancy endorsement and collection of drinking data may have contained a context related shift in the relationship It could be, for example, that these student participants, having completed the AEQ ea rly in the semester and the drinking items later, had meanwhile adjusted their quantity but not frequency in response to academic and other demands, whereas expectancies remained relatively unchanged
37 Mean r esponse accuracy across tasks ranged from .57 to .71, which is consistent with research on similar tasks using the response window technique ( Draine & Greenwald,1998 ). I t was expected that positive expectancy endorsement would be posititively correlated with each of the affective conditions (word and pic ture) in which alcohol primes were paired with positively valenced targets T his relationship was born out only for the AEQ Tension Reduction scale, only in the picture condition, and only among males T hat is, male drinkers with higher positive expectancies for the tension reduction properties of alcohol responded with greater accuracy to positive pictures when they were preceded by pictures of alcohol. Contrary to hypotheses, a similar relationship was not evident in the corresponding word condi tion. This somewhat confusing and counterintuitive result may suggest that in this sample of student drinkers, men sensitive to ed more s trongl y to pictures of alcohol as a function of their current environment, which in this case was a potentially stress inducing laboratory task in an academic setting a possibility indirectly supported by the overall downward trend of positive affect across tasks As for the absence of such an effect in the word condition, it may be that real world representations (i.e., pictures) of alcohol were salient enough to overcome contextual interference to influence response accuracy for these drinkers while language based representations (i.e., words) were not. This possibility runs counte r, however, to the preponderence of previous research demonstrating affetcive priming with words and much less with pictures.
38 Relationships between drinking variables and response accuracy were also examined Although correlations between these variables r evealed no significant relationships, univariate ANOVA suggested a possible, non linear drink quantity group effect Quadratic regressions indicated such effects in both congruent and incongruent conditions of Affective/Alcohol Positive cue conditions, a r esult which was not anticipated and is not theoretically supported The ambiguity of these results likely reflects experimental artifact, rather than anything related to hypothesized effects. It is likely that the design of this study was ill suited to its purpose. Specifically, the effects demonstrated by many other studies of affective priming were achieved within very constrained experimental space; that is, what constituted several conditions within a single study here might have made up several indepen dent studies in the affective priming literature. Future studies should take more care in determing the limits of the methodology and variables of interest involved and incorporate these caveats accordingly. Finally, decreased positive affect across the ex perimental procedure suggests that participants may have become fatigued or at least bored with the tasks to a degree that associative activation failed to engage beyond predominantly cognitive processing. Future research should focus on building into the proc e dure a means of engaging and maintaining sufficient affective activation There is a burgeoning interest in alcohol research regarding the ways in which the complementary roles of affect and cognition interact to affect the
39 operation of alcohol expect anci es. Though the relationship betw een these two basic (and perhaps ultimately inseparable) processes is complex and difficult to examine, recent methodological advances have shown promise in this area. This study utilized one of these methods (i.e., the response window technique of affective priming), in an attempt to demonstrate the interactive relationship between affect and alcohol expectancies. Due to methodological limitations, no conclusions can be made about the role of affect in expectancy operati on based on the findings reported in this study. It does, however, highlight the elusive nature of affect as a psychological construct outside of tightly constrained experimental settings, raising several important points regarding its study in relation to real world phenomena, such as drinking and expectancies.
40 References Arnett, J. (1994). Sensation seeking: A new conceptualization and a new scale. Personality and Individual Differences, 16(2), 289 296. Avero, P. & Calvo, M.G. (2006). Affective Priming with Pictures of Emotional Scenes: The Role of Perceptual Similarity and Category. The Spanish Journal of Psychology, 9 (1), 10 18. Banse, R. (1999). Automatic evaluation of self and significant ot hers: Affective priming in close relationships. Journal of Social and Personal Relationships 16 803 821. Bargh, J.A., Chen, M. & Burrows, L. (1996). Automaticity of social behavior: Direct effects of trait construct and stereotype activation on action. Journal Of Personality And Social Psychology, 71(2), 230 244. Berridge, K.C. & Robinson, T.E. (2003). Parsing reward. Trends in Neuroscience, 26 (9), 507 513. Berridge, K.C. (2004). Pleasure, unfelt affect, and irrational desire. In Manstead, A.S.R., Frijda N.H., & Fischer, A.H. (Eds.) Feelings and Emotions: The Amsterdam Symposium (pp. 243 262). Cambridge, U.K: Cambridge University Press. Berridge, K.C., & Winkielman, P. (2003). What is an unconscious emotion? (The Cognition and Emotion, 17(2), 181 211. Bower, G. H. (1991). Mood congruity of social judgements. In J. P. Forgas (Ed.), Emotion and social judgements (pp. 31 53). New York: Pergamon. Bradley, M. M., & Lang, P. J. (1999). Affective norms for English words ( ANEW) Gainesville, FL: The National Institute of Mental Health Center for the Study of Emotion and Attention, University of Florida. Brown, S.A. (1985b) Reinforcement expectancies and alcoholism treatment outcome after a one year follow up. Journal Of Stu dies On Alcohol, 46, 304 308.
