UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reinforcement sensitivity and alcohol use : the role of depression, hopelessness, anxiety sensitivity… Maclean, A. Michael 2003

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2003-860647.pdf [ 8.03MB ]
Metadata
JSON: 831-1.0091248.json
JSON-LD: 831-1.0091248-ld.json
RDF/XML (Pretty): 831-1.0091248-rdf.xml
RDF/JSON: 831-1.0091248-rdf.json
Turtle: 831-1.0091248-turtle.txt
N-Triples: 831-1.0091248-rdf-ntriples.txt
Original Record: 831-1.0091248-source.json
Full Text
831-1.0091248-fulltext.txt
Citation
831-1.0091248.ris

Full Text

Reinforcement Sensitivity and A l c o h o l Use: The Role o f Depression, Hopelessness, Anxiety Sensitivity and Trait Anxiety by A . Michael Maclean B . A . , Simon Fraser University, 1998 M . A . , University o f British Columbia, 2000 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF D O C T O R OF P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department o f Psychology) We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A October 2003 © 2003 Angus Michael N e i l Maclean  In  presenting this  degree at the  thesis in  University of  partial  fulfilment  of  of  department  this or  requirements  British Columbia, I agree that the  freely available for reference and study. I further copying  the  his  or  her  representatives.  an advanced  Library shall make it  agree that permission for extensive  thesis for scholarly purposes may be granted by  for  It  is  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  |  - ' - ' V - . v, -i ;  The University of British Columbia Vancouver, Canada ...  Date  DE-6 (2/88)  /  f yC Tfjt^i  ... / •-- i  11 Abstract The current research investigated the role of depression, hopelessness, anxiety sensitivity, and trait anxiety in alcohol use and problems, using hierarchical regression and structural equation modelling (SEM).  Using adolescent sample in Study 1, coping drinking motives  mediated the relations of both depression and hopelessness with alcohol problems, whereas conformity motives mediated the relationship between anxiety sensitivity and alcohol problems. Trait anxiety did not significantly contribute to the prediction of alcohol use once other predictors were accounted for. The structural model was cross-validated on a second group of adolescents.  Using a young adult sample, Study 2 further investigated distinct motivational  pathways to alcohol use, distinguishing between anxiety-related and depression-related drinking motives.  Depressive symptomatology had an indirect impact on alcohol dependence through  depression drinking motives, whereas trait anxiety had an indirect impact on alcohol dependence through anxiety motives and in turn alcohol frequency. characterized  by  an underactive  Depression is hypothesized to be  Behavioural Activation System  (BAS; i.e.,  reward-  responsiveness) and an overactive Behavioural Inhibition System (BIS; i.e., punishmentsusceptibility), whereas anxiety is said to be characterized by an overactive BIS only. Self-report BIS/BAS scales and a reinforcement contingency paradigm Responsivity Objective Test (CARROT) -  Card Arranging Reward  were used to examine the extent to which  reinforcement sensitivity was able to account for relations among personality/symptom variables and alcohol use. A sip of alcohol was used to examine whether cues for alcohol could increase reward-responsiveness and diminish punishment-susceptibility.  Generally, findings confirmed  hypotheses concerning B A S and BIS activity. However, depressive symptomatology was not characterized by low B A S activity using self-report, and neither hopelessness nor depressive symptomatology exhibited low B A S on the CARROT. Anxiety sensitivity was characterized by high BIS with self-report, but not the CARROT.  Individuals scoring high on depressive  Ill  symptomatology, hopelessness, and trait anxiety all exhibited high self-reported B I S activity and diminished punishment-susceptibility on the C A R R O T after presentation o f alcohol cues.  The  current research provides preliminarily evidence for depression and anxiety playing independent roles i n motivation for alcohol use, with both reflecting a desire to diminish BIS activity (i.e., negative affect). The current as well as past findings are integrated into a model o f depression and anxiety-related alcohol use.  iv Table o f Contents Abstract  ii  List o f Tables  vii  List o f Figures  x  Acknowledgements  xi  CHAPTER  I  Overview o f Current Research  1  CHAPTER  II  Study 1  7  2.1 Introduction  7  2.1.2 Present Study  9  2.1.3 Hypotheses  10  2.2 Method  12  2.3 Results  20  2.3.1 A g e and Gender  20  2.3.2 Drinking Motives and A l c o h o l Problems  20  2.3.3 Hypothesized and Modified Models  27  2.3.4 Cross-validation o f M o d e l  32  2.3.5 Depression Items, Alcohol Dependence and Depression Drinking Motives 2.4 Summary  CHAPTER  HI  36 40  2.4.1 Depressive Symptomatology and Hopelessness  42  2.4.2 Anxiety Sensitivity and Trait Anxiety  43  2.4.3 Limitations  46  2.3.4 Sources o f Motivation to Use A l c o h o l to Cope  47  Study 2  49  V  3.1 Introduction  49  3.1.2 Reinforcement Sensitivity Theory  50  3.1.3 Reinforcement Sensitivity: Depression and Hopelessness .  54  3.1.4 Reinforcement Sensitivity and Substance Use  57  3.1.5 Present Study  61  3.1.6 Hypotheses  65  3.2 Method  68  3.3 Results  76  3.3.1 A g e and Gender  76  3.3.2 Drinking Motives  76  3.3.3 A l c o h o l Abuse and Dependence  83  3.3.4 Relation o f BDI-II Items to A l c o h o l Dependence and Depression Drinking Motives 3.3.5 Behavioural Inhibition and Behavioural Activation System  91 96  3.3.6 C A R R O T Reward-Responsiveness and PunishmentSusceptibility CHAPTER  TV  General Discussion  98 117  4.1. Summary o f Study 2  117  4.2 Drinking Motives and A l c o h o l Use  119  4.3 Reinforcement Sensitivity: Personality and Symptom Variables  125  4.3.1 B A S Activity and Personality/Symptom Variables  125  4.3.2 B I S Activity and Personality/Symptom Variables  128  4.4 Reinforcement Sensitivity and Alcohol Cues 4.4.1 B A S Activity and Alcohol Cues  131 132  vi 4.4.2 BIS Activity and A l c o h o l Cues 4.5 Limitations and Future D i r e c t i o n s . . 4.7 Conclusion  ;  132 135 147  Footnotes  149  References  150  Appendix  177  V  vii List o f Tables Table 1. Study and Cross-validation Sample Descriptives  13  Table 2. Study Sample: Bivariate correlations among variables  21  Table 3. Hierarchical Multiple Regression Predicting Drinking Motives from Demographic and Personality/Symptom Variables  22  Table 4. Hierarchical Multiple Regression Predicting A l c o h o l Problems from Personality/Symptom Variables and Drinking Motives  25  Table 5. Hypothesized and Modified Models: Summary o f Goodness-of-Fit Indices .  29  Table 6. Cross-validation Sample: Bivariate correlations among variables  33  Table 7. Cross-validation o f Models: Summary o f Goodness-of-Fit Indices  34  Table 8. Study Sample: Correlations o f Depression items with A l c o h o l Problems and Coping motives  37  Table 9. Study Sample: Depression Items regressed on A l c o h o l Problems and Coping motives Table 10. Cross-validation Sample: Correlations o f Depression items with A l c o h o l  38 39  Problems and Coping motives Table 11. Descriptive Statistics for Study 2 Variable  77  Table 12. Correlations among Personality/Symptom Variables and Drinking Motives  78  Table 13. Hierarchical Multiple Regression Predicting Drinking Motives  80  Table 14. Self-reported A l c o h o l Use Frequency  84  Table 15. Number o f A l c o h o l Abuse and Dependence Criteria Endorsed  85  Table 16. Correlations for Personality/Symptom Variables with A l c o h o l Abuse, Dependence and Frequency Table 17. Hierarchical Multiple Regression Predicting A l c o h o l Dependence from  86  viii A l c o h o l Frequency, Personality/Symptom Variables and Drinking Motives  88  Table 18. Summary o f Goodness-of-Fit Indices  91  Table 19. Depression Items with A l c o h o l Problems and Depression Motives  93  Table 20. Stepwise Regression o f Depression Items regressed on A l c o h o l Problems .  94  Table 21. Stepwise Regression o f Depression Items regressed on Depression Motives  95  Table 22. Hierarchical Regression for Depression Motives regressed on BDI-II Items  96  Table 23. Correlations between B I S / B A S Scales and the Depressive Symptomatology, Hopelessness, Anxiety Sensitivity, Trait Anxiety, and the Eysenck Personality Questionnaire-Revised  97  Table 24. Three-way A N O V A for A l c o h o l Groups by Hopelessness by Condition on C A R R O T Reward-responsiveness  99  Table 25. Three-way A N O V A for Alcohol Groups by Hopelessness by Condition on C A R R O T Punishment-susceptibility  100  Table 26. Three-way A N O V A F O R Alcohol Groups by Hopelessness Split at 1/2 SD by Condition on C A R R O T Reward-responsiveness  102  Table 27. Three-way A N O V A for A l c o h o l Groups by Hopelessness split at Plus/Minus 1/2 S D by Condition on C A R R O T Punishment-susceptibility.  103  Table 28. Three-way A N O V A for Alcohol Groups by Depressive Symptomatology by Condition on C A R R O T Reward-responsiveness  105  Table 29. Three-way A N O V A for Alcohol Groups by Depressive Symptomatology by Condition on C A R R O T Punishment-susceptibility Table 30. Three-way A N O V A for A l c o h o l Groups by Anxiety Sensitivity by  106  ix Condition on C A R R O T Reward-responsiveness  109  Table 31. Three-way A N O V A for A l c o h o l Group by Anxiety Sensitivity by Condition on C A R R O T Punishment-susceptibility  110  Table 32. Three-way A N O V A for Alcohol Groups by Trait Anxiety by Condition on C A R R O T Reward-responsiveness  Ill  Table 33. Three-way A N O V A for A l c o h o l Groups by Trait Anxiety by Condition for C A R R O T Punishment-susceptibility  112  Table 34. Three-way A N O V A for A l c o h o l Groups by Trait Anxiety Split at 1/2 S D by Condition on C A R R O T Reward-responsiveness  114  Table 35. Three-way A N O V A for A l c o h o l Groups by Trait Anxiety Split at 1/2 S D by Condition C A R R O T Punishment-susceptibility  115  X  List o f Figures  Figure 1. Hypothesized M o d e l for Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity  11  Figure 2. Hypothesized M o d e l : Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity.  30  Figure 3. Modified M o d e l : Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity  31  Figure 4. Replication M o d e l : Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity  35  Figure 5. Hypothesized M o d e l : Influence o f Personality/Symptom Variables on Depression Motives, Coping Motives, and A l c o h o l Dependence  90  Figure 6. Condition by Hopelessness Interaction for Punishment-Susceptibility . . . .  104  Figure 7. Condition by BDI-II Interaction for Punishment-Susceptibility  107  Figure 8. Punishment-susceptibility at the Neutral Condition for BDI-II split at a Cut-off o f 14  108  Figure 9. Trait Anxiety by Condition Interaction for Punishment-Susceptibility . . .  113  Figure 10. L o w , Moderate and H i g h Trait Anxiety ( T A ) groups by Condition  116  Interaction for Punishment-Susceptibility Figure 11. Theoretical M o d e l o f the Role o f Reinforcement Sensitivity i n Depression and Anxiety Motivated Drinking  144  xi Acknowledgements I would like to thank Patricia Conrod for her research in substance use risk typologies in laying the theoretical foundation for the current proposal, as well as for all o f her many helpful comments, direction and support from the formulation to completion o f this research. Thanks also go to E k i n Blackwell for all o f her helpful input while I was working out the methodology for Study 2. I would also like to thank my other core dissertation committee members, Larry Walker and Lynn Alden, and additional members that served as examiners for taking the time out o f their busy schedules to read and provide feedback on my research. Larry Walker has played an especially important role i n my tenure at U B C , to which I am indebted. Finally, I thank Trina Maclean for all o f her support throughout my long hours o f reading and writing and ignoring much else. Study 1 was supported by a grant from the A l c o h o l Beverage Medical Research Foundation and Study 2 by a predoctoral fellowship from the Social Sciences and Humanities Research Council.  1 CHAPTER I  Overview o f Current Research A number o f large epidemiologic and longitudinal studies investigating the prevalence and comorbidity o f psychiatric disorders have found substance use to be related to both mood and anxiety disorders (Grant, 1995; Grant & Harford, 1995; Helzer & Pryzbeck, 1988; Kessler, Nelson, McGonagle, Edlund, Frank, & Leaf, 1996; Kandel, Huang, & Davies, 2001). For instance, individuals with either a depressive or substance use disorder are at a two- to five-fold greater risk for developing the other disorder (Grant & Harford, 1995; Kandel et al., 2001; Merikangas et al., 1998). Similar odds ratios have been found for anxiety disorders and substance dependence (Kandel et al., 2001; Merikangas et al., 1998). A m o n g adolescents with substance use disorders the prevalence o f depressive disorders is estimated to be between 24% and 50% and anxiety disorders to be between 7% and 40% (Birmaher et al., 1996; Kaminer, 1999). However, the extent to which depression occurs prior to the onset o f substance abuse/dependence is a contentious issue (Schuckit et al., 1997; Schuckit, Tipp, Bergman, & Riech, 1997). A m o n g adolescents, substance use often predates the onset o f depressive and anxiety symptoms (Boyle et al., 1992, 1993). Brook, Cohen, and Brook (1998) found no evidence for prior psychopathology to be predictive o f later substance use, but that both tobacco and illicit drug use in adolescence predicted adult diagnosis o f major depression and anxiety disorders after controlling for adolescent psychopathology, especially among heavy alcohol users. A s well, only mixed results have been found when treating substance use with antidepressants, perhaps challenging the possibility that substance use is caused by depression (Nunes et al., 1995; Petrakis, Carroll, N i c h , Gordon, Kosten, & Rounsaville, 1998; Schmitz et al., 2001). The order of onset, however, may depend on the type o f substance being abused. For example, depression is more common among individuals with alcohol dependence than those without, and among  2 such individuals depression appears more likely to precede alcohol dependence (Abraham & Fava, 1999). Longitudinal research suggests that there are two pathways to comorbidity between alcohol dependence and major depression, with either primary disorder increasing the risk for the other (Gilman & Abraham, 2001). Additionally, individuals with double depression or early onset depression are at greater risk for recurrent depression, as well as substance abuse or dependence occurring subsequent to a major depressive episode (Alpert, Maddocks, Rosenbaum, & Fava, 1994; Kashani, 1985; K l e i n , Schwartz, Rose, & Leader, 2000). Although there is a great deal o f research individually examining mood and anxiety in relation to substance use, very little research has addressed the relative contribution o f each (Kandel et al., 2001). One study looked at the relative contribution o f sensation-seeking, depression, state anxiety and trait anxiety in predicting alcohol use among adolescents, finding sensation-seeking to be the only significant predictor after controlling for age and previous alcohol use (Teichrnan, Barnea, & Ravav, 1989). Because sensation-seeking appears to be one o f the strongest predictors o f substance use (Schuckit, 1998; Zuckerman & Kuhlman, 2000), this study does not directly speak to the relative contribution o f mood and anxiety, nor address the extent to which mood and anxiety are predictive o f alcohol problems. The present research is an attempt to address to what extent anxiety and depression can explain unique variance in alcohol problems. In addition, hopelessness, a specific subtype or risk factor for depression (Abramson, Metalsky, & A l l o y , 1989; A l l o y , Reilly-Harrington, Fresco, Whitehouse & Zechmeister, 1999; Abramson et al., 1998, 1999; Beck, 1987; Beck, Riskind, Brown, & Steer, 1988), appears to be especially associated with substance dependence (Conrod, Pihl, Stewart, & Dongier, 2000), and the psychological treatment o f which has been found to have an attenuating effect on substance use (Conrod, Pihl, Stewart, Cote et al., 2000). There is also a great deal o f evidence for the cooccurrence o f hopelessness and substance use among individuals at risk for suicide (Balon, 1987; H a l l , Piatt, & H a l l , 1999; Lolas, Gomez, & Suarez, 1991; Overholser, Freiheit, & DiFilippo,  3 1997; Suominen, Isometsa, Henriksson, Ostamo, & Loennqvist, 1997). The relations between hopelessness and alcohol use were also, therefore, examined i n the present study. There appears to be a great deal o f overlap between anxiety and depression (Mehrabian & Bernath, 1991; Sanderson, DiNardo, Rapee, & Barlow, 1990). Both anxiety and depression appear to be characterized by high negative affect (e.g., upset, distressed etc.), whereas depression is distinguished by low positive affect (e.g., happy, excited etc.; Clark & Watson, 1991; Gencoz, 2002) and anxiety distinguished by hyper-arousal and the fear o f these physiological sensations (Clark & Watson, 1991). The fear o f potential negative consequences (i.e., negative evaluation, loss o f mental control, physical illness) as a result o f anxiety symptoms is referred to as Anxiety Sensitivity (McNally, 1996; Reiss, Peterson, Gursky, & M c N a l l y , 1986). Anxiety sensitivity is predictive o f individuals' panic responses to a biological challenge (e.g., carbon dioxide, hyperventilation) in clinical (Rapee, Brown, Antony, & Barlow, 1992) and non-clinical samples (Zvolensky, Feldner, Eifert, & Stewart, 2001) and predictive o f future panic attack and panic disorder, even after controlling for trait anxiety (Eke & M c N a l l y , 1996; Schmidt, Lerew, & Jackson, 1997, 1999). In addition, individuals high in anxiety sensitivity appear to be at increased risk for substance use disorders (Stewart, Peterson, & Pihl, 1995; Woicik, Conrod, Stewart, P i h l , & Dongier, 2002). Anxiety sensitivity and trait anxiety were used to examine the relations between anxiety and alcohol use in the current investigation. The use o f both personality (i.e., hopelessness, anxiety sensitivity, trait anxiety) and symptom (i.e., depression) variables appears necessary to explore the roles o f anxiety and depression i n alcohol use. There is a great deal o f variability i n the extent to which different anxiety disorders are related to substance use (van B a l k o m et al., 2000), whereas anxiety sensitivity and trait anxiety have consistently been found to be related to substance use (e.g., Stewart & Zeitlin, 1995). Anxiety sensitivity was chosen for its close relationship to the physiological arousal component o f anxiety, whereas both anxiety sensitivity and trait anxiety  4  appear to be closely associated with anxiety disorder diagnoses (Muris, Schmidt, Merckelbach, & Schouten, 2001). Although anxiety sensitivity and trait anxiety are related, they are generally accepted as reflecting conceptually distinct constructs (Muris et al., 2001; Taylor, 1995) and are able to account for unique variance in anxiety disorders. For example, anxiety sensitivity is more strongly related to panic disorder and agoraphobia than is trait anxiety (Eke & McNally, 1996; Schmidt, Lerew, & Jackson, 1997, 1999). Anxiety sensitivity appears to reflect the tendency to react with anxiety toward bodily sensations, whereas trait anxiety reflects the tendency to react anxiously toward anxiety-provoking stimuli in general (Muris et al., 2001; Taylor, 1995). Although some have argued that anxiety sensitivity is merely a lower-order factor of trait anxiety (see Lilienfeld, Turner, & Jacob, 1993), an equally adequate explanation is that anxiety sensitivity is the cause of trait anxiety (Taylor, 1995). Nevertheless, there seems to be good reason to include both anxiety sensitivity and trait anxiety as predictors of alcohol use in the current research. Hopelessness and depression appear to be strongly related as well and, to some extent, hopelessness may be indistinguishable from the cognitive symptoms of depression in depressed individuals (Joiner et al., 2001). However, hopelessness was chosen because it also appears to be a personality risk factor predisposing individuals to both depression and substance use disorders (Conrod et al., 2000), and is present even when individuals are not currently depressed c  (Alloy, Abramson, Hogan, et al., 1999). Personality and symptom levels of mood and anxiety have implications for the treatment of substance use disorders as well. Symptom level psychopathology appears to be relevant to the current treatment of substance use disorders, whereas personality risk can be assessed and prevention programmes implemented among individuals not yet experiencing problematic use of alcohol and other drugs (e.g., adolescents). A second emphasis of the current research is to investigate the application of Reinforcement Sensitivity Theory (RST) to models of drug abuse vulnerability by examining the  5 extent to which reinforcement sensitivity (i.e., reward-responsiveness & punishmentsusceptibility) can account for the relations between personality/symptom variables and substance use. Consistent use o f tobacco or alcohol appear to produce changes in the brain reward system that lead to decreases i n an individual's ability to experience reward from reinforcers, which has been implicated in the maintenance o f substance abuse ( A l - A d a w i & Powell; 1997; Kambouropoulos & Staiger, 2001; Powell, Dawkins, & Davis, 2002). Similarly, it has been posited that depression is characterized by a decreased capacity to find reinforcing stimuli rewarding (Costello, 1972). Depressed individuals appear to underestimate the frequency of pleasant events (Buchwald, 1977), and exhibit greater recognition memory for unpleasant stimuli than pleasant stimuli, whereas the opposite is found in nondepressed individuals (Dunbar & L i s h m a n , 1984). Depue and Iacono (1989) suggest that depressed individuals engage in less approach behaviour due to an underactive behavioural facilitation (or activation; Gray, 1994) system—a hypothesized brain system responsible for positive emotions, motoric motivation, and rewardresponsiveness. Indeed, dysphoric mood appears to result in lower susceptibility to reward and greater susceptibility to punishment (Henriques, Glowacki, & Davidson 1994; Henriques & Davidson, 2000; Lewinsohn, Lobitz, & Wilson 1973). Whether due to chronic substance use or depression per se, an underactive behavioural facilitation system may result in a motivational vulnerability (i.e., apathy, anhedonia, low positive affect), i n which individuals experience less reward from their environment and therefore use alcohol and other drugs for their ability to stimulate the brain reward system (Balfour & Ridley, 2000; DiChiara, Acquas, & Tanda, 1996; Reid, 1996). In sum, the current research proposes to investigate the relations between depression, hopelessness, anxiety sensitivity, trait anxiety, motivations for alcohol use and alcohol use/problems, as well as the degree to which R S T is able to account for variance in these  6  relations. The first goal o f this investigation is to establish, i n Study 1, the unique relations o f mood (i.e., hopelessness, depression) and anxiety (i.e., anxiety sensitivity, trait anxiety) and alcohol problems, and the extent to which motivations for alcohol use mediate this relationship. Study 2 is an attempt to extend the findings o f Study 1, as well as examine the degree to which depression and hopelessness are predictive o f reward-responsiveness and punishmentsusceptibility. Reinforcement sensitivity was assessed i n two ways, using both self-report scales and a behavioural measure. Study 2 also employs a cue-reactivity paradigm, in which participants were exposed to a cue for alcohol to establish whether alcohol can have a normalizing effect on reward-responsiveness and punishment-susceptibility.  7 CHAPTER n STUDY 1  2.1 Introduction Individuals' motivation for substance use appears to be helpful i n understanding the relations between co-occurring psychopathology and substance use. Cooper and colleagues propose that individuals are motivated to use alcohol to regulate positive and negative affect (Cooper, 1994; Cooper, Frone, Russell, & Mudar, 1995; Cooper, Russell, Skinner, & Windle, 1992). Specifically, individuals drinking to attain a euphoric state are said to be drinking for enhancement reasons, whereas individuals drinking to alleviate negative emotions, such as anxiety and depression, are said to be drinking for coping reasons. In addition, conformity motives are used to avoid perceived or real social disapproval and social motives are found among individuals drinking for normative or affiliative reasons, such as on special occasions. Coping and enhancement motives have been associated with greater alcohol consumption (Cooper, 1994). In addition, coping, enhancement, and conformity motives have been associated with a greater number o f alcohol problems, even after controlling for usual alcohol consumption (Bradizza, Reifman, & Barnes, 1999; Carey & Correia, 1997; Cooper et al., 1992; Cooper, 1994) . Although not always the case (e.g., Bradizza et al., 1999), social motives appear less likely to be associated with problematic drinking (Cooper et al., 1992). Although drinking motives overlap within a person, there seems to be significant individual differences in the motivations individuals report, particularly in relatively inexperienced drinkers (Cooper et al., 1995) . Personality and symptom variables appear to be very important determinants o f the type and degree o f motivation an individual w i l l report for using alcohol. A m o n g female substance users, Conrod et al. (2000) found personality risk groups to further distinguish specific motivations for substance use based on patterns o f dependence on specific classes o f drugs (e.g., sedatives, stimulants, analgesics). Individuals high in anxiety sensitivity reported using more substances with anxiolytic properties (e.g., benzodiazepines) and  8 individuals high in hopelessness reported using more substances with analgesic properties (e.g., opiates). Subsequent studies have demonstrated that individuals scoring high on anxiety sensitivity or hopelessness report greater frequency and quantity o f alcohol use (Stewart, Peterson, & P i h l , 1995) and more alcohol-related problems in young adulthood (Woicik, Conrod, Stewart, Pihl, & Dongier, 2002). Moreover, hopelessness appears to relate to problematic alcohol use as strongly as do impulsivity and sensation-seeking (Woicik et al., 2002), the latter two personality factors known to be strong predictors o f alcohol use'(Cloninger, 1987; Conrod et al., 2000; Schuckit, 1998; Zuckerman & Kuhlman, 2000). Anxiety sensitivity, state and trait anxiety, and anxiety symptoms all appear to be positively related to negative reinforcement drinking (i.e., coping and conformity motives; Kushner, Thuras, Abrams, & Brekke, 2001; Stewart, Zvolensky, & Eifert, 2001, 2002; Comeau, Stewart, & Loba, 2001; Stewart & Zeitlin, 1995). Stewart et al. (2001) found the association between anxiety sensitivity and frequency o f alcohol use to be mediated by coping motives among females and the association between anxiety sensitivity and binge drinking to be mediated by conformity motives among males. A m o n g adolescents, anxiety sensitivity has also been shown to be a significant predictor o f conformity motives (Comeau et al., 2001). However, it may be that conformity motives take the form o f coping with negative emotions in social situations (Stewart et al., 2002). Trait anxiety, although able to significantly contribute to the prediction o f coping motives, appears to only account for a small portion o f the variance in coping motives after taking into account anxiety sensitivity (Stewart & Zeitlin, 1995). However, among adolescents, trait anxiety may be a significant predictor o f coping motives, especially in the presence o f high anxiety sensitivity (Comeau et al., 2001). Although much less research has been conducted on personality variables that put individuals at risk for depressive symptomatology, there is some evidence emerging that individuals experiencing hopelessness may use alcohol for negative reinforcement reasons  9 (Conrod, Stewart, & P i h l , 2002; Blackwell, Conrod, & Hanson, 2002). A s well, some evidence has been found for a relationship between drinking to cope and depressive symptomatology (Cooper et al., 1995; Peirce, Frone, Russell, & Cooper, 1994), whereas others have failed to find this relationship (Nishith, Resick, & Mueser, 2001). Cooper and colleagues found coping motives to mediate the relations between depression and both alcohol use and alcohol problems using path-analysis (Cooper et al., 1995; Peirce, Frone, Russell, & Cooper, 1994). O n the whole, drinking motives appear to be the final common pathway for alcohol use through which other variables influencing alcohol use are mediated (Cooper et al., 1995). Considering the extent to which depression- and anxiety-related variables overlap in accounting for psychopathology, further research is needed to explore the extent to which depressive symptomatology is able to account for unique variability in alcohol problems after accounting for anxiety sensitivity and trait anxiety, as well as determining the role that hopelessness may play. In addition, no research to date appears to have explored the mediational role played by coping and conformity motives in the relationship between personality/symptom variables and drinking behaviour, with respect to the relative contribution o f each o f these personality/symptom predictors (i.e., depression symptoms, hopelessness, anxiety sensitivity, trait anxiety) to alcohol problems i n a single model. The relative contribution o f each o f these personality/symptom variables to drinking motives and alcohol problems w i l l further elucidate the functional role o f alcohol among individuals at risk for alcohol use disorders. 2.1.2 Present Study Study 1 investigated the relations between depression, hopelessness, anxiety sensitivity, trait anxiety, and drinking motives i n an adolescent sample. In light o f evidence that negative reinforcement drinking is associated with anxiety sensitivity, trait anxiety, and depression (Cooper et al., 1995; Frone et al., 1994; Stewart et al., 2001, 2002; Stewart & Zeitlin, 1995), and that there is significant overlap between anxiety and depression (Clark & Watson, 1991; Gencoz,  10 2002), as well as evidence suggesting separate motivational processes for alcohol use among anxious and depressed individuals (Conrod et al., 2000), the current study examined the extent to which each o f these predictors, as well as hopelessness, can independently account for variance in coping and conformity motives, which in turn were expected to evince a relationship with alcohol problems, with the relationship between conformity motives and alcohol problems likely mediated by coping motives. 2.1.3 Hypotheses 1. Depression symptoms, hopelessness, anxiety sensitivity, and trait anxiety are positively related to coping and conformity drinking motives. Personality/symptom predictors are unrelated or negatively related to social and enhancement motives. 2. Depression symptoms and hopelessness account for unique variance i n coping drinking motives after taking anxiety sensitivity and trait anxiety into account. Anxiety sensitivity accounts for unique variance i n conformity motives after taking into account depression symptoms, hopelessness, and trait anxiety. 3. Depression symptoms, hopelessness, anxiety sensitivity, and trait anxiety are positively related to alcohol problems. 4.  Coping and conformity drinking motives mediate relations between personality/symptom predictors and alcohol problems as shown in Figure 1.  11 Figure 1. Hypothesized M o d e l for Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity  e  Note. A l l paths are expected to be positive. Solid lines indicate hypothesized pathways. Dashed lines indicate possible modifications that would lead to better model fit.  12 2.2 Method 2.2.1  Participants A sample o f 595 high school students was surveyed from a rural high school in N o v a  Scotia. The present sample consists o f 393 adolescents who reported drinking alcohol within the last 4 months. Another 21 (3.5%) participants were removed due to missing data (less than the 5% maximum recommended deletion o f missing cases; Roth, 1994), leaving a sample o f 381 (46% male) high school students between the ages o f 13 and 20 years ( M = 15.97, SD 1.25). Participants identified themselves as predominantly White (93%), with a minority o f other ethnicities represented: A s i a n (0.5%), East Indian (0.5%), Black (2%), First Nations (1%), and Other (3%). A second sample o f 612 adolescents was taken from three urban high schools in Vancouver, British Columbia. O f these, 311 reported drinking within the last 4 months. Another 5 (0.1 %) participants had missing data, leaving a total sample o f 306 (48% male) between the ages o f 15 and 19 years o f age ( M = 16.20, SD 0.66). The Vancouver sample appeared to be a much more ethnically diverse sample: White (37%), A s i a n (51 %), East Indian (6%), Black (0.5%), First Nations (0.5%), and Other (5%). A s it was the larger o f the two samples, the N o v a Scotia sample was used to examine the main hypotheses o f Study 1 i n order to provide adequate power for the initial structural equation analysis and to avoid problems due to non-normally distributed variables (see section 2.2.4 Statistical Analysis below). The Vancouver sample was used to cross-validate the structural model. The geographic (East vs. West Coast) and demographic (rural vs. urban) differences between these two samples provides a good test o f the generalizability o f the hypothesized relations among variables. A M A N O V A , carried-out to examine differences between the two samples, was significant, F(10, 675) = 11.87, p < .01. Follow-up A N O V A s revealed significant  differences in depressive symptomatology, anxiety sensitivity, trait anxiety, enhancement motives, alcohol problems and age (See Table 1). Table 1 Study and Cross-validation  Sample Descriptives  N o v a Scotia Variables  (H  = 381)  Vancouver (n = 306)  Depression  14.55(6.52)  15.72(5.81)  Hopelessness  16.88(4.70)  17.33(4.34)  CASI  27.27(5.86)  30.27(5.27)  b  STAIC-T  35.41(7.09)  37.73(6.51)  b  Coping  8.95(4.20)  9.16(4.33)  Conformity  7.31(3.74)  7.45(3.54)  Social  13.91(5.75)  13.76(3.54)  Enhancement  13.96(6.44)  13.01(5.26)  38.25(17.26)  31.77(12.14)  15.98(1.25)  16.20(.66)  Alcohol problems Age  a  a  b  b  Note. Hopelessness = Substance Use Risk Profile Scale-Hopelessness subscale; C A S I = Childhood Anxiety Sensitivity Index; S T A I C - T = State-Trait Anxiety Inventory for ChildrenTrait Subscale; Coping, Conformity, Social, Enhancement = Drinking Motives QuestionnaireRevised scales; A l c o h o l Problems = Rutger's Alcohol Problems Index. a  Follow-up A N O V A for differences between sites significant atp< .05  b  Follow-up A N O V A for differences between sites significant atp < .01  14  2.2.2 Measures Hopelessness.  