Substance Use & Misuse, Early Online:1–9, 2015 C 2015 Informa Healthcare USA, Inc. Copyright  ISSN: 1082-6084 print / 1532-2491 online DOI: 10.3109/10826084.2015.1005315 ORIGINAL ARTICLE Reward and Cognition: Integrating Reinforcement Sensitivity Theory and Social Cognitive Theory to Predict Drinking Behavior Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. Penelope Hasking, Mark Boyes and Barbara Mullan School of Psychology and Speech Pathology, Curtin University, Bentley, Australia Theory (SCT; 1986, 1997) have been applied to better understanding of the initiation and maintenance of problematic drinking, in an effort to inform optimal prevention, early intervention, and treatment initiatives. Although each are well-supported, and offer unique contributions to furthering our understanding of drinking behavior, we propose that theoretical relationships between these models support an integrated approach to understanding drinking behavior. Speciically, we propose that the relationships between variables proposed in RST and drinking behaviors are mediated by variables central to SCT. Background: Both Reinforcement Sensitivity Theory and Social Cognitive Theory have been applied to understanding drinking behavior. We propose that theoretical relationships between these models support an integrated approach to understanding alcohol use and misuse. Objectives: We aimed to test an integrated model in which the relationships between reward sensitivity and drinking behavior (alcohol consumption, alcohol-related problems, and symptoms of dependence) were mediated by alcohol expectancies and drinking refusal self-efficacy. Methods: Online questionnaires assessing the constructs of interest were completed by 443 Australian adults (M age = 26.40, sd = 1.83) in 2013 and 2014. Results: Path analysis revealed both direct and indirect effects and implicated two pathways to drinking behavior with differential outcomes. Drinking refusal self-efficacy both in social situations and for emotional relief was related to alcohol consumption. Sensitivity to reward was associated with alcohol-related problems, but operated through expectations of increased confidence and personal belief in the ability to limit drinking in social situations. Conversely, sensitivity to punishment operated through negative expectancies and drinking refusal self-efficacy for emotional relief to predict symptoms of dependence. Conclusions: Two pathways relating reward sensitivity, alcohol expectancies, and drinking refusal self-efficacy may underlie social and dependent drinking, which has implications for development of intervention to limit harmful drinking. Reinforcement Sensitivity Theory In Gray’s (1970, 1982) original RST, the Behavioral Inhibition System (BIS), mediated by serotonergic projections, was proposed to be sensitive to punishment and nonrewarding cues and regulate avoidant behavior leading to negative emotion, such as anxiety. The Behavioral Approach System (BAS) was proposed to underpin differences in reward sensitivity, and to be associated with impulsivity and positive emotion. Commensurate with its focus on reward, the BAS is thought to be mediated by dopimanergic pathways (Carver & White, 1994). More recent conceptualizations incorporate three biologically mediated motivational systems: (i) the BAS, (ii) the Fight, Flight, Freeze System (FFFS), which is involved in threat sensitivity, and (iii) the BIS, which detects and resolves conlict between the BAS and the FFFS. In situations that are neither clearly rewarding nor threatening, but contain elements of both, the BIS inhibits ongoing behavior, arousal is increased, and an assessment of potential risk versus reward is conducted which then drives behavioral outcomes (Corr, 2002, 2004; McNaughton & Corr, 2004; Smillie, Pickering, & Jackson, 2006). While this later conceptualization of RST is theoretically sophisticated, most measures of RST assess the Keywords alcohol, BIS/BAS, alcohol expectancies, self-eficacy Heavy alcohol use presents a signiicant public health concern with consumption causally related to more than 200 health conditions (World Health Organisation, 2014). Both Gray’s Reinforcement Sensitivity Theory (RST; 1970, 1982) and Bandura’s Social Cognitive Address correspondence to Penelope Hasking, School of Psychology and Speech Pathology, Curtin University, Bentley, Western Australia, 6102, Australia; Email: penelope.hasking@curtin.edu.au 1 Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. 2 P. HASKING ET AL original BIS/BAS components. Consequently, much of the empirical support for the model is based on the twofactor approach. Not surprisingly, heavy drinkers consistently record higher scores on measures of reward sensitivity than lighter drinkers (Franken & Muris, 2006; Hamilton, Sinha, & Potenza, 2012; Hundt, Williams, Mendelson, & Nelson-Graym, 2013; Loxton & Dawe, 2001, 2006, 2007). BAS also differentiates individuals with a DSM IV diagnosis of alcohol dependence (Johnson, Turner, & Iwata, 2003; Perry et al., 2013) and those with a family history of alcohol dependence (Yarosh et al., 2014) from controls. The role of BIS in drinking behavior is less clear, with some researchers observing higher BIS scores among hazardous compared to nonhazardous drinkers (Hamilton, Hasking, & Allen, 2012), some reporting weak-negative relationships