41 Brown, S.A., Christiansen, B.A., & Goldman, M.S. (1987). The Alcohol Expectancy Questionnaire: An instrument for the assessment of adolescent and adult expectancies. Journal Of Studies On Alcohol, 48, 483 491. Brown, S.A., Goldman, M.S., & Christiansen, B.A. (1985) Do alcohol expectancies mediate drinking patterns of adults? Journal Of Consulting And Clinical Psychology, 53, 512 519. Brown, S.A., Goldman, M.S., Inn, A. & Anderson, L. (1980). Expectations of reinforcement fro m alcohol: Their domain and relation to drinking patterns. Journal Of Consulting And Clinical Psychology, 48, 419 426. Carver, C.S. & White, T.L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishme nt: The BIS/BAS Scales. Journal Of Personality And Social Psychology, 67, 319 333. Christiansen, B.A., Smith, G.T., Roehling, P.V., & Goldman, M.S. (1989). Using alcohol expectancies to predict adolescent drinking behavior after one year. Journal Of Consul ting And Clinical Psychology, 57, 93 99. Comeau, N., Stewart, S. H., & Loba, P. (2001). The relations of trait anxiety, alcohol, cigarette, and marijuana use. Addictive Behaviors, 26(6), 803 825. Darkes, J. & Goldman, M.S., (1993). Expectancy challenge and drinking reduction: Experimental evidence for a mediational process. Journal Of Consulting And Clinical Psychology, 61, 344 353. De Houwer, J., & Eelen, P. (1998). An affective va riant of the Simon paradigm. Cognition and Emotion, 12 45 61. De Houwer, J., Hermans, D., & Eelen, P. (1998). Affective and identity priming with episodically associated stimuli. Cognition and Emotion, 12 145 169. De Houwer, J., Hermans, D., Rothermund, K., & Wentura, D. (2002). Affective priming of semantic categorisation responses. Cognition & Emotion, 16 (5), 643 666. Dimberg, U. & Karlsson, B. (1997). Facial reactions to different emotionally relevant stimuli. Scandinavian Journal Of Psychology, 38, 29 7 303. Dimberg, U. & Thunberg, M. (1998). Rapid facial reactions to emotional facial expressions. Scandinavian Journal Of Psychology, 39, 39 45.
42 Draine, S.C., & Greenwald, A. G. (1998). Replicable unconscious semantic priming. Journal of Experimental Psych ology: General 127 286 303. Dunn, M.E., Lau, H.C., & Cruz, I.Y. (2000). Changes in Activation of Alcohol Expectancies in Memory in Relation to Changes in Alcohol Use After Participation in an Expectancy Challenge Program. Experimental and Clinical Psycho pharmacology, (8) 566 575. Fazio, R.H. (2001). On the automatic activation of associated evaluations: an overview. Cognition and Emotion, 15 115 41. Fazio, R.H., Sanbonmatsu, D.M., Powell, M.C., & Kardes, F.R. (1986). On the automatic activation of attitudes. Journal of Personality and Social Psychology, 50 (2), 229 238. Fridlund, A.J. & Cacioppo, J.T. (1986). Guidelines for human electromyographic research. Psychophysiology, 23, 567 589. Giner Soro lla, R., Garcia, M.T., & Bargh, J.A. (1999). The automatic evaluation of pictures. Social Cognition 17 76 96. Glaser, J., & Banaji, M. R. (1999). When fair is foul and foul is fair: Reverse priming in automatic evaluation. Journal of Personality and Soci al Psychology, 77 669 687. Goldman, M.S. (1999). [Project Advance, Year 1]. Unpublished raw data. Goldman, M.S. (2002). Expectancy and risk for alcoholism: The unfortunate exploitation of a fundamental characteristic of neurobehavioral adaptation. Alcohol ism: Clinical And Experimental Research, 26(5), 737 746. Goldman, M.S., Darkes, J. & Del Boca, F.K. (1999). Expectancy mediation of biopsychosocial risk for alcohol use and alcoholism. In I. Kirsch (Ed.), How Expectancies Shape Experience (pp. 233 262). Wa shington, D.C: American Psychological Association. Goldman, M.S., Greenbaum, P.A. & Darkes, J. (1997). A confirmatory test of hierarchical expectancy structure and predictive power: Discriminant validation of the Alcohol Expectancy Questionnaire. Psycholog ical Assessment, 9, 145 157. Greenwald, A.G., Draine, S.C., & Abrams, R.L. (1996). Three cognitive markers of unconscious semantic activation. Science 273 1699 1702. Ham, L.S. & Hope, D.A. (2003). College students and problematic drinking: A review of th e literature. Clinical Psychology Review, 23, 719 759.