The Substance Use Risk Profile Scale ( S U R P S ; W o i c i k et al., 2002)  assesses four personality profiles at-risk for substance abuse/dependence: hopelessness, anxiety sensitivity, impulsivity, and sensation-seeking (see Appendix). U n l i k e single scales, measuring each o f these factors separately, the S U R P S is designed to have non-overlapping items between the risk factors, to the extent that they can be discriminated. Overlap between scales can therefore be thought to represent overlapping profiles within an individual. W o i c i k et al. (2002) found items loaded as expected onto four distinct factors in both exploratory and confirmatory factor-analysis. Agreement with each o f the items is rated on a 4-point Likert scale ranging from 1 ("strongly disagree") to 4 ("strongly agree"). Each brief subscale appears to have sufficient Cronbach coefficient alphas (hopelessness a = .87, anxiety sensitivity a = .64, impulsivity a = .64, sensation-seeking a = .71) and has been shown to have good convergent and discriminant validity (Conrod et al., 2000; W o i c i k et al., 2002). The S U R P S scale has been used in both substance abusing (Conrod et al., 2000) and non-substance abusing undergraduate populations (Blackwell et al., 2002; W o i c i k et al., 2002), and has good test-retest reliability (Woicik et al., 2002). The S U R P S hopelessness scale is positively related to neuroticism, trait anxiety, low self-esteem, introversion and the Beck Hopelessness Scale (Beck, Weissman, Lester, & Trexler, 1974). Anxiety Sensitivity. The Childhood Anxiety Sensitivity Index ( C A S I ; Silverman, Fleisig, Rabian, & Peterson, 1991) is an 18-item scale designed to assess anxiety sensitivity in children and adolescents. Agreement with items is rated on a 3-point Likert scale ranging from 1 ("none") to 3 ("a lot"). The C A S I has good convergent and discriminant validity, and test-retest reliability (Silverman et al., 1991). The C A S I was used to measure anxiety sensitivity in the current sample due to its exhibiting better internal consistency ( a =.83) than the S U R P S anxiety sensitivity scale.  15 Drinking Motives.  Cooper's (1994) 20-item Revised Drinking Motives Questionnaire  ( D M Q - R ) provides a measure o f four drinking motives: coping motives (e.g., " T o forget about my worries", "Because it helps me when I feel depressed or nervous"), conformity motives ("To fit in with a group"), social motives ("As a way to celebrate"), and enhancement motives ("To get a high"). Participants indicate their frequency o f drinking for each o f the reasons on a 5point Likert scale ranging from 1 ("almost never/never") to 5 ("almost always/always"). The D M Q - R has good internal consistency and has undergone confirmatory factor analyses, validating the four-factor structure (Cooper, 1994). Depressive Symptomatology  The depression scale o f the B r i e f Symptom Inventory (BSI;  Derogatis & Melisaratos, 1983), a brief form o f the Symptom Checklist 90 (Derogatis, 1993), consisting o f seven items assessing general depressive symptomatology (i.e., suicidal ideation, loneliness, feeling blue, loss o f interest, hopelessness, worthlessness, feelings easily hurt), was used to assess depressive symptomatology. Participants rate items according to how much distress it has caused them during the past 7 days on a 5-point Likert scale ranging from 1 (not at all) to 5 (extremely). The Depression scale o f the B S I has good test-retest reliability and internal consistency ( a = .85), as well as good convergent, discriminant, and predictive validity (Derogatis, 1993). Alcohol Problems. Rutgers A l c o h o l Problems Index ( R A P I ; White & Labouvie, 1989) is a 23-item unidimensional self-report measure o f drinking consequences. The R A P I is able to assess alcohol problems i n populations where quantity and frequency o f alcohol use may not correspond to drinking consequences (e.g., adolescents; Thombs & Beck, 1994). Participants respond by indicating the frequency to which they have experienced each consequence o f alcohol use (e.g., "Passed out or fainted suddenly"), using categories 1 (never), 2 (1-2 times), 3 (3-4 times), 4 (5-6 times), 5 (more than 6 times). The R A P I has been shown to have good internal consistency ( a = .92).  16 Trait Anxiety. The State-Trait Anxiety Inventory for Children-Trait Subscale (Spielberger, Edwards, Montouri, & Lushene, 1973), a 20-item self-report questionnaire, was used to assess characterological or trait anxiety (e.g., "I get a funny feeling in my stomach", I worry too much."). Participants rate the frequency o f their experience " i n general" on a 3-point scale: 1 (almost never), 2 (sometimes), 3 (often). The S T A I C - T has been found to be a reliable ( a = .91) and valid measure o f trait anxiety among adolescents (Comeau et al., 2001; Muris, Merckelbach, Ollendick, K i n g , & Bogie, 2002). 2.2.3 Procedure Passive consent was obtained from parent(s) to allow their child to take part in the survey through a letter sent home with the students. One school also used a newsletter and one an automated calling system to inform parents o f the study. Parents were provided with a telephone number and an email address in order to withdraw their child from participation in the survey. Only four parents asked for their child to be excluded from the survey. Active consent was obtained from adolescents before completing the questionnaire. Adolescents were given the option to withdraw from the study. The exact number o f adolescents who chose not to take part in the survey was not recorded, but the number was very small. Adolescents completed the questionnaire during school time. Participants from the Vancouver site filled out the survey in their respective classrooms, whereas at the Halifax site participants filled out the survey in an assembly in the gymnasium. Participants were assured o f confidentiality and asked to be honest and accurate with their answers. However, not all participants were assured o f anonymity as some o f them provided their contact information so they could be contacted to take part in an intervention focus group (not reported here). 2.2.4 Statistical  Analyses  The relations among personality/symptom predictors (i.e., depression symptoms, hopelessness, anxiety sensitivity, and trait anxiety) and drinking motives (coping, conformity,  17 social, and enhancement motives, and alcohol problems) were examined using a combination o f Pearson's product moment correlation coefficients and multiple regression, in order to explore the independent variance accounted for by each o f the predictors. Several calculations involved exploring mediating variables (e.g., coping motives mediating between depressive symptomatology and alcohol frequency). The procedure established by Baron and Kenny (1986) was followed: (a) There must be a significant zero-order correlation between the predictor variable and the criterion variable, (b) There must be a significant zero-order correlation between the mediators and the criterion variable, (c) The relationship between the predictor and criterion variables should disappear once the mediators are entered into the regression equation. A form o f Sobel test o f mediation, the Goodman (I) test (Baron & Kenny, 1986; M a c K i n n o n , Warsi, & Dwyer, 1995), was used to test the extent to which the presence o f a mediator decreased the relations between the predictor and the criterion. A z-statistic higher than 1.65 indicates significance at an alpha ofp < .05 on the Goodman (I) test. According to D a v i d Kenny (personal communication, June 2003), the Goodman (I) test can be used with multiple predictors using hierarchal multiple regression. The relations outlined i n Figure 1 were examined using hierarchical multiple regression and structural equation modelling ( S E M ) . A M O S 4.0 (Arbuckle & Wothke, 1999) structural equation modeling program was used to conduct latent variable modelling. S E M followed a two-step approach, first fitting the hypothesized model to the study group (« = 381) and making any modifications, then using the second Vancouver sample (n - 306) to cross-validate the model (Kaplan, 2000). The most common estimates o f model fit depend upon minimization o f weighted squared discrepancies between the sample covariance matrix and the hypothesized covariance matrix implied by the model. The model outlined in Figure 1 was tested against the independence model, i n which exogenous and endogenous variables were assumed to be unrelated. In contrast, the saturated model, in which all variables were allowed to covary,  18 provided a benchmark to which the hypothesized model can be compared. A number o f absolute fit indices were used to examine model fit. The chi-square goodness-of-fit statistic and the % /df discrepancy ratio are reported, as well as Root Mean Square Error o f Approximation ( R M S E A ) , the Standardized Root M e a n Square Residual ( R M R ) and the Goodness-of-Fit Index (GFI). The % is a measure o f the null hypothesis that the hypothetical and sample covariance matrices are 2  the same. Thus, a significant yf indicates poor model fit. A % /df ratio less than 5 can be 2  considered a reasonable fit o f the model. R M S E A provides an approximation o f the population variance-covariance matrix in relation to the degrees o f freedom, with an R M S E A o f less than or equal to .05 indicating a close fit o f the model and greater than 0.1 a poor fit (Browne & Cudeck, 1993). R M R is a measure o f the square root o f the mean squared standardized residuals from the sample variances and covariances obtained under the model, and considered a better fit the closer to 0 it is. G F I is a comparison o f predicted and actual squared residuals, with values greater than .90 considered adequate model fit. T w o incremental indices o f fit were also used to evaluate model fit: the Comparative Fit Index (CFI) and the Tucker-Lewis Fit Index (TLI), both o f which provide a comparison o f the proposed and null models, adjusted for degrees o f freedom (Bentler, 1990; Bentler & Bonett, 1980). Each o f them ranges between 0 and 1, with greater than .90 considered an adequate fit o f the model. Modification indices were inspected for possible changes to the model to provide better fit with the data. When dealing with distributions that depart from multivariate normality, such as those found in substance abuse research, estimations derived from the sample covariance matrix and/or the optimal weight matrix may be incorrect (West, Finch, & Curran, 1995). In such cases, the % goodness-of-fit test can be subject to Type I error and fit indices underestimated (Chou & Bentler, 1995). However, as sample size increases (i.e., n > 200), nonnormality becomes less o f a threat, even with highly non-normally distributed variables (Anderson & Gerbing, 1988; H u &  19 Bentler, 1995; West et al., 1995). Moreover, several % goodness-of-fit statistics (e.g., M a x i m u m 2  Likelihood, Generalized Least Squares, S C A L E D statistic) have been shown to be robust in the face o f non-normality and, i n fact, over-estimate the % and underestimate path coefficients (Chou, 1995; H u & Bentler, 1995). Given that previous research has shown there to be gender differences with respect to both internalizing disorders and substance use (Kessler, McGonagle, Swartz, Blazer, & Nelson, 1993; Weissman, Bland, Joyce, Newman, Wells, & Wittchen, 1993; Deshmukh et al., 2003), it would seem prudent to examine gender in the structural model. However, the small sample size in the current study does not allow such hypotheses to be adequately examined.  20 2.3 Results 2.3.1 Age and Gender A g e was not significantly related to any o f the variables o f interest (i.e., depressive symptomatology, hopelessness, anxiety sensitivity, trait anxiety, drinking motives, and alcohol problems). Differences due to gender among variables were examined using M A N O V A s . For the N o v a Scotia sample, F(9, 360) = 5.27, p < .01, females scored higher than males on depression (15.50 vs. 13.42,/? < .01), anxiety sensitivity (28.15 vs. 26.29,/? < .01), and trait anxiety (33.88 vs. 36.89,/? < .01), whereas males scored higher than females on conformity motives (7.89 vs. 6.82,/? < .01) and alcohol problems (40.00 vs. 36.06,/? < .05). In the Vancouver sample, F(l, 443) = 2.49,/? < .01), similar results were found for females on depression (16.29 vs. 15.10,/? < .10), anxiety sensitivity (30.99 vs. 29.43,/? < .05), and trait anxiety (38.59 vs. 29.43,/? < .01), and for males on conformity motives (7.84 vs. 7.07,/? < .10) and alcohol problems (32.99 vs. 30.40,/? < .01). 2.3.2 Drinking Motives and Alcohol  Problems  Bivariate correlations among personality/symptom variables, alcohol motives, and alcohol problems are presented i n Table 2 (unless otherwise stated, the following analyses were carried out on the N o v a Scotia sample). Depression, anxiety sensitivity, and trait anxiety were all significantly related to coping and conformity drinking motives. Interestingly, hopelessness, anxiety sensitivity, and trait anxiety were positively correlated with enhancement motives. Anxiety sensitivity and trait anxiety evinced low positive correlations with social drinking motives. There appears to be significant overlap among the personality/symptom variables. Depressive symptomatology and trait anxiety were highly correlated, consistent with previous research findings on trait anxiety (Bieling, Antony, & Swinson, 1998; Nelson & Novy, 1997). L o w to moderate correlations were also found among drinking motives, with social and enhancement motives appearing to be highly related.  21 Table 2 Study Sample: Bivariate correlations among variables 1. 1. Depression 2. Hopelessness 3. C A S I 4.  .89  2.  3.  4.  5.  6.  .47**  .47**  .64**  .33**  .75  .26**  .32**  .84  STAIC-T  5. Coping 6. Conformity 7. Social 8.  Enhancement  9. Alcohol problems  7.  8.  .19**  .11*  .17*  .33*  .35**  .23**  .09  .14*  .37*  .65**  .33**  .33**  .16**  .17*  .30**  .85  .28"  .24"  .13"  .16"  .22*  .77  .51**  .52**  .53**  .55**  .85  .28**  .25**  .39**  .84  .74**  .32**  .90  .41**  **  **  **  **  _ _**  .94  Note. Hopelessness = Substance Use Risk Profile Scale-Hopelessness subscale; C A S I = Childhood Anxiety Sensitivity Index; S T A I C - T = State-Trait Anxiety Inventory for ChildrenTrait Subscale; Coping, Conformity, Social, & Enhancement = Drinking Motives QuestionnaireRevised scales; A l c o h o l Problems = Rutgers Alcohol Problems Index. Values on diagonal are coefficient alphas from the N o v a Scotia sample. *p < .05, **p< .01 two-tailed correlations  22 Depression, hopelessness, anxiety sensitivity, and trait anxiety were used to predict drinking motives using hierarchical multiple regression. Given the shared variance among drinking motives, residual scores for each participant were created by predicting each o f the drinking motives from the remaining motives using multiple regression, providing a measure o f the unique variance for each drinking motive without the influence o f the other motives (Comeau et al., 2001). For example, coping, social, and conformity motives were regressed on enhancement motives and the residual was saved to the data file. A series o f individual hierarchical regression analyses were then conducted in which demographics (i.e., age, gender) were entered i n Step 1 and personality/symptom predictors (i.e., depression, hopelessness, anxiety sensitivity, and trait anxiety) were entered in Step 2 (see Table 3). In the final model, demographic and personality/symptom variables were able to account for 9% o f the variance in the coping motives residual, F ( 6 , 363) = 6.81, p < .01, 6% o f the variance i n the conformity motives residual, F(6, 363) = 5.09, p < .01, 2% o f the variance in the social motives residual, F(6, 363) = 2.08, p < .10, and 2% o f the variance in the enhancement motives residual, F(6, 363) = 2.01,/? < .10. In Step 1, demographic variables were predictive o f conformity motives, F ( 2 , 367) = 5.39,/? < .01, social motives, F ( 2 , 367) = 3.49,/? < .05, and enhancement motives, F ( 2 , 367) = 5.19,/? < .01, but not coping motives, F ( 2 , 367) = 1.74,/? > .05. In Step 2, the personality/symptom variables were significant predictors o f coping and conformity motives, AF(4, 363) = 9.27,/? < .01 and AF(4, 363) = 4.83,/? < .01, respectively, but not social motives or enhancement motives, AF(4, 363) = .24, p > .05 and AF(4, 363) = .78,/? > .05, respectively. In the final model, age remained a significant predictor o f social and enhancement drinking motives, indicating that older adolescents were more likely to be drinking for social reasons and younger individuals more likely to drink for enhancement reasons. Among the personality/symptom variables, depression and hopelessness were significant predictors o f coping motives and anxiety sensitivity was predictive o f conformity motives.  22  vo  CO.  in o  vo  o  Tf  o  o o  S3 CD  s o §3  GO CD  "S  >  CQ  W oo  o  PH  ro  o  - H  o o o  ^H  o  I  00  03  c o GO l-l  <U PH  .13 •  .2 oo CD >  'o o oo  e  <D >  w  Tf  TT  o  o  00  0\  CO.  Q  <u GO  o  ^ H  o  ro  IT)  ON  O  CN  o  o  00  Q o  O  PH  PH  00  o  VO  CN O  C2L  O  o  PH  W o  Tl-  o  00  U  o  • I-H  ro  —'  o  oo o  oo ro  o o  r-  o  vo o  CN CN  Tf  O  -4—»  o c CQ.  00  •c Q  oo  13  'a,  PH  W  Tt-  U  00  o  PH  o  oo  S3  o  o  VO  o  o  <—  —'  CN O  uo O  T3  <D I-  Pi  53  ro  o  O  • »—I  GO oo CD 00 CD  P< CD  S3  "-4—»  O CD  oo  ro CD  1  o o >H  CD  >  CD  a, CD  oo  oo  a  o  CN OH CD  oo  00  <  •a a  CD  o  Q  CD  00 GO CD  S3 CO 00  !H  OH  <D  O  CX  (N O  ft*  CD  13 CD  o  .2  M  o  s  o  in CD  o  a  CD -*-»  >-  .2 o  CD  o  §  rt CD I  (H  _S o  •>-  a o CD >  c  C/3 CO  CD  OH O  O  IT)  0,  O  00 >-  A  CD <  o  H  OO  CU  O  OO VH  Vi  v in O  O  <2 m o  o o  rt i-  Hi  CD » rt -*-> OO  II H I  O  I—I  <OH  o  CD  «s (3  00  CN OH  O  -*-» 00  s  VH  o U  CD  l-l  <2 o  o  H  oo X CD  •a  O  N  -y rt CD  s  "§  CD  i/l  ">  CD  rt  a<  ^  'rt  CD OO CD  1) 11  ,o  i fD  -ri  i  .rt  CO  rt <y Vi  CD >  U II  OO  00  o c/> 13  CD  OH  OH  CD  13 CD  V Os  Vi  <2 < O  60  rt  •a ^  A  ti  'a  <rt2 u H OO  o U  60 g  o O 50  OH  JS  CD  'T  w  rt  o c«  oo°  60 g 'OH  o U  **  >-> / p V  V Si,  24 A l l four personality/symptom variables were significantly correlated with alcohol problems (see Table 2). Additionally, coping, conformity, social, and enhancement drinking motives were moderately related to alcohol problems. However, social motives were not a significant predictor o f alcohol problems (P = -.12, p < .10) once the variance due to the other three drinking motives had been taken into account, all o f which remained significant predictors of alcohol problems, F(4, 376) = 49.33, p < .01. Hierarchical multiple regression was used to demonstrate the relations among each o f the personality/symptom variables, coping and conformity drinking motives, and alcohol problems. Hopelessness and anxiety sensitivity were entered in the first step to examine the extent to which hopelessness could account for variability in alcohol problems in the presence o f a personality level anxiety variable. G i v e n the overlap between trait anxiety and depressive symptomatology in the present sample, they were entered in Step 2 to demonstrate the ability o f depressive symptomatology to explain variance not accounted for by the other three personality/symptom variables and to examine whether depressive symptomatology may be mediating between the other personality/symptom variables (e.g., hopelessness) and alcohol problems. Conformity motives were entered i n Step 3 to examine their ability to mediate between anxiety sensitivity and alcohol problems. Finally, coping motives were entered i n Step 4 to demonstrate their ability to mediate between personality/symptom variables and alcohol problems, and between conformity motives and alcohol problems. Coping and conformity motives were allowed to overlap instead o f using residuals. The results o f the regression analysis are provided in Table 4.  25 Table 4 Hierarchical Multiple Regression Predicting Alcohol Problems from Personality/Symptom Variables and Drinking Motives A l c o h o l Problems B  SEB  P  Hopelessness  1.14  .18  .31**  Anxiety sensitivity  .64  .14  .22**  Hopelessness  .97  .19  .26**  Anxiety sensitivity  .68  .18  .23  Trait anxiety  -.29  .17  -.12  Depressive symptomatology  .47  .17  .18**  Hopelessness  .80  .19  .22**  Anxiety sensitivity  .42  .18  .14*  Trait anxiety  -.32  .16  Depressive symptomatology  .49  .16  .19**  Conformity  1.34  .22  .29**  Hopelessness  .56  .18  .15**  Anxiety sensitivity  .31  .17  .10  Trait anxiety  -.29  .15  -.12  Depressive symptomatology  .35  .15  Variables Step 1  Step 2  Step 3  -.13  Step 4  .13*  26 Conformity  -62  .23  .13  Coping  1-56  .21  .38  Note. R = .14 for Step 1; Ai? = .02 (p < .01) for Step 2; Ai? = .08 (p < .01) for Step 3; Ai? = .10 2  2  2  2  (p < .01) for Step 4. Hopelessness = Substance Use Risk Profile Scale-hopelessness subscale; Anxiety sensitivity = Childhood Anxiety Sensitivity Index; Trait anxiety = State-Trait Anxiety Inventory for Children-Trait Subscale; Coping and Conformity = Drinking Motives Questionnaire-Revised scales; Alcohol Problems = Rutger's A l c o h o l Problems Index. *p< .05, **p< .01  In the final model, personality/symptom variables and drinking motives were able to account for 36% o f the variance i n alcohol problems, F(6, 374) = 35.64, p < .01. Personality/symptom variables were predictive o f alcohol problems in Step 1 and Step 2, F(2, 378) = 41.12,/? < .01 and AF(2, 376) = 3.95,p < .05, respectively. Both hopelessness and anxiety sensitivity were predictive o f alcohol problems in Step 1. In Step 2, depressive symptomatology was able to account for a significant portion o f the variance i n alcohol problems, and appeared to be acting as a partial mediator between hopelessness and alcohol problems, z = 3.50. Hopelessness and anxiety sensitivity remained significant predictors o f alcohol problems i n Step 2. Conformity motives were a significant predictor o f alcohol problems in Step 3, A F ( 1 , 375) = 37.87,/? < .01, and appeared to be a partial mediator between anxiety sensitivity and alcohol problems, z = 1.86. In the final model, coping motives were able to account for a significant portion o f the variance in alcohol problems, A F ( 1 , 374) = 55.51,/? < .01, and appeared to be mediating between the relations o f depressive symptomatology and hopelessness with alcohol problems, z = 3.68 and z = 3.30, respectively, and between conformity motives and alcohol problems, z = 7.31. Depressive symptomatology and hopelessness remained  27 significant predictors o f alcohol problems in the presence o f other personality/symptom variables and drinking motives. 2.3.3 Hypothesized and Modified  Models  Structural equation modeling was used in order to further demonstrate the relations among personality/symptom variables, drinking motives and alcohol problems in a single causal model. Each o f the scales (i.e., depression, hopelessness, anxiety sensitivity, coping motives, conformity motives, and alcohol problems) was split into two smaller scales to provide two observed indicators for each latent variable in the model (Arbuckle & Wothke, 1999). Given that trait anxiety was unable to significantly contribute to the prediction o f either drinking motives or alcohol problems in the presence o f the other three personality/symptom variables, it was dropped from the model. Overall the hypothesized model provided adequate fit i n comparison to the independence model, as indicated by high GFI, CFI, and T L I , as well as low R M R and R M S E A values (see Table 5). Figure 2 provides the results for the paths o f the predicted relations in the hypothesized model. The model appeared to account for a significant portion o f variability in each of the endogenous variables, accounting for 15% o f the variance i n conformity motives, 52% o f coping motives, and 47% o f alcohol problems. Although the hypothesized model resulted in a significant chi-square, the x /df ratio was less than 5. Finally, the direction and size o f the paths 2  were in line with previous research (e.g., Cooper et al., 1995). Given the significant chi-square, however, it seemed prudent to examine the modification indices for possible improvements in the model. Only theoretically based changes were made due to the atheoretic nature o f using modification indices to bring about better model fit (Anderson & Gerbing, 1988). A direct path from anxiety sensitivity to alcohol problems was able to significantly add to the model, x (1) 2  =  14.76, p < .01, and led to a non-significant chi-  square, x (45) = 52.62,/? < .05. A s well, a significant covariance path from the error terms o f 2  28 D P 2 (i.e., depression) and C O l (i.e., conformity motives), indicating that there is overlap in the variance unaccounted for by the depression and conformity motives measures, was able to further add to the model, % (1) = 7.14, p < .01 (see Figure 3). The non-significant chi-square for 2  this modified model, as well as the other absolute and incremental indices o f fit, indicates that the model is a good approximation Of population (see Table 5).  29 Table 5 Hypothesized and Modified Models: Summary o f Goodness-of-Fit Indices Fit Indices Model  I  Saturated  0.00  l /df 2  0 **  RMR  RMSEA  GFI  TLI  1.00  0.00  CFI 1.00  66  10.18  14.70  .15  .70  0.00  0.00  67.38*  46  1.46  1.28  .03  .98  .95  .95  45.47  44  1.05  .63  .01  .98  1.00  1.00  Independence  671.93  Hypothesized Modified * p < .05, **p < .01  df  2  30 Figure 2. Hypothesized M o d e l : Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity.  e  e  C02 .88~~^ e  '  t .73 e  Note. Depression = B r i e f Symptom Inventory-Depression subscale; Hopelessness = Substance Use Risk Profile Scale-Hopelessness subscale; Anxiety Sensitivity = Childhood Anxiety Sensitivity Index; Coping motives and Conformity motives = Drinking Motives Questionnaire-Revised scales; Alcohol Problems = Rutgers A l c o h o l Problems Index.  * p < .05, ** p < .01  31 Figure 3. Modified M o d e l : Depressive Symptomatology, Hopelessness, and Anxiety Sensitivity. e  e  Note. Depression = B r i e f Symptom Inventory-Depression subscale; Hopelessness = Substance Use Risk Profile Scale-Hopelessness subscale; Anxiety Sensitivity = Childhood Anxiety Sensitivity Index; Coping motives and Conformity motives = Drinking Motives Questionnaire-Revised scales; Alcohol Problems = Rutgers A l c o h o l Problems Index. /X.05,  **p<.0\  32 2.3.4 Cross-validation  of Model  The modified model was then cross-validated on the Vancouver sample (n = 306). Although none o f the personality/symptom variables were directly correlated with alcohol problems in the Vancouver sample (see Table 6), the model appeared to fit the sample data well, as indicated by a non-significant chi-square, and low % /df ratio well below 5 (see Table 7). In 2  addition, high G F I , T L I , and C F I values, as well as low R M R and R M S E A values indicate that the model fit the Vancouver sample data very well. Figure 4 illustrates the results for the paths of cross-validation model. Generally, many o f the path coefficients were relatively stable across the two samples, with the exception o f the path from anxiety sensitivity to alcohol problems and the covariance path between D P 2 and C O l error terms being non-significant in the Vancouver sample and significantly different from the N o v a Scotia sample, z = -2.86, p < .01 and z = -2.16, p < .01, respectively. N o other paths significantly differed between the two samples.  33 Table 6 Cross-validation Sample: Bivariate correlations among variables 1.  2.  3.  4.  skA  1. Depression 2. Hopelessness 3. Anxiety sensitivity 4. Trait anxiety  .85  5.  6.  7.  8.  9.  A  4ok  .63  .37  .71  .35  .13  -.04  -04  .07  .82  .23**  .52**  .35**  .15**  -.12*  -.08  .08  .78  .55**  .23**  .27**  .02  -.03  -.04  .83  .26**  .17**  -.07  -.03  -.02  .82  .41  .43  .39  .44  .83  .39**  .22**  .28**  .86  .67  .35  .88  .40**  5. Coping 6. Conformity 7. Social 8. Enhancement 9. Alcohol problems  .93  Note. Hopelessness = Substance Use Risk Profile Scale-hopelessness subscale; Anxiety sensitivity = Childhood Anxiety Sensitivity Index; Trait anxiety = State-Trait Anxiety Inventory for Children-Trait Subscale; Coping, Conformity, Social, & Enhancement = Drinking Motives Questionnaire-Revised scales; A l c o h o l Problems = Rutger's A l c o h o l Problems Index. Values on diagonal are Coefficient alphas from the Vancouver Sample. +  < .10, * p < . 05, ** p< . 01 two-tailed correlations  34 Table 7 Cross-validation o f Models: Summary o f Goodness-of-Fit Indices Fit Indices Model  ^  Saturated  0.00  Independence  1243.47**  Cross-validation  107.04  *p < .05, ** p <  .01  df  /df  2 x  RMR  RMSEA  ooo  0  GFI  TLI  CFI  1.00  LOO  132  9.42  11.70  .11  .70  0.00  0.00  88  1.22  .68  .02  .97  .97  .98  35 Figure 4. Cross-validation M o d e l : Depressive Symptomatology, Hopelessness, Anxiety Sensitivity, and Trait Anxiety. e  e  Note. Depression = B r i e f Symptom Inventory-Depression subscale; Hopelessness = Substance Use Risk Profile Scale-Hopelessness subscale; Anxiety Sensitivity = Childhood Anxiety Sensitivity Index; Coping motives and Conformity motives = Drinking Motives Questionnaire-Revised scales; A l c o h o l Problems = Rutgers A l c o h o l Problems Index. \p<  .10,p< .05, **p< .01  36 2.3.5 Depression Items, Alcohol Dependence and Depression Drinking  Motives  To begin to explore whether specific symptoms o f depression are particularly linked to drinking behaviour, correlational analyses were used to examine the extent to which individual depression items were related to alcohol problems and coping drinking motives (see Table 8). A l l seven depression items were significantly related to both alcohol problems and coping motives in the N o v a Scotia sample. Step-wise multiple regression was used to explore which o f the depression items were most strongly predictive o f alcohol problems and coping motives (see Table 9). Depression items were able to account for 2 1 % o f the variance i n alcohol problems and 18% o f the variance in coping motives, F(4, 376) = 25.04, p< .01 and F ( 4 , 376) = 22.35, p < .01, respectively. Suicidal ideation appeared to be the strongest predictor o f both alcohol problems and coping motives, entering in Step 1, F(l, 379) = 67.91,/? < .01 and F ( l , 379) = 56.72, p < .01, respectively, followed by hopelessness about the future in Step 2, A F ( 1 , 378) = 14.22,/? < .01 and A F ( 1 , 379) = 14.69,/? < .01, respectively. Interestingly, loneliness was inversely predictive o f both alcohol problems and coping drinking motives in Step 3, AF(1, 377) = 8.87,/? < .01 and A F ( 1 , 377) = 6.58,/? < .01, respectively. In Step 4, loss o f interest was a significant predictor o f alcohol problems, AF(1, 379) = 4.22,/? < .05, and worthlessness a significant predictor o f coping drinking motives, A F ( 1 , 376) = 6.98,/? < .01. Only those items found to be significant predictors o f alcohol problems and coping motives in the N o v a Scotia sample were correlated with alcohol problems and coping motives in the Vancouver sample (see Table 10). A s expected, given the lack o f direct relationship between depression and alcohol problems, no depression items were significantly individually related to alcohol problems in the Vancouver sample. Similar to the N o v a Scotia sample, suicidal ideation, hopelessness, loneliness, and worthlessness were all significantly related to coping motives in the Vancouver sample (see Note 1).  37 Table 8 Study Sample: Correlations  of Depression  items with Alcohol Problems and Coping motives Alcohol  Coping  Depression Items  problems '  motives  Suicidal ideation  .39"  .36"  Lonely  .17**  .17**  Blue  .21**  .22**  Loss o f interest  .29**  .25**  Hopeless about future  .35**  .34**  Worthlessness  .29**  .35**  Feelings easily hurt  .12*  .14**  *p<.05,  ** p<  .01  38 Table 9 Study Sample: Depression Items regressed on Alcohol Problems and Coping motives Alcohol problems  Coping motives  B  SE o f B  P  B  SE o f B  P  5.96  .72  .39"  1.34  .18  .36"  Suicidal ideation  4.4  .82  .29"  .95  .20  .26"  Hopeless about future  2.98  .79  .20"  .75  .19  .21"  Suicidal ideation  5.49  .89  .36"  1.18  .22  .32"  Hopeless about future  3.91  .84  .27"  .95  .21  .26"  Lonely  -2.59  .87  -.18"  -.55  .21  -.16"  Suicidal ideation  5.38  .89  .35"  1.05  .22  .28"  Hopeless about future  3.14  .92  .21"  .62  .24  .28"  Lonely  -3.07  .90  -.21"  -.73  .22  -.21"  Loss o f interest  1.89  .92  .12*  -  -  -  Worthlessness  -  -  -  .67  .25  .19"  Depression Items Step 1 Suicidal ideation Step 2  Step 3  Step 4  Note. R = .15 i n Step 1; Ai? = .03 (p < .01) in Step 2; Ai? = .02 (p < .01) in Step 3; Ai? = .01 (p 2  2  2  2  < .05) in Step 4 i n A l c o h o l Problems; i ? = .13 i n Step 1; AR = .03(p< .01) in Step 2; Ai? = .01 2  2  2  (p < .05) in Step 3; Ai? = .01 (p < .01) in Step 4 for Coping Motives. Line appears where no 2  relationship was found for that variable in Step 4. *p< .05, ** p< .01  39  Table 10 Cross-validation  Sample: Correlations  of Depression  items with Alcohol Problems and Coping  Motives Alcohol  Coping  Depression Items  problems  motives  Suicidal ideation  .08  .23  Hopeless about future  .08  .28"  Lonely  .08  .27"  Loss o f interest  -.05  (.18*)  Worthlessness  (.07)  .29**  Note. Parentheses indicate no a priori relationship specified based on the N o v a Scotia study sample. ** p < .01  40 2.4 Summary Study 1 was designed to explore the relative contribution o f hopelessness, depressive symptomatology, anxiety sensitivity, and trait anxiety to drinking motives and alcohol problems among adolescents. Previous research has found a relationship between each o f the four personality/symptom variables and both drinking motives and alcohol problems (e.g., Blackwell et al., 2002; Comeau et a l , 2001; Conrod et al., 2000; Kandel et al., 2001; Teichman et a l , 1989). Although Cooper and colleagues (e.g., Cooper et al., 1995) have investigated the role o f depression in drinking motives and alcohol problems, and Stewart and colleagues (e.g., Comeau et al., 2001; Stewart et al., 2001) the role o f anxiety sensitivity and trait anxiety, there appears to be no research to date exploring the relative contribution o f each o f these predictors to drinking motives and alcohol problems. In addition, hopelessness is emerging as an important predictor of substance use and problems (Conrod et al., 2000). To summarize, Study 1 provided preliminary evidence for depressive symptomatology, hopelessness, and anxiety sensitivity being able to account for independent variance in drinking motives and alcohol problems, after considering variance in alcohol problems accounted for by other predictors. Hypothesis 1 appears to have been supported by the current research in that all four personality/symptom variables were significantly related to negative reinforcement drinking. Consistent with previous research among adolescents and adults (Cooper et al., 1995; Peirce et al., 1994; Conrod et al., 2000), hopelessness and depressive symptomatology were significantly related to coping drinking motives. Anxiety sensitivity and trait anxiety were also positively related to both coping and conformity motives, providing further evidence that adolescents endorse similar drinking motives as adults (Comeau et al., 2001). Hypotheses 2 and 3, for the most part, were also supported. Regressing the personality/symptom predictors on drinking motive residuals provided a conservative test o f these relations. Depressive symptomatology and hopelessness were significant predictors o f  41 coping motives residuals, whereas anxiety sensitivity was the only significant positive personality/symptom predictor o f conformity motives residual after accounting for the other demographic and personality/symptom predictors i n the regression analysis. None o f the personality/symptom variables were predictive o f either social or enhancement motives in the regression analysis. The current study extends previous findings (Cooper et al., 1995) in that depression was able to account for independent variance i n coping motives after consideration o f anxiety sensitivity and trait anxiety, as well as variance due to other drinking motives. Although all four personality/symptom variables were significantly related to alcohol problems in the N o v a Scotia sample, they were not directly related to alcohol problems in the Vancouver sample. The lack o f direct relationship between personality symptom variables and alcohol problems in the Vancouver sample may be due to these adolescents not being as far along on the substance abuse cycle. Evidence for this comes from the N o v a Scotia sample scoring higher than the Vancouver sample on alcohol problems. A s w e l l , in comparison to the Vancouver sample, a greater proportion o f the N o v a Scotia sample consisted o f older adolescents. In accordance with Hypothesis 4, hopelessness, depressive symptomatology and anxiety sensitivity were all significant predictors o f alcohol problems in the N o v a Scotia sample, with these relations mediated by negative reinforcement drinking motives. Using structural equation modelling to examine these causal relations, the hypothesized model was comparable to the saturated model with respect to the relative fit indices and, therefore, a good approximation o f the population, but did not necessarily fit the data well, given the significant chi-square. The modified model provided a better fit in the study sample and also appeared to fit the crossvalidation sample data well, with no loss o f explanatory power in the modified model and a small loss o f variance accounted for in the Vancouver sample with respect to alcohol problems, likely due to variation between the two samples (Kaplan, 2000) and the lesser degree o f alcohol  42 problems in the Vancouver sample. Depression and hopelessness appeared to have a direct relationship with coping drinking and an indirect impact on alcohol problems, whereas anxiety sensitivity had both a direct impact on alcohol problems in the N o v a Scotia sample and an indirect impact through conformity motives. Both coping and conformity motives also directly influenced alcohol problems. Generally, depressive symptomatology, hopelessness, and anxiety sensitivity appear to be predictive o f alcohol problems through negative reinforcement drinking motives (i.e., coping and conformity motives). 