between BIS and drinking (Loxton & Dawe, 2001) and others inding no direct relationship (Feil & Hasking, 2008; Hasking, 2006). Social Cognitive Theory While traditional learning theories argue that an individual learns by behavioral association, social cognitive theorists argue that cognitions mediate the inluence of environmental stimuli in determining behavior (Bandura, 1986, 1997). Speciically, the cognitive outcome expectancies that are formed as a result of the ability to foresee the consequences of actions allow appraisal of potential consequences, and the formation of informed decisions about volitional behavior. Yet, intention to act and a desired outcome are not suficient to enact behavior, rather individuals must possess self-regulatory mechanisms which allow them to exert control over their behavior. Self-eficacy is the key factor underlying the human agency central to SCT (Bandura, 1989, 1991). Outcome expectancies and self-eficacy have long been implicated in drinking behavior (Bandura, 1999). Consistent with operant principles, in community-based samples positive alcohol outcomes expectancies are associated with greater drinking and resultant alcohol-related problems (Jones, Corbin, & Fromme, 2001; Ham & Hope, 2003; Hasking & Oei, 2004, 2008). Conversely negative expectancies are related to lower levels of consumption (Jones et al., 2001; Jones & McMahon, 1992). However, the opposite is often observed in clinical samples, and both alcohol expectancies and self-eficacy differentiate samples of social and dependent drinkers (Hasking & Oei, 2002; Oei, Hasking, & Young, 2005). While this could be an artifact of the predominantly crosssectional, self-report designs used to relate expectancies to drinking behavior, theoretical models assert the relative importance of expectancies early in a drinking career and the conditioned inability to refuse a drink among clinical samples (Oei & Baldwin, 1994). Further, general self-eficacy and alcohol-speciic self-eficacy appear to exert different effects on drinking behavior, with alcoholspeciic self-eficacy being more salient in nonclinical samples and general self-eficacy important in treatment seeking samples (Oei, Hasking, & Phillips, 2007). Drinking refusal self-eficacy refers to the perceived ability to refuse a drink in a variety of situations (e.g., when upset, when friends are drinking). High drinking refusal self-eficacy is related to less alcohol consumption among community-based drinkers (Hasking & Oei, 2004), and prospectively mediates the relationship between alcohol expectancies and drinking behavior (Baldwin, Oei, & Young, 1993; Bandura, 1999; Connor, George, Gullo, Kelly, & Young, 2011; Young & Oei, 1993). Not only are expectancies formed prior to considerations about the ability to resist drinking (Christiansen, Smith, Roehling, & Goldman, 1989; Miller, Smith, & Goldman, 1990), but also expectancies may inform refusal self-eficacy. For example, the outcome expectancy that alcohol use will reduce tension is likely to minimize an individual’s belief in their ability to resist drinking when tense. An integrated Approach to Understanding Drinking Behavior An individual who responds to cues of reward, according to RST, is driven to seek rewarding environments and situations that stimulate dopaminergic reward pathways (Gray, 1970). Thus, an individual with high reward sensitivity is more likely to engage in behaviors that stimulate these dopaminergic pathways, including drinking. However, the individual who seeks out rewarding environments (e.g., pubs), also has a greater opportunity to learn behavioral consequences in this environment.1 Further, reward sensitive drinkers show a greater tendency to notice and attend to alcohol-related cues (Kambouropoulos & Staiger, 2001). Thus, this individual is likely to have strong outcome expectancies that have been formed over numerous experiences, and lowered self-eficacy for refusing the temptations of reinforcing behaviors. As noted above, the relationship between BIS and drinking is less clear. However, given that BIS is thought to underlie anxiety, it is possible that people sensitive to punishment are predisposed to the anxiolytic effects of alcohol, and more likely to drink in the context of expectations of tension reduction (e.g., Booth & Hasking, 2009). Consequently, social cognitive constructs, such as expectancies and selfeficacy, may mediate the relationship between sensitivity to punishment and drinking behavior. The Current Study We aimed to test the proposition that outcome expectancies and self-eficacy mediate the relationships between sensitivity to punishment, sensitivity to reward, and drinking behavior. Given the ability of these variables to differentiate dependent and nondependent drinkers, it is plausible that the variables differentially relate to particular indices of drinking. Speciically, while 1 Dopamine is also activated in response to salient events (Franken, Booij, & van den Brink, 2005), supporting Gray’s assertion that individuals highly sensitive to reward are likely to seek out such environments. REWARD AND COGNITION sensitivity to reward, positive expectancies and reduced self-eficacy might increase alcohol consumption per se, alcohol expectancies (for example) might be less salient in predicting symptoms of alcohol dependence. As such, we explored whether