43 Hatfield, G. (2002). Psychology, philosophy, and cognitive science: Reflections on the history and philosophy of experimental psychology. Mind & Language 17, 207 232. Hermans D, Baeyens F, & Eelen P. (2003). On the acquisition of evaluative information in memory: the study of evaluative learning and affective priming combined. Hermans, D., De Houwer, J., & Eelen, P. (1994). The affective priming effect: Automatic activation of evaluative information in memory. Cognition and Emotion, 8 515 533. Hermans, D., De Houwer, J., & Eelen, P. (2001). A time course analysis of the affective priming effect. Cognition and Emotion, 15, 143 165. Hermans, D., Spruyt, A., De Houwer, J., & Eelen, P. (2003). Affective Pr iming With Subliminally Presented Pictures. Canadian Journal of Experimental Psychology, 57 (2), 97 114. Holland PC, & Gallagher M. (2004). Amygdala frontal interactions and reward expectancy. Current Opinion in Neurobiology 14 ,148 155. Jorm, A.F., Christe nsen, H., Henderson, A.S., Jacomb, P.A., Korten, A.E., Rodgers, B. (1999). Using the BIS/ BAS scales to measure behavioural activation: Factor structure, validity and norms in a large community sample. Personality & Individual Differences 26 49 58. Klaue r, K.C. & Musch, J. (2003). Affective Priming: Findings and theories. In J. Musch and K.C. Klauer (Eds.), The psychology of evaluation: Affective processes in cognition and emotion Mahwah, NJ: Lawrence Erlbaum. Klauer, K.C., & Musch, J. (2002). Goal depend ent and goal independent effects of irrelevant evaluations. Personality and Social Psychology Bulletin, 28 802 814. Klauer, K.C., & Stern, E. (1992). How attitu des guide memory based judgments: A two process model. Journal of Experimental Social Psychology, 28 186 206. Klinger, M.R., Burton, P. C., & Pitts, G. S. (2000). Mechanisms of unconscious priming I: Response competition not spreading activation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26 441 455. Kramer, D.A. & Goldman, M.S. (2003). Using a modified Stroop task to implicitly discern the cognitive organization of alcohol expectancies. Journal Of Abnormal Psychology, 112(1), 17 1 175.
44 Lang, P.J. (1980). Behavioral treatment and bio behavioral treatment: Computer applications. In J. B. Sidowski, J. H. Johnson & T. A. Williams (Eds.), Technology in mental health care delivery systems (pp. 119 137). Norwood, NJ: Ablex. Lang, P.J., hman, A., & Vaitl, D. (1988). The international affective picture set (photographic slides). Gainesville, FL: Center for Research in Psycholophysiology, University of Florida. Leone, L., Perugini, M., Bagozzi, R.P., Pierro, A., Mannetti, L. (2 001). Construct validity and generalizability of the Carver White Behavioural Inhibition System / Behavioural Activation System Scales. European Journal of Personality 15 373 390. Maddux, J.E. (1999). Expectancies and the Social Cognitive Perspective: Ba sic Priciples, Processes, and Variables. In I. Kirsch (Ed.), How Expectancies Shape Experience (pp. 17 40). Washington, D.C: American Psychological Association. Murphy, S.T. & Zajonc, R.B. (1993). Affect, cognition, and awareness: Affective priming with op timal and suboptimal stimulus exposures. Journal of Personality & Social Psychology, 64, 723 739. Musch, J., & Klauer, K. C. (2003). The psychology of evaluation : Affective processes in cognition and emotion Mahwah, NJ: Lawrence Erlbaum. National Institut es of Health: Task Force of the National Advisory Council on Alcohol abuse and Alcoholism A Call to Action: Changing the Culture of Drinking at U.S. Colleges. Available online: www.collegedrinkingprevention.gov 2004. Neely, J. H. (1991). Semantic priming effects in visual word recognition: A selective review of current findings and theories. In D. Besner, & G. W. Humphreys (Eds.), Basic processes in reading (pp. 264 336). Hillsdale, NJ: Lawrence Erlbaum Associates. Nelson, D. L., McEvoy, C. L., & Schreibe r, T. A. (1998). The University of South Florida word association, rhyme, and word fragment norms. http://www.usf.edu/FreeAssociation/. Niedenthal, P. M. (1990). Implicit perception of affective information. Journal of Experimental Social Psychology, 26 5 05 527. Niedenthal, P.M. (2007). Embodying Emotion, Science, 316 1002 1005).