2,4.1 Depressive Symptomatology and  Hopelessness  Both depressive symptomatology and hopelessness were significantly related to coping motives in the model. The personality factor o f hopelessness was able to account for independent variance in alcohol problems. That hopelessness and depressive symptomatology only evinced a small relationship with coping motives may be due to coping motives focusing mainly on anxiety related motivations for alcohol use. Blackwell et al. (2002) recently investigated the role o f depression specific drinking motives, finding a moderate relationship with hopelessness. Whether hopelessness would remain a significant predictor o f depression drinking motives in the presence o f depressive symptomatology is yet to be investigated, but the use o f depression motives in predicting depression and/or hopelessness related drinking would certainly provide a clearer picture o f why such individuals use alcohol to cope. Hopelessness also evinced a low positive correlation with enhancement motives in the N o v a Scotia sample, but was not a significant predictor o f enhancement motives in the regression analysis after accounting for demographic and other personality/symptom predictors. W i t h respect to depressive symptomatology, all seven depression items were significantly related to coping drinking motives and alcohol problems in the study sample. In the stepwise regression analysis B S I suicidal ideation and hopelessness items were significant predictors o f both alcohol problems and coping motives. Additionally, loss o f interest was a significant  f  43  predictor o f alcohol problems but not coping motives, suggesting that alcohol use may be leading to anhedonic symptoms o f depression. O n the other hand, worthlessness was a significant predictor o f coping motives but not alcohol problems, suggesting that it may be a motive for drinking, but less affected by drinking behaviour. Although none o f these depression items were associated with alcohol problems in the Vancouver sample, B S I suicidal ideation, hopelessness, and worthlessness depression items continued to be significantly associated with coping motives in the Vancouver sample. Nevertheless, recent evidence suggests that alcohol problems may be as much a result as a cause o f suicidal ideation (Light, Grube, Madde, & Gover, 2003), and it is worth considering that this relationship may also reflect the influence o f alcohol use on depression symptoms. It is interesting that the B S I hopelessness about the future item was a significant predictor o f both coping motives and alcohol problems in the study sample, and was related to coping motives in the Vancouver sample. This may also shed some light on the small path coefficients from S U R P S hopelessness to coping motives in the path model. It would seem that some o f the variance in coping motives due to hopelessness would have already been accounted for by B S I hopelessness and suicidal ideation depression items (Overholser et al., 1997). Certainly, there is a great deal o f overlap between the personality construct o f hopelessness and the symptom manifestation o f hopelessness in depression (Joiner et al., 2001). Nevertheless, whether at the personality or symptom level, hopelessness and its behavioural correlate, suicidal ideation, appear to be strong predictors o f motivation to use alcohol to cope and resultant alcohol problems. This was especially the case in the N o v a Scotia sample, considering that depressive symptomatology and hopelessness continued to be significant predictors o f alcohol problems in the regression analysis even after accounting for other personality/symptom variables and drinking motives. That loneliness was an inverse predictor of alcohol problems may indicate that within the context o f suicidal ideation, hopelessness and  44 drinking, reporting loneliness is a sign o f less severe disturbance (i.e., social isolation) and more that the individual is desirous o f social contact. Overall, depressive symptomatology and hopelessness were able to account for a moderate amount o f variance in coping motives and alcohol problems, as demonstrated by hierarchical regression analyses and structural equation modelling. Depression appears to play a significant role i n adolescents' use o f alcohol for coping reasons, especially considering that hopelessness was the only other personality/symptom variable with a significant path coefficient to coping motives. Moreover, it appears to be the hopelessness component o f depressive symptomatology that is accounting for variance in drinking motives and alcohol problems. Accordingly, hopelessness should be assessed when faced with an alcohol dependent individual with comorbid anxiety and depression. That hopelessness and hopelessness symptoms were also related to coping drinking motives in the Vancouver sample, though evincing less severe alcohol problems, suggests that hopelessness should be addressed i n prevention programs for individuals not yet experiencing problems from alcohol use. 2.4.2 Anxiety Sensitivity and Trait Anxiety Although adults appear to readily endorse coping motives i n relation to anxiety sensitivity (e.g., Kushner, Thuras, Abrams, Brekke, & Stritar, 2001; Stewart et al., 2001), the present study provides further confirmation that this may not be the case among adolescents (Comeau et al., 2001). Anxiety sensitivity was not a significant predictor o f coping motives in the regression analysis when other demographic and personality/symptom predictors and conformity motives were taken into account. O n the other hand, anxiety sensitivity consistently exerted significant influence on alcohol problems through conformity motives in the hierarchical regression analysis, as well as a direct path to alcohol problems in the N o v a Scotia sample model. The assumption by Stewart and colleagues concerning the mediation o f conformity motives through coping motives was also  45 borne out in the present sample (Comeau et al., 2001; Stewart et al., 2002). Anxiety sensitive individuals are especially concerned with social disapproval (Peterson & Reiss, 1992; Reiss, Peterson, Gursky, & M c N a l l y , 1986), and in turn appear to use alcohol as a means o f coping in social situations (Stewart et al., 2001). What we see among such adolescents then is drinking to conform because o f the anxiety associated with social situations. Thus, an adolescent who is high in anxiety sensitivity w i l l be more likely to perceive the possibility o f negative evaluation by peers as fear-provoking and therefore drink to cope with anxiety i n social situations (e.g., party). Comeau et al. (2001) suggest that fear o f anxiety symptoms may follow a developmental progression, in which the use o f substances to cope with anxiety sensitivity, not so readily seen in adolescence, may become more prominent in adulthood. The present research suggests that anxiety-motivated drinking can begin by using alcohol to cope with anxiety i h s b c i a l situations and then generalizes to other distressing situations at later stages o f development. Traditional treatment approaches to alcohol use disorders that do not address social anxiety w i l l be less effective with high anxiety sensitivity individuals. Furthermore, it would appear that prevention programs among adolescents need to address social anxiety and conformity motivations to drink. In opposition to previous research (Comeau et al., 2001; Stewart et al., 2001; Stewart & Zeitlin, 1995), trait anxiety did not remain a predictor o f coping motives residuals i n the presence of the other demographic and personality/symptom predictors and was not predictive o f alcohol problems. However, Comeau et al. (2001) found anxiety sensitivity moderated the relationship between trait anxiety and coping motives, wherein adolescents high i n anxiety sensitivity, in addition to being high trait anxiety, were more likely to endorse coping motives than those high in trait anxiety alone (see Note 2). However, previous studies on the relations o f trait anxiety and drinking motives have not included depressive symptomatology. It may be that depressive symptomatology was able to account for the variance in coping motives that had previously been  46 attributed to trait anxiety. Evidence for this comes from the fact that there was significant overlap between trait anxiety and the depression scale in the present sample. Indeed, trait anxiety as measured by the State-Trait Anxiety Inventory (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983) appears to assess a higher order factor of negative affect, consisting of one lower-order depression factor and one lower-order anxiety factor (Bieling et al., 1998). Overall, predictions concerning the role of anxiety sensitivity were found. Anxiety sensitivity was able to account for a significant portion of the variance in conformity motives, and evinced significant paths to both conformity motives and alcohol problems in the structural model. Conformity motives appear to be a significant mediator between anxiety sensitivity and alcohol problems among adolescents, given that the path from anxiety sensitivity to alcohol problems was not found in the Vancouver sample. In opposition to some previous research with adolescents, trait anxiety was not found to have a direct impact on negative reinforcement drinking in the presence of other personality/symptom variables. 2.4.3 Limitations The variance in alcohol problems accounted for in the hierarchical regression analysis and the modified and cross-validation models was smaller than that found in Cooper et al. (1995), even though the current model included more personality/symptom predictors. Given that age is a significant predictor of alcohol use (O'Malley, Johnston, & Bachman, 1998), it may be due to the mean age of the present samples being on average at least 1 year younger than Cooper et al.'s group of adolescents and, therefore, likely to have had less experience with alcohol. The lack of significant correlations between personality/symptom variables and alcohol problems in the Vancouver sample is also likely due to less experience with alcohol. It is interesting that age range of the Vancouver adolescents was lower and they reported less problematic use of alcohol, and, yet, depressive symptomatology and hopelessness were still related to negative reinforcement drinking motives, which in turn were related to alcohol  47 problems. Specific aspects o f neurotic personality and psychopathology, associated with alcohol use (Cooper, 1994; Cooper et al., 1995), appear to put adolescents at greater risk for developing substance use disorders as they acquire expectations regarding the negative reinforcement effects o f alcohol (see Smith & Goldman, 1994). Over time, depressive symptomatology may become more proximal to alcohol problems, resulting in it having a direct influence on alcohol problems. Intervention strategies are needed that attempt to change adolescents' personality/symptom risk (Stewart, Comeau, Loba et al., 2002) or their expectations concerning alcohol (Kraus, Smith, & Ratner, 1994) early on in this developmental process. Additionally, previous research has included alcohol frequency in the prediction o f alcohol problems, which appears to be able to account for a significant portion o f direct variance in alcohol problems (Cooper et al., 1995). The present study is limited by the exclusion o f a measure o f alcohol frequency in the model. Because alcohol use has been found to be predictive of depression (Schuckit et al., 1997; Schuckit, Tipp et al., 1997), it is not known from Study 1 i f depressive symptomatology would remain a significant predictor o f coping drinking motives and alcohol problems after accounting for alcohol frequency. 2.3.4 Sources of Motivation to Use Alcohol to Cope Interestingly, depressive symptomatology and hopelessness had a low positive correlation with enhancement motives i n the Vancouver sample, indicating that individuals high in hopelessness, i f anything, may be more likely to be drinking to enhance positive mood. Given that depressed or hopeless individuals evince low positive affect (Clark & Watson, 1991; Gencoz, 2002), it seems conceivable that such individuals would report using alcohol to increase positive affect. Indeed, Cooper et al. (1995) also found a low positive relationship between depression and enhancement motives, as well as with expectations that alcohol would enhance positive mood. However, no such relationship was found when other predictors, such as sensation-seeking, were considered. A n inspection o f enhancement motive items (e.g., "Because  48 it's exciting." "To get high.") reveals that the focus is on enhancing positive affect, rather than restoring positive affect, the former desired by sensation-seekers (Comeau et al., 2001) and the latter likely desired by depressed or hopeless individuals. It is interesting to note that Cooper et al. (1995) used depressive symptomatology as an analogue of negative affect, though depression appears to consist of both high negative affect and low positive affect (Clark & Watson, 1991; Gencoz, 2002). Given the inability of enhancement motives to reflect a desire for a normalizing effect of positive affect from alcohol, it is unclear whether alcohol use among depressed individuals is a result of a desire to restore/increase positive affect or a desire to numb/decrease negative affect. Moreover, individuals high in anxiety sensitivity and/or trait anxiety would also evince high negative affect and report drinking for the purpose of coping. The current and previous findings, taken together, indicate similar relations with respect to coping motives among individuals scoring high on depression and among those scoring high on anxiety sensitivity or trait anxiety. It is therefore difficult to distinguish whether these relations are due to overlapping constructs, comorbid psychopathology, or perhaps distinct motivations with the similar result of using alcohol to reduce negative affect. That is, it is unclear whether depressed individuals are drinking to reduce anxiety that tends to co-occur with depression, or if their depression symptoms represent specific motivation for alcohol use that is independent of a sedative effect. What's more, the present research offers no way of establishing the extent to which depressed individuals may be using alcohol to restore positive affect, a feature of depression that may provide an additional source of motivation for alcohol use.  49 CHAPTER  III Study 2  3.1 Introduction Study 2 was an attempt to address some o f the limitations in Study 1, as well as explore the possibility o f distinct motives for alcohol use among depressed versus anxious individuals. The first step was to establish depressive symptomatology and hopelessness as uniquely predictive o f drinking to cope with depression and anxiety sensitivity and/or trait anxiety as uniquely predictive o f drinking to cope with anxiety. Study 1 used a measure o f drinking motives that lumps depression- and anxiety-related motives for alcohol use together into one coping motives scale (Cooper, 1994; Cooper et al., 1992). However, motivations for drinking appear to be distinguishable on the basis o f what differentiates depression and anxiety. A s previously mentioned, both depression and anxiety appear to consist o f high negative affect, while only depression is characterized by low positive affect (Clark & Watson, 1991; Gencoz, 2002). Thus, depressed individuals could be drinking to increase positive affect and/or to decrease negative affect, while anxious individuals would likely only report drinking to decrease negative affect. O f note, Blackwell et al.'s (2002) depression drinking motives include items focusing on coping with negative affect associated with depression (e.g., " T o numb my pain"), as well as restoring positive affect ("To help me feel more positive about things in my life", "To cheer me up when I'm in a bad mood"), whereas their revised anxiety coping motives scale only consists o f items focusing on dealing with anxiety-related coping (e.g., " T o reduce my anxiety"). Study 2 used the depression motives scale and the revised version o f the anxiety coping motives scale. Interestingly, drinking for the purpose o f improving positive affect and drinking for the purpose o f reducing negative affect can be mapped onto brain reward and punishment systems, respectively (Gray, 1990, 1994). Evidence is emerging to suggest that depressed individuals evince an underactive brain reward system and perhaps also an overactive brain punishment  50 system (Johnson, Turner, & Iwata, 2003; Meyer, Johnson, & Winters, 2001; Depue & Iacono, 1989; Henriques et a l , 1994; Kasch, Rottenberg, Arnow, & Gotlib, 2002). However, to date, no research appears to have investigated the role o f brain reward and punishment systems with respect to alcohol use among depressed individuals and only one study has investigated hopelessness and brain reward and punishment systems among a group o f female alcohol users (Conrod, Stewart, & P i h l , 2002). A second purpose o f Study 2 was to attempt to separate negative affect and positive affect motivations for alcohol use among low versus high depression and hopelessness individuals using both self-report and experimental means. I w i l l begin with a discussion o f Reinforcement Sensitivity Theory (RST), which appears to have established the existence o f separate but interacting brain reward and punishment systems (Gray, 1970). Next, I w i l l discuss the evidence for the role o f these two brain systems in anxiety and depression, as well as emerging evidence for the role o f R S T i n hopelessness. Finally, where available, evidence for the relations o f these brain systems with substance use w i l l be presented. . 3.1.2 Reinforcement  Sensitivity Theory  Gray (1970,1982) has argued that there are two fundamental dimensions o f personality, representing individual differences in the strength o f functioning o f two different brain reinforcement systems, namely the behavioural activation system ( B A S ) , said to be responsible for impulsivity, and the behavioural inhibition system (BIS), said to be responsible for anxiety. The behavioural inhibition and activation systems are distinct in their response to reinforcement and therefore the theory is referred to as Reinforcement Sensitivity Theory (Pickering & Gray, 1999). The behavioural activation system responds to conditioned signals o f reward or nonpunishment (Gray, 1994). Dopamine appears to play a central role in B A S functioning, particularly in the substantia niagara or ventral tegmental area (Cloninger, 1987; Depue &  51 Iacono, 1989; Depue, Luciana, Arbisi, Collins, & Leon, 1994; Pickering & Gray, 1999). Rewarding stimuli, whether unconditioned or conditioned, by way o f striatal neurons, are suggested to increase dopaminergic output to the prefrontal cortex, which involves activation, selection, and control o f behaviours (Pickering & Gray, 1999). Increased motivational output in the presence o f reward and increased susceptibility to incentive reward has been found on behavioural tests (Arnett & Newman, 2000; Powell, Dawkins, & Davis, 2002), with the greatest increase in motivation found in individuals high i n impulsive/sensation-seeking personality traits and individuals with long D 4 D R alleles, hypothesized to be associated with increased striatal output in impulsive/sensation-seeking individuals (Pickering & Gray, 1999). Interestingly, alcohol and tobacco appear to interact with the brain reward system to produce a similar effect to that o f conventional reinforcers (e.g., food, sex) by increasing extracellular dopamine in the nucleus accumbens ( D i Chiara, Acquas, & Tanda, 1996; Reid, 1996). Gray and McNaughton (2000) postulate that the septo-hippocampal system and posterior cingulate cortex are largely responsible for BIS activity, which occurs in response to signals o f punishment, nonreward, novel stimuli and innate fear stimuli, resulting in inhibition or interruption o f ongoing behaviour, increased attention and arousal. They distinguish between BIS activity, resulting i n passive avoidance or avoidance o f frustrative nonreward (i.e., negative reward), and activity o f the fight or flight system (FFS), resulting in active avoidance or the avoidance o f punishment (i.e., negative punishment). BIS activity results in behavioural inhibition (i.e., focusing o f attention, environmental and internal scanning or risk assessment) in situations with conflicting approach-avoidance stimuli—that is, situations requiring an individual to move toward a source o f danger in order to reach a goal (Gray & McNaughton, 2000). Thus, BIS activity occurs i n situations requiring entry into, and the F F S system i n situations requiring exit from, threatening situations. Evidence for this distinction comes from studies with animals and humans using anxiolytic drugs, indicating that anxiolytics are effective in reducing BIS  52 behaviours, but ineffective i n reducing F F S behaviours (Gray, 1982; Gray & McNaughton, 2000). Gray (1990, 1994) attributes positive emotions, such as anticipation, hope, elation or happiness to B A S activity, and negative emotions such as fear, frustration, and sadness to BIS activity. Gray (Pickering, Corr, & Gray, 1997) has postulated that a combination o f neuroticism and extraversion contribute to mood, with negative affect positively related to neuroticism and negatively to extraversion, and positive affect positively correlated with both (see also Tellegen, 1985). Some studies have found positive affect to relate to extraversion, but not neuroticism, and negative affect to relate to neuroticism, but not extraversion (David, Green, Martin, & Suls, 1997; Larsen & Ketelaar, 1989, 1991; Watson & Clark, 1992). Certainly, studies o f mood correspond to two orthogonal factors o f positive and negative affect, which can be used to differentiate between different forms o f psychopathology (Clark & Watson, 1991). Heubeck, Wilkinson, and Cologon (1998) argue that a lack o f accepted operationalization of RST-related constructs is "the single most important issue which has held back" research on Gray's theory (p. 787). Predictions made by Gray's theory are often tested using personality measures developed to investigate quite different constructs (i.e., extraversion and neuroticism) that only minimally correspond to anxiety and impulsivity. It should be noted that Gray's dimensions are not synonymous with neuroticism and extraversion, but rather neuroticism and extraversion are viewed as derived personality dimensions. B I S and B A S are said to be found at an approximate 4 5 ° rotation to extraversion and introversion in personality space, such that high and low BIS run from the introverted-neurotic quadrant to the extraverted-stable quadrant, respectively, whereas high and low B A S run from the extraverted-neurotic quadrant to the introverted-stable quadrant, respectively (Gray, 1970,1982). Using factor analysis, Carver and White (1994) developed self-report B I S / B A S scales to directly assess reinforcement sensitivity. They derived one BIS scale, assessing anxiety in the  53 face o f impending punishment, and three B A S scales that load onto one higher order factor, assessing pursuit o f appetitive goals (Drive), responsiveness to rewards (Rewardresponsiveness), and the tendency to seek out new potentially rewarding experiences (Funseeking). Replications o f the principal components analysis in college and community samples and confirmatory factor analysis in both English and non-English speaking samples indicate that a four-factor model provides the best fit to variance on the scale (Heubeck, Wilkinson, & Cologon, 1998; Jorm, Christiansen, Henderson, Jacomb, Korten, & Rodgers, 1999; Leone, Perugini, Bagozzi, Pierro, & Mannetti, 2001). A l l three B A S scales positively relate to extraversion, positive affect, and positive temperament, and are unrelated to trait anxiety, negative affect, or negative temperament (Carver & White, 1994; Heubeck et al., 1998; Jorm et al., 1999). Each B A S scale appears to measure different aspects o f B A S functioning, however. For example, reward-responsiveness and funseeking individually positively relate to reward dependence and novelty seeking, respectively, while both negatively relate to harm avoidance (Carver & White, 1994). In various samples the BIS scale has been related to trait and state anxiety, depression, neuroticism, negative affect, negative temperament, and harm avoidance, and either negatively related or unrelated to optimism, extraversion, and positive affect (Carver & White, 1994; Heubeck et al., 1998; Jorm et al., 1999). The implications o f BIS and B A S functioning appear to be wider than simply the constructs o f anxiety and impulsiveness. Rather, as originally proposed by Gray (1970), BIS and B A S represent orthogonal dimensions corresponding to individual differences in sensitivity to reward and punishment. However, the implications for anxiety and impulsive/sensation-seeking have been the focus o f much o f the research on Reinforcement Sensitivity Theory. This is likely due to the implied correspondence between anxiety and B I S activity, and impulsivity and B A S  54 activity. Only recently has there begun to emerge research concerning the relations o f R S T and other personality and symptom variables.  3.1.3 Reinforcement Sensitivity: Depression and Hopelessness Gray (1994) has made some specific predictions concerning B I S and B A S activity in relation to depression, specifically that there is a continuum o f depression, with neurotic depression at one end and psychotic depression at the other. Neurotic depression would be seen in individuals with mixed anxiety and depression, whereas psychotic depression would be seen in individuals with lowered reward-responsiveness and motivation, but little anxiety. Through the process o f reciprocal inhibition between the BIS and B A S (Gray, 1987; Pickering & Gray, 1999; Gray & MacNaughton, 2000) both types o f depression would exhibit high BIS and low B A S activity. Neurotic depression would result from reciprocal inhibition o f BIS on B A S activity, resulting in high negative affect and low positive affect (see also Watson & Clark, 1991; Gencoz, 2002). Whereas, low B A S activity in psychotic depression would result in a lack o f inhibition on the B I S , and in turn low positive affect and high negative affect (Gray, 1994). Depression then could result from a lack o f reward-responsiveness due to an underactive B A S or from increased vulnerability to punishment due to an overactive B I S , either way resulting in high low positive affect and negative affect. Both hopelessness and depression are associated with high neuroticism and low extraversion (Saklofske, K e l l y , & Janzen, 1995; Tanaka, Sakamoto, Ono, Fujihara, & Kitamura, 1998; Velting, 1999), which corresponds with elevated BIS activity. A s previously mentioned, the BIS scale is positively associated with depression (Jorm et al., 1999), and the BIS and B A S scales appear to correspond to positive and negative affect, respectively (Carver & White, 1994; Heubeck et al., 1998; Jorm et al., 1999). There is some evidence that hopelessness is positively correlated with the B I S scale and negatively correlated with the B A S scale (Woicik, unpublished  55 data). Interestingly, introverts appear to be less susceptible to positive mood induction procedures and stable individuals less susceptible to negative mood induction (Larsen & Ketelaar, 1989, 1991), suggesting that individuals scoring high on hopelessness m a y b e less susceptible to rewarding stimuli and more susceptible to punishing stimuli. Meyer, Johnson, and Carver (1999) found relations between B I S / B A S scales and depression symptoms among a group o f college students at risk for bipolar spectrum symptomatology, with B I S positively and B A S negatively associated with depression. Hierarchical regression revealed that BIS and B A S reward-responsiveness accounted for 44% o f the variance i n depression. Given that they did not assess anxiety, it is difficult to distinguish the roles o f BIS and B A S activity in depression versus depression mixed with anxiety. Furthermore, all o f their measures were self-report, which may have led to artificial correlations due to method variance. Interestingly, Johnson et al. (2000) found no relationship between goal attainment, a real-life measure o f B A S activity (i.e., motivation), and depressive symptoms in a group of bipolar inpatients. In order to overcome problems with using only self-report, the present study employed both self-report and behavioural measures o f BIS and B A S activity. Further evidence o f the role o f the behavioural activation system in depression can be found in resting frontal E E G studies. Relative left frontal E E G activity has been related to approach behaviour and the expression o f positive emotions and relative right frontal activity to withdrawal behaviour and the expression o f negative emotions (Davidson, 1992, 2001). A s well, positive and negative mood induction procedures have been associated with greater relative left and right activation, respectively (Davidson, Ekman, Saron, Senulis, & Friesen, 1990; Hagemann, Naumann, Becker, Maier, & Bartussek, 1998). The relationship between depressed mood and frontal activation appears to be such that depressed individuals show lower relative left frontal E E G activity (Henriques & Davidson, 1990,1997), even when asymptomatic (Gotlib, Ranganath, & Rosenfeld, 1998). Likewise, individuals scoring high on the B A S scales show  56 greater relative left than right frontal activity, whereas individuals scoring low on B A S show little asymmetry (Harmon-Jones & Allen, 1997). Given the behaviours associated with relative left frontal activity, it is likely that depressed individuals would engage in less approach behaviour and evince lower reward-responsiveness, and thus lower B A S activity. Certainly, reduced motivation and lessened behavioural activity correspond with the clinical presentation o f depressed individuals (American Psychological Association, 1994), and especially to that o f the hopelessness subtype o f depression (Abramson et al., 1989; A l l o y & Clements, 1998), which may also include lowered reward-responsiveness (Metalsky & Joiner, 1997; Whisman & Pinto, 1997). Henriques et al. (1994) used a memory task, in which they asked dysphoric and nondysphoric individuals, selected on the basis o f high and low BDI-II scores, to identify words they had been presented with earlier. In three separate conditions, individuals were a) simply asked to identify target words, b) received a monetary reward for correctly identifying target words, c) or lost a monetary reward for failing to identify target words. Since there was no loss of monetary reward for incorrectly identifying a nontarget word, a more liberal strategy in both reward and punishment trials would have been to their advantage. Although they found no differences between dysphoric and nondysphoric individuals' ability to remember target words, the dysphoric group was more conservative during the reward trial than the nondysphoric group. Furthermore, dysphorics' responding did not differ between the neutral and reward condition and they took a much more liberal approach on the punishment trial. That is, in comparison to the nondysphoric group, the dysphoric group's performance appeared to be unaffected by the presence o f a reward, and more avoidant in the face o f punishment. Henriques and Davidson (2000) replicated the finding o f reduced reward-responsiveness among a group o f depressed individuals i n comparison to nondepressed controls. They were, however, unable to replicate the finding o f a more avoidant stance i n response to punishment among depressed individuals, but  57 rather found that this stance was taken by a subset o f individuals with a comorbid past or current anxiety disorder, whereas nonanxious depressed individuals were more conservative (i.e., less avoidant) during the punishment condition. Their inability to find differences between depressed and nondepressed individuals i n their punishment condition could also havebeen due to lack o f power. Only one study has examined hopelessness using a behavioural measure o f reinforcement sensitivity. Conrod, Stewart, and Pihl (2002) found a small relationship between hopelessness and reinforcement sensitivity among a group o f substance abusing/dependent women. Individuals low on hopelessness evinced a disinhibiting effect i n a mixed reward/punishment condition that was not found among high hopelessness individuals. However, this was only the case when the mixed condition was presented prior to a punishment condition, but disappeared when the punishment condition was presented prior to the mixed condition. Interestingly, high hopelessness individuals did not show any differences due to the presence o f reward or for the order o f conditions. Considering that behavioural reward and punishment procedures can affect mood (see Gomez & McLaren, 1997; Gomez et al., 2000), presenting punishment prior to the mixed incentive condition may have induced a dysphoric mood among low hopelessness individuals, perhaps leaving them similar to the high hopelessness individuals in terms o f reward-responsiveness. Conrod et al.'s research provides some novel findings with respect to differences in reward-responsiveness and punishment susceptibility between low and high hopelessness groups. In order to further examine the interaction o f hopelessness with reward and punishment, the present research examined reward and punishment separately, as opposed to the use o f a mixed incentive condition. 