these relationships differed for alcohol consumption, alcohol-related problems, and symptoms of dependence (Figure 1), but did not propose speciic hypotheses regarding these differential effects. METHOD Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. Participants This project was approved by the University Human Research Ethics Committee. A total of 443 participants (83.3% female; mean age = 26.40, sd = 1.83) completed anonymous self-report questionnaires assessing the constructs of interest. Of the sample, 77.7% identiied as White/Caucasian, and a further 18.5% as Asian. A large proportion were undergraduate students (44%) and a further 29.8% had completed an undergraduate degree. All participants were aged over 18 years (the legal drinking age in Australia, where this study was conducted) and only 4.3% had never consumed alcohol.2 Materials Alcohol Use Disorders Identiication Test (AUDIT; Degenhardt, Conigrave, Wutzke, & Saunders, 2001). The AUDIT is comprised of ten items and is used to screen for excessive alcohol use and harmful drinking. While a total score gives an indication of global risk, three subscales assess alcohol consumption (three items), alcoholrelated problems (three items), and symptoms of alcohol dependence (four items). The AUDIT demonstrates high internal consistency and adequate test–retest reliability with nonclinical and alcohol-dependent samples (Reinert & Allen, 2007). Reliability was acceptable in the current sample: consumption α = .73; alcohol-related problems α = .69; and dependence α = .71. Sensitivity to Punishment/Sensitivity to Reward Questionnaire (SPSRQ; Torrubia, Avila, Molto, & Caseras, 2001). The SPSRQ was developed to assess BIS and BAS through two scales of Sensitivity to Punishment (SP) and Sensitivity to Reward (SR). Respondents indicate whether they engage in each of 24 behaviors (no = 0; yes = 1) assessing each construct (48 items in total; e.g., “Are you often afraid of new or unexpected situations?” “Do you sometimes do things for quick gains?”). In the initial psychometric evaluation, the two scales were observed to be independent, with acceptable reliability, convergent, and discriminant validity (Torrubia et al., 2001). Reliabilities of the two subscales were excellent in the current study: SP α = .80; SR α = .86. 3 Drinking Expectancy Questionnaire-Revised (DEQ-R; Lee, Oei, Greeley, & Baglioni, 2003). The ive DEQ-R subscales (cognitive change, tension reduction, increased conidence, sexual enhancement, and negative expectancies) measure anticipated positive and negative outcomes of alcohol consumption3 . The test-retest reliability (r = .72–88) and discriminant validity are well established (Lee et al., 2003). In the current sample reliabilities ranged from α = .67 for tension reduction to α = .91 for increased conidence. Drinking Refusal Self-Eficacy Questionnaire-Revised (DRSEQ-R; Oei et al., 2005). The DRSEQ-R is comprised of three subscales (social pressure, emotional relief, and opportunistic self-eficacy) assessing whether a person believes they can resist drinking in certain situations or moods (e.g., “When I am at a pub or club,” “When I feel down”). Higher scores indicate a greater belief in the ability to resist drinking. Alpha reliabilities for the DRSEQR range from .87 to .94, and test-retest reliabilities range from .84 to .93 (Oei et al., 2005). Reliabilities were excellent in the current sample: social pressure α = .91, emotional relief α = .96, and opportunistic α = .90. Procedure Participants self-selected into the study in response to advertisements posted around a large metropolitan university, community notice boards, and via social media. Advertisements included the study URL, and participants completed the online questionnaire in their own time (30–40 minutes). A Participant Information Sheet was presented prior to the survey, and informed participants of the study aims, participation requirements, voluntary nature of participation, and the right to withdraw. To reimburse participants for their time, all were offered the opportunity to enter a rafle to win an Apple IPad Mini, or 2 Gold Class movie tickets. All participants were provided with mental health information and contact details for crisis lines should they wish to further discuss any issues raised in the questionnaire. Data Analysis AUDIT data were positively skewed, as expected in a nonclinical sample. Mean replacement was utilized to replace less than 1% of data on the AUDIT and less than 3% on the other variables. To partially protect against nonnormal data, our mediational hypotheses were tested using Bias-Corrected Bootstrap procedures in AMOS 22 (Efron & Tibshirani, 1993; MacKinnon, Lockwood, & Williams, 2004). Given the correlations between our variables (Table 1), residuals were free to correlate. Model it was determined by it indices (CFI, NFI, GFI) greater than .95, a χ 2 /df less than 4, and Root Mean Square Error of Approximation (RMSEA) less than .06 (Hu & Bentler, 1999; Kline, 2010). 2 As low reward sensitivity, weak positive expectancies and high levels of refusal self-eficacy may predict abstinence, and we were interested in the continuum of drinking behavior, these 19 participants were included in analyses. 3 The negative expectancies subscale assesses anticipated negative outcomes; all other scales are considered positive expectancies. 