45 Ohman, A., Flykt, A., & Lundqvist, D. (2000). Unconscious emotion: Evolutionary perspectives, psychophysiological data and neuropsychological mechanisms. In R.D. Lane, L. Nadel, & G. Ahern (Eds.), Cognitive neuroscience of emotion (pp. 296 327). New York: Oxford University Press. Rather, B. C., Goldman, M. S., Roehrich, L., & Brannick, M. (1992). Empirical modeling of an alcohol expectancy memory network using multidimensional sc aling. Journal of Abnormal Psychology, 101 174 183. Rather, B.C. and Goldman, M.S., 1994. Drinking related differences in the memory organization of alcohol expectancies. Experimental and Clinical Psychopharmacology, 2 167 183. Ray, J.M., Darkes, J., & Goldman, M.S. (2007). Implicit affect and alcohol expectancies. Alcoholism: Clinical and Experimental Research,31 (S2), p. 46A. Reich, R.R., Goldman, M.S. & Noll, J.A. (2004). Using a false memory paradigm to test two key elements of alcohol expectancy theo ry. Experimental And Clinical Psychopharmacology, 12 (2), 102 110. Robinson, T.E. & Berridge, K.C. (1993). The neural basis of drug craving: an incentive sensitization theory of addiction. Brain Research Reviews, 18 247 291. Robinson, T.E. & Berridge, K. C. (2000). Animal models in craving research: The psychology and neurobiology of addiction: an incentive sensitization view. Addiction 95(S2), S91 117. Roehrich, L., & Goldman, M.S. (1995). Implicit priming of alcohol expectancy memory processes and subseq uent drinking behavior. Experimental And Clinical Psychopharmacology, 3, 402 410. Rotteveel, M., DeGroot, P., Geutskens, A., & Phaf, R.H. (2001). Stronger suboptimal than optimal affective priming? Emotion, 1 (4), 348 364. Sher K.J., Grekin E.R. & William s N.A. ( 2005 ). The development of alcohol use disorders. Annual Review of Clinical Psychology, 1, 493 523. Stein, K.D., Goldman, M.S.& Del Boca, F.K. (2001). The influence of alcohol expectancy priming and mood manipulation on subsequent alcohol consumption. Journal Of Abnormal Psychology,109 (1), 106 115 Tolman, E.C. (1932). Purposive behavior in animals and men. New York: Appleton Centu ry Crofts.
46 Watson, D., Clark, L.A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality & Social Psychology, 54, 1063 1070. Winkielman, P. & Berridge, K.C. (2003). Ir rational wanting and subrational liking: How rudimentary motivational and affective processes shape preferences and choices. Political Psychology, 24 (4), 657 680. Winkielman, P. & Berridge, K.C. (2004). Unconscious Emotion. Current Directions In Psychologi cal Science, 13 (3), 120 123. Winkielman, P. & Cacioppo, J.T. (2001). Mind at ease puts a smile on the face: Psychophysiological evidence that processing facilitation elicits positive affect. Journal Of Personality And Social Psychology, 81 (6), 989 1000 Winkielman, P., Berridge, K.C. & Wilbarger, J.L. (2005). Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality and Social Psychology Bulletin, 31(1), 121 135 Winkielman, P., K nutson, B., Paulus, M.P. & Tujillo, J.T. (2007). Affective influence on decisions: Moving towards the core mechanisms Review of General Psychology, 11 179 192. Winkielman, P., Zajonc, R.B. & Schwarz, N. (1997). Subliminal affective priming resists attrib utional interventions. Cognition And Emotion, 11 (4), 433 465 Wise, R. A. (2002 ). Brain reward circuitry: insights from unsensed incentives. Neuron, 36 (2), 229 240. Wittenbrink, B. (2007). Measuring attitudes with priming. In Wittenbrink, B., & Schwarz, N. (Eds.). (2007). Implicit measures of attitudes New York: Guilford Press. Wittenbrink, B., & Schwarz, N. (Eds.). (2007). Implicit measures of attitudes New York: Guilford Press. Zajonc, R.B. (1980). Feeling and thinking: Preferences need no inferences. Am erican Psychologist, 35, 151 175. Zuckerman, M., Kuhlman, D.M., Joireman, P.T., Teta, P. & Kraft, M. (1993). A comparison of three structural models of personality: The Big Three, the Big Five, and the Alternative Five. Journal Of Personality And Social Ps ychology, 65, 757 768.