3.1.4 Reinforcement Sensitivity and Substance Use Although little research has been done directly predicting alcohol use from B A S and BIS activity, the BIS scale appears to be positively related to alcohol misuse (Loxton & Dawe, 2001)  58  and high scores on the BAS-drive and fun-seeking scales appear to be predictive of excessive alcohol use (Loxton & Dawe, 2001; Johnson et al., 2003) and alcohol craving (Franken, 2002). High scores on the BAS-reward-responsiveness and drive scales have also been found to be related to the belief that alcohol will relieve craving (Franken, 2002). In a comparison between heavy and light drinkers, only BAS reward-responsiveness was predictive of craving in the presence of alcohol cues (Kambouropoulos & Staiger, 2001). Whether assessed using the BIS/BAS scales or extraversion and neuroticism, it appears that a personality profile of high BAS (i.e., impulsivity) is predictive of drinking. There appears to be no available evidence that low BAS is predictive of alcohol use. However, recent research using a behavioural measure indicates that individuals with lowered reward-responsiveness may engage in substance use to restore reward-responsiveness. Powell and her colleagues have investigated reward-responsiveness using the Card Arranging Reward Responsivity Objective Test (CARROT) in brain-injured patients and abstaining smokers (Powell, Al-Adawi, Morgan, & Greenwood, 1996; Powell, Dawkins, & Davis, 2002). The C A R R O T is a card-sorting task in which a reward trial is compared to nonreward trials as a measure of susceptibility to incentive reward. In a group of brain-injured patients with motivational impairments, severity of impairment was highly related to rewardresponsiveness and, among 11 poorly motivated patients, bromocriptine, a dopamine agonist, significantly increased reward-responsiveness (Powell et al., 1996). Al-Adawi and Powell (1997) administered the C A R R O T to a group of Muslim smokers abstaining for Ramadan. Although they found no difference between non-smokers and smokers who had been abstaining from cigarettes overnight on non-reward trials, non-smokers sorted much faster on reward trials than abstaining smokers. After smoking a cigarette, previously abstinent smokers continued to resemble non-smokers on the non-rewarded trial, but also showed a similar increase in card-sorting rate on the reward trial, indicating that reward-  59 responsiveness had been restored. Furthermore, smokers who were abstaining from cigarettes for all o f Ramadan (i.e., at least 10 days at the time o f the study) showed little increase in reward-responsiveness subsequent to consuming a small amount o f food, indicating that the mere presence o f appetitive stimuli was not enough to restore reward-responsiveness.  Rather,  overnight abstaining smokers' improved performance on the reward trial after smoking was likely due to nicotine stimulating dopamine activity ( A l - A d a w i & Powell, 1997). Powell, Dawkins, and Davis (2002) replicated these findings with smokers abstaining overnight, with the increase i n reward-responsiveness subsequent to smoking remaining significant after withdrawal symptoms were taken into account. Moreover, abstaining smokers indicated receiving less pleasure from activities than both current smokers and non-smokers, who did not differ from one another. Powell and her colleagues' research indicates that rewardresponsiveness can be' dampened by chronic nicotine use, and that the use o f drugs may be motivated by the restoration o f reward-responsiveness. Moreover, an association between the drug and its restorative effect may be established to the point where only a cue for the drug is enough to produce this restorative effect. Kambouropoulos and Staiger (2001) assessed heavy and light drinkers using the C A R R O T , both before and after being exposed to an alcohol cue. Heavy drinkers were defined as engaging in binge drinking (i.e., four or more drinks for females, five or more for males) at least once a week for the 2 previous weeks. The alcohol cue consisted o f taking a sip o f beer and focusing on it for 3 minutes. Similar to Powell's ( A l - A d a w i & Powell, 1997; Powell et al., 2002) findings, light drinkers showed normal reward-responsiveness, whereas heavy drinkers showed virtually no reward-responsiveness on the reward trial in the neutral condition (i.e., sip of water). Following the alcohol cue, card-sorting speed on the reward trial increased among heavy drinkers, but had no effect on light drinkers (Kambouropoulos & Staiger, 2001). Although heavy and light drinkers did not appear to differ on the B A S scales, the differences in  60 card-sorting speed on the reward trial between light and heavy drinkers suggests lower rewardresponsiveness in heavy drinkers, which was then restored by an alcohol cue (Kambouropoulos & Staiger, 2001). It may be that the C A R R O T task is more sensitive to B A S activity than selfreport, in that it is able to detect variability in B A S activity where the B A S scales are not. The direct physiological effect o f such a small amount o f alcohol was unlikely to have been responsible for the restoration o f reward-responsiveness, especially when one takes into account alcohol tolerance among heavy drinkers. Therefore, the process o f restored rewardresponsiveness may be due to secondary conditioning o f alcohol cues to the primary reward (i.e., alcohol intoxication), resulting in an alcohol cue being enough to increase B A S activity (Pickering & Gray, 1999). However, restoration o f reward-responsiveness maybe mediated through higher cognitive functions, such as individuals' expectations concerning the physiological effects o f alcohol. There is a large body o f research on alcohol expectancies (Brown, Christiansen, & Goldman, 1987; Brown, Goldman, Inn, & Anderson, 1980; Goldman, Greenbalm, Darkes, 1997; Leigh & Stacey, 1993; Sheafer & Leigh, 1996; Natvig & Leigh, 1998; Williams & Clark, 1998). That reduced responsivity to reward and vulnerability to threat can act as motivators to consume alcohol and other drugs may seem intuitive. Drinking i n order to cope with negative emotions or to conform with others' expectations are synonymous with negative reinforcement drinking, whereas drinking to enhance positive emotions or to be sociable are synonymous with positive reinforcement drinking (Cooper, 1994; Cooper, Russell, Skinner, & Windle, 1992 ; Cooper, Frone, Russell, & Mudar, 1995). Stewart and Divine (2000) found neuroticism to be positively related and extraversion negatively related to drinking for coping reasons. However, neuroticism was the only significant predictor in a multiple regression analysis, with depression and self-consciousness facets o f neuroticism related to coping motives (Stewart & Divine, 2000). Moreover, the relationship between neuroticism and drinking problems appears to be mediated  61 by coping motives (Stewart, Laughlin, & Rhyno, 2001). Grau and Ortet (1999), on the other hand, found extraversion but not introversion to be predictive o f frequency o f alcohol consumption. Collectively, these studies indicate that high neuroticism or depression and high extraversion are predictive o f excessive alcohol use. Motivation for substance use among hopeless or depressed individuals is complicated by the fact that such individuals have been found to use substances that are traditionally understood to have depressant effects (Conrod et al., 2000). Alcohol, for example, would be expected to exacerbate feelings o f hopelessness. Then again, alcohol has been found to have biphasic effects at different points on the blood alcohol curve, which interact with individual differences to influence reinforcement sensitivity (Brunelle et al., 2003). When first ingested, on the ascending limb o f the blood alcohol curve, alcohol appears to have a stimulant effect, whereas on the descending curve it appears to have a sedating effect (Conrod, Peterson, Pihl, & Mankowski, 2001; Earleywine, 1994a; Papineau, Roehrs, Petrucelli, Rosenthal, & Roth, 1998). Furthermore, such effects appear to be dose related, with low doses o f alcohol producing behavioural facilitation and higher doses behavioural inhibition (Varlinskaya, Spear, & Spear, 2001). Alcohol also seems to act on multiple brain systems producing both stimulant and sedative effects (Di Chiara et al., 1996; Reid, 1996), likely interacting with individual differences in personality (Conrod, P i h l , & Vassileva, 1998; Gianoulakis, De Waele, & Thavundayil, 1996; Pihl & Peterson, 1995). The differential effects o f alcohol raise the question as to whether hopeless or depressed individuals are motivated to drink to achieve positive, stimulant effects (e.g., increase positive affect, socialization behaviour, etc.) or to achieve sedative effects (e.g., decrease negative affect, numbing, etc.). 3.1.5 Present Study The purpose o f the second study is to extend the findings o f Study 1 i n a young adult population and to establish a relationship between reinforcement sensitivity and both  62 hopelessness and depression. Because positive and negative affect appear to be orthogonal (Clark & Watson, 1991), individuals experiencing hopelessness or depression may drink to increase positive affect and/or to decrease negative affect. The construct o f coping drinking motives, though helpful, does not provide clarity with respect to the particular motivations to drink among depressed or hopeless individuals. The present research therefore explored the extent to which coping drinking motives, split into depression coping motives and anxiety coping motives (Blackwell, 2003; Blackwell et al., 2002), can better account for negative reinforcement drinking. Previous research finding a relationship between depressive symptomatology or hopelessness and the B I S / B A S scales, with individuals high in depression or hopelessness scoring higher on BIS and lower on B A S (Carver & White, 1994; Heubeck et al., 1998; Jorm et al., 1999; Meyer et al., 1999; Woicik, unpublished data) has begun to confirm Gray's (1994) hypotheses concerning depression. Pickering and Gray (1999) postulate that high introversion should be associated with response slowing i n the face o f punishment, since introverts are likely to exhibit low B A S and therefore little reciprocal inhibition on BIS activity. Performance differences on behavioural tasks, however, have provided mixed results, with dysphoric or depressed individuals showing less reward-responsiveness than nondepressed individuals, and little difference between the two groups on punishment susceptibility (Henriques et al., 1994; Henriques & Davidson, 2000). Similarly, both depression and B A S activity have been related to frontal asymmetry (Davidson et al., 1990; Hagemann et a l , 1998; Henriques & Davidson, 1990, 1997), suggesting less approach behaviour, but no differences in avoidance behaviour (see Davidson, 1992). Because positive and negative emotions have been said to correspond to B A S and BIS activity, respectively (Gray, 1990,1994), and appear to differentiate between depression and anxiety, in that depression is uniquely related to low positive affect (Clark & Watson, 1991; Gencoz, 2002), Study 2 was used to investigate the extent to which hopelessness and depression  63 are related to low B A S and high BIS activity, and the extent to which these motivational features mediate drinking behaviour and drinking problems. Reinforcement sensitivity was assessed using a modified version of the CARROT that includes both a reward and punishment condition. It seemed prudent to modify the CARROT, rather than using two different tasks to assess reward-responsiveness and punishment susceptibility. One advantage of using the CARROT is that individuals receive reinforcement throughout the task, as opposed to only for certain responses. Non-reward may inadvertently activate the BIS due to frustrative nonreward, resulting in slowed responding (Arnett & Newman, 2000; Corr, 2002). In this case, it would be unclear whether slowed responding was due to low BAS activity or to inhibition of BAS activity by the BIS. Another advantage is that there is no significant learning component to this task, thus precluding any influence of cognitive factors on task performance. Although seen in psychotic disorders, a general slowing of reaction time does not appear to be characteristic of affective disorders, without psychotic features (Schwartz et al., 1989). Therefore, any differences between individuals scoring high and low on hopelessness or depressive symptomatology are likely to be seen on rewarded trials rather than on nonrewarded trials. In addition to the C A R R O T , the self-report BIS/BAS scales was used. Although no specific data were given, Pickering and Gray (1999) reported finding the expected relations between Carver and White's (1994) BIS/BAS scales and the C A R R O T . The BIS/BAS scales appear to have received sufficient support for their inclusion as a self-report measure of BIS and BAS activity. Overall, the BIS scale appears to correspond to predictions better than the BAS scales. It seems prudent, therefore, to include a measure of extraversion and neuroticism, in addition to the BIS/BAS scales. Evidence suggests that alcohol and other drugs may be able to restore rewardresponsiveness among individuals reporting chronic or heavy use of such substances (Al-Adawi  64 & Powell, 1997; Kambouropoulos & Staiger, 2001; Powell, Dawkins & Davis, 2002). However, there does not appear to be any research establishing the ability o f alcohol to restore rewardresponsiveness among hopeless or depressed individuals. The present study used an experimental cue-reactivity procedure to investigate restoration o f reward-responsiveness and diminishment o f punishment-susceptibility in response to an alcohol cue among individuals scoring high and low on hopelessness and depressive symptomatology. A cue for alcohol appears to be enough to lead to alcohol craving and physiological changes associated with alcohol (Carter & Tiffany, 1999; Kambouropoulos & Staiger, 2001; M c K a y & Schare, 1989). Restored reward-responsiveness and/or diminished punishment susceptibility is expected only among high hopelessness individuals who report drinking. Those who score high on hopelessness, but do not drink are unlikely to receive a restorative effect from alcohol cues, as such an association is unlikely to have been established. A l l four personality/symptom variables (i.e., hopelessness, depression, anxiety sensitivity, and trait anxiety) have been associated with alcohol use (Cooper et al., 1995; Mehrabian, 2001; Stewart et al., 1995; W o i c i k et al., 2002). In addition, both coping and conformity motives are associated with frequency o f alcohol use (Cooper et al., 1995; Peirce et al., 1994; Stewart et al., 2001). Therefore, a measure o f frequency o f alcohol use was included in order to further examine the ability o f the personality/symptom variables to predict drinking motives and alcohol problems after accounting for alcohol use. A number o f the limitations o f Study 1 were addressed i n the current study. First, personality/symptom predictors only being able to account for a modest amount o f variance in alcohol problems may have been due to limited experience with alcohol among an adolescent population. Therefore, the current study used a young adult sample in order to allow sufficient variability with respect to alcohol use and to extend the findings o f Study 1 to an adult population. In line with previous research concerning personality/symptom predictors and  65 alcohol behaviours, Study 2 employed a measure o f alcohol frequency i n prediction o f alcohol problems. In addition, alcohol frequency was used to examine the ability o f personality/symptom predictors to account for variance in drinking motives after controlling for frequency o f alcohol use. This is o f special importance given the lack o f clear directional causality between depression and alcohol use—that is, it appears just as likely that alcohol use causes depressive symptomatology as it does depressive symptomatology leads to alcohol use (Schuckit et al., 1997a; Schuckit et al., 1997b). Finally, in order to place the current line o f research into the wider arena o f alcohol use disorders, it seemed appropriate to incorporate a measure o f D S M - I V alcohol abuse and dependence (American Psychological Association, 1994). Finally, the relationship between depression and conformity motives error terms in Study 1 may have been due to the use o f only a 7-item measure o f depression. The 21-item Beck Depression Inventory-II (Beck, Steer, & Brown, 1996) was used to assess depression in the current study. The following is a list o f the primary hypotheses under investigation in the present research. Secondary hypotheses, such as the relationship between B I S and B A S activity, as assessed by behavioural and self-report measures and other self-report measures are reported, but are not listed below. Predictions concerning gender differences on B I S and B A S activity are difficult, as few studies have found such differences (Diaz & Pickering, 1993; Pickering, Diaz & Gray, 1995). Therefore, no predictions concerning gender are made. 3.1.6 Hypotheses 1. After controlling for frequency o f alcohol use, depressive symptomatology is expected to be predictive o f depression drinking motives, whereas anxiety sensitivity and/or trait anxiety are expected to be predictive o f anxiety drinking motives, and anxiety sensitivity predictive o f conformity motives. The extent to which hopelessness is able to account for significant variance in depression motives in the presence o f depressive symptomatology is unknown,  66 but this hypothesis was examined. The personality/symptom predictors are expected to be either nonsignificantly or negatively predictive o f social and enhancement motives. 2. Based on previous research (Blackwell et al., 2002; Cooper et al., 1995; Pierce et al., 1994) depressive symptomatology and hopelessness are either positively related or unrelated to frequency o f alcohol use and positively related to alcohol dependences and/or abuse. Again, based on previous research (Stewart et al., 2001), anxiety sensitivity is positively related to alcohol frequency, alcohol dependence and/or alcohol abuse. Anxiety motives, depression motives, conformity motives, and enhancement motives are all expected to be positively related to alcohol abuse, dependence and alcohol frequency, whereas social motives are only associated with alcohol frequency and not alcohol problems. 3. The modified model from Study 1 is expected to be able to account for the relations among personality/symptom variables, drinking motives, and alcohol problems. However, the paths from depressive symptomatology and hopelessness to alcohol problems are expected to be mediated by depression motives, whereas the path from anxiety sensitivity to alcohol problems is expected to continue to be mediated by conformity motives. Previous research (Comeau et al., 2002) also indicates that there are indirect paths from anxiety sensitivity and trait anxiety to alcohol problems through anxiety motives. There is also expected to be significant covariance between depression and anxiety motives. 4. The B A S scale and extraversion are expected to be negatively related to depression, hopelessness, and either unrelated or negatively related to anxiety sensitivity and trait anxiety. The B I S scale and neuroticism are expected to be positively related to depression, hopelessness, anxiety sensitivity, and trait anxiety. CARROT  Hypotheses.  5. Individuals scoring high on hopelessness and depressive symptomatology in comparison to their low scoring counterparts are expected to evince less reward-responsiveness ( B A S ) and  67 more punishment-susceptibility (BIS) on the C A R R O T . Additionally, individuals scoring high on anxiety sensitivity and trait anxiety in comparison to their low scoring counterparts are expected to evince more punishment-susceptibility (BIS) on the C A R R O T . That is, individuals scoring high on personality/symptom variables are expected to show pre-existing differences in B A S and BIS activity. 6. In terms o f the effect o f alcohol cues on reward-responsiveness, individuals scoring high on hopelessness and depressive symptomatology are expected to evince restoration (or increased) reward-responsiveness after taking a sip o f alcohol. However, this effect is expected to be qualified by drinking experience. Regular drinkers scoring high on hopelessness or depressive symptomatology are expected to show greater restoration of reward-responsiveness from alcohol cues than light drinkers, whereas both categories of drinkers low on hopelessness or depressive symptomatology are expected to show normal reward-responsiveness and therefore little restorative effect upon receiving alcohol cues. 7. In terms o f the effect o f alcohol cues on punishment-susceptibility, individuals scoring high on hopelessness, depressive symptomatology, anxiety sensitivity, and trait anxiety are expected to evince diminished (or decreased) punishment-susceptibility after taking a sip o f alcohol. However, this effect is expected to be qualified by drinking experience. Regular drinkers scoring high on personality/symptom variables are expected to show greater diminishment o f punishment-susceptibility from alcohol cues than light drinkers, whereas both categories o f drinkers low on personality/symptom variables are expected to show normal punishment-susceptibility and therefore little diminished effect upon receiving alcohol cues.  68 3.2 Method 3.2.1 Participants Participants (n = 207) consisted o f undergraduates recruited from the University o f British Columbia psychology subject pool, with 34 males (16%) and 172 females (1 individual did not indicate gender). Their ages ranged from 17 years to 41 years with a mean age o f 19.70 years (SD = 2.64). The University o f British Columbia's psychology subject pool is representative o f the multicultural undergraduate population, with an over-representation o f individuals o f self-identified Asian decent in comparison to the general North American population: White (33%), Asian (60%), East Indian (4%), Black (1%), and Other (2%). Although 42.5% o f self-identified Asians indicated their birth place to be a non-Western country, their average number o f years in Canada was 9 years, ranging from 1 year to 19 years (Mode =11 years), with 7% o f these individuals claiming English as their first-language. In some ways, the diversity o f the sample provides an advantage i n terms o f generalizability of the current research. M a n y o f the measures used i n the present study have been validated crossculturally (e.g., B I S / B A S scales), though other measures have not (e.g., C A R R O T ) . A subset o f participants (n = 103) from Study 2 participated i n the C A R R O T task (15% males). A M A N O V A indicated there were no significant differences on any o f the variables o f interest (i.e., age, B I S / B A S Scales, S U R P S , D M Q - R , BDI-II, E P Q - R , S T A I , alcohol abuse, alcohol dependence, alcohol frequency) between those who did and did not participate in the C A R R O T portion o f the study, F(23, 157) = \.43,p > .05. Participants were split into two drinking groups based on their self-reported frequency o f drinking occasions. Light drinkers (n = 44) range between those indicating having "tried" alcohol to those reporting drinking alcohol "more than 5 times in the past year." Regular drinkers (n = 59) range between those indicating drinking more than 20 times in the past year to those drinking weekly. Participants were excluded from the C A R R O T portion o f the study i f they were under the age o f 19 years (i.e.,  69 legal drinking age in British Columbia), abstaining from alcohol, currently taking medication, pregnant or suspected pregnancy. Hopelessness and depression can occur at clinical as well as a subsyndromal levels and therefore hypotheses concerning hopelessness and depression can be tested in nonclinical samples (Abramson et al., 1989; A l l o y , Abramson, Whitehouse et al., 1999), and are perhaps best examined among individuals evincing a wide range o f symptoms (Abramson, A l l o y , & Hogan, 1997; A l l o y & Clements, 1998). 3.2.2 Measures Hopelessness and Anxiety Sensitivity.  Hopelessness and anxiety sensitivity were assessed  using the Substance Use Risk Profile Scale ( S U R P S ; Woicik et al., 2002). Agreement with each of the items is rated on a 4-point Likert scale ranging from 1 ("strongly disagree") to 4 ("strongly agree"). Cronbach alphas in the current sample for hopelessness ( a = .86) and anxiety sensitivity (a = .78) were sufficient. See Study 1 for further details on the S U R P S . The S U R P S hopelessness scale is positively related to neuroticism, trait anxiety, low self-esteem, introversion and the Beck Hopelessness Scale (Beck, Weissman, Lester, & Trexler, 1974). The S U R P S anxiety sensitivity scale is positively related to neuroticism, trait anxiety and the Anxiety Sensitivity Index (Peterson & Reiss, 1992). See Study 1 for further details o f the S U R P S . Depression.  The Beck Depression Inventory-II (BDI-II; Beck et al., 1996) is a 21-item  scale assessing cognitive, behavioural, and somatic symptoms o f depression experienced in the past two weeks. Items are rated on a scale ranging from 0 to 3, with higher scores indicating greater severity o f depression. The BDI-II has good internal consistency ( a = .73 to .92) and is a valid measure o f depression in a university population (Dozois, Dobson, & Ahnberg, 1998; Oliver & Burkham, 1979). Trait Anxiety. The trait scale o f the State-Trait Anxiety Inventory (STAI-T; Spielberger et al., 1983), a 20-item self-report questionnaire, was used to assess characterological or trait  70 anxiety (e.g., "I feel nervous and restless"). Participants rate the frequency of their experience on a 4-point scale (1 = almost never to 4 = almost always). The STAI-T has demonstrated good test-retest reliability (r = .73-.86) and correlates with other measures of anxiety (r = .73-.85). The trait scale evinced good internal consistency in the present sample (a = .91). Behavioural Activation and Inhibition Systems. Carver and White's (1994) BIS/BAS scales consist of 7 BIS items and 13 BAS items. Participants were asked to endorse items on a 5-point Likert scale, ranging from 1 (strongly agree) to 5 (strongly disagree; see Leone et al., 2001) . Moderate Cronbach coefficient alphas have been reported for each of the scales: BIS (a = .74-.78; current sample a = .75), Reward-responsiveness (a = .65-.73; current sample a = .59), Drive (a = .76-.83; current sample a = .74), and Fun-seeking (a = .66-.70; current sample a = .71; Carver & White, 1994; Heubeck et al., 1994; Jorm et al., 1999). The total BAS scale appeared to have adequate internal consistency in the current sample (a = .78). As well, the scales appeared to show satisfactory test-retest reliability among normal (r = .59 to .92) and depressed individuals (r = .62 to .75; Carver & White, 1994; Kasch et al., 2002). Extraversion  and Neuroticism.  Participants completed the Eysenck Personality  Questionnaire-Revised (EPQ-R; Eysenck et al., 1985), which yields Extraversion (E), Neuroticism (N), Psychoticism (P), and Lie (L) scales. Reliability coefficient alphas appear to be high: Ea = .86-.90, A^a-.85-.88, P .78-.81, L a = .82 (Eysenck et al., 1985; Heubeck et al., 1994) and the factor structure has been shown to be similar in 34 different countries (Barrett, Petrides, Eysenck, & Eysenck, 1998). Alcohol Frequency.  A six-category system (Woicik, Conrod, Stewart, Pihl, & Dongier,  2002) was used to assess alcohol frequency. Participants were instructed to endorse each category ("ever tried", "used in the past year", "used more than 5 times in the past year", "use more than 20 times in the past year", "used on a weekly basis") that is true of them and to leave  71 blank those not true o f them. Self-report has been shown to be a valid measure o f alcohol use when confidentiality is ensured (Sobell & Sobell, 1986, 1990). Substance Abuse and Dependence.  Alcohol abuse and dependence were assessed using a  self-report measure o f D S M - I V ( A P A , 1994) abuse and dependence adapted from the SCED-IV (First, Spitzer, Gibbon, & Williams, 1998). The number o f symptoms for abuse (four criteria) and dependence (seven criteria), respectively, were added to produce abuse and dependence indices. Abuse and dependence indices showed adequate internal consistency ( a = .76 and .63, respectively). Drinking Motives.  Blackwell's (2003) 28-item Modified Drinking Motives  Questionnaire ( D M Q - M ) , consisting o f Cooper's (1994) conformity motives, social motives, and enhancement motives subscales, and the new depression motives and anxiety motives subscales, was used to assess motivation for alcohol use. In order to assess the extent to which individuals are motivated to use alcohol to alleviate hopeless/depressive symptomatology, six additional depression specific items were added (i.e., "To numb my pain", "To stop me from feeling so hopeless about m y future", "To turn off negative thoughts about m y s e l f , "To stop me from dwelling on things", "To help me feel more positive about things i n m y life", "To forget painful memories"). In addition, three o f the coping motives items were moved to the depression motives scale (i.e., "To cheer me up when I'm in a bad mood", "To relax", "To forget my worries") and one item was split into "Because it helps me when I am feeling nervous" and "Because it helps me when I am feeling depressed." Blackwell (2003) found a five-factor model fit well. Each o f the drinking motives scales evinced good internal consistency in the present sample: Anxiety ( a = .87), Depression ( a = .92), Conformity ( a = .93), Social ( a = .92), Enhancement ( a = .86). Reward-responsiveness  and Punishment-Susceptibility.  Powell's (Powell, A l - A d a w i ,  Morgan, & Greenwood, 1996) Card Arranging Reward Responsivity Objective Test ( C A R R O T )  72 was used to measure reward-responsiveness. O n each card is a series o f five numbers (e.g., 52849, 87641, 96357). Participants are asked to sort the cards according to a certain number that appears within the series o f five numbers (i.e., 1, 2, or 3). Trial 1 ( T l ) consists o f participants sorting 60 cards as fast as they can in order to obtain an individualized time limit, which is then used for all subsequent trials and allows for possible differences i n psychomotor ability that could otherwise confound inter-individual comparison o f sorting speed. Trial 2 (T2) is a baseline trial, in which participants sort a stack o f 100 cards as quickly as possible. During trial 3 (T3) participants are rewarded (i.e., 25 cents) for every five cards they sort within the time limit. The average increase o f the reward trial among a group o f normal participants is approximately 4% (Powell et al., 1996). Following the procedure used by A l - A d a w i and Powell (1997), reward-responsiveness was derived by comparing T3 with T 2 . A second baseline trial (T4) follows T 3 . Trial 5 (T5) was added to assess vulnerability to punishment (i.e., negative reinforcement) by having participants sort the cards as fast as possible, but taking 25 cents away every time a card with red numbers appears. Punishment susceptibility is derived by comparing T5 to T4. Piloting showed that most individuals show a general slowing on T5 in comparison to T4. Alcohol Cue. Participants chose their preferred alcoholic beverage (i.e., beer, red & white wine, rum/mix, rye/mix, vodka/mix). Available mixes were gingerale, 7-Up, and coke. M i x e d drinks were mixed to a 1:4 ratio and presented in a tumbler. Beer was served in a beer glass and wine in a wine glass.  3.2.3 Procedure Participants were given a questionnaire package containing the self-report measures listed above to be completed at home and then returned it to the same location. The questionnaire took approximately 60-70 minutes to complete. A l l participants were required to sign a consent form prior to completing the questionnaire package or taking part in the card sorting task. Participants  73 provided a contact number and/or e-mail address and were contacted to set up two mutually convenient times for them to perform the neutral and alcohol conditions o f the C A R R O T task. Upon arriving for the cracker conditions, participants were asked to take a bite o f a whole-wheat cracker and focus on it for 3 minutes, in order to provide a neutral stimulus control prior to beginning the C A R R O T . After the 3 minutes, participants were administered the first three trials o f the C A R R O T . Following the reward trial o f the C A R R O T , participants were again asked to take a bite o f a whole-wheat cracker and focus on it for 3 minutes, and then administered the final two C A R R O T trials. In light o f Conrod et al.'s (2002) finding that punishment presented prior to reward interfered with reward-responsiveness, but reward presented prior to punishment did not seem to differentially affect reward or punishment trials, all participants received the reward version o f the C A R R O T prior to punishment version to avoid possible interference o f the punishment trial with the reward trial. Although interference between punishment and reward is an interesting finding in itself, the reward version o f the C A R R O T has been established as a valid measure o f reward-responsiveness, whereas the punishment version has yet to be established as a valid measure o f punishment-susceptibility. Upon arriving for the alcohol condition, participants were asked to select their preferred alcoholic beverage (i.e., beer, wine, rum/mix, rye/mix, vodka/mix). Participants were asked to take a sip o f the alcoholic beverage and focus on it for 3 minutes. Next, participants were administered the first three trials o f the C A R R O T , followed by the alcohol cue a second time and the final two C A R R O T trials. Cracker and alcohol conditions o f the C A R R O T were counterbalanced. Once all five trials o f the C A R R O T in both conditions had been completed, participants were debriefed as to the purposes o f the research. In order to abide by subject-pool policy stating that participants are not able to receive both course credit and monetary compensation, the punishment trial was also used to deplete monetary rewards received. If any money was left over participants were informed o f the policy  74 and given the choice o f the monies or course credit. N o participant chose the money (e.g., $1.25) over course credit. 