4 P. HASKING ET AL Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. FIGURE 1. Hypothesized model: Alcohol expectancies and drinking refusal self-eficacy mediate the relationship between reward sensitivity and different indices of drinking behavior. RESULTS Participants reported a pattern of drinking consistent with similar Australian samples (M AUDIT score = 7.77, SD = 5.86; Halim & Hasking, 2012). Scores on the consumption, dependence, and alcohol-related problems subscales were also consistent with previous reports in young adult samples (Table 1; Lyvers, Hasking, Hani, Rhodes, & Trew, 2010). The majority of the sample (74.58%) reported drinking four or fewer standard drinks on a typical drinking occasion; approximately 10% of the sample reported consuming six or more drinks on a single occasion at least monthly, somewhat fewer than national igures suggest (AIHW, 2011). The majority of variables were correlated in the expected direction (Table 1). Neither age (r = .04, p = .46) nor gender [t(441) = 1.76, p = .08], were related to drinking behavior, and therefore, were not included in the models reported.4 The full model in which all direct and indirect paths between the variables (including all DRSEQ and DEQ subscales) were speciied, provided good it to the data; however, RMSEA and χ 2 /df did not meet accepted guidelines for good model it: χ 2 (7, N = 443) = 48.183, p < .001, χ 2 /df = 9.21, CFI = .98, NFI = .98, GFI = .98, RMSEA = .12 (Hu & Bentler, 1999; Kline, 2010). We, therefore, tested a reduced model in which nonsigniicant paths were removed.5 The model it of the reduced model improved, and with the exception of χ 2 (which is sensitive to sample size) all it statistics met accepted guidelines for good model it: χ 2 (29, N = 443) = 85.32, p < .001, χ 2 /df = 2.94, CFI = .98, NFI = .96, GFI = .97, RMSEA = .06 (Figure 2). The inal model accounted for 52% of the variance in alcohol consumption, 31% of the variance in alcoholrelated problems, and 35% of the variance in symptoms of dependence. There were direct effects of sensitivity to reward on both alcohol consumption and alcohol-related 4 As a check, we did include age and gender in the model and there were no changes to the signiicance of any pathways or the percentage of variance explained. 5 All variables remained in the model, but nonsigniicant pathways removed. Where remaining paths did not assist in interpretation of the proposed mediation, variables were excluded from Figure 2. problems, and indirect effects on consumption and symptoms of alcohol dependence. There were also signiicant negative direct effects of sensitivity to punishment on alcohol consumption and symptoms of alcohol dependence, and indirect effects on all indices of drinking (Table 2). As expected, alcohol expectancies were negatively related to drinking refusal self-eficacy, which in turn was negatively related to the drinking measures. Additional significant paths, not depicted in Figure 2, were observed between sensitivity to punishment and expectations of sexual enhancement (β = −.15, 95% CI −.23 to −.08, p = .003), sensitivity to reward and expectations of cognitive enhancement (β = .09, 95% CI .01 to .16, p = .05) and between cognitive enhancement and opportunistic drinking refusal self-eficacy (β = −.04, 95% CI −.15 to −.04, p = .01). DISCUSSION We proposed that the relationship between Gray’s (1970, 1982) original conceptualization of the RST and drinking behavior could be mediated by elements of Bandura’s (1986, 1997) SCT. Speciically, we tested the proposition that relationships between both sensitivity to reward and sensitivity to punishment and drinking behavior operate indirectly through alcohol expectancies and drinking refusal self-eficacy. Predicted direct effects between reward sensitivity and drinking behavior were found, as were predicted indirect effects. Evidence of indirect effects demonstrates the utility of integrating these theoretical frameworks in predicting drinking behavior and has implications for prevention and treatment. The inal model accounted for a substantial proportion of the variance in drinking behavior (R2 = .31–.52); however, reward sensitivity, alcohol expectancies, and self-eficacy were differentially associated with indices of drinking. Speciically, sensitivity to reward was associated with more alcohol consumption and alcohol-related problems, both directly and indirectly via expectations of increased conidence and lower drinking refusal self-eficacy in social situations. In contrast, sensitivity to punishment was negatively related to less alcohol 5 REWARD AND COGNITION Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. TABLE 1. Bivariate correlations between variables of interest Variable Mean (sd) 1. Consumption 2. Alcohol-related problems 3. Dependence 4. Sensitivity to punishment 5. Sensitivity to reward 6. Sexual enhancement AE 7. Increased conidence AE 8. Cognitive enhancement AE 9. Tension reduction AE 10. Negative expectancies 11. Social pressure DRSE 12. Opportunistic DRSE 13. Emotional relief DRSE 4.65 (2.54) .57∗∗∗ .56∗∗∗ −.22∗∗∗ .25∗∗∗ 2.25 (2.93) – .67∗∗∗ −.01 .26∗∗∗ .96 (1.43) 13.44 (5.47) 2 3 – 4 −.07 – 5 6 7 8 9 10 .16∗∗ −.06 .42∗∗∗ .34∗∗∗ −.07 .05 .52∗∗∗ .37∗∗∗ .05 .40∗∗∗ −.63∗∗∗ −.47∗∗∗ −.54∗∗∗ −.30∗∗∗ −.36∗∗∗ −.43∗∗∗ .19∗∗∗ −.07 .33∗∗∗ .15∗∗ −.16∗∗ .10∗ .06 .04 .39∗∗∗ −.08 .40∗∗∗ .14∗∗ −.31∗∗∗ −.41∗∗∗ −.49∗∗∗ .09 −.02 −.10∗ .04 .29∗∗∗ .12∗ .17∗∗∗ .13∗∗ −.23∗∗∗ −.24∗∗∗ −.26∗∗∗ − .06 10.34 (4.39) − 11.19 (2.33) 40.12 (8.86) − 5.75 (1.97) 8.62 (2.94) 11 −.24∗∗∗ .18∗∗∗ −.39∗∗∗ −.19∗∗∗ .10∗ .30∗∗∗ .25∗∗∗ − .04 .39∗∗∗ – .11∗ 30.49 (9.43) − 14.45 (5.57) 12 −.04 13 −.03 −.44∗∗∗ −.19∗∗∗ −.35∗∗∗ .00 −.18∗∗∗ −.10∗ −.50∗∗∗ −.53∗∗∗ −.62∗∗∗ .12∗ − 34.59 (6.18) 31.03 (8.42) −.22∗∗∗ −.27∗∗∗ .52∗∗∗ .59∗∗∗ – .75∗∗∗ – AE, alcohol expectancies; DRSE, drinking refusal self-eficacy. ∗ p < .05, ∗∗ p < .01, ∗∗∗ p < .001. FIGURE 2. The relationship among sensitivity to punishment (SP), sensitivity to reward (SR), and drinking is mediated by alcohol expectancies and drinking refusal self-eficacy (DRSE). ∗ p < .05 ∗∗ p < .01. Note: Signiicant direct paths are also evident between: increased conidence, tension reduction, and negative expectancies with each of the three drinking variables (all positively related; all p < .05). Expectations of cognitive enhancement were also directly related to drinking (all negatively related; all p < .05). To aid interpretation, these paths are not illustrated in the igure. TABLE 2. Summary of all standardized indirect effects predicting drinking variables and drinking refusal self-eficacy. Drinking variables Sensitivity to punishment Sensitivity to reward Increased conidence Tension reduction Negative expectancies Drinking refusal self-eficacy Sensitivity to punishment Sensitivity to reward Consumption (95% CI) Alcohol-related problems (95% CI) Dependence (95% CI) −.01∗ (−.02 to −.001) .06∗∗ (.03 to .10) .15∗∗ (.11 to .19) .23∗∗ (.18 to .27) −.05∗ (−.10 to −.01) Social pressure (95% CI) −.02∗∗ (−.04 to −.01) −.09∗∗ (−.13 to −.04) .04∗∗ (.020 to .07) .02 (−.002 to .05) .05∗∗ (.03 to .09) .06∗∗ (.03 to .09) −.04∗∗ (−.06 to −.02) Opportunistic (95% CI) −.02∗∗ (−.04 to −.01) −.01∗ (−.02 to −.001) .04∗∗ (.02 to .07) .03∗ (.01 to .06) .04∗∗ (.02 to .06) .14∗∗ (.08 to .20) .04∗∗ (.02 to .06) Emotional relief (95% CI) −.02∗∗ (−.04 to −.01) −.03∗∗ (−.06 to −.02) Note: All 95% CIs are bias corrected and estimated from 1,000 bootstrapped resampling draws. ∗ p < .05, ∗∗ p < .01. Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. 6 P. HASKING ET AL consumption and dependence; however, positive indirect effects operating via negative expectancies and drinking refusal self-eficacy both in social situations and for emotional relief were observed. Taken together, the results suggest that changing alcohol expectancies and self-eficacy may interrupt well-established pathways between reward sensitivity and drinking behavior, and that interventions can target individuals according to reward sensitivity. Reduced drinking refusal self-eficacy in both social situations and for emotional relief is related to greater alcohol consumption, suggesting increasing self-eficacy can reduce the amount of alcohol consumed. However, although preliminary, the indings implicate two pathways, mediated by different internal states, in differentially predicting social drinking and symptoms of dependence. Speciically, for people who are sensitive to reward, the social aspects of drinking (expectations of increased conidence, resisting drinking in social situations) appear to be key and interventions might target socially salient cues. In contrast, punishment sensitivity appears to be indirectly linked to symptoms of dependence via a pathway that is characterized by emotion regulation (e.g., negative expectancies, drinking refusal self-eficacy for emotional relief). Similar relationships have previously been noted, with avoidant coping mediating the relationship between BIS and drug and alcohol use (Hundt et al., 2013). Intervention for individuals high in punishment sensitivity might, therefore, target improving selfregulatory processes including emotion regulation, coping, and stress reduction (Feil & Hasking, 2008; Lyvers, Hasking, Albrecht, & Thorberg, 2012). Future testing of these proposed pathways and resultant outcomes is warranted. A number of unexpected indings emerged. First, opportunistic drinking refusal self-eficacy was not related to drinking. Arguably in a nonclinical sample, social factors are more salient and opportunities to drink are not necessarily cues to drink. Future testing of whether resistance to drinking when the opportunity presents itself is an important predictor of drinking in clinical samples is warranted. Second, alcohol expectancies were directly related to symptoms of alcohol dependence, contradicting theoretical and empirical accounts that expectancies are less salient to alcohol dependence (Oei & Baldwin, 1994; Oei, Fergusson, & Lee, 1998; Oei & Morawska, 2004). This may be a function of the current community-based sample; however, it does suggest that the relationship between alcohol expectancies and alcohol dependence may be on a continuum and change as drinking behaviors/problems get more severe. Of particular interest, there was no relationship between sensitivity to punishment and beliefs that alcohol consumption would reduce tension. Given the extensive literature relating anxiety, particularly social anxiety, to drinking through expectations of tension reduction (e.g., Booth & Hasking, 2009; Burke & Stephens, 1999), it may have been expected that punishment sensitivity would op- erate in a similar manner. One explanation might relate to Gray’s original