3.2.4 Statistical  Analyses  Relations among self-report scales were examined using Pearson's product moments. Further analyses were conducted using hierarchical regression. Specifically, depressive symptomatology, hopelessness, anxiety sensitivity, and trait anxiety were regressed on drinking motives in a similar hierarchical regression as that o f Study 1, entering age and gender first, followed by alcohol frequency, and then depressive symptomatology, hopelessness, anxiety sensitivity and trait anxiety. A s well personality/symptom variables were regressed on alcohol dependence, entering alcohol frequency first, followed by hopelessness and anxiety sensitivity, and then trait anxiety and depressive symptomatology, and finally negative reinforcement drinking motives. A sample o f n = 207 would provide sufficient power (i.e., > .80) to find statistical significance with a small effect size for both zero order correlations (r = .18) and multiple regression (P = .08) using eight predictors. Baron and Kenny's (1986) procedure, as outline in Section 2.2.4, was used to demonstrate mediation and the Goodman (I) test (Baron & Kenny, 1986; M a c K i n n o n , Warsi, & Dwyer, 1995) was used to test the extent to which the presence o f a mediator decreased the relations between the predictor and the criterion. A z-statistic higher than 1.65 indicates significance at an alpha o f p < .05 on the Goodman (I) test. A M O S 4.0 (Arbuckle & Wothke, 1999) was again used to carry-out structural equation modelling in order to examine the causal relations among personality/symptom variables, drinking motives and alcohol problems. Personality variables (i.e., hopelessness, anxiety sensitivity and trait anxiety) were split at the sample mean to produce high and low groups. The BDI-II was split at the cut-off for mild depression (i.e., 9) to produce high and low BDI-II groups. Some follow-up analyses involved splitting personality/symptom variables at the mean plus 1/2 standard deviation and the mean  75 minus 1/2 standard deviation to create low, moderate and high groups. Three-way Alcohol Groups (light drinkers, regular drinkers) by Personality/symptom variable (low, high) by Condition (cracker, alcohol) mixed model A N O V A s were performed individually on rewardresponsiveness and punishment susceptibility for each of the four personality/symptom variables. A sample size of n = 103 provides sufficient power (i.e., .80) to find statistical significance for the three-way repeated measures A N O V A s (dfl, 89) given a small to moderate effect size. As with Study 1, although previous research has shown there to be gender differences with respect to the variable being investigated (Kessler et al., 1993; Weissman et al., 1993; Deshmukh et al., 2003), the small sample size does not allow such hypotheses to be adequately examined with respect to the structural model.  76  3.3 Results 3.3.1 Age and Gender Table 11 provides descriptive statistics for each o f the main variables i n the current study. None o f these variables o f interest were significantly related to age. A M A N O V A performed to examine differences due to gender was significant, F(20, 159) = 2.42,/? < .01. Subsequent follow-up A N O V A s revealed that males scored higher than females on conformity drinking motives (7.03 vs. 5.86), F ( l , 178) = 9.46,/? < .01, whereas females scored higher than males on the BIS scale (24.77 vs. 27.15), F ( l , 178) = 8.57,p< .01, and slightly higher on the B A S reward-responsiveness subscale (20.66 vs. 21.56), F(l, 178) = 5.13, p < .05. 3.3.2 Drinking  Motives  Correlations between drinking motives and depressive symptomatology, hopelessness, anxiety sensitivity and trait anxiety are provided in Table 12. Both depressive symptomatology and hopelessness were positively correlated with anxiety, depression, and conformity drinking motives. Only depressive symptomatology was significantly related to social drinking motives. Interestingly, anxiety sensitivity was also positively correlated with depression drinking motives, but unrelated to either anxiety or conformity motives. Trait anxiety was significantly positively correlated with anxiety, depression and conformity motives. Not surprisingly there appeared to be significant overlap among the personality/symptom variables. The five drinking motives were also significantly inter-correlated.  77 Table 11 Descriptive Statistics for Study 2 Variables N  M  SD  Depression  207  9.72  8.38  Hopelessness  207  15.29  4.30  Anxiety sensitivity  207  16.53  4.11  Trait anxiety  207  42.14  9.77  Anxiety motives  193  4.39  2.03  Depression motives  193  12.67  4.51  Conformity motives  193  6.05  1.97  Social motives  193  11.39  3.92  Enhancement motives  193  9.82  4.43  BIS Scale  207  26.75  4.21  B A S Scale  207  48.68  5.66  Neuroticism  207  13.52  5.23  Extraversion  207  13.07  3.83  Age  207  19.70  2.64  Note, ns for drinking motives are smaller due to 14 individuals not filling out the D M Q . Depression = Beck Depression Inventory-II; Hopelessness and Anxiety Sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = StateTrait Anxiety Inventory-Trait Subscale; Anxiety motives, Depression motives, Conformity motives, Social motives, and Enhancement motives = Drinking Motives Questionnaire-Revised scales; B A S = Behavioural Activation System; BIS = Behavioural Inhibition System.  78 Table 12 Correlations among Personality/Symptom Variables and Drinking Motives Variables 1. Depression 2. Hopelessness  .91  2.  3.  4.  5.  6.  7.  .60"  .38"  .72"  .18*  .39  .18  .86  .38"  .78"  .14*  .27**  .78  .47  .07  3. Anxiety sensitivity 4. Trait anxiety 5. Anxiety motives 6. Depression motives 7. Conformity motives 8. Social motives  .91  **  .19 .88  *  8.  9.  .20  .12+  .20  .02  -.02  .21"  .11  .07  .03  .23"  .17*  .09  .03  .74**  .16*  .60"  .69**  .92  .14*  .48**  .56"  .83  .35"  .18*  .86  .78**  ?  9. Enhancement motives  .92  Note. Depression = Beck Depression Inventory-II; Hopelessness and Anxiety Sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = State-Trait Anxiety Inventory-Trait Subscale; Anxiety motives, Depression motives, Conformity motives, Social motives, and Enhancement motives = Drinking Motives Questionnaire-Revised scales. Values on diagonal are coefficient alphas for each scale. f p < .10, *p < .05, **p < .01 two-tailed  Hierarchical multiple regression was used to determine the extent to which depressive symptomatology, hopelessness, anxiety sensitivity, and trait anxiety are able to account for unique variance in drinking motives. Given the shared variance among drinking motives, residual scores for each participant were created by predicting each o f the drinking motives from the remaining motives using multiple regression. For example, anxiety, depression, social, and  79 conformity motives were regressed on enhancement motives and the residuals were saved to the data file. These residuals then provided a measure o f the unique variance for each drinking motive without the influence o f the other drinking motives, providing a somewhat conservative test o f the unique relations between personality/symptom predictors and specific motivations to drink alcohol. A series o f individual hierarchical regression analyses were then conducted i n which age and gender were entered in Step 1, followed by alcohol frequency in Step 2 and finally depressive symptomatology, hopelessness, anxiety sensitivity, and trait anxiety in Step 3 (see Table 13). Demographic variables, alcohol frequency and personality/symptom variables were able td account for 7.7% o f the variance in anxiety motives, 14.4% o f the variance i n depression motives, 7.5% o f the variance in conformity motives, 7.3% o f the variance in social motives, and 6.3% o f the variance i n enhancement motives. Gender was a significant predictor o f conformity drinking motives, indicating that males were more likely to endorse drinking for conformity reasons, F(2, 180) = 3.15, p < .05. Alcohol frequency was a significant predictor o f both social, F ( l , 175) = 5.58,/? < .05, and enhancement motives,  F ( l , 175) = 7.22,/? < .01.  The block o f personality/symptom variables was a significant predictor o f both anxiety motives,  F(4, 175) = 2.73,p < .05, and depression motives,  F ( l , 175) = 6.54,/? < .01. A s  expected, trait anxiety was a significant predictor o f anxiety motives and likely as a result o f removing depression motives from anxiety motives, low BDI-II scores were predictive o f high anxiety motives. Somewhat counter to expectations, anxiety sensitivity was not a significant predictor o f anxiety motives or conformity motives. A s expected, depression motives were significantly predicted by depressive symptomatology and again due to the removal o f variance accounted for by anxiety motives, negatively predicted by Trait Anxiety. A l c o h o l frequency and, surprisingly, depressive symptomatology were significant predictors o f social motives. Finally, alcohol frequency was the only significant predictor o f enhancement motives.  80 CN O  ca c CD  s CD o  PQ  W oo  ca  .2 o o  o  Tf  o  co O  o CN  o  Tf  o  cn  in o  os  o  PQ  CO  Tf  o  CO  o  W oo  co  o  Os  CO  PQ  o  ^  r-H  oa  CN O  00  CN  co O  os  CO  o  o CN  o CN  o  VO  o  00  CO  CD >  60  •a Q  PQ  cS  W oo  U  PQ  c o  _o 'oo Vi  CD  i-  OH  o  in oo o  o  o  CN  o  oo  ca  CN  r-o  m  o  CN  m  o  2  ©  CN O  co  o  CO  o  o  PQ  W oo  o  o  CN  CN  CD  bo  Q  K a,  cu . <a  Tf  o  oa  e o to ?>  Tf  PQ  o  Tf  o CD  PQ  W oo  co  o  PQ  o  CO  o  CN  o  Tf  o  CN  CN O  m O  Os  o  ro  CN CO  »—»  CD  co  I  CD  CO  ca  oo o cr  Q  .CJ  CD  o  CD  >  ID 00  a  CD  -*—»  00  CD  •a a  CD  o  CD 60  CN OH CD  H—*  00  <  T3 0  CD  a  o o CO  OH CD  H—» 00  CN O ro o  ro o  vo o  o CN  o CN  Tf  o  ro o  o  Tf  VO  o  ro O  o  CN O  o o  VO  m CN  CN o  GO  Tf  G  00  Tf  ON  ro O  VO  OH  o  OO  CN CN  ro O  Tf  o  o o  o o  o  V  >  Tf  CN O  00  t-o  CN O  CN  Tf  o  vo o  OH  CD H—»  <a CD  o  CD O  CD  Q ro  IH  •n  CD  OH  OO  \<B  Q  o  IH  00  IH  CD  CD  CD •*->  ro O  00  oo  PH  ^H  o  o o  in  o  CD  CD  ro O  CD  .Q  CQ  o  >  !H OH  A  Tf  CD  CD  B  <D  00 o  IH  <2  00  00 00  M o  OH  O  A  00  CD  a _o  o  O  l-H  ro  HJ  G O  oo  OH  in © lo CD _> '-*->  81  IH  OH CD H—»  00 IH  a  00  a  CD  co  .£>  O  o CN  ro  'x  .2  V  A  ^  00  ro O  CN O  o  g oo 00 <D  c  CN  CO  o  O  CN  TT  vo o  o o  O  OH CD  00  J D  O  CN o  -*-»  in  00  o  IH  OH CD  00  "CD  OH CD  00  IH  in  o  o ro  00 ro  A  ro O  in  o  ON  ro O  VO  o Tf  CN  in  o  in  o  Tf  o  CN o ro O  ro  CN ro  GO <D > • I—I +-»  IH  ,o  CN OH  o o 00  B  ro  a  O0  00  OH  CD  in ©  ro o  ©  CN  r-o  ro O  CN  CN O  Lo Tf  o  00 CN  ON O  Tf  o  o o  CN O  H—»  CD  CD T 3  00  <  CD  o  .SH  <  o o  '-*-»  cx o  A  3-  H  ca  13 o O0 o IH  PH co  c o 'oo co <D i— OH <D  CO CD  > H-»  co CD  OH CD  oo IH  ,o o o  od CD >  03  o  CO  oo  oo  <  u  -O  =3 O0 03 IH  H  h o  o  e  S  CD >  •a  CD  '>  co co  O  OH  IH  CD H—>  ,o  OH  00  C  CO  a) 00  CD  o  CD O  CN  u  •4—»  CD  o3  CD  £ 00  co  co co  e  >^ PH  00 <  I  _CD  Hi—*  ro O  s  CD  Q CQ  o  in ©  H-» CD  *o o o  _>  A  o  cr s-  ro O  CD  o  Tf  a 00  o ro  V  OH  c o O  o  O  H—»  o  ro  ro  CD  o  o  CN  a  2  00  o  V CN O  a,  CD H—»  <2  in  CN O  CN  kH  o ro  §  in  in ©  00  ro  m  -a  o s CD >  ni l-H  H  CD •4—»  00  co  82  p V  * #  in p V # V +—  83  3.3.3 Alcohol Abuse, Dependence, and Frequency There was a great deal o f variability in self-reported frequency o f alcohol use (see Table 14), with only a small proportion o f people indicating never having tried alcohol and a large proportion having used alcohol within the last year (87%). The majority o f participants indicated experiencing little or no problems associated with alcohol use (see Table 15). A total o f 69 (33%) individuals met D S M - I V criteria ( A P A , 1994) for alcohol abuse (endorsed one or more o f the four criteria) and 23 (11%) individuals met criteria for alcohol dependence (endorsed three or more o f the seven criteria). Between-subjects M A N O V A s were individually performed for each of abuse and dependence diagnoses using the depressive symptomatology, hopelessness, anxiety sensitivity and trait anxiety as the dependent variables. Neither abuse nor dependence M A N O V A s were significant, indicating that there was no difference in terms o f any o f the dependent variables between those with and without a D S M - I V alcohol abuse or dependence diagnosis, F(4, 190) = 1.25, p > .05 andF(4, 190) = .66, p > .05, respectively. The following analyses, therefore, use alcohol abuse and alcohol dependence criteria as an interval scale. The mean number o f abuse criteria was .78 (SD = 1.01) and dependence criteria was 1.01 (SD = 1.28). Zero-order correlations among personality/symptom variables, drinking motives and alcohol abuse and dependence and alcohol frequency are shown in Table 16. Interestingly, among the personality variables, only depressive symptomatology was significantly positively correlated with alcohol dependence. None o f the other personality variables were related to alcohol abuse or alcohol frequency. Anxiety, depression, and enhancement motives were all significantly positively correlated with both alcohol abuse and dependence. Counter to previous research, social motives, but not conformity motives, were also positively correlated with alcohol abuse and dependence, but did not remain so once enhancement motives were taken into  84 account, (3 = .06,p > .05 and p = .1 l,p > .05, respectively. A s well, anxiety, depression, social, and enhancement motives were all significantly correlated with frequency o f alcohol use. Although social motives remained a significant predictor o f alcohol frequency, p = .36, p < .01, the relationship was significantly reduced after taking into account enhancement motives, z = 5.25, p < .01, indicating that this relationship is at least partially mediated by drinking for enhancement reasons.  Table 14 Self-reported Alcohol Use Frequency Frequency Alcohol Use Never Ever  17 9  Within the last year  41  More than 5 times i n the last year  52  More than once per month  55  Weekly  33  Total  207  Table 15 Number of Alcohol Abuse and Dependence Criteria  Endorsed Frequency  Number o f Criteria Endorsed Alcohol Abuse 0  126  1  19  2  23  3  20  4  .  7  Alcohol Dependence 0  92  1  49  2  31  3  12  4  7  5  3  6  0  7  1  86 Table 16 Correlations for Personality/Symptom  Variables with Alcohol Abuse, Dependence and  Frequency Variable Depression  Alcohol abuse .07  Alcohol dependence .18*  Alcohol frequency .02  Hopelessness  -.12  .14*  -.04  Anxiety sensitivity  -.09  .03  -.08  Trait anxiety  -.05  .13  +  .01  Anxiety motives  .42"  .54"  .58"  Depression motives  .39*  .54*  .47"  .09  .13  Conformity motives  -.03  +  Social motives  .48"  .47**  .69**  Enhancement motives  .58"  .55"  .70"  Note. Depression = Beck Depression Inventory-II; Hopelessness and Anxiety sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = State-Trait Anxiety Inventory-Trait Subscale; Anxiety motives, Depression motives, Conformity motives, Social motives, and Enhancement motives = Drinking Motives Questionnaire-Revised scales. t/><.10, *p<.Q5, **/?<.01  Similar to Study 1, hierarchical multiple regression was used to predict alcohol dependence based on alcohol frequency, personality/symptom variables, and drinking motives (see Table 17). A l c o h o l frequency was entered in Step 1, in order to examine at subsequent steps the extent to which personality/symptom variables were able to account for variance in  87 alcohol dependence over and above alcohol use. Hopelessness and anxiety sensitivity were entered in Step 2, followed by trait anxiety and depressive symptomatology i n Step 3, in order to demonstrate the ability o f depressive symptomatology to explain variance in alcohol dependence in the presence o f other predictors and mediate between other personality/symptom variables and alcohol dependence. Since conformity motives were not significantly related to alcohol dependence, negative reinforcement drinking motives (i.e., depression motives, anxiety motives, and conformity motives) were entered as a block i n Step 4 to examine whether they mediated between personality/symptom variables and alcohol dependence. Drinking motives were allowed to overlap instead o f using residuals. In the final model, personality/symptom variables and drinking motives were able to account for 39% o f the variance in alcohol dependence, F(8, 172) = 14.04,/? < .01. Alcohol frequency was a significant predictor o f alcohol dependence i n Step 1, F(l, 179) = 58.22,/? < .01. Personality/symptom variables were able to contribute to the prediction o f alcohol problems in Step 2, but not i n Step 3, AF(3, 176) = 3.29,/? < .10 and A F ( 1 , 175) = 1.35,/? > .05. Hopelessness was a significant predictor o f alcohol dependence in Step 2. Depressive symptomatology was not a significant predictor o f alcohol dependence i n Step 3, but appeared to partially mediate between hopelessness and alcohol dependence, z = 1.89. Trait anxiety was not a significant predictor o f alcohol dependence. Depression motives and anxiety motives were a significant predictor o f alcohol dependence in Step 4, A F ( 3 , 172) = 10.57,/? < .01, and led to a reduction in betas between personality/symptom variables and alcohol dependence i n the final model, indicating that the relations between personality/symptom variables and alcohol dependence are mediated by drinking motives. Conformity motives were not a significant predictor o f alcohol dependence.  88 Table 17 Hierarchical  Multiple Regression Predicting Alcohol Dependence from Alcohol  Personality/Symptom  Variables and Drinking  Frequency,  Motives A l c o h o l Problems  Variables  B  SE B  P  .58  .08  .49"  Alcohol frequency  .59  .08  .50"  Hopelessness  .05  .02  .17*  Anxiety sensitivity  .00  .02  -.01  Alcohol frequency  .59  .08  .50  Hopelessness  .05  .03  .16  Anxiety sensitivity  -.01  .02  -.02  Trait anxiety  -.01  .02  -.10  Depression  .03  .02  .14  Alcohol frequency  .32  .09  .27*  Hopelessness  .04  .03  .14  Anxiety sensitivity  -.01  .02  -.04  Trait anxiety  -.01  .02  -.10  Depression  .02  .01  .10  Depression motives  .07  .03  .24**  Step 1 Alcohol frequency Step 2  Step 3  +  +  Step 4  89 Anxiety motives  .13  .06  .20**  Conformity motives  -.04  .04  -.06  Note. R ='.24 for Step 1; A i ? = .03 (p < .10) for Step 2; Ai? = .01 (p < .10) for Step 3; Ai? = .11 2  2  2  2  (p < .01) for Step 4. Depression = Beck Depression Inventory-II; Hopelessness and Anxiety Sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = State-Trait Anxiety Inventory-Trait Subscale; Anxiety motives, Depression motives, Conformity motives, Social motives, and Enhancement motives = Drinking Motives Questionnaire-Revised scales, t /? < .10, * p < . 0 5 , **/?<.01  Structural equation modelling was used to demonstrate the causal relations among each of the predictors (i.e., depressive symptomatology, hopelessness, anxiety sensitivity, trait anxiety) and alcohol dependence, taking into account the relationship between trait anxiety and anxiety motives, as well as the distinction and inter-relationship between depression motives and anxiety motives (see Figure 5). A l s o , given that alcohol frequency was found to be significantly related to alcohol dependence after taking personality/symptom variables into account, it was included as a predictor in the current path diagram. However, previous research (Cooper, Frone, Russell, & Mudar, 1995; Stewart, Zvolensky, & Eifert, 2001) indicates the direction o f the relationship to be from anxiety motives to alcohol frequency. Depression and anxiety drinking motives, as opposed to residuals, were used and allowed to covary in the model in order to allow the personality/symptom variables to account for as much variance i n drinking motives and in turn alcohol problems as possible. Again, scale items were divided into two smaller scales to provide at least two indicators for each latent variable.  90 Figure 5. Hypothesized Model: Influence of Personality/Symptom Variables on Depression Motives, Anxiety Motives, and Alcohol Dependence  Note. Depression = Beck Depression Inventory-II; Hopelessness and Anxiety Sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = State-Trait Anxiety Inventory-Trait Subscale; Anxiety motives, Depression motives, Conformity motives = Drinking Motives Questionnaire-Modified. j p< .\0,p<  .05,  **p<.0\  91 The model provided an excellent fit i n comparison to the independence model (see Table 18), as indicated by a non-significant chi-square Value and a % / d f ratio o f less than 5. A l s o , the model evinced very high C F I , G F I , and T L I and low R M S E A and R M R values. Moreover, the model appeared to account for a significant portion o f variability i n alcohol frequency (35%) and alcohol dependence (57%) and a small portion o f the variance i n anxiety motives (5%) depression motives (7%), and conformity motives (7%). The direction and size o f the paths are similar to those found in Study 1. Exceptions to this are the lack o f direct path from anxiety sensitivity to alcohol dependence and the non-significant path from conformity motives to anxiety motives.  Table 18 Summary o f Goodness-of-Fit Indices Fit indices Model  X "  Saturated  O00  1  df  ~yfldf 0  RMR  RMSEA  .00  GFI  TLI  CFI  1.00  1.00  Independence  438.12**  136  3.22  4.13  .11  .72  .00  .00  Hypothesized  105.77  104  1.02  .38  .01  .93  .99  .99  * p < .05, **p < .01  3.3.4 Relation of BDI-II Items to Alcohol Dependence and Depression Drinking  Motives  Correlations between BDI-II items and alcohol dependence and depression drinking motives are presented in Table 19. Unlike Study 1, suicidal ideation was not related to alcohol dependence. O f the depression items, self-criticism, worthlessness, loss o f energy, and changes in sleep pattern were all significantly related to both alcohol dependence and depression drinking  92 motives. A l l BDI-II items, except loss o f pleasure, loss o f interest, and irritability, were significantly correlated with depression drinking motives. Stepwise regression also painted a different picture than that o f Study 1 (see Table 20 and Table 21). Depression items were able to account for 6.8% o f the variance in alcohol dependence after two iterations, and 16% o f the variance in depression motives after four iterations, F ( 2 , 192) = 7.02, p < .01 and F(4, 188) = 9.87, p < .01, respectively. Changes in sleep pattern was the strongest predictor o f alcohol dependence, A F ( 1 , 193) = 8.58,/? < .01, and remained a significant predictor i n the final model, along with self-criticism, A F ( 1 , 192) = 5.27,/? < .01. Guilt was the strongest predictor o f depression motives entering in Step 1, A F ( 1 , 191) = 18.04,/? < .01, followed by sadness in Step 2, AF(1, 190) = 8.36,/? < .01, and changes in sleep pattern in Step 3, A F ( 1 , 189) = 6.94,/? < .01. Interestingly, loss o f pleasure was a significant inverse predictor o f depression motives in Step 4, A F ( 1 , 188) = 4.04,/?<.01. Five items from the BDI-II are similar to depression items used in Study 1 (i.e., sadness, pessimism/hopelessness, suicidal ideation, loss o f interest, worthlessness). The only significant correlation between these depression items and alcohol dependence i n the present sample was worthlessness (see Table 19). Since all five o f the depression items from Study 1 were significantly correlated with depression motives, a stepwise regression analysis was performed in order see i f similar results to Study 1 would be found (Table 22). The final model accounted for 11 % o f the variance in depression motives after two iterations, F(2, 190) = 11.74, p < .01. Sadness was the strongest predictor depression motives, A F ( 1 , 191) = 17.26,/? < .01, followed, interestingly, by suicidal ideation in Step 2, A F ( 1 , 190) = 5.79,/? < .05.  93 Table 19 Depression Items with Alcohol Problems and Depression Depression Items  Motives  A l c o h o l dependence  Depression motives  Sadness  .11  .29"  Pessimism  .04  .22"  Past failure  -.01  .20  Loss o f pleasure  .10  .04  Guilt  .08  .29  Punishment feelings  .06  .23"  Self-dislike  .05  .26"  Self-criticism  .19"  .17"  Suicidal ideation  .07  .28"  Crying  .10  .18"  Agitation  .06  .25"  Loss o f interest  .10  .12*  Indecisiveness  .11  .18"  Worthlessness  .14"  .26"  Loss o f energy  .16"  .10  Changes in sleep pattern  .21"  .26"  Irritability  .08  .12  Changes in appetite  .12*  .23"  Concentration difficulty  .19"  .21"  Tiredness/fatigue  .05  .17*  Loss in interest in sex  .13  t  .19"  94 f p < .10, * p < .05, ** p < . 0 1  Table 20 Stepwise Regression of Depression Items regressed on Alcohol  Problems A l c o h o l Problems B  SE o f B  P  .39  .13  .21"  Changes i n sleep pattern  .34  .13  .18"  Self-criticism  .27  .12  .16*  Depression Items Step 1 Changes i n sleep pattern Step 2  Note. R = .04 i n Step 1; AR = .03 (p < .05) i n Step 2. 2  *p< .05, **p< .01  2  95 Table 21 Stepwise Regression of Depression Items regressed on Depression  Motives Depression Motives  Depression Items  B  SE o f B  P  Step 1 Guilt  2.09  .49  .29"  Guilt  1.56  .52  .22"  Sadness  1.58  .55  .21"  Guilt  1.31  .52  .18"  Sadness  1.43  .54  .19"  Changes in sleep pattern  1.16  .44  .18"  Guilt  1.47  .52  .21"  Sadness  1.83  .57  .24"  Changes in sleep pattern  1.20  .44  .19"  -1.11  .55  -.15*  Step 2  Step 3  Step 4  Loss o f pleasure  Note. R = .09 in Step 1; Ai? = .04 (p < .01) in Step 2; Ai? = .03 (p < .01) in Step 3; Ai? = .02 (p 2  2  < .05) in Step 4. *p<.05,  ** p< .01  2  2  96  Table 22 Hierarchical  Regression for Depression Motives regressed on BDI-II Items Depression Motives  Depression Items  B  SE o f B  P  Step 1 Sadness  2.17  .52  .29"  Sadness  1.51  .58  .20  Suicidal ideation  2.29  .95  .19  Step 2  *  Note. R = .08(p< .01) for Step 1; AR = .03 (p < .05) for Step 2. 2  *p<.05,  2  ** p<  .01  5.5.5 Behavioural Inhibition and Behavioural Activation System Scales Table 23 shows the correlations between the B I S / B A S scales and hopelessness, anxiety sensitivity, trait anxiety, neuroticism, and extraversion. A s predicted, hopelessness was positively related to the B I S scale and negatively related to the B A S scale. However, depression was positively correlated with the BIS scale and evinced no significant relationship with the B A S scale. In line with Gray's theory, the BIS scale was significantly positively correlated with neuroticism and negatively correlated with extraversion. A s has been found in previous research concerning B A S activity, the B A S scale was significantly positively related to extraversion, but unrelated to neuroticism. Both anxiety sensitivity and trait anxiety were positively correlated with the BIS scale. Interestingly, trait anxiety was also negatively correlated with the B A S scale.  97 Table 23 Correlations between BIS/BAS Scales and the Depressive Symptomatology, Anxiety Sensitivity, Trait Anxiety, and the Eysenck Personality  Hopelessness,  Questionnaire-Revised N  E  BIS  BAS  Depression  .24  -.06  .53"  -.06  Hopelessness  .28  -.18"  .51"  -.26  Anxiety sensitivity  .37  .14*  .51"  -.03  Trait anxiety  .37"  -.18"  .68"  -.27*  Neuroticism (N)  .53"  .03  Extraversion (E)  -.18"  .44  A  A**  -  -  -  -  *  Note. Depression = Beck Depression Inventory-II; Hopelessness and Anxiety Sensitivity = Substance Use Risk Profile Scale-Hopelessness and Anxiety sensitivity subscales; Trait Anxiety = State-Trait Anxiety Inventory-Trait Subscale; B A S = Behavioural Activation System; BIS = Behavioural Inhibition System. * p < .05, ** p < .01  Given that individuals scoring higher on the B A S scale appear to report more alcohol abuse and dependence criteria, r = .11, p < .05 and r = .23, p < .01, respectively, in order to examine the extent to which low B A S activity is also predictive o f alcohol problems, a residual was created by predicting the B A S scale from impulsivity and sensation-seeking, thus removing the effect o f these personality predictors. Although the B A S residual continued to be negatively correlated with hopelessness, r - -.22, p < .01, it was unrelated to alcohol abuse or dependence, r = .03, p > .05 and r = .09, p > .05, respectively. N o r did it correlate with alcohol frequency, r = .11, p> .05. Neither was the BIS scale related to alcohol abuse, dependence or frequency, r = .09, p > .05, r = .01,p > .05 and r = -.01,p > .05, respectively.  98 3.3.6 CARROT Reward-Responsiveness  and  Punishment-Susceptibility  There were no main or interaction effects due to order o f cracker and alcohol conditions. Six participants evinced considerable slowing on the punishment trial o f the C A R R O T (i.e., > 3 SD). These individuals also affirmed that they had slowed on purpose in order to try to hang on to their "winnings" from the reward trial. A s these participants were clearly outliers as well, their data was not used in the punishment-susceptibility analyses. Paired samples t-tests, specified a priori, were used to examine the extent to which reward (i.e., T2 vs. T3) and punishment (i.e., T4 vs. T5) resulted in the desired rewardresponsiveness and punishment-susceptibility effects. Indeed, reward resulted in a general speeding up o f card-sorting, r(103) = 3.59, p < .01, and punishment a general slowing down o f card sorting, ^(103) = 7.29, p < .01, providing evidence that the modified C A R R O T appears to measure reward-responsiveness and punishment-susceptibility within the current sample. Two 2 (Alcohol Group) x 2 (Hopelessness) x 2 (Condition) mixed model A N O V A s were individually performed for each o f reward-responsiveness and punishment-susceptibility. There were no main or interaction effects for hopelessness on reward-responsiveness (see Table 24). There was also no main effect for hopelessness or condition on punishment-susceptibility (see Table 25). There was a small nonsignificant main effect for alcohol group on punishmentsusceptibility, with light drinkers evincing more punishment-susceptibility than regular drinkers. There was also a trend for the three-way Condition by Hopelessness by alcohol group interaction on punishment-susceptibility.  Table 24 Three-way ANOVA for Alcohol Groups by Hopelessness by Condition on CARROT  Reward-  responsiveness  Source  df  F  P  Between subjects Hopelessness ( H O )  1  1.65  .02  .20  Alcohol group ( A G )  1  .27  .00  .60  HO x A G  1  .44  .01  .51  M S error  99  (20.31) Within subjects  Condition  1  1.37  .01  .24  Condition x H O  1  .21  .00  .65  Condition x A G  1  .289  .00  .59  Condition x H O x A G  1  2.09  .02  .15  Error(Condition)  99  (16.33)  Note. Values in parentheses represent mean square errors.  100 Table 25 Three-way ANOVA for Alcohol Groups by Hopelessness by Condition on CARROT susceptibility  Source  df  F  P Between subjects  Hopelessness (HO)  1  .96  .01  .331  Alcohol Groups ( A G )  1  3.13  .03  .08  HOx A G  1  .02  .00  .88  Error  91  (15.89) Within subjects  Condition  1  .00  .00  .99  Condition x H O  1  .91  .01  .34  Condition x A G  1  .38  .00  .54  Condition x H O x A G  1  2.30  .02  .13  Error(Condition)  91  (18.64)  Note. Values in parentheses represent mean square errors.  Punishment-  101 In order to better reflect high versus low hopelessness, the mean plus/minus a 1/2 standard deviation was used to split hopelessness into low, moderate, and high hopelessness groups, and the three-way mixed model A N O V A s were performed again. There were no significant main or interaction effects for hopelessness on reward-responsiveness (see Table 26). However, punishment-susceptibility evinced a significant Condition by Hopelessness interaction (see Table 27 & Figure 6). Simple main effects analyses revealed that neither individuals scoring low on hopelessness, F ( 2 , 89) = .44,/? > .05, nor those scoring moderately on hopelessness, F(2, 89) = 1.17,/? > .05, differed between cracker and alcohol conditions in terms of punishment-susceptibility. However, individuals scoring high on hopelessness were significantly slower i n the neutral condition in comparison to the alcohol condition, F(2, 55) = 5.77, /? < .01, indicating less punishment-susceptibility after taking a sip o f alcohol. A n a priori one-way A N O V A performed across low, moderate, and high hopelessness groups to examine differences in punishment-susceptibility in the neutral condition was not significant, F(2, 92) = 1.54,/? > .05, indicating that low, moderate and high hopelessness groups did not differ in terms of punishment-susceptibility in the neutral condition.  Table 26 Three-way ANOVA FOR Alcohol Groups by Hopelessness Split at 1/2 SD by Condition on CARROT  Reward-responsiveness  Source  df  F  ii  P  Between subjects Hopelessness (HO)  2  .66  .01  .52  Alcohol Groups ( A G )  1  .21  .00  .65  HOx AG  2  .70  .01  .50  M S error  97  (20.03) Within subjects  Condition  1  1.97  .02  .16  Condition x H O  2  .60  .01  .55  Condition x A G  1  .32  .00  .57  Condition x H O x A G  2  .82  .02  .44  Error(Condition)  97  (16.10)  Note. Values in parentheses represent mean square errors. Hopelessness= hopelessness split into 3 groups based on 1/2 S D above and below the mean.  103 Table 27 Three-way ANOVA for Alcohol Groups by Hopelessness split at Plus/Minus 1/2 SD by Condition on CARROT  Punishment-susceptibility  Source  df  F  r|  p  Between subjects Hopelessness (HO)  2  .19  .00  .82  Alcohol Group ( A G )  1  2.84  .03  .10  HO x A G  2  .52  .01  .60  Error  89  (16.19) Within subjects  Condition  1  .20  .00  .65  Condition x H O  2  3.85  .08  .02  Condition x A G  1  .84  .01  .36  Condition x H O x A G  2  .08  .00  .92  Error(Condition)  89  (18.04)  Note. Values in parentheses represent mean square errors.  104 Figure 6. Condition by Hopelessness Interaction for Punishment-Susceptibility  0 -1 o  -2  CD O  00  i  -3  "S CD  • Cracker Condition  -4  H Alcohol Condition  c  O-i  -5 -6 LoHO  Mod HO  Hi HO  Note. L o H O , M o d H O and H i H O = Substance Use Risk Profile Scale hopelessness scale split into low, moderate and high hopelessness groups at the mean plus/minus 1/2 SD. * indicates that the group differs between neutral and alcohol conditions at p < .05  BDI-II scores split at the cut-off for mild depression (i.e., 9; Beck et al., 1996) showed no significant main or interaction effects on reward-responsiveness (see Table 28). There was, however, a significant Condition by Depression interaction for punishment-susceptibility (see Table 29 & Figure 7). Similar to hopelessness, there were no differences between the cracker and alcohol conditions for punishment-susceptibility among individuals scoring low on depression, F(l, 91) = 2.32, p > .05. Those scoring high on depression, however, evinced a significant decrease in punishment-susceptibility between the cracker and alcohol conditions, F(l, 91) = 4.70, p < .05. A n a priori independent samples Mest, performed to examine the extent to which individuals with low depression scores evince less punishment-susceptibility than those scoring high on depression in the cracker condition, was not significant, t(95) = 1.47, p > .05. Setting the cut-off for the BDI-II at 14, half way between the cut-off for mild and  105 moderate depression, i n order to better reflect dysphoric mood (see Dozois, Dobson, & Ahnberg, 1998) while still maintaining adequate power (i.e., n = 22 in the high BDI-II group) revealed a significant difference in punishment-susceptibility between low and high BDI-II scorers in the neutral condition (see Figure 8), t(93) = 2.0$>, p< .05, indicating that individuals high in depressive symptomatology showed more punishment-susceptibility than those low in depressive symptomatology.  Table 28 Three-way ANOVA for Alcohol Groups by Depressive Symptomatology by Condition on CARROT  Reward-responsiveness  Source  df  F  r\  p  Between subjects Depressive Symptomatology (BDI)  1  .27  .00  .60  Alcohol Groups ( A G )  1  .62  .01  .43  BDI x A G  1  .82  .01  .37  M S error  99  (20.06)  Within subjects Condition  1  2.61  .03  .11  Condition x B D I  1  .76  .01  .39  Condition x A G  1  .28  .00  .60  Condition x B D I x A G  1  .06  .00  .80  Error(Condition)  99  (16.06)  Note. Values i n parentheses represent mean square errors.  106 Table 29 Three-way ANOVA for Alcohol Groups by Depressive Symptomatology CARROT  by Condition on  Punishment-susceptibility  Source  df  F  r\  p  Between subjects Depressive Symptomatology (BDI)  1  .26  .00  .61  Alcohol Group ( A G )  1  5.55  .03  .11  BDI x A G  1  .04  .00  .85  Error  91  (16.06)  Within subjects Condition  1  .63  .01  .43  Condition x B D I  1  7.99  .08  .006  Condition x A G  1  .003  .00  .96  Condition x B D I x A G  1  2.32  .02  .13  Error(Condition)  91  (17.47)  Note. Values in parentheses represent mean square errors.  107 Figure 7. Condition by BDI-II Interaction for Punishment-Susceptibility  o.  CD. O co  •3  1 a CD  • Cracker Condition 11 Alcohol Condition  CM  Lo BDI  Hi BDI  Note. L o B D I and H i H O = Beck Depression Inventory split into low and high depression groups at the mild depression cut-off score o f 9 (Beck et al., 1996). * indicates that the group differs between neutral and alcohol conditions atp < .05  108  Figure 8. Punishment-susceptibility at the Neutral Condition for BDI-II split at a Cut-off o f 14  o O  00  • Lo BDI  OJ  • Hi BDI rt rt  Note. L o B D I and H i H O = Beck Depression Inventory split into low and high depression groups at the dysphoric mood cut-off score o f 14 (Dozois et al., 1998). * indicates that the group differs between neutral and alcohol conditions at p < .05  Anxiety sensitivity and trait anxiety were split at the mean in order to examine changes in reward-responsiveness and punishment-susceptibility from alcohol cues. Anxiety sensitivity evinced no main or interaction effects on reward-responsiveness (see Table 30). Anxiety sensitivity also showed no main effect for punishment-susceptibility (see Table 31). There were also no differences between individuals scoring low and high on anxiety sensitivity on punishment-susceptibility (-3.28 vs. -4.04), t(93) = .96,p > .05.  109 Table 30 Three-way ANOVA for Alcohol Groups by Anxiety Sensitivity by Condition on CARROT responsiveness  Source  df  F  Tl  P  Between subjects Anxiety Sensitivity ( A S )  1  .01  .00  .92  Alcohol Groups ( A G )  1  .25  .00  .62  ASx AG  1  1.34  .01  .25  M S error  99  (19.99) Within subjects  Condition  1  1.88  .02  .17  Condition x A S  1  .52  .01  .47  Condition x A G  1  .40  .00  .41  Condition x A S x A G  1  2.07  .02  .15  Error(Condition)  99  (15.84)  Note. Values i n parentheses represent mean square errors.  Reward-  110 Table 31 Three-way ANOVA for Alcohol Group by Anxiety Sensitivity by Condition on CARROT Punishment-susceptibility  Source  df  F  *1  P  Between subjects Anxiety sensitivity ( A S )  1  ,11  .00  .74  Alcohol Group ( A G )  1  3.03  .03  .09  ASx AG  1  1.24  .01  .27  Error  91  (15.81) Within subjects  Condition  1  .02  .00  .89  Condition x A S  1  2.72  .03  .10  Condition x A G  1  .35  .00  .56  Condition x A S x A G  1  .45  .00  .50  Error(Condition)  91  (18.51)  Note. Values in parentheses represent mean square errors.  Trait anxiety showed no significant main effect or interaction with alcohol group or condition on reward-responsiveness (see Table 32). There was a significant two-way Trait Anxiety by Condition interaction on punishment-susceptibility (see Table 33), the source o f which was found in a small nonsignificant decrease in punishment-susceptibility among high trait anxious individuals between cracker and alcohol conditions, F ( l , 91) = 2.86,p < .10 (see  Ill  Figure 9). N o difference was found among low trait anxious individuals, F ( l , 91) = 2.17,/) > .05.  Table 32 Three-way ANOVA for Alcohol Groups by Trait Anxiety by Condition on CARROT  Reward-  responsiveness Source  df  F  P  Between subjects Trait anxiety ( T A )  1  1.68  .02  .20  Alcohol Groups ( A G )  1  .32  .00  .57  TA x A G  1  .46  .01  .50  M S error  99  (19.77) Within subjects  Condition  1  2.15  .02  .14  Condition x T A  1  .38  .00  .54  Condition x A G  1  .28  .00  .60  Condition x T A x A G  1  2.66  .03  .11  M S residual  99  (15.67)  Note. Values i n parentheses represent mean square errors.  112  Table 33 Three-way ANOVA for Alcohol Groups by Trait Anxiety by Condition for CARROT susceptibility  Source  df  F  *1  P  Between subjects Trait anxiety ( T A )  1  3.04  .03  .09  Alcohol Groups ( A G )  1  2.92  .03  .09  TAx AG  1  1.59  .02  .21  Error  91  (15.18) Within subjects  Condition  1  .00  .00  .96  Condition x T A  1  5.55  .06  .02  Condition x A G  1  .31  .00  .58  Condition x T A x A G  1  1.14  .01  .29  Error(Condition)  91  (18.01)  Note. Values in parentheses represent mean square errors.  Punishment-  113 Figure 9. Trait Anxiety by Condition Interaction for Punishment-Susceptibility  -2.31  P. CD O  GO CO i  -3.76  CD  • Cracker  PH  • Alcohol Lo T A  HiTA  Note. L o T A and H i T A = State-Trait Anxiety Inventory for Children-Trait Subscale split into low and high trait anxiety groups at the mean. t Indicates that the group differs between neutral and alcohol conditions atp < .10  Trait anxiety was split into low, moderate and high groups using the mean plus/minus a 1/2 standard deviation (STAI1/2&D). Trait anxiety showed no significant main or interaction effects on reward-responsiveness (see Table 34). A s before, there was a significant Condition by Trait Anxiety interaction for punishment-susceptibility (see Table 35 & Figure 10). N o differences were found in terms o f punishment-susceptibility for low or moderate trait anxiety scorers, F ( l , 91) = .60, p > .05 and F (2, 89) = .76, p > .05, respectively, but high trait anxiety scorers evinced a significant decrease i n punishment-susceptibility from the cracker to the alcohol condition, F (2, 89) = 4.43,p > .05. There were no significant differences between low and high trait anxiety, t(93) = .32, p > .05), or among low, moderate and high trait anxiety (F(2, 92) = .37, p > .05) in terms o f punishment-susceptibility i n the neutral condition.  Table 34 Three-way ANOVA for Alcohol Groups by Trait Anxiety Split at 1/2 SD by Condition on CARROT  Reward-responsiveness  Source  df  F Between subjects  Trait Anxiety ( T A )  2  .42  .01  .66  Alcohol Groups ( A G )  1  .26  .00  .61  TAx AG  2  .12  .00  .89  M S error  97  (21.42) Within subjects  Condition  1  1.36  .02  .25  Condition x T A  2  1.45  .03  .24  Condition x A G  1  .11  .00  .74  Condition x T A x A G  2  1.33  .03  .27  M S residual  97  (15.83)  Note. Values in parentheses represent mean square errors.  Table 35 Three-way ANOVA for Alcohol Groups by Trait Anxiety Split at 1/2 SD by Condition Punishment-susceptibility  Source  df  F  P Between subjects  Trait Anxiety ( T A )  2  1.44  .03  .24  Alcohol Groups ( A G )  1  1.44  .03  .24  TAx AG  2  2.38  .03  .13  Error  89  (15.13) Within subjects  Condition  1  .03  .00  .87  Condition x T A  2  3.12  .07  .04  Condition x A G  1  .10  .00  .75  Condition x T A x A G  2  1.31  .03  .27  Error(Condition)  89  (18.01)  Note. Values i n parentheses represent mean square errors.  CARROT  116 Figure 10. L o w , Moderate and H i g h Trait Anxiety ( T A ) groups by Condition Interaction for Punishment-Susceptibility  0 -1  -1.05  P  CD O wi  rt 00  -2\ -3.14  I  rt CD  -3  -3.81  -4.03  -4.00  -4  • Alcohol  -4.04  OH  • Cracker  -5 Lo T A  Mod T A  HiTA  Note. L o T A , M o d T A , and H i T A = State-Trait Anxiety Inventory for Children-Trait Subscale split into low, moderate, and high trait anxiety groups at the mean plus/minus 1/2 SD. * indicates that the group differs between neutral and alcohol conditions atp < .05  Given the significant Condition by Trait Anxiety interaction, it seemed prudent to examine the repeated measures analyses for hopelessness and the BDI-II using trait anxiety as a covariate. The two-way Hopelessness by Condition interaction was no longer significant, F ( l , 89) = 2.73, p > .05, whereas the two-way BDI-II by Condition interaction remained marginally significant, F ( l , 90) = 2.97,p < .10. A s before, low BDI-II scorers showed no significant change in punishment-susceptibility across conditions (-3.10 vs. -4.30), F ( l , 90) = .54,/? > .05, whereas high BDI-II scorers showed decreased punishment-susceptibility after taking a sip o f alcohol (2.32 vs. -4.39), F ( l , 90) = 3.51,/? < .10.  117 C H A P T E R TV  General Discussion 4.1 Summary of Study 2 Study 2 set out to expand the structural model o f Study 1 in a young adult population by examining the extent to which individuals high in depressive symptomatology or hopelessness report drinking for different reasons than people high i n anxiety sensitivity or trait anxiety. Additionally, the second study began to explore negative affect and positive affect motivations for alcohol use by examining B A S and BIS activity through self-report and a behavioural measure o f reward-responsiveness and punishment-susceptibility. There was partial support for Hypothesis 1, in that depressive symptomatology and hopelessness were predictive o f depression drinking motives and trait anxiety was predictive of anxiety drinking motives. Surprisingly, anxiety sensitivity was neither related to, nor was a significant predictor of, either anxiety motives or conformity motives in the regression analysis in Study 2. Hypothesis 2 also received partial support. Depression was the only personality/symptom variable significantly related to alcohol dependence. Concerning negative reinforcement drinking, anxiety motives and depression motives were both strongly related to alcohol abuse, dependence, and frequency. A s was the case here, enhancement motives have consistently been found to relate to alcohol use and problematic drinking, but not to depressive symptomatology or anxiety sensitivity (Comeau et al., 2001; Cooper, 1994; Cooper et al., 1995). Moreover, alcohol frequency was the only significant predictor o f enhancement motives. After modifying the model to reflect the findings o f the young adult sample in Study 2 with regard to the differentiation between depression- and anxiety-related drinking motives, as well as there being a significant relationship between trait anxiety and anxiety drinking motives, Hypothesis 3 appeared to have adequate support. Indeed, depressive symptomatology was a significant predictor o f depression motives. Although trait anxiety was not significantly  118 predictive o f anxiety motives in the structural model, the small effect size was comparable to that of depressive symptomatology to coping motives in Study 1, with alcohol frequency appearing to mediate between anxiety motives and alcohol dependence i n Study 2. The paths from depression motives and alcohol frequency to alcohol dependence were significant. Finally, conformity motives were not related to anxiety motives in the structural model. Expectations concerning the B I S / B A S scales were, for the most part, met. In line with Hypothesis 4, B A S was significantly positively related to extraversion and negatively related to hopelessness and trait anxiety. The BIS scale was positively related to neuroticism and all four personality/symptom variables and negatively related to extraversion. Similarly, neuroticism was positively related to each o f the four personality/symptom variables and extraversion negatively related to hopelessness and trait anxiety. There was partial support for hypotheses concerning C A R R O T reward-responsiveness and punishment-susceptibility as well. Generally, there appeared to be a significant increase in the number o f cards sorted from the baseline trial to the reward trial and a significant decrease in the number o f cards sorted from the baseline trial to the punishment trial, demonstrating the C A R R O T as a valid measure o f reward responsiveness and punishment-susceptibility. Anecdotally, participants appeared to hesitate considerably after two red cards appeared in close contiguity to one another on the punishment trial. Expectations concerning interactions between personality/symptom variables and punishment-susceptibility were partially met. Although individuals scoring high on hopelessness and anxiety sensitivity did not differ in punishmentsusceptibility in comparison to low scorers, differences in punishment-susceptibility were found for depressive symptomatology and trait anxiety. Additionally, personality/symptom variables interacted with condition (i.e., neutral vs. alcohol cue) in terms o f punishment-susceptibility. H i g h scorers on depressive symptomatology, hopelessness and trait anxiety, but not anxiety sensitivity, evinced diminished punishment-susceptibility in the alcohol condition in comparison  119 to the neutral condition. Conversely, no differences in reward-responsiveness were found for any o f the personality/symptorn variables; nor was there an increase in reward-responsiveness found between neutral and alcohol conditions for any o f the personality/symptom variables. Additionally, there was no significant effect for frequency o f alcohol use or interaction with personality/symptom variables concerning reward-responsiveness or punishment-susceptibility across conditions. To follow is an integration o f the results o f both studies and the theoretical implications of these findings. The present research has also produced a number o f anomalous findings with respect to prior research, such as the lack o f relationship between anxiety sensitivity and negative reinforcement drinking in Study 2. Explanations for conflicting results between the two studies and for findings that are in opposition to previous research are offered. A s well, the limitations of the present research are discussed. 4.2 Drinking Motives and Alcohol Use Anxiety sensitivity was found to be a significant predictor o f negative reinforcement drinking in Study 1, trait anxiety a predictor o f negative reinforcement drinking in Study 2, and depressive symptomatology and hopelessness significant predictors o f negative reinforcement drinking in both studies. However, personality/symptom variables were not related to alcohol frequency or alcohol abuse symptoms, whereas depressive symptomatology and hopelessness were related to alcohol dependence in Study 2, indicating that depressed individuals may exhibit psychological dependence on alcohol, but not necessarily drinking at an increased frequency. Alcohol dependence criteria ( A P A , 1994) are not necessarily based on the amount o f alcohol consumed, but rather on the recognition that the use o f alcohol has become problematic (e.g., difficulty cutting down, drinking more than intended). Individuals scoring high on personality/symptom variables may simply differ in their reasons (i.e., drinking motives) for using alcohol in comparison to their low scoring counterparts, and in turn high scorers'  120 motivations are then directly related to alcohol frequency and alcohol dependence. It is interesting to note the similarity between the lack o f direct relationship between personality/symptom variables and alcohol problems i n Study 2 and i n the Vancouver sample in Study 1. Individuals drinking for negative reinforcement reasons may exhibit greater frequency of use at later stages o f the dependence cycle. A s well, the present study used a measure o f the alcohol frequency with somewhat restricted variability, given that the highest category was drinking "weekly." Grouping together once per week drinkers with daily drinkers, and all those in between, limited the likelihood o f Study 2 finding a relationship between alcohol frequency and other variables o f interest. The results o f both studies provide support for the notion that depressive symptomatology and hopelessness play significant and independent roles in alcohol use and alcohol problems. Study 1 presented preliminary evidence for the predictive power o f depression symptoms even after accounting for anxiety sensitivity and trait anxiety, previously found to be significant predictors o f alcohol use and problems (Stewart et al., 2001; Teichman et al., 1989). Study 2 was able to extend these findings, in that depressive symptomatology was exclusively predictive of depression drinking motives. Nevertheless, there were some important differences in the structural model between Study 1 and Study 2. While anxiety sensitivity was predictive o f conformity motives in Study 1, anxiety sensitivity and conformity motives were not significantly related in Study 2. Similarly, conformity motives and anxiety motives were not significantly related Study 2. Anxiety sensitivity and conformity motives played no significant role in predicting alcohol dependence in the structural model in Study 2. Rather, the path from trait anxiety to anxiety motives i n Study 2 accounted for a small portion o f the variance in anxiety motives. The extent to which developmental issues impact drinking motives is unknown, but the lack o f effect for conformity motives in Study 2 may reflect the different kinds o f developmental issues faced by young adults versus adolescents. Such developmental issues may actually impact  121 the type o f motivations they experience. For example, one o f the socio-developmental priorities of adolescence is individuation from one's parental figures, as well as vying for independent adult status among one's peers. Whereas, young adults are more concerning with exploring different roles in order to establish their identities (Erikson, 1959). The imaginary audience o f adolescents—that is, the belief that everyone is looking at you—may result in more social concerns and therefore drinking for conformity reasons. A l s o , recent research with adolescent smokers suggests a developmental pathway moving from social motives to conformity, through enhancement-like motives to emotion regulation and finally using nicotine to feel normal (Johnson et al., 2003). Y o u n g adults may be more interested i n using alcohol for its pleasurable or enhancement qualities. Although hopelessness was able to account for variance in coping motives in Study 1, hopelessness was not a significant predictor o f depression drinking motives i n the presence o f depressive symptomatology the Study 2. This may be partially accounted for by the overlap between hopelessness and hopelessness-related depression items. Depressive symptomatology may be more proximal to alcohol use and alcohol problems than the personality construct o f hopelessness (c.f, Kushner et al., 2001). Although hopelessness is related to alcohol problems, not surprisingly depression symptoms per se appear to predict drinking to alleviate these symptoms. Whereas alcohol may numb negative affect associated with hopelessness (i.e., depression) in the short term, it is unlikely that someone would be motivated to use alcohol to brighten up the future. A m o n g adolescents, suicidal ideation was a predictor o f coping motives and alcohol problems in the N o v a Scotia sample, although there is likely a bidirectional relationship between suicidal ideation and alcohol problems (Light et al., 2003). In the adult sample, however, "changes in sleep pattern" was the only consistent predictor o f depression motives and alcohol dependence. Interestingly, when symptoms similar to those used in the adolescent samples were  122 used in Study 2, suicidal ideation was a significant predictor o f depression drinking motives. However, sadness was the strongest predictor o f depression drinking motives in the latter analysis. That suicidal ideation was a significant predictor o f negative reinforcement drinking among adolescents and only partially found to be so among adults is likely due to a higher prevalence o f suicide among adolescents in comparison to adults (Thorpe, Whitney, Kutcher, Kennedy, et al., 2001). However, it is unknown i f suicidal ideation would continue to be a significant predictor o f coping motives and alcohol problems among adolescents with the inclusion o f "changes in sleep pattern," as this item was not assessed in the adolescent sample. Given the high rate o f suicide among elderly adults (Thorpe et al., 2001), it would be interesting to see i f suicidal ideation is a significant predictor o f alcohol use in that population as well. The limited ability o f suicidal ideation to significantly predict alcohol dependence may also have been due to differences in methodology between the two studies. While Study 1 allowed individuals to endorse the extent to which they experience suicidal ideation, the BDI-II, used in Study 2, assesses the presence or absence o f suicidal ideation and then the extent to which the individual intends to on act on these thoughts. Thus, there is limited variability o f measurement with respect to suicidal ideation when using the BDI-II. There were a number o f other discrepancies between the findings o f Study 1 and Study 2 worth noting. First, conformity motives were unrelated to alcohol abuse or dependence in Study 2, whereas they correlated with alcohol problems in Study 1. Furthermore, unlike Study 1, no relationship between anxiety sensitivity and conformity motives was found in Study 2. Fewer studies have used the revised version o f Cooper's drinking motives questionnaire and therefore less research exists regarding conformity motives. A t any rate, conformity motives appear to be the least likely to be endorsed by university undergraduates (Kairouz et al., 2002). The relationship between anxiety sensitivity and conformity motives in Study 1 is likely due to social concerns among adolescents. Adolescents appear to be especially susceptible to their perception  123 that their peers are using alcohol and other drugs (Iannoti, Bush, & Weinfurt, 1996). It is interesting to note that Stewart et al. (2001) found the social concerns subscale o f anxiety sensitivity to be the only significant predictor o f both alcohol frequency and the number o f binge drinking episodes. The significant path from conformity motives to coping motives in Study 1 provides further evidence for the idea that adolescents are using alcohol as a means o f coping in social situations (Comeau et al., 2001). Although social motives are likely to be positively related to frequency o f alcohol use, it is unlikely that individuals drinking for social reasons would evince problematic drinking (Bradizza et al., 1999; Cooper, 1994; Copper et al., 1992). Therefore, it was somewhat surprising to find a significant relationship between social motives and alcohol abuse and dependence in Study 2. The inconsistency between the present and prior research may be due to the age o f the sample. Only one previous study has found alcohol problems to be significantly related to social motives, and similar to Study 2, they used an undergraduate sample; however, an inspection o f their social motive items reveals that their social motives scale was confounded with conformity motive items (Carey & Correia, 1997). Certainly, university students are known to drink excessively relative to their adolescent and adult counterparts (Wechsler, Lee, K u o , & Lee, 2000; Kairouz, Gliksman, Demiers, & Adlaf, 2002), and therefore more likely to develop alcohol problems even though they are drinking for "normal" social reasons. Excessive alcohol use among social drinkers is likely related to physical dependence (i.e., tolerance, withdrawal), as opposed to psychological dependence. It should also be noted that Study 2 consisted o f primarily a female sample. Generally, due to differences in body-mass, females need to consume less alcohol than males to reach intoxication, which may result in university age female social drinkers building up tolerance and experiencing withdrawal much more readily. It would be interesting to see i f male university students' excessive social drinking also leads to a significant relationship between social drinking motives and alcohol dependence, and whether  124 there are in fact differences i n the degree to which social motives and alcohol dependence are related among males and females. A second difference between the two studies was that depressive symptomatology was a significant predictor o f social drinking motives in Study 2. The relations between depressive symptomatology and social motives found i n the regression analysis may be due to the effects o f alcohol use, especially considering that alcohol frequency was the only other significant predictor o f social motives. There is evidence that in many cases alcohol use disorders precede depressive symptomatology (Schuckit et al., 1997a; Schuckit et al., 1997b). Thus, drinking for social reasons may result i n depression symptoms in individuals drinking large quantities o f alcohol at a high enough frequency, as seems to be the case among university undergraduates (Wechsler et al., 2000; Kairouz et al., 2002). Alcohol-related problems resulting from excessive alcohol use, such as depressed affect, among individuals drinking for social reasons may be addressed by a simple informational intervention program, rather than such individuals being misdiagnosed with depression. Neither would such individuals necessarily require full-scale alcohol use disorder treatment i f they are not showing signs o f psychological dependence. Taken together, the two studies provide preliminary evidence for the specific role o f depressive symptomatology and hopelessness in predicting substance use disorders. A s well, anxiety sensitivity and trait anxiety appear to have independent roles i n alcohol use and alcohol problems, perhaps depending on an individual's age and motivation for using alcohol. Depressive symptomatology appears to be independently predictive o f drinking to cope with depression symptoms, whereas trait anxiety appears to be predictive o f drinking to cope with anxiety symptoms. Additionally, anxiety sensitivity appears to be predictive o f drinking to reduce anxiety associated with the possibility o f negative evaluation i n social situations, especially among adolescents  125 4.3 Reinforcement  Sensitivity: Personality and Symptom Variables  A second aim o f Study 2 was to investigate the extent to which B I S and B A S activity are able to account for variance i n hopelessness, depressive symptomatology, anxiety sensitivity and trait anxiety using self-report and a reinforcement contingency paradigm. Individuals scoring high on depressive symptomatology and hopelessness were expected to evince an underactive behavioural activation system and an overactive behavioural inhibition system (Gray, 1994). Anxiety sensitivity and trait anxiety, on the other hand, were expected to be exclusively related to behavioural inhibition system activity (Gray & McNaughton, 2000). 4.3.1 BAS Activity and Personality/Symptom  Variables  The current study was able to provide further evidence for the hypothesis that hopelessness is associated with low B A S activity. Hopelessness was negatively correlated with the B A S scale. Conversely, depressive symptomatology was not significantly related to the B A S scale. Although positive affect is inversely related to both depressive symptomatology (Clark & Watson, 1991; Gencoz, 2002) and the B A S scale (Carver & White, 1994; Heubeck et al., 1998; Jorm et al., 1999), previous research has often failed to find an inverse relationship between the B A S scale and depressive symptomatology (Jorm et al., 1999; Meyer et al., 2001; Johnson et al., 2003). Only two previous studies have found such a relationship. One used a measure reflecting anhedonic depression, "a subset o f symptoms that are especially likely to be influenced by the B A S , " as well as more severe depression (Beevers & Meyer, 2002, p. 562). The other used a clinical sample o f individuals diagnosed with Major Depressive Disorder (Kasch et al., 2002). The lack o f relationship between B A S and depression in the current research is likely due to limited variability with respect to depressive symptomatology, given the use o f an undergraduate sample. A s well, the BDI-II appears to be more heavily weighted toward negative affect than lack o f positive affect (Morley, Williams, Black, 2002). Given the relationship between negative  126 affect and BIS activity, it is not surprising that the BDI-II was related to the B I S and not B A S scale. There were also no differences on the C A R R O T reward-responsiveness task with respect to either hopelessness or depressive symptomatology. This is in opposition to previous research using a reinforcement contingency paradigm (Henriques & Davidson, 2000; Henriques et al., 1994). The lack o f relationship between depressive symptomatology and reward-responsiveness should not necessarily be interpreted as a lack o f relationship between depression and B A S activity. There are two possible explanations for a relationship between depressive symptomatology and low B A S activity. It may be that low B A S activity results in less motivation to seek out rewarding experiences and i n turn a paucity o f positive experiences leads to depression (Lewinsohn, Lobitz, & Wilson 1973). Alternatively, low B A S activity may result in an information-processing bias that interferes with recall o f past positive events and lowers anticipation o f future positive experiences (MacLeod, Tata, Kentish, & Jacobson, 1997; Stober, 2000). Likely, it is a combination o f the two. L o w B A S activity results in low recall of previous positive experiences and biased expectations regarding future positive experiences, and consequently such individuals are less motivated to seek out future rewarding experiences (Gable, Reis, & Elliot, 2000). Indeed, Beevers and Meyer (2002) found that self-reported expectations regarding future positive experiences mediated between the B A S scales and anhedonic depression symptoms in an undergraduate sample. Moreover, retrospective report o f positive experiences did not directly mediate between the B A S scales and anhedonic depression symptoms, but rather had an indirect relationship through expectations regarding future positive experiences, indicating that biased expectations interfered with recall o f past positive experiences as well. The present research may not have been able to establish an inverse relationship between depressive symptomatology and reward-responsiveness due to the salience o f the reward in the  127 C A R R O T reward trial. It was obvious to participants that they truly would be rewarded, or became so after the first five cards, and, therefore, provided less room for biased expectations regarding reward. Previous research finding a relationship between reward-responsiveness and depressive symptomatology, on the other hand, may have activated reward expectation biases. For example, Henriques and Davidson (2000) required individuals to remember target words in order to receive a reward. Similar to the current study, participants likely trusted the experimenter that they would indeed be rewarded upon remembrance o f target words (i.e., no deception). However, their participants also needed to hold the expectation that they would be able to remember the words properly, as opposed to the biased expectation that this would be another unrewarding experience (i.e., "I never w i n anything."). Depressed individuals' bias regarding anticipated reward would likely only interfere with motivation on tasks that activate these cognitive biases. O n the other hand, motivational deficits would not be seen when individuals believe that their efforts indeed result in reward. The C A R R O T is unlikely to activate reward expectation biases and is therefore a task that assesses reward-responsiveness per se, whereas tasks relying on individuals' belief that reward is contingent upon their abilities would better assess the interaction between positive expectation biases and rewardresponsiveness or motivation. The current null findings concerning the relationship between depressive symptomatology and B A S activity are also inconsistent with E E G research indicating a relationship between depression and relative left frontal hypoactivity, postulated as the neurological correlate o f the behavioural approach/activation system (Gotlib et al., 1998; Henriques & Davidson, 1990, 1991; Harmon-Jones, Abramson, Sigelman, Bohlig, Hogan, & Harmon-Jones, 2002). A s well, the current research suggests that there is a relationship between depression and BIS activity, whereas E E G research has found little evidence for a relationship between depression and its counterpart to BIS activity, hyperactive right frontal activity (Gotlib  128 et al., 1998; Henriques & Davidson, 1990, 1991). Counter to the information-processing interpretation presented above, Gotlib et al. (1998) did not find an association between E E G alpha asymmetry and cognitive bias among depressed individuals using an emotion Stroop task—words with emotional content (e.g., afraid). However, like the present research, their findings are not able to directly speak to biased expectations regarding rewarding experiences. A n alternative explanation for the lack o f relations among hopelessness, depressive symptomatology and reward-responsiveness m a y b e found in Gray's (1994) hypothesis that depressive symptomatology can arise from high BIS functioning having a reciprocal effect on B A S activity. It may be that activation o f the BIS is required i n order to see lower levels o f B A S among depressed individuals. Evidence for this comes from previous research finding differences between low and high hopelessness when using a mixed reward/punishment incentive design (Conrod et a l , 2002), as opposed to the reward only and punishment only conditions i n the C A R R O T , which were specifically chosen so as to not lead to reciprocal inhibition o f B I S on B A S activity, and thus provide an individual measure o f each behavioural reinforcement system separately. Additionally, trait anxiety was negatively related to the B A S scale. However, the S T A I T showed a great deal o f overlap with depression. A confirmatory factor analysis o f the S T A I - T found anxiety and depression factors to be subsumed under a higher order negative affect factor (Bieling et al., 1998). It is therefore likely that the relationship between trait anxiety and B A S activity is reflective o f the overlap with hopelessness and depressive symptomatology, especially given that a number o f items on the BDI-II correspond to items on the S T A I - T (Bieling et al., 1998). 4.3.2 BIS Activity and Personality/Symptom  Variables  A s expected, depressive symptomatology was positively correlated with the self-report BIS scale. In addition, individuals scoring high on depressive symptomatology evinced more  129 punishment-susceptibility on the C A R R O T than low scorers. The current research is in line with previous findings with respect to there being a positive relationship between B I S activity and depressive symptomatology using self-report (Beevers & Meyer, 2002; Jorm et al., 1999; Kasch et al., 2002; Meyer et al., 2001; Johnson et al., 2003) and an experimental reinforcement contingency paradigm (Henriques et al., 1994). In contrast to Henriques and Davidson (2000), who attributed this relationship to the presence o f anxiety among depressed individuals, the current study found a marginal difference in punishment-susceptibility with respect to depressive symptomatology after controlling for trait anxiety. Johnson et al. (2003) found individuals with a lifetime diagnoses o f either depression or anxiety disorders to be higher on the BIS scale. Moreover, the relationship between depression and BIS was not qualified b y a depression by anxiety interaction. It is interesting to note that they found high B I S to be associated with as great a likelihood o f being diagnosed with depression as with being diagnosed with an anxiety disorder, even after controlling for other psychopathology, including substance abuse/dependence.  The B I S scale was also able to  distinguish among nondepressed, previously depressed, and currently depressed individuals. Thus, the current research adds to the mounting evidence suggesting that B I S activity plays a significant role in depressive symptomatology, distinct from its role in anxiety. Also consistent with previous research, hopelessness was positively correlated with the BIS scale (Woicik, unpublished data). This relationship may be due to the overlap between depressive symptomatology and hopelessness. Alternatively, high hopelessness individuals may evince a personality trait that is associated with high B I S activity, putting them at risk for depression and other psychopathology (see Conrod et al., 2000). Unexpectedly, high scorers on hopelessness did not evince more punishment-susceptibility than low scorers using a reinforcement contingency paradigm. Although this finding is in line with previous research (e.g., Conrod et al., 2002), it may seem, on the surface, somewhat incongruous with the  130 relationship between hopelessness and the BIS scale. However, hopeless individuals appear more likely to see the future as less rewarding than they are to see the future as more punishing (Abramson et al., 1989; Macleod, Pankhania, & M i c h e l l , 1997; Steed, 2001; Tanaka et al., 1998). It may be a lack o f inhibition o f BIS by B A S that leads to high B I S activity among hopeless individuals. Moreover, beliefs about the future reflect cognitive schema and do not necessarily reflect poor B A S motivation, but rather would be accessible through an experimental paradigm involving mixed incentives or involving decision-making (e.g., Conrod et al., 2002). Alternatively, the present research provides evidence for a punishment-sensitive motivational deficit when hopelessness manifests at the symptom level in depression. The relationship between self-report BIS activity and hopelessness may also be due to the large overlap with trait anxiety found in Study 2 (see Table 12 & Figure 5). Anxiety sensitivity and trait anxiety were significantly positively related to self-reported BIS activity. A s characterological or trait anxiety is postulated to closely correspond to BIS activity (Gray & McNaughton, 2000; Zinbarg & Mohlman, 1998), the relationship between trait anxiety and punishment-susceptibility on the C A R R O T was likely due to the expected BIS freezing response when red cards appeared. The current study was the first to examine the relationship between anxiety sensitivity and BIS activity. Although hypotheses concerning selfreported B I S activity were borne out, high anxiety sensitivity individuals were not found to be more susceptible to punishment on the C A R R O T . Most, but not all (Kasch et al., 2002), previous studies have found a positive relationship between the B I S scale and current anxiety symptoms (Beevers & Meyer, 2002; Johnson et a l , 2003; Jorm et al., 1999). Considering the relationship between anxiety sensitivity and anxiety disorders (Taylor, K o c h , & M c N a l l y , 1992), it is somewhat surprising that anxiety sensitivity and punishment susceptibility were unrelated. However, anxiety sensitivity consists o f fear o f negative evaluation, loss o f mental control, and physical illness ( M c N a l l y , 1996; Peterson & Reiss, 1992; Reiss et al., 1986) due to the  131 misinterpretation o f bodily sensations concomitant with anxiety (Clark & Salkovskis, 1997; M c N a l l y , 2002). The C A R R O T punishment-susceptibility task likely did not trigger a physiological anxiety response and therefore was unable to evaluate punishment-susceptibility among anxiety sensitive individuals. L i k e hopelessness, anxiety sensitivity is a personality variable involving cognitions about anxiety symptoms and may be more readily assessed using a paradigm that primes for these cognitions. Generally, Study 2 provided evidence for the relationship between depressive symptomatology and B I S activity using self-report and experimental means, but no support for the hypothesized relations between depressive symptomatology and B A S activity. O n the other hand, there was mixed evidence for high hopelessness individuals to manifest lower B A S activity and higher B I S activity, depending on whether self-report or the C A R R O T was used to assess BIS. Although hypotheses concerning trait anxiety and B I S activity were borne out, there was mixed evidence for the relations between BIS activity and anxiety sensitivity. 4.4 Reinforcement  Sensitivity and Alcohol Cues  The final goal o f Study 2 was to examine whether reward-responsiveness could be restored and/or punishment-susceptibility dampened using a cue-reactivity paradigm. It was hypothesized that individuals scoring high on hopelessness and depressive symptomatology are motivated to use alcohol to increase/restore positive affect and to decrease/numb negative affect. Given the hypothesized correspondence between positive affect and B A S activity and between negative affect and B I S activity (Gray 1990,1994), it was expected that high hopelessness and depressive symptomatology scorers would show more reward-responsiveness and less punishment-susceptibility on the C A R R O T after an alcohol cue (i.e., sip o f alcohol) in comparison to a neutral cue (i.e., nibbling a cracker). In contrast, individuals scoring high on anxiety sensitivity and trait anxiety were expected be motivated to drink i n order to decrease  132 negative affect, and therefore show a decrease in punishment-susceptibility after an alcohol cue in comparison to a neutral cue. 4.4.1 BAS Activity and Alcohol Cues Although none o f the personality/symptom variables were associated with lower rewardresponsiveness using the C A R R O T , it was possible that hopelessness and depressive symptomatology could still have led to personality specific sensitivities to the effects of alcohol on reward-responsiveness. However, this was not the case. Neither hopelessness nor depressive symptomatology resulted in significant increases in reward-responsiveness following an alcohol cue. Consistent with previous research using physiological correlates o f reward from alcohol (e.g., increased heart-rate; Conrod et al., 1998), high anxiety sensitivity individuals showed no increase in reward-responsiveness in the alcohol condition in comparison to the neutral condition. The current research also found no differences for reward-responsiveness among alcohol groups and therefore was unable to replicate the findings o f Kambouropoulos and Staiger (2001). Whereas they found less reward-responsiveness among heavy drinkers i n comparison to light drinkers, the current results suggest no differences due to frequency o f alcohol use (see Note 3). However, Kambouropoulos and Staiger used binge drinking within the previous 2 weeks versus no binge drinking in the previous 2 weeks as a definition o f heavy versus light drinkers, respectively, whereas the current study used frequency o f alcohol use to define light versus regular drinkers. The difference between the two studies regarding selection criteria, along with possible cultural differences in drinking between Canadian and Australian undergraduates, may explain the current null findings. 4.4.2 BIS Activity and Alcohol Cues Individuals scoring high on both hopelessness and depressive symptomatology indeed showed less punishment induced slowing o f card-sorting in the alcohol condition than they did in  133 the neutral condition, indicating that they experienced diminished punishment-susceptibility  after  taking a sip o f alcohol. Similarly, individuals high in trait anxiety appeared to experience less punishment-susceptibility after sipping alcohol than in the neutral condition. Individuals scoring high on anxiety sensitivity, however, showed no diminishment o f punishment susceptibility in the alcohol condition. A l s o , frequency o f alcohol use did not interact with personality/symptom variables across neutral and alcohol conditions, indicating that experience with alcohol was not necessary to produce a diminishment of punishment-susceptibility from alcohol. This finding is interesting i n light o f previous research indicating that, among individuals at high risk for developing alcohol disorders, punishment dampening was only found at moderate to high doses of alcohol (Stewart, Finn, & Pihl, 1992). Alternatively, the measure o f alcohol frequency used in Study 2 may not have allowed for enough variability i n alcohol use. Given that there was a general slowing of card-sorting among all individuals on the punishment trial, it is interesting that only those high in personality/symptom variables evinced diminished punishment-susceptibility in the alcohol condition. The current findings correspond to previous research establishing differential alcohol reinforcement effects among individuals at risk and not at risk for alcohol use disorders (Conrod, Peterson, & Pihl, 1997; Conrod, Pihl, & Vassileva, 1998; Pihl & Peterson, 1995). For example, Conrod et al. (1998) found that all individuals show increased heart rate and skin conductance level i n anticipation of an electrical shock, but only high anxiety sensitivity individuals and those with multigenerational family history o f alcoholism showed a slowing of heart rate in response to alcohol intoxication. Furthermore, anxiety sensitivity appears to result in a distinct pattern o f physiological reactivity, in that only high anxiety sensitivity was associated with diminished skin conductance level with alcohol intoxication, even though alcohol appeared to lessen self-reported pain in all groups. Conrod et al. also found a trend for alcohol to result i n less self-reported anxiety in anticipation of shock among high anxiety sensitivity individuals. This was similarly the case here, with both  134 low and high personality/symptom groups evincing punishment-susceptibility on the C A R R O T , but only high scoring individuals experiencing punishment dampening from alcohol. Thus, there appear to be personality specific idiosyncratic responses to alcohol's reinforcement properties. Although everyone may attend to punishing stimuli i f it is o f a large enough magnitude (Mathews & M a c L e o d , 2002), only certain individuals are expected experience negative reinforcement from alcohol. Since the current research relied on an alcohol cue rather than the physiological effects o f alcohol, it would appear that expectations regarding the reinforcement properties o f alcohol are enough to produce the observed differences between personality/symptom groups. Although predictive o f alcohol use (McKee, Hinson, W a l l & Spriel, 1998), experience with alcohol does not appear necessary for an individual to possess alcohol expectancies (Kraus et al., 1994; Smith & Goldman, 1994), providing some insight into the lack o f effect for alcohol groups in the present study. Furthermore, expectancies may be prior to motivations for alcohol use (Cooper et al., 1995; Williams & Clark, 1998). Relevant to the current study are tension reduction or negative emotion regulation alcohol expectancies (Goldman et al., 1997; Kassel, Jackson & Unrod, 1999; Leigh & Stacy, 1993; W i l l i a m & Clark, 1998), which appear to be related to drinking to reduce negative affect (Kushner, Sher, Wood, & W o o d , 1994; Leonard & Blane, 1988). Personality and psychopathology appear to interact with alcohol expectancies to predict frequency o f alcohol use and problematic drinking (Kushner et a l , 1994; Lee, Greely, & Oei, 1999; Leonard & Blaine, 1988). For example, social anxiety is associated with the belief that alcohol increases social assertiveness (Leonard & Blaine, 1988). A s well, negative emotion regulation alcohol expectancies are related to state anxiety and depressive symptomatology, and contribute to the prediction o f problem drinking after accounting for drinking motives (Kassel et al., 1999). A l c o h o l expectancies may have interacted with personality/symptom variables in the current sample to reduce punishment-susceptibility, whereby those scoring high on negative  135 affect regulation expectancies and personality/symptom variables are those who would have evinced the least punishment-susceptibility in the alcohol condition. Although punishment-susceptibility in depression and anxiety are both theoretically the result o f BIS activity, and an alcohol cue was able to reduce punishment-susceptibility in both, depressive symptomatology continued to result in some diminishment o f punishmentsusceptibility in the alcohol condition even when covarying out trait anxiety, postulated to be a strong predictor o f BIS activity (Gray & McNaughton, 2000). Thus, a single brain system appears to be associated with two distinct personality/symptom profiles that result in a similar behavioural response (i.e., behavioural inhibition), which can i n turn be interrupted by an alcohol cue. Likewise, the structural model indicates that individuals high in depressive symptomatology drink for different reasons than those high in trait anxiety. Although trait anxious individuals appear to experience distress concerning threat o f punishment and depressed individuals experience distress concerning threat o f loss (see Beck & Clark, 1997), the results for the structural model and the C A R R O T task indicate that both personality/symptom profiles are associated with drinking to decrease negative affect. In general, the current results support previous findings with respect to a relationship between depressive symptomatology and BIS activity. Moreover, depressive symptomatology was able to account for variability in punishment susceptibility, even after controlling for trait anxiety. Study 2 was unable to establish a difference in reward-responsiveness among individuals scoring low and high on the BDI-II (cf, Henriques & Davidson, 2000; Henriques et al., 1994). A s well, there appears to be evidence for a relationship between trait anxiety and BIS activity, whereas the relationship between anxiety sensitivity and B I S activity was mixed. 4.5 Limitations and Future  Directions  The present research has a number o f limitations worth noting. First, there are significant differences between the Vancouver samples (i.e., second sample in Study 1 and Study 2 sample)  136 and the N o v a Scotia sample i n Study 1. Although the differences in the ethnic make-up o f these samples provides wider generalizability, the Vancouver samples were ethnically overly representative o f individuals o f Asian decent in comparison to the N o v a Scotia sample in Study 1. There are some differences in terms o f the effects o f alcohol on individuals o f differing racial backgrounds worth noting. For example, the flushing response (i.e., reddening o f the skin, feeling hot etc.), which has been referred to as a protective factor against alcohol disorders, is seen more among Asians than non-Asians (Park et al., 1992; Yamamoto, Lee, L i n , & Cho, 1987). A s well, undergraduates o f Asian decent report lower alcohol frequency, binge drinking, and problem drinking (Hansen, Blackwell, & Conrod, 2002). Moreover, conformity motives are not predictive o f alcohol use among Asians, even though they appear to score higher on conformity motives (Hansen et al., 2002). The differences in ethnic makeup o f samples may account for the lack o f relationship between conformity motives and problematic drinking in Study 2. Ethnicity may also have contributed to the lack o f suicidal ideation as a significant predictor o f depression motives and alcohol dependence in the Vancouver samples (see Note 1), given that Asians appear to be less likely than North Americans to report suicidal ideation (Zhang & Jin, 1996). Future research w i l l need to establish the extent to which the structural model and the findings regarding punishment-susceptibility and personality hold up in Asian and non-Asian samples. Additionally, males consisted o f only a small proportion o f the sample in Study 2. A s well, all three samples were two small to allow hypotheses concerning gender to be examined with respect to the structural models. Given that females score higher on anxiety sensitivity (Stewart, Taylor, & Baker, 1997), trait anxiety (King, Bernardy, & Hauner, 2003), and depressive symptomatology (Bertakis, Helms, Callahan, A z a r i , Leigh, & Robbins, 2001; NoelenHoeksema, 1990), the current findings may not generalize well to adult males. Moreover, there appear to be gender differences with respect to these personality/symptom variables and drinking  137 motives, alcohol use and alcohol problems (Deshmukh et al., 2003; Kessler et al., 1993; K i n g , Bernardy, & Harner, 2003; Stewart et a l , 2001; Stewart & Zeitlin, 1995 Weissman et al., 1993). For example, as was the case i n the present research, males consistently endorse more conformity motives than females (Comeau et al., 2001; Stewart et al., 2001). The extent to which these findings would differ between males and females is an empirical question in need o f further research. The structural models i n Study 1 and 2 were achieved using a correlational methodology. However, a longitudinal methodology w i l l need to be used i n order to truly establish causal pathways with respect to personality/symptom predictors and substance use. Although BIS and B A S activity is thought to be relatively stable (Kasch et al., 2002), there is evidence that it may fluctuate over time (Jorm et al., 1999). Given that increases and decreases in B A S activity are hypothesized to account for unipolar and bipolar depression (Depue & Iacono, 1989), a longitudinal design may best capture BIS and B A S motivations for substance use. There were some minor inconsistencies between the results using traditional regression analysis and structural equation modeling. Structural equation modeling, using latent variables, is a powerful tool for demonstrating casual pathways between variables, more so than hierarchical regression because it is able to disattenuate path coefficients for unreliability o f measurement. However, the decision to use latent-variable modeling in the current research was post data collection and, therefore, empirically unidimensional measures were divided in order to provide two indicators o f each latent variable. Future research using more than one statistically sound measure to assess each latent variable w i l l be needed i n order to confirm the current findings. Both studies used a non-clinical sample, Study 1 using high school students and Study 2 university undergraduates. Although the use o f such samples provides wider generalizability, it may be that the relationships under investigation would be more salient within a clinical sample.  138 In order to provide further evidence for the unique roles o f BIS and B A S motivations among depression and anxiety motivated substance users, the next step would be to attempt to establish the current findings using a clinical sample o f heavier drinking or alcohol-disordered individuals diagnosed with depression without comorbid anxiety and individuals diagnosed with anxiety disorders without comorbid depression. Also, anhedonic depression symptoms appear to be more strongly related to lower B A S activity (Beevers & Meyer, 2002). Although the BDI-II has a number o f items reflecting anhedonia (Beck et al., 1996), a measure o f depression more heavily loaded with anhedonia items appears appropriate. A lab setting may not have been the optimal environment for a cue-reactivity experiment, and may explain the small effects for dampening o f punishment-susceptibility. However, a recent meta-analysis indicates that the desired physiological and behavioural effects from an alcohol cue can be achieved in a lab setting, whereas converting one's lab into a bar can lead to opposite effects to those desired, with the optimal setting being a naturalistic environment (e.g., sitting in easy-chair; M c K a y & Schare, 1999). The presence o f taste and odour o f alcohol appear to lead to the largest physiological and behavioural changes associated alcohol ( M c K a y & Schare, 1999). Because the current research mixed and poured drinks in front o f participants and had them take a sip o f real alcohol, it is more likely to have produced the effects associated with alcohol use. However, in addition to the physiological effects associated with alcohol, cues appear to more readily lead to alcohol craving (Carter & Tiffany, 1999). Since no measures o f physiological reactivity or alcohol craving were taken, the effects o f alcohol cue in the present research, and the extent to which cue-reactivity interacted with personality is unknown. However, previous research using the same design found increased alcohol craving in response to alcohol cue (Kambouropoulos & Staiger, 2001). Individuals hypothesized to show high BIS activity may have experienced less dampening o f punishment-susceptibility, due to alcohol craving, than they would have to the actual physiological effects o f alcohol.  139 The present research may be further limited by the use o f alcohol cues rather than intoxication when one considers that previous research has found little relationship between hopelessness and alcohol use (Blackwell et al., 2002; Conrod et al., 2000). Furthermore, alcohol intoxication is likely necessary to induce affective change (Van Tilburg & Vingerhoets, 2002). However, given that dampened punishment-susceptibility was found among high hopelessness individuals, it may be expectancy concerning alcohol that is at work, and therefore evidence that conditioned alcohol cue is what led to the effect. The lack o f interaction between personality and alcohol use with respect to C A R R O T results indicates that the current research would have benefited from an alcohol expectancy measure. Considering that alcohol can become associated with alcohol behaviours and the drinking environment (Carter & Tiffany, 1999), individuals may have developed secondary conditioning of alcohol to the research lab. Powell et al. (2002) found differences among smokers due to order o f presentation between abstaining and smoking conditions, which may have been the results o f secondary conditioning with the lab environment. Although there was no significant effect due to order o f cue presentation in the current research, presentation o f the alcohol condition prior to the neutral condition may have introduced noise, which may further explain the small effects for diminished punishment susceptibility and a non-significant effects for reward-responsiveness. That some participants purposefully slowed during the C A R R O T punishment trial does not necessarily mean that their behaviour is not BIS-related. Rather, this strategy was reportedly employed to avoid punishment (i.e., loss o f money) and therefore likely reflects BIS activity mediated through conscious decision-making, again indicative o f an information-processing interpretation o f R S T . In order to achieve the BIS freezing response on the punishment C A R R O T trail, a cue for punishment not directly linked to card-sorting speed would have been  140 better. For example, as opposed to red numbers on the cards, a red light could be turned on to signify punishment. Unexpectedly, the B I S / B A S scales and reinforcement sensitivity on the C A R R O T were not related. This may be due to differences between assessing B I S and B A S activity using selfreport as opposed to a reinforcement contingency task. The B I S / B A S scales merely reflect emotional and behavioural consequences o f these constructs, whereas the C A R R O T is designed to be a direct measure o f approach and avoidance behaviour associated with B A S and BIS activity in the face o f reward and punishment, respectively. Items on the BIS scale reflect anxiety and fear, as opposed to behavioural avoidance o f punishment and B A S items reflect impulsivity and excitement regarding reward, as opposed to approach behaviour. Previous research with individuals exhibiting low approach behaviour and the restorative effect of a dopamine agonist on approach behaviour indicates the C A R R O T is a valid measure o f this aspect o f B A S activity. Although previous research has reported a positive relationship between the C A R R O T reward-responsiveness and B A S scales (Pickering & Gray, 1999), the current study indicates that B I S / B A S scales and the C A R R O T may be assessing different aspects o f reinforcement sensitivity. Further research w i l l be needed to establish the punishment portion o f the modified C A R R O T as valid measure o f BIS activity. Henriques and Davidson (2000) proposed that depressed individuals would likely evince increasing reward-responsiveness as the magnitude o f reward increased and, at certain levels of reward, display normal B A S activity. This is also likely the case concerning the salience or likelihood o f reward. That is, as the likelihood o f reward increases, individuals with low B A S activity would display normal reward-responsiveness. The finding o f normal rewardresponsiveness among individuals high in depressive symptomatology and hopelessness in Study 2 may have been due to depressed individuals' biased expectations regarding reward (Beevers & Meyer, 2002). Reward on the C A R R O T may have exceeded the expectations o f individuals high  141 in depressive symptomatology leading to normal reward-responsiveness. It is easy to conceive of an experiment i n which both the size and probability o f the occurrence o f reward would vary between conditions to test whether it is the magnitude or the likelihood o f reward that has differential impact on depressed and non-depressed individuals. Gray (1982; Gray & McNaughton, 2000) has suggested two alternative interpretations o f R S T . The first and more popular version is the motivational interpretation, i n which BIS activity corresponds to avoidance motivation and B A S activity corresponds to approach motivation. The second interpretation can be referred to as an associational or information-processing version, wherein individuals high in B I S activity allocate more attentional and associational resources to aversive stimuli and individuals high in B A S activity allocate more attentional and associational resources to appetitive stimuli (Zinbarg & Revelle, 1989). Whereas the latter interpretation is able to account for differences i n speed o f associational acquisition found among trait anxious and impulsive individuals, the former interpretation predicts that there should be no such differences (Zinbarg & Revelle, 1989; Zinbarg & Mohlman, 1998). Using a daily diary, Gable et al. (2000) found the B A S scale to be related to daily positive affect and positive life events and the B I S scale to be related to daily negative affect and daily negative events. Furthermore, people high in BIS reported more negative affect when negative events occurred, but not greater occurrence or importance o f negative life events. In other words, someone high i n B I S finds the same events more distressing than someone low on BIS. Although there is some support for an associational or information-processing bias among high B A S individuals, given that extraverts appear to be more susceptible to positive mood induction (Larsen & Ketelaar, 1989, 1991; Rusting & Larsen, 1997), Gable and colleagues (2000) found support for a motivational interpretation o f B A S activity. Individuals high in B A S reported greater occurrence and more importance o f positive life events, which in turn mediated between B A S and positive affect, as opposed to high B A S individuals reporting greater positive  142 affect or more pleasure when positive events occurred. Furthermore, higher B A S activity seemed to increase the likelihood o f future positive events—that is, today's B A S activity increases the number o f positive events experienced tomorrow. In general, there appears to be support for both the motivational and informationprocessing interpretations o f R S T (Gable et al., 2000; Zinbarg & M o h l m a n , 1998). H i g h BIS activity appears to influence the interpretation o f negative events, wherein negative events are interpreted more negatively, whereas B A S appears to influence the number o f positive events experienced. Negative events are not usually sought, but rather simply occur in everyday life, with BIS activity determining the emotional reaction to such events, whereas individuals must be motivated to seek out positive events, with previous positive events determining positive affect, and in turn motivation to seek out future positive events (Gable et al., 2000). A combination o f the information-processing and motivational interpretations o f R S T would appear able to account for variation i n positive affect (i.e., B A S ) and negative affect (i.e., BIS) that, in turn, correspond to anxiety and depression. The results o f the current and previous research allow for some speculation concerning the role o f B A S and B I S activity in depression and anxiety motivated drinking (see Figure 11). Through reciprocal inhibition high BIS activity or low B A S activity could produce depressive or anxious symptomatology (Gray, 1994). H i g h BIS activity, among both anxious and depressed individuals, would likely lead to heightened expectations concerning future negative events (Beevers & Meyer, 2000; Gable et al., 2000), and therefore hypervigilance toward negative events, resulting in negative affect, manifesting through threat o f loss i n depression and threat o f danger in anxiety (Beck & Clark, 1997). The role o f negative affect in depression and anxiety (Clark & Watson, 1991; Gencoz, 2002) and alcohol use (McCreary & Sadava, 2000; W i l l s , Sandy, Shinar, & Yaeger, 1999) has been well established.  143 Generally, high trait anxious individuals appear to orient toward threatening information more than low trait anxious individuals (Bradley, M o g g , Falla, & Hamilton, 1998; Mathews & MacLeod, 2002; Schmidt, Lerew, & Trakowskim, 1997). Evidence is somewhat mixed for depressed individuals evincing an information-processing bias toward negative information (e.g., A l l o y , Abramson, Murray, Whitehouse, & Hogan, 1997; Segal & Ingram, 1994; Watkins, 2002; MacLeod & Byrne, 1996). For example, depressed individuals take longer to name negative words on an emotion Stroop task (Gotlib & M c C a n n , 1984) and recall more negative selfreferent information than non-depressed individuals (MacLeod & Mathews, 1991), but do not consistently showing a bias toward future negative information (MacLeod et al., 1997; Stober, 2000). However, individuals scoring high on depressive symptomatology showed greater punishment-susceptibility than low scorers in Study 2, indicating that they experienced behavioural inhibition when faced with negative (i.e., punishing) information.  144 Figure 11. Theoretical M o d e l o f the Role o f Reinforcement Sensitivity in Depression and Anxiety Motivated Drinking.  ftBAs|  UBIS  •U Inhibition o f BIS High BIS  Low B A S  ft Inhibition o f B A S Hypervigilance toward negative events  Less Motivation to seek out positive experiences  Few positive experiences  Threat of Danger  Less Motivation to seek out positive experiences (hopelessness)  Memory Bias  Anxiety  I  Threat o f Loss  Negative Affect  | ti G A B A  Il.BAS j ft Dopamine  Anxiety motives  Less Positive Affect  r Alcohol Use  zn  Alcohol Expectancies  Depression  r Depression motives  Alcohol Problems Note. L o w B A S activity leads to an increase in BIS activity and vice versa, which in turn contribute to less motivation to seek out positive experiences and hypervigilance toward negative experiences, respectively. Biased expectations regarding future events negative and positive events lead to low positive and high negative affect, and in turn depression and anxiety symptoms. Experiences with alcohol in these affective states are stored i n memory in the form of expectations and are recalled when individuals are motivated to drinking to cope with anxiety or depression symptoms, which in turn lead to alcohol problems, including tolerance and withdrawal and further alcohol use. The B A S and BIS reinforcement sensitivity portion o f the model is based on Gray and Pickering (1999).  145 Depressed individuals consistently show a bias regarding future positive events (Beevers & Meyer, 2000), leading to a decrease in positive affect (Gable et al., 2000; Stober, 2000; M a c L e o d & Salaminiou, 2001). The current research provides evidence for the relationship between depressive symptomatology and depression motivated alcohol use, while previous research has established the role o f low positive affect in both depression (Clark & Watson, 1991; Gencoz, 2002) and alcohol use (Wills, Sandy, Shinar, & Yaeger, 1999). It is interesting to note that most o f the depression items that were predictive o f drinking motives and alcohol problems i n both studies have a large cognitive component to them (e.g., hopelessness, guilt, worthlessness). Moreover, the role o f hopelessness, at the personality and symptom level, in the current research suggests that biases toward the future play a significant role in negative reinforcement drinking. Indeed, depressed individuals consistently show a bias regarding less expectation o f future positive events (MacLeod & Byrne, 1996; M a c L e o d & Slaminiou, 2001; MacLeod et al., 1997; Stober, 2000). Strober (2000) found that anxiety and depression were equally related to the number o f anticipated negative events, but only depression was related to the number o f anticipated positive events, a relationship that remains significant even after controlling for verbal fluency (MacLeod & Slaminiou, 2001). In addition to the number o f anticipated positive events, depressed individuals also rate such events as likely holding less pleasure (MacLeod & Salaminiou, 2001). A n individual's personality/symptom profile appears to lead to different expectations regarding alcohol, based on his/her experience with the effects o f alcohol on the brain (e.g., G A B A , dopamine) and/or through modeling or informational sources o f alcohol's effects (Goldman, Darkes, & D e l Boca, 1999; Kushner et al., 1994; Leonard & Blaine, 1988). Due to the biphasic effects o f alcohol, increased dopamine activity on the ascending limb o f the Blood Alcohol Curve ( B A C ) would likely increase B A S activity, whereas increased G A B A activity on the descending limb o f the B A C would likely lower BIS activity as well as inhibit the dopamine  146 system, leading to lower B A S activity (Conrod et a l , 2001; Earleywine, 1994a; Papineau et al., 1998). A n individual's experiences with alcohol would be stored in alcohol expectancy memory networks that are activated when an individual drinks, influencing the affective and behavioural response to alcohol (Goldman, 1999; Goldman et a l , 1999). A n individual's expectations with regard to the effects o f alcohol influence their motivations to use alcohol, which are then directly related to alcohol frequency and problems (Cooper et al., 1995). Recent research provides some insight into the link between information-processing biases and alcohol expectancies. Zack, Poulos, Fragopoulos, and M a c L e o d (2003) found negative mood phrases to prime for alcohol concepts, with individuals reporting drinking to regulate negative emotions more likely to show stronger priming o f alcohol concepts following negative cues. Stewart, H a l l , Wilkie, and B i r c h (2002) found that drinking motives result i n a similar priming affect. They presented coping motivated drinkers with positive (e.g., joyful), negative (e.g., afraid, rejection) and neutral words followed by alcohol (e.g., beer) and control words. Coping motivated drinkers showed greater priming for alcohol words than control words following negative primes, but did (  y  not differ between alcohol and control words following a neutral prime. Given that coping motivated drinkers also showed a stronger priming o f alcohol words following a positive prime, they may also associate alcohol with increases in positive affect. Unfortunately, the current research was not able to provide evidence for a relationship between depressive symptomatology and positive affect drinking on the C A R R O T . Future research w i l l need to use an experimental paradigm that activates distinct motivational and information-processing biases relevant to depression and anxiety i n order to demonstrate the attenuation o f these biases following alcohol cues or intoxication. Finally, the current status o f drinking motives needs to be re-addressed.  Blackwell  (2003; Blackwell et al., 2002) has begun along these lines by differentiating between depression motives and anxiety motives. Yet, this distinction does not address differential motivation for  147 alcohol arising from the multiple manifestations o f BIS and B A S activity. For example, depression motives could consist o f drinking to increase motivation or happiness, or coping with negative cognitions (i.e., rumination) and loss, whereas anxiety motives could consist o f drinking to attenuate hypervigilance to threatening or fearful stimuli, or coping with physiological arousal and negative social evaluation. The number o f drinking motives that would be needed in order to represent the large variety o f motivations for alcohol use indicates that the current line o f research regarding personality motivated drinking may be a more practical indicator o f individuals' reasons for drinking (see also Conrod et al., 2000). 4.6 Conclusion A number o f previous investigations have established a relationship between negative reinforcement drinking and both depression and anxiety. The current research has extended those findings by showing that each o f these predictors is able to account for independent variance in negative reinforcement drinking and in turn alcohol problems. Furthermore, depression and anxiety appear to be predictive o f distinct motives for alcohol use. Current and previous findings also indicate that individuals low on B A S and high in B I S are more likely to experience depressive symptomatology (Henriques & Davidson, 2000; Kasch et al., 2002), whereas those high in BIS only are more likely to be at risk for anxiety disorders (Johnson et al., 2003; Jorm et al., 1999). Moreover, individuals high on depressive symptomatology, hopelessness, and trait anxiety all evinced a decrease in punishment-susceptibility following an alcohol cue, indicating that such individuals may be drinking to decrease negative affect. The present investigation is in line with previous research suggesting that alcohol use is not a unitary phenomenon (Conrod et al., 2000; Cooper et al., 1995; P i h l & Peterson, 1995), but rather consists o f multiple pathways to addiction. Treatment o f substance disorders need to be just as variegated. Most treatment matching studies, attempting to match personality characteristics to generalized substance use treatment (e.g., Alcoholics Anonymous), have  148 yielded few significant results (e.g., Babor, Miller, DiClemente, & Longabaugh, 1999). Recent research focusing on personality features which predispose individuals to substance use and other psychopathology have had more success (Conrod et al., 2000; Stewart et al., 2002). Findings with respect to B I S and B A S activity and their relations with drinking motives and drinking problems may provide information necessary to further narrow the focus o f such interventions. A functional analysis o f individuals' drinking behaviour is necessary in order to apply an appropriate treatment strategy. Treating individuals for the negative affect associated with depressive symptomatology would consist o f addressing negative cognitions (Beck, Rush, Shaw, & Emery, 1979). O n the other hand, depressed individuals drinking to increased positive affect would likely benefit from behavioural activation (e.g., activity scheduling; Jacobson, Dobson, Truax, & Addis 1996; Lejuez, Hopko, & Hopko, 2001), whereas anxiety sensitive individuals drinking for conformity reasons would likely benefit from treatment o f social anxiety. Finally, higher scores on the B A S drive scale have been found to be predictive o f clinical improvement in depression, and B A S reward-responsiveness and B A S drive predictive o f future depression, even after controlling for previous depressive symptomatology (Kasch et al., 2002). B A S and BIS activity then appear relevant to treatment outcome o f depressed individuals. It would be interesting to see i f similar results are found in the treatment o f anxiety disorders. A s well, pre- and post-treatment levels o f BIS and B A S activity may be predictive o f alcohol treatment outcome.  149 Footnotes A stepwise regression analysis o f all seven B S I depression items was conducted on the  1  Vancouver sample predicting coping motives. Depression items explained 12.5% o f the variance in coping motives after 3 iterations, with worthlessness ((3 = .11), feelings easily hurt (p = .17), and hopelessness (P = .15) significant predictors i n the model, F(3, 302) = 13.65, p < .05. A n anxiety sensitivity/trait anxiety interaction term was created and regressed on the anxiety  2  motives residual, along with the other demographic and personality/symptom predictors, but it was not a significant predictor (P = -.67, p > .05). Regressing age, gender, sensation-seeking, anxiety sensitivity and trait anxiety on the anxiety motives residual, trait anxiety was the only significant predictor, thus replicating Comeau et al. (2001). However, unlike their findings, the moderator analysis did not show the anxiety sensitivity/trait anxiety interaction term to be a significant predictor. 3  In light o f Kambouropoulos and Staiger's (2001) finding that heavy alcohol users are  susceptible to the restorative effects o f reward-responsiveness from an alcohol cue, it was thought that the effect o f punishment sensitivity could be due to the presence o f binge drinkers (i.e., 4 or more drinks for females, 5 or more for males). Therefore, binge drinkers were removed from the sample (n = 8) and the analyses for hopelessness, BDI-II, and trait anxiety were re-examined. The two-way Hopelessness by Condition, BDI-II by Condition, and Trait Anxiety by Condition interactions remained significant, F(2, 81) = 3.46,p < .05, F ( l , 81) = 6.28, p < .05, and F(l, 81) = 3.97,p < .05 respectively. In order to attempt to replicate the findings o f Kambouropoulos and Staiger (2001), a third alcohol group was formed from regular drinkers who indicated binge drinking (n = 8). However, there were no main or interaction effects for reward-responsiveness, likely due to low power with so few binge drinkers in the current sample, but also differences in alcohol group selection criteria.  150 References Abraham, H . D . , & Fava, M . (1999). Order o f onset o f substance abuse and depression in a sample o f depressed outpatients. Comprehensive Psychiatry, 40, 44-50. Abramson, L . Y . , A l l o y , L . B . , Hogan, M . E . , Whitehouse, W . G . , Cornette, M . , Akhavan, S., & Chiara, A . (1998). Suicidality and cognitive vulnerability to depression among college students: A prospective study. Journal of Adolescence, 21, 473-487. Abramson, L . Y . , A l l o y , L . B . , Hogan, M . E . , Whitehouse, W . G . , Donovan, P., Rose, D . T., Panzarella, C , & Raniere, D . (1999). Cognitive vulnerability to depression: Theory and evidence. Journal of Cognitive Psychotherapy,  13, 5-20.  Abramson, L . Y . , Metalsky, G . I., & A l l o y , L . B . (1989). Hopelessness depression: A theorybased subtype o f depression. Psychological Review, 96, 358-372. A l - A d a w i , S., & Powell, J. (1997). The influence o f smoking on reward-responsiveness and cognitive functions: A natural experiment. Addiction, 92, 1773-1782. Alloy, L . B . , Abramson, L . Y . , Murray, L . A . , Whitehouse, W . G . , & Hogan, M . E . (1997). Selfreferent information-processing in individuals at high and low cognitive risk for depression. Cognition and Emotion, 11, 539-568. A l l o y , L . B . , Abramson, L . Y . , Whitehouse, W . G . , Hogan, M . E . , Tashman, N . A . , Steinberg, D . L . , Rose, D . T., & Donovan, P. (1999). Depressogenic cognitive styles: Predictive validity, information-processing and personality characteristics, and developmental origins. Behaviour Research and Therapy, 37, 503-531. Alloy, L . B . , Abramson, L . Y . , Hogan, M . E . , Whitehouse, W . G . , Rose, D . T., Robinson, M . S., K i m , R. S., & Lapkin, J. B . (1999). Temple-Wisconsin cognitive vulnerability to depression project: Lifetime history o f A x i s I psychopathology in individuals at high and low cognitive risk for depression. Journal of Abnormal Psychology,  109, 403-418.  A l l o y , L . D . , & Clements, C . M . (1998). Hopelessness theory o f depression: tests o f the symptom component. Cognitive Therapy and Research, 22, 303-335. Alpert, J. E , Maddocks, A . , Rosenbaum, J. F , & Fava, M . (1994). Childhood psychopathology retrospectively assessed among adults with early onset major depression. Journal of Affective Disorders, 31, 165 171 American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4 ed.). Washington, D C : Author. th  Anderson, J. C , & Gerbing, D . W . (1988). Structural equation modelling in practice: A review and recommended two-step approach. Psychological Bulletin, 103, 411-423. Arbuckle, J. L . , & Wothke, W . (1999). Amos 4.0 user's guide. Chicago, I L : SmallWaters Corporation. Arnett, P . A . , & Newman,' J.P. (2000). Gray's three-arousal model: A n empirical investigation. Personality and Individual Differences, 28, 1171-1189. Babor, T. F., M i l l e r , W . R., DiClemente, C , & Longabaugh, R. (1999). A study to remember: Response o f the project M A T C H research group. Addiction, 94, 66-69. Balfour, D . J K . , & Ridley, D . L . (2000). The effects o f nicotine on neural pathways implicated in depression: A factor i n nicotine addiction? Pharmacology,  Biochemistry and Behavior,  66, 79-85. Balon, R. (1987). Suicide: Can we predict it? Comprehensive Psychiatry, 28, 236-241. Baron, R. M . , & Kenny, D . A . (1986). The moderator-mediator variable distinction i n social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Barrett, P. T., Petrides, K . V . , Eysenck, S. B . G . , & Eysenck, H . J. (1998). The Eysenck Personality Questionnaire: A n examination o f the factorial similarity o f P, E , N , and L across 34 countries. Personality and Individual Differences, 25, 805-819.  152 Beck, A . T. (1987). Cognitive models o f depression. Journal of Cognitive Psychotherapy,  1, 5-  37. Beck, A . T., & Clark, D . A . (1997). A n information-processing model o f anxiety: Automatic and strategic processing. Behavior Research and Therapy, 35, 49-58. Beck, A . T., Riskind, J. H . , Brown, G . K . , & Steer, R. A . (1988). Levels o f hopelessness in D S M - m disorders: A partial test o f content specificity in depression. Cognitive Therapy and Research, 12, 459-469. Beck, A . T., Rush, A . J., Shaw, B . F., & Emery, G . (1979). Cognitive therapy for  depression.  N e w Y o r k : Guilford. Beck, A . T., Steer, R. A . , & Brown, G . K . (1996). Beck Depression Inventory manual (2nd ed.). San Antonio, T X : Psychological Corporation. Beck, A . T., Weissman, A . , Lester, L . , & Trexler, L . (1974). The measurement o f pessimism: The Hopelessness Scale. Journal of Consulting and Clinical Psychology, 42, 861-865. Beevers, C . G . , & Meyer, B . (2002). Lack o f positive experiences and positive expectancies mediate the relationship between B A S responsiveness and depression. Cognition and Emotion, 16, 549-564. Bertakis, K . D . , Helms, J. H . , Callahan, E . J., Azari, R., Leigh, P., & Robbins, J. A . (2001). Patient gender differences in the diagnosis o f depression in primary care. Journal Of Women's Health and Gender-Based Medicine, 10, 689-698. Bieling, P. J., Antony, M . M . , & Swinson, R. P. (1998). The State-Trait Anxiety Inventory, Trait version: Structure and content re-examined. Behaviour Research and Therapy, 36, 777788. Birmaher, B . , Ryan, N . D . , Williamson, D . E . Dougles, E . Brent, D . A . , Kaufman, J., Dahl, R E . , Perel, J., & Nelson, B . (1996). Childhood and adolescent depression: A review o f the  153 past 10 years. Part I. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 1427-1439. Blackwell, E . (2003). A Five-Dimensional  Measure of Drinking Motives. Unpublished  manuscript, University o f British Columbia. Blackwell, E . , Conrod, P. J., & Hansen, N . M . (2002). Introversion/hopelessness,  negative  cognitions, and drinking motives. Paper presented at the 2002 meeting o f the Society for Research on Alcoholism, San Francisco, C A . Boyle, M . H . , Offord, D . R., Racine, Y . A . , Fleming, J. E . , Szatmari, P., & Links, P. S. (1993). Predicting substance use in early adolescence based on parent and teacher assessment o f childhood psychiatric disorder: Results from the Ontario C h i l d Health Study follow-up. Journal of Child Psychology and Psychiatry and Allied Disciplines,  34, 535-544.  Boyle, M . H . , Offord, D . R., Racine, Y . A . , Szatmari, P., Fleming, J. E . , & Links, P. S. (1992). Predicting substance use i n late adolescence: Results from the Ontario C h i l d Health Study follow-up. American Journal of Psychiatry, 149, 761-767'. Bradizza, C . M . , Reifman, A . , & Barnes, G . M . (1999). Social and coping reasons for drinking: predicting alcohol misuse i n adolescents. Journal of Studies on Alcohol, 60, 491-499. Bradley, B . P., M o g g , K . , Falla, S. J., Hamilton, L . R.. (1998). Attentional bias for threatening facial expressions in anxiety: manipulation o f stimulus duration. Cognition and Emotion, 12, 789-808. Brook, J. S., Cohen, P., & Brook, D . W . (1998). Longitudinal study o f co-occurring psychiatric disorders and substance use. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 322-330. Brown, S. A . , Christiansen, B . A . , & Goldman, M . S. (1987). The A l c o h o l Expectancy Questionnaire: A n instrument for the assessment o f adolescent and adult alcohol expectancies. Journal of Studies on Alcohol, 48, 483-491.  Brown, S. A . , Goldman, M . S., Inn, A . , & Anderson, L . (1980). Expectations o f reinforcement from alcohol: Their domain and relation to drinking patterns. Journal of Consulting and Clinical Psychology, 48, 419-426. Brunelle, C , Assaad, J., Barrett, S. P., A v i l a , C , Conrod, P. J., Tremblay, R. E . , & Pihl, R. (2003). Heightened Heart Rate Response to A l c o h o l Intoxication is Associated with a Reward-Seeking Personality Profile. Manuscript i n preparation. Buchwald, A . M . (1977). Depressive mood and estimates o f reinforcement frequency.  Journal  of Abnormal Psychology, 86, 443-446. Carey, K . B . , & Correia, D . J. (1997). Drinking motives predict alcohol-related problems in college students. Journal of Studies on Alcohol, 58, 100-105. Carter, B . L . , & Tiffany, S. T. (1999). Meta-analysis o f cue-reactivity i n addiction research. Addiction, 94, 327-340. Carver, C . S., & White, T. L . (1994). Behavioral inhibition, behavioral activation, and affective responses to impending reward and punishment: The B I S / B A S scales. Journal of Personality and Social Psychology, 67, 319-333. Chou, C , & Bentler, P. M . (1995). Estimates and tests a structural equation modeling. In R. H . Hoyle (Ed.), Structural equation modeling: concepts, issues, and applications,  (pp. 37-  55). Thousand Oaks, C A : Sage. Clark, D . M . , & Salkovskis, P. M.(1997). Misinterpretation o f body sensations i n panic disorder. Journal of Consulting and Clinical Psychology, 65, 203-213. Clark, L . A . , & Watson, D . (1988). M o o d and the mundane: Relations between daily life events and self-reported mood. Journal of Personality and Social Psychology, 54, 296-308. Clark, L . A . , & Watson, D . (1991). Tripartite model o f anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology,  100, 316-336.  155 Cloninger, C . R. (1987). Neurogenetic adaptive mechanisms i n alcoholism. Science, 236, 410416. Comeau, N . , Stewart, S.H., & Loba, P. (2001). The relations o f trait anxiety, anxiety sensitivity, and sensation-seeking to adolescents' motivations for alcohol, cigarette, and marijuana use. Addictive Behaviors, 26, 803-825. Conrod, P. J., Peterson, J. B . , Pihl, R. O., & Mankowski, S. (1997). Biphasic effects of alcohol on heart rate are influenced by alcoholic family history and rate o f alcohol ingestion. Alcoholism:  Clinical and Experimental Research, 21, 140-149.  Conrod, P. J., Pihl, R . O., Stewart, S. H . , & Dongier, M . (2000). Validation o f a system o f classifying female substance abusers on the basis o f personality and motivational risk factors for substance abuse. Psychology of Addictive Behaviors, 14, 243-256. Conrod, P., Pihl, R . O , & Vassileva, J. (1998). Differential sensitivity to alcohol reinforcement in groups o f men at risk for distinct alcoholic syndromes. Alcoholism:  Clinical  Experimental Research, 22, 585-597.  x  and  Conrod, P.J., Peterson, J.B., Pihl, R . O . (1997). Disinhibited personality and sensitivity to alcohol reinforcement: Independent correlates o f drinking behavior i n sons o f alcoholics. Alcoholism:  Clinical and Experimental Research, 21, 1320-1332.  Conrod, P.J., Pihl, R. O., & Vassileva, J. (1998). Differential sensitivity to alcohol reinforcement in groups o f men at risk for distinct alcoholism subtypes. Alcoholism:  Clinical  and  Experimental Research, 22, 585-597. Conrod, P.J., Pihl, R . O . , Stewart, S.H., & Dongier, M . (2000). Validation o f a system o f classifying female substance abusers based on personality and motivational risk factors for substance abuse. Psychology of Addictive Behaviors, 14, 243-256.  Conrod, P.J., Stewart, S. H . , & Pihl, R. O. (2002). Anxiety sensitivity, sensation-seeking  introversion/hopelessness,  an impulsivity: different patterns of reinforcement sensitivity and drug  abuse susceptibility.  Unpublished manuscript, University o f British Columbia.  Conrod, P.J., Stewart, S.H., Pihl, R . O . , Cote, S., Fontaine, V . , & Dongier, M . (2000). Efficacy o f brief coping skills interventions that match different personality profiles o f female substance abusers. Psychology of Addictive Behaviors, 14, 231-242. Cooper, M . L . (1994). Motivations for alcohol use among adolescents: Development and validation o f a four-factor model. Psychological Assessment, 6, 117-126. Cooper, M . L . , Russell, M . , Skinner, J. P., & Windle, M . (1992). Development and validation o f a three-dimensional measure o f drinking motives. Psychological Assessment, 4, 123-132. Cooper, M . L . , Frone, M . R . , Russell, M . , & Mudar, P. (1995). Drinking to regulate positive and negative emotions: A motivational model o f alcohol use. Journal of Personality and Social Psychology,  69, 990-1005.  Costello, C . G . (1972). Depression: Loss o f reinforcers or loss o f reinforcer effectiveness? Behaviour Therapy, 3, 210-247'. David, J. P., Green, P. J., Martin, R., & Suls, J. (1997). Differential roles o f neuroticism, extraversion, and event desirability for mood i n daily life: A n integrative model o f top down and bottom up influences. Journal of Personality and Social Psychology,  73, 149-  159. Davidson, R. J. (1992). Anterior cerebral asymmetry i n the nature o f emotion. Brain and Cognition, 20, 125-151. Davidson, R. J., Ekman, P., Saron, C , Senulis, J., & Friesen, W . V . (1990). Approach withdrawal and cerebral asymmetry: Emotional expression and brain physiology: I. Journal of Personality and Social Psychology, 58, 330-341.  Davidson, R . J . (2001). Toward a biology o f personality and emotion. In A . R . Damasio & A . Harrington (Eds.), Annals of the New York Academy of Sciences:  Vol. 935. Unity of  knowledge: The convergence of natural and human science (pp. 191-207). N Y : N e w Y o r k Academy o f Sciences. Depue, R. A . , & Iacono, W . G . (1989). Neurobehavioral aspects o f affective disorders. Annual Review of Psychology,  40, 457-492.  Depue, R . A . , Luciana, M . , Arbis, P., Collins, P., & Leon, A . (1994). Dopamine and the structure of personality: Relation o f agonist-induced dopamine activity to positive emotionality. Journal of Personality and Social Psychology, 67, 485-498. Derogatis, L . R. (1993). Brief Symptom Inventory (BSI). Administration,  scoring, and procedures  manual (3rd ed.). Minneapolis, M N : National Computer Systems, Inc. Derogatis, L . R., & Melisaratos, N . (1983). The brief symptom inventory: an introductory report. Psychological  Medicine, 13, 595-605.  Deshmukh, A . , Rosenbloom, M . J., Sassoon, S., O'Reilly, A . , Pfefferbaum, A . , & Sullivan. E . V . (20003) Alcoholic men endorse more D S M - I V withdrawal symptoms than alcoholic women matched in drinking history. Journal o f Studies on A l c o h o l , 64, 375-380. D i Chiara, G . , Acquas, E . , & Tanda, G . (1996). Ethanol as a neurochemical circuit o f conventional reinforcers: the dopamine-opioid link. Alcohol, 13, 13-17. Diaz, A . , & Pickering, A . D . (1993). The relationship between Gray's and Eysenck's personality spaces. Personality and Individual Difference,  7,81-88.  Dozois, D . J. A . , Dobson, K . S., & Ahnberg, J. L . (1998). A psychometric evaluation o f the Beck Depression Inventory-II. Psychological Assessment, 10, 83-89. Dunbar, G . C , & Lishman, W . A . (1984). Depression, recognition-memory and hedonic tone: a signal detection analysis. British Journal of Psychiatry, 144, 376-382.  158 Earleywine, M . (1984). Anticipated biphasic effects o f alcohol vary with risk for alcoholism: A preliminary report. Alcoholism:  Clinical and Experimental Research, 18, 711-714.  Eke, M . , & M c N a l l y , R . J. (1996): Anxiety sensitivity, suffocation fear trait anxiety, and breathholding duration as predictors o f response to carbon dioxide challenge. Behaviour Research and Therapy, 34, 603-607. Erikson, E . H . (1959). Identity and the life cycle. Psychological  Issues, 1, 53-177.  Eysenck, S. B . , Eysenck, H . J., & Barrett, P. (1985). A revised version o f the Psychoticism scale. Personality and Individual Differences, 6, 21-29. First, M . B . , Spitzer, R. L , Gibbon M . , & Williams, J. B . W . (1996). Structured Clinical Interview for D S M - I V A x i s I Disorders, Clinician Version ( S C J D - C V ) . Washington, D C : American Psychiatric Press. Franken, I. H . A . (2002). Behavioural approach system ( B A S ) sensitivity predicts alcohol craving. Personality and Individual Differences, 32, 349-355. Gable, S. L . , Reis, H . T , & Elliot, A . J. (2000). Behavioral Activation and Inhibition in ;  Everyday Life. Journal of Personality and Social Psychology,  78, 1135-1149.  Gencoz, T. (2002). Discriminate validity o f low positive affect: is it specific depression? Personality and Individual Differences, 32, 991-999. Gianoulakis, C , de Waele, J., & Thavundayil, J. (1996). Implications o f the endogenous opioid system i n excess o f ethanol consumption. Alcohol, 13, 19-23. Gilman, S. E . , & Abraham,'H. D . (2001). A longitudinal study o f the order o f onset o f alcohol dependence and major depression. Drug and Alcohol Dependence, 63, 277-286. Goldman, M . S., D e l Boca, F. K . , & Darkes, J (1999). A l c o h o l expectancy theory: The application o f cognitive neuroscience. In K . E . Leonard & H . T. Blane (Eds.), Psychological  theories of drinking and alcoholism (2nd ed.). N e w Y o r k : Guilford Press.  159 Goldman, M . S., Greenbaum, P. E . , & Darkes, J. (1997). A confirmatory test o f hierarchical expectancy structure and predictive power: Discriminant validation o f the alcohol expectancy questionnaire. Psychological Assessment, 9, 145-157. Goldman, M . S . (1999). Expectancy operation: Cognitive and neural models and architectures. In I. Kirsch (Ed.), How expectancies shape experience (pp. 41-63). Washington, D C : A P A Books. Goldman, M . S . , Darkes, J., & D e l Boca, F . K . (1999). Expectancy mediation o f biopsychosocial risk for alcohol use and alcoholism. In I. Kirsch (Ed.), How expectancies shape experience (pp. 233-262). Washington, D C : A P A Books. Gomez, R., & McLaren, S. (1997). The effects o f reward and punishment on response disinhibition, moods, heart rate, and skin conductance level during instrumental learning. Personality and Individual Differences, 23, 305-316. Gotlib, I. FL, & M c C a n n , C . D . (1984). Construct accessibility and depression: A n examination o f cognitive and affective factors. Journal of Personality and Social Psychology, 47, 427439. Gotlib, I. H . , Ranganath C , & Rosenfeld, J. P. (1998). E E G alpha asymmetry, depression, and cognitive functioning. Cognition and Emotion, 12, 449-478. Grant, B . F. (1995). Comorbidity between D S M I V drug use disorders and major depression: Results o f a national survey o f adults. Journal of Substance Abuse, 7, 481-497. Grant, B.S., & Harford, T. C . (1995). Comorbidity between D S M - I V alcohol use disorders in major depression: Results o f a national survey. Drug and Alcohol Dependents, 39, 197206. Grau, E . , & Ortet, G . (1989). Personality traits and alcohol consumption in a sample of nonalcoholic women. Personality and Individual Differences, 27, 1057-1066.  160 Gray, J. A . (1970). The psychophysiological basis o f introversion-extraversion. Behavior Research and Therapy, 8, 249-266. Gray, J. A . (1982). The neuropsychology  of anxiety: An enquiry into the functions of the septo-  hippocampal system. Oxford: Oxford University Press. Gray, J. A . (1987). The psychology of fear and stress. N e w Y o r k : Cambridge University Press. Gray, J. A . (1988). Anxiety and personality. But the In M . Roth, R. Noyles Jr., & G . D . Burrows, (Eds.), Handbook of anxiety: Vol. 1. Biological, clinical, and cultural  perspectives.  Amsterdam: Elsevier Science Publishers. Gray, J. A . (1990). Brain systems to mediate both emotion and cognition. Cognition and Emotion, 4, 269-288. Gray, J. A . (1991). The neuropsychology o f temperament. In J. Strelau & A . Angleitner (Eds.), Explorations  of temperament: International perspectives on theory and measurement (pp.  105-128). N e w Y o r k : Plenum. Gray, J. A . (1994). Framework for a taxonomy o f psychiatric disorder. In S. H . M . van Goozen, N . E . V a n de P o l l , & Sergeant, J. A . (Eds.), Emotions: Essays on emotion theory, (pp. 2959). Hillsdale, N J : Erlbaum. Gray, J. A . , & MacNaughton, N . (2000). The neuropsychology  of anxiety. N e w York: Oxford  University Press. Hagemann, D . , Naumann, E . , Becker, G., Maier, S., & Bartussek, D . (1998). Frontal brain asymmetry and affective style: A conceptual replication. Psychophysiology,  35, 372-388.  Hall, R. C . W . , Piatt, D . E . , & H a l l , R. C . W . (1999) Suicide risk assessment: A review o f risk factors for suicide i n 100 patients who made severe suicide attempts: Evaluation o f suicide risk in a time o f managed care! Psychosomatics,  40, 18-27  Hansen, N . M . , Blackwell, E . , & Conrod, P. J. (2002, June). Personality and drinking motives: a cross-cultural  comparison. Paper presented at the meeting o f the Society for Research on  Alcoholism, San Francisco, C A . Harmon-Jones, E . , & A l l e n , J. J. B . (1997). Behavioral activation sensitivity and resting frontal E E G asymmetry: Covariation o f putative indicators related to risk for mood disorders. Journal of Abnormal Psychology, 106, 159-163. Harmon-Jones, E . , Abramson, L . Y . , Sigelman, J., Bohlig, A . , Hogan, M . E . , & Harmon-Jones, C. (2002). Proneness to hypomania/mania symptoms or depression symptoms and asymmetrical frontal cortical responses to an anger-evoking event. Journal of Personality and Social Psychology,  52,610-618.  Helzer, J. E . , & Pryzbeck, T. R . (1988). The co-occurrence o f alcoholism with other psychiatric disorders in the general population and its impact on treatment. Journal of Studies on Alcohol, 49, 219-224. Henriques, J. B . , & Davidson, R. J. (1990). Regional brain electrical asymmetries discriminate between previously depressed and healthy control subjects. Journal of Abnormal Psychology, 99, 22-31. Henriques, J. B . , & Davidson, R. J. (1991). Left frontal hypoactivation in depression. Journal of Abnormal Psychology,  100, 535-545.  Henriques, J. B . , & Davidson, R. J. (1997). Brain electrical asymmetries during cognitive task performance in depressed and nondepressed subjects. Biological Psychiatry, 42, 10391050. Henriques, J. B . , & Davidson, R. J. (2000). Decreased responsiveness to reward in depression. Cognition and Emotion, 14, 711-724. Henriques, J. B . , Glowacki, J. M . , & Davidson, R. J. (1994). Reward fails to alter response bias in depression. Journal of Abnormal Psychology, 103, 460-466.  Heubeck, B . G , Wilkinson, R. B , & Cologon, J. (1998). A second look at Carver and White's (1994) B I S / B A S scales. Personality and Individual Differences, 25, 785-800. H u , L . , & Bentler, P. M . (1995). Evaluating model fit. In R. H . Hoyle (Ed.), Structural equation modeling: Concepts, issues, and applications,  (pp. 76-99). Thousand Oaks, C A : Sage.  H u , L . , & Bentler, P. M . (1999). Cut-off criteria for fit indiexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modelling, 6, 1-55. Iannoti, R. J., Bush, P. J., & Weinfurt, K . P. (1996). Perception o f friends' use o f alcohol, cigarettes, and marijuana among urban schoolchildren: A longitudinal analysis. Addictive Behaviors, 21, 615-632. Johnson, J. L . , Bottorff, J. L . , Moffat, B . , Ratner, P. A . , Shoveller, J. A . , & Lovato, C . (2003). Tobacco dependence: Adolescents' perspectives on the need to smoke. Social Science andMedicine,  56, 1481-1492.  Jacobson, N . S., Dobson, K . S., Truax, P. A . , Addis, M . E . , Koerner, K . , Gollan, J. K . , Gortner, E., & Prince, S. (1996). A component analysis o f cognitive-behavioral treatment for depression. Journal of Consulting and Clinical Psychology, 64, 295-304 Johnson, S. L . , Sandrow, D . , Meyer, B . , Winters, R., M i l l e r , I., Solomon, D . , & Keitner, G . (2000). Increases in manic symptoms after life events involving goal attainment. Journal of Abnormal Psychology,  109, 721-727.  Johnson, S. L . , Turner, R. J., Iwata, N . (2003). B I S / B A S levels and psychiatric disorder: A n epidemiological study. Journal of Psychopathology  and Behavioural Assessment, 25, 25-  36. Jorm, A . F., Christensen, H . , Henderson, A . S., Jacomb, P. A . , Korten, A . E , & Rodgers, B . (1999). U s i n g the B I S / B A S scales to measure behavioural inhibition and behavioural activation: Factor structure, validity and norms in a large community sample. Personality and Individual Differences, 26, 49 58.  Kairouz, S., Gliksman, L . , Demiers, A . , & Adlaf, E . M . (2002). For all these reasons, I do... drink: A multilevel analysis o f contextual reasons for drinking among Canadian undergraduates. Journal of Studies on Alcohol, 63, 600-608. Kambouropoulos, N . , & Staiger, P . K . (2001). The influence o f sensitivity to reward on reactivity to alcohol-related cues. Addiction, 96, 1175-1185. Kaminer, Y . (1999). Addictive disorders in adolescents. Addictive Disorders, 22, 275-287'. Kandel, D . B . , Huang F. Y . , & Davies, M . (2001). Comorbidity between patterns o f substance use dependence and psychiatric syndromes. Drug and Alcohol Dependence, 64, 233-241. Kaplan, D . (2000). Structural equation modeling: Foundations and extensions. Thousand Oaks, Thousand Oaks, C A : Sage. Kasch, K . L . , Rottenberg, J., Arnow, B . A . , & Gotlib, I. H . (2002). Behavioral activation and inhibition systems and the severity and course o f depression. Journal of Abnormal Psychology, 111, 589-597. Kashani, J. H . , Keller, M . B . , Solomon, N . , Reid, J. C , & Mazzola, D . (1985). Double depression i n adolescent substance users. Journal of Affective Disorders, 8, 153-157. Kassel, J. D . , Jackson, S. I., & Unrod, M . (2000). Generalized expectancies for negative mood regulation and problem drinking among college students. Journal of Studies on Alcohol, 61, 332-340. Kessler, R . C , McGonagle, K . A . , Swartz, M . , Blazer, D . G . , Nelson, C . B . , 1993. Sex and depression in the National Comorbidity Survey. I: lifetime prevalence, chronicity, and recurrence. Journal o f Affective. Disorders, 29, 85-96. Kessler, R . C , Nelson, C . B . , McGonagle, K . A . , Edlund, M . J., Frank, R. G . , & Leaf, P. J. (1996). The epidemiology o f co-occurring addictive and mental disorders: Implications for prevention and service utilization. American Journal of Orthopsychiatry,  66, 17-31.  K i n g , A . C , Bernardy, N . C , Hauner, K . (2003). Stressful events, personality, and mood disturbance: Gender differences i n alcoholics and problem drinkers. Addictive  Behaviors,  28, 171-187. K l e i n , D . N . , Schwartz, J. E . , Rose, S., & Leader, J. B . (2000). Five-year course and outcome o f dysthymic disorder: A prospective, naturalistic follow-up study. American Journal of Psychiatry, 157, 931 939. Kraus, D . , Smith, G . T., & Ratner, H . H . (1994). Modifying alcohol-related expectancies in grade-school children. Journal of Studies on Alcohol, 55, 535-542. Kushner, M . G . , Sher, K . J., Wood, M . , & Wood, P. K . (1994). Anxiety and drinking behavior: Moderating effects o f tension-reduction expectancies. Alcoholism:  Clinical  and  Experimental Research, 18, 852-860. Kushner, M . G . , Thuras, P., Abrams, K . , Brekke, M . , & Stritar, L . (2001). Anxiety mediates the association between study sensitivity and coping related drinking motives in alcoholism treatment patients. Addictive Behaviours, 26, 869-885. Larsen, R. J., & Ketelaar, T. (1989). Extraversion, neuroticism and susceptibility to positive and negative mood induction procedures. Personality and Individual Differences,  10, 1221-  1228. Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to positive and negative emotional states. Journal of Personality and Social Psychology,  61, 132-140.  Ledoux, J. E . (1994). Emotion, memory and brain. Scientific American, 270, 50-59. Lee N . K . , Greely J., Oei T. P. (1999). The relationship o f positive and negative alcohol expectancies to patterns o f consumption o f alcohol in social drinkers. Addictive Behaviors, 24, 359-369.  165 Leigh, B . C , & Stacy, A . W . (1993). A l c o h o l outcome expectancies: Scale construction and predictive utility i n higher order confirmatory models. Psychological  Assessment, 5, 216-  229. Lejuez, C . W . , Hopko, D . R., & Hopko, S. D . (2001). A brief behavioural activation treatment for depression. Behaviour Modification, 25, 255-286. Leonard, K . E . , & Blane, H . T. (Eds.). (1999). Psychological  theories of drinking and alcoholism  (2nd ed.). N e w Y o r k : Guilford Press. Leone, L . , Perugini, M . , Bagozzi, R. P, Pierro, A . , & Mannetti, L . (2001). Construct validity and generalizability o f the carver white behavioural inhibition system/behavioural activation system scales. European Journal of Personality, 15, 373-390. Lewinsohn, P. M . , Lobitz, W . C , & Wilson, S. (1973). "Sensitivity" o f depressed individuals to aversive o f stimuli. Journal of Abnormal Psychology,  57,259-263.  Light, J . M . , Grube, J. W . , Madden, P. A . , & Gover, J. (2003). Adolescent alcohol use and suicidal ideation: A nonrecursive model. Addictive Behaviors, 28, 705-724. Lilienfeld, S. O., Turner, S. M . , & Jacob, R. G . (1993). Anxiety sensitivity: A n examination o f theoretical and methodological issues. Advances in Behaviour Research and Therapy, 15, 147-183. Lolas, F., Gomez, A . , & Suarez, L . (1991