conceptualization of RST. While BIS is thought to underlie anxiety, within the revised model it is portrayed as incorporating both BIS and threat sensitivity (FFFS). Studies that have assessed FFFS show results consistent with ours, with the relationship between FFFS and drinking mediated by self-regulatory processes, such as avoidant and emotion-focused coping (Ivory & Kambouropoulos, 2012). Also of interest, while there are direct negative relationships between sensitivity to punishment and drinking outcomes, there are indirect positive effects on alcohol-related problems and symptoms of dependence. This appears to be driven by a negative indirect effect on drinking refusal self-eficacy, through alcohol expectancies. Speciically, sensitivity to punishment per se appears to be protective, but when this lowers the ability to refuse a drink then it ultimately increases drinking behavior and the associated consequences. This indirect effect may explain the inconsistencies in work relating BIS to alcohol-related outcomes, and suggests more complex relationships than typically assessed. More work is needed to explore how the revised RST can best be measured (Jackson, 2009; Smillie et al., 2006) and to investigate the exact mechanisms by which threat sensitivity relates to drinking and alcohol expectancies. These unexpected indings might also relate to the cross-sectional self-report nature of the study. Speciically, people who drink less may have fewer opportunities to drink and thus opportunistic self-eficacy would not be a precursor to drinking. This would also explain the inverse relationship between negative outcome expectancies and drinking refusal self-eficacy (i.e., people who consume more alcohol report more negative outcomes). Similarly, social desirability bias may mask associations among reward sensitivity, alcohol expectancies, and drinking refusal self-eficacy. Unfortunately, our cross-sectional design precludes the ability to draw conclusions regarding the temporal relationships between the variables. Both prospective (i.e., longitudinal) studies and more objective behavioral outcomes measures (e.g., including ecological momentary assessment) would provide further evidence for these arguments. For example, assessing reward sensitivity and the social cognitive variables before a drinking occasion, and subsequently counting the number of alcoholic drinks consumed, would provide prospective data in an ecologically valid context. CONCLUSION The outcomes of the current study highlight the utility of integrating RST and SCT in understanding factors thought to initiate and maintain drinking behavior. Importantly, differential pathways from reward sensitivity to speciic drinking indices were evident. While prospective work testing whether these pathways lead to different drinking outcomes over time is REWARD AND COGNITION still needed, this research highlights important avenues to explore for the development of optimal prevention, early intervention, and treatment efforts. Declaration of Interest The authors report no conlicts of interest. The authors alone are responsible for the content and writing of the article. Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. THE AUTHORS Penelope Hasking’s research focus is on high-risk behaviors, particularly alcohol abuse and nonsuicidal self-injury (NSSI), exhibited by young people. Her focus has been on the sociocognitive and emotional factors underlying these behaviors among adolescents and young adults. Mark Boyes’ interests span the domains of health, developmental, and clinical psychology, with the overarching aim of understanding how both individual difference and social/community variables are related to psychological, social, and educational outcomes across the life-span. Barbara Mullan is an Associate Professor with the Health Psychology & Behavioral Medicine Research Group, in the School of Psychology and Speech Pathology at Curtin University. She has worked in health psychology for more than 15 years. Her research interests include social cognition models in predicting and intervening to improve health, particularly food-related behaviors, physical health, and addiction. She has over 100 peer reviewed publications. GLOSSARY Alcohol expectancies: Beliefs about the anticipated consequences of consuming alcohol. Behavioral Approach System (BAS): Biologically mediated motivational system proposed to underpin individual differences in sensitivity to reward. Behavioral Inhibition System (BIS): Biologically mediated motivational system proposed to underpin individ- 7 ual differences in sensitivity to punishment and situations of nonreward. Drinking refusal self-eficacy: An individual’s belief in their own ability to resist consuming alcohol in a given situation. REFERENCES AIHW, Australian Institute of Health and Welfare (2011). 2010 National drug strategy household survey report, Drug Statistics Series 25, Australian Institute of Health and Welfare, Canberra. Baldwin, A. R., Oei, T. P. S., & Young, R. (1993). To drink or not to drink: The differential role of alcohol expectancies and DRSE in quantity and frequency of alcohol consumption. Cognitive Therapy and Research, 17, 511–530. Bandura, A. (1986). Social foundations of thought and action: A Social Cognitive Theory. NJ: Prentice Hall. Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175–1184. Bandura, A. (1991). Social cognitive theory of self-regulation. Organizational Behavior and Human Decision Processes, 50, 248–287. Bandura, A. (1997). Self-eficacy: The exercise of control. New York: W. H. Freeman and Co. Bandura, A. (1999). A sociocognitive analysis of substance abuse: An agentic perspective. Psychological Science, 10, 214–217. Booth, C., & Hasking, P. A. (2009). Social anxiety and alcohol consumption: The role of alcohol of alcohol expectancies and reward sensitivity. Addictive Behaviors, 34, 730–736. doi:10.1016/j.addbeh.2009.04.010 Burke, R. S., & Stephens, R. S. (1999). Social anxiety and drinking in college students: A social cognitive theory analysis. Clinical Psychology Review, 19, 513–530. Carver, C. S., & White, T. L. (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: the BIS/BAS scales. Journal of Personality and Social Psychology, 67, 319–33. Christiansen, B. A., Smith, G. T., Roehling, P. V., & Goldman, M. S. (1989). Using alcohol expectancies to predict adolescent drinking behaviour after one year. Journal of Studies on Alcohol, 57, 93–99. Connor, J. P., George, S. M., Gullo, M. J., Kelly, A. B., & Young, R. McD. (2011). A prospective study of alcohol expectancies and self-eficacy as predictors of young adolescent alcohol misuse. Alcohol and Alcoholism, 46, 161–169. doi:10. 1093/alcalc/agr004 Corr, P. J. (2002). Gray’s RST: tests of the joint subsystems hypothesis of anxiety and impulsivity. Personality and Individual Differences, 33, 511–32. Corr, P. J. (2004). Reinforcement sensitivity theory and personality. Neuroscience & Biobehavioral Reviews, 28, 317–332. Degenhardt, L. J., Conigrave, K. M., Wutzke, S. E., & Saunders, J. B. (2001). The validity of an Australian modiication of the AUDIT questionnaire. Drug and Alcohol Review, 20, 143–154. doi: 10.1080/09595230124592. Efron, B., & Tibshirani, R. (1993). An introduction to the bootstrap. New York: Chapman & Hall/CRC. Feil, J., & Hasking, P. A. (2008). The relationship between personality, coping strategies and substance use. Addiction, Research & Theory, 16, 526–537. doi:10.1080/16066350802025714 Franken, I. H. A., Booij, J., & van den Brink, W. (2005). The role of dopamine in human addiction: From reward to motivated Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. 8 P. HASKING ET AL attention. European Journal of Pharmacology, 526, 199–206. doi:10.1016/j.ejphar.2005.09.025 Franken, I. H. A., & Muris, P. (2006). BIS/BAS personality characteristics and college students’ substance use. Personality and Individual Differences, 40, 1497–1503. Gray, J. A. (1970). The psychophysiological basis of introversionextraversion. Behaviour Research and Therapy, 8, 249–266. Gray, J. A. (1982). The neuropsychology of anxiety: An equiry into the functions of the septo-hippocampal system. Oxford: Clarendon Press. Halim, A., Hasking, P., & Allen, F. (2012). The role of social drinking motives in the relationship between social norms and alcohol consumption. Addictive Behaviours, 37, 1335–1341. doi:10.1016/j.addbeh.2012.07.004 Ham, L. S., & Hope, D. A. (2003). College students and problematic drinking: A review of the literature. Clinical Psychology Review, 23, 719–759. doi:10.1016/S0272-7358(03)00071-0 Hamilton, K. R., Sinha, R., & Potenza, M. N. (2012). Hazardous drinking and dimensions of impulsivity, behavioural approach and inhibition in adult men and women. Alcoholism: Clinical and Experimental Research, 36, 958–966. doi:10.1111/j.15300277.2011.01708.x Hasking, P. A. (2006). Reinforcement sensitivity, coping, disordered eating and drinking behavior among adolescents. Personality and Individual Differences, 40, 677–688. doi:10.1016/j.paid.2005.07.017 Hasking, P. A., & Oei, T. P. S. (2002). The differential role of alcohol expectancies, drinking refusal self-eficacy and coping resources in predicting alcohol consumption in community and clinical samples. Addiction Research & Theory, 10, 465–494. doi:10.1080/1606635021000034049 Hasking, P. A., & Oei, T. P. S. (2004). The complexity of drinking: Interactions between the cognitive and behavioural determinants of alcohol consumption. Addiction, Research & Theory, 12, 469–488. doi:10.1080/16066350410001713240 Hasking, P. A., & Oei, T. P. S. (2008). Incorporating coping into an expectancy framework for explaining drinking behaviour. Current Drug Abuse Reviews, 1, 20–35. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for it indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modelling, 6, 1–55. doi:10.1080/10705519909540118 Hundt, N. E., Williams, A. M., Mendelson, J., & NelsonGraym R. O. (2013). Coping mediates the relationship between reinforcement sensitivity and symptoms of psychopathology. Personality and Individual Differences, 54, 726–731. http://dx.doi.org/10.1016/j.paid.2012.11.028 Ivory, N. J., & Kambouropoulos, N. (2012). Coping mediates the relationships between revised reinforcement sensitivity theory and alcohol use. Personality and Individual Differences, 52, 822–827. doi:10.1016/j.paid.2012.01.013 Jackson, C. J. (2009). Jackson-5 scales of revised Reinforcement Sensitivity Theory (r-RST) and their application to dysfunctional real world outcomes. Journal of Research in Personality, 43, 556–569. doi:10.1016/j.jrp.2009.02.007 Johnson, S. L., Turner, R. J., & Iwata, N. (2003). BIS/BAS levels and psychiatric disorder: An epidemiological study. Journal of Psychopathology and Behavioral Assessment, 25, 25–36. doi:0882-2689/03/0300-0025/0 Jones, B. T., Corbin, W., & Fromme, K. (2001). A review of expectancy theory and alcohol consumption. Addiction, 96, 57–72. http://dx.doi.org/10.1046/j.1360-0443.2001.961575.x Jones, B. T., & McMahon, J. (1992). Negative and positive expectancies in lone and group problem drinkers. British Journal of Addiction, 87, 929–930. Kambouropoulos, N., & Staiger, P. K. (2001). The inluence of sensitivity to reward on reactivity to alcohol-related cues. Addiction, 96, 1175–1185. Kline, R. B. (2010). Principles and practice of structural equation modeling (3rd ed.). New York: Guilford Press. Lee, N. K., Oei, T. P. S., Greeley, J. D., & Baglioni, A. J. (2003). Psychometric properties of the Drinking Expectancy Questionnaire: A review of the factor structure and a proposed new scoring method. Journal of Studies on Alcohol, 64, 432–436. Loxton, N. J, & Dawe, S. (2001). Alcohol abuse and dysfunctional eating in adolescent girls: The inluence of individual differences in sensitivity to reward and punishment. International Journal of Eating Disorders, 29, 455–462. doi: 0.1002/eat.1042 Loxton, N. J., & Dawe, S. (2006). Reward and punishment sensitivity in dysfunctional eating and hazardous drinking women: Associations with family risk. Appetite, 47, 361–371. doi: 10.1016/j.appet.2006.05.014 Loxton, N. J., & Dawe, S. (2007). How do dysfunctional eating and hazardous drinking women perform on behavioural measures of reward and punishment sensitivity? Personality and Individual Differences, 42, 1163–1172. http://dx.doi.org/10.1016/j.paid.2006.09.031 Lyvers, M., Hasking, P. A., Albrecht, B., & Thorberg, F. (2012). Alexithymia and alcohol: The roles of punishment sensitivity and drinking motives. Addiction, Research and Theory, 20, 348–357. doi:10.3109/16066359.2011.636494 Lyvers, M., Hasking, P. A., Hani, R., Rhodes, M., & Trew, E. (2010). Drinking motives, drinking restraint and drinking behaviour among young adults. Addictive Behaviors, 35, 116–122. doi:10.1016/j.addbeh.2009.09.011 MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Conidence limits for the indirect effect: Distribution of the product and resampling methods. Multivariate Behavioral Research, 39, 199–128. McNaughton, N., & Corr, P. J. (2004). A two-dimensional neuropsychology of defense: fear/anxiety and defensive distance. Neuroscience & Biobehavioral Reviews, 28, 285–305. doi:10.1016/j.neubiorev.2004.03.005 Miller, P. M., Smith, G. T., & Goldman, M. S. (1990). Emergence of alcohol expectancies in childhood: A possible critical period. Journal of Studies on Alcohol, 51, 343–349. Oei, T. P. S., & Baldwin, A. R. (1994). Expectancy theory: A twoprocess model of alcohol use and abuse. Journal of Studies on Alcohol, 55, 525–534. Oei, T. P. S., Fergusson, S., & Lee, N. K. (1998). The differential role of alcohol expectancies and DRSE in problem and nonproblem drinkers. Journal of Studies on Alcohol, 59, 704–711. Oei, T. P. S., Hasking, P. A., & Phillips, L. (2007). A comparison of general self-eficacy and drinking refusal self-eficacy in predicting drinking behaviour. The American Journal of Drug and Alcohol Abuse, 33, 833–841. doi:10.1080/00952990701653818 Oei, T. P. S., Hasking, P. A., & Young, RMcD. (2005). Drinking refusal self-eficacy questionnaire - revised (DRSEQR): A new factor structure with conirmatory factor analysis. Drug and Alcohol Dependence, 78, 297–307. doi:10.1016/j.drugalcdep.2004.11.010 Oei, T. P., & Morawska, A. (2004). A cognitive model of binge drinking: The inluence of alcohol expectancies and drinking refusal self-eficacy. Addictive Behaviors, 29, 159–179. doi:10.1016/S0306-4603(03)00076-5 Perry, R. I., Krmpotich, T., Thompson, L. L., Mikulich-Gilbertson, S. K., Banich, M. T., & Tanabe, J. (2013). Sex modulates REWARD AND COGNITION Subst Use Misuse Downloaded from informahealthcare.com by Curtin University of Technology on 02/23/15 For personal use only. approach systems and impulsivity in substance dependence. Drug and Alcohol Dependence, 133, 222–227. doi:10.1016/j.drugalcdep.2013.04.032 Reinert, D. F., & Allen, J. P. (2007). The alcohol use disorders identiication test: An update of research indings. Alcoholism: Clinical and Experimental Research, 31, 185–199. doi:10.1111/j.1530-0277.2006.00295.x Smillie, L. D., Pickering, A. D., & Jackson, C. J. (2006). The new Reinforcement Sensitivity Theory: Implications for personality measurement. Personality and Social Psychology Review, 10, 320–335. Torrubia, R., Avila, C., Molto, J., & Caseras, X. (2001). The Sensitivity to Punishment and Sensitivity to Reward Questionnaire (SPSRQ) as a measure of Gray’s anxiety and impul- 9 sivity dimensions. Personality and Individual Differences, 31, 837–862. World Health Organisation. (2014). Global status report on alcohol and health 2014. Geneva: WHO Press. Yarosh, H. L., Hyatt, C. J., Meda, S. A., Jiantonio-Kelly, R., Potenza, M. N., Assaf, M., & Pearlson, G. D. (2014). Relationships between reward sensitivity, risk-taking, and family history of alcoholism during an interactive competitive fMRI task. PLOSOne, 9. doi:10.1371/journal.pone.0088188 Young, R. McD., & Oei, T. P. S. (1993) Grape expectations: The role of alcohol expectancies in the understanding and treatment of problem drinking. International Journal of Psychology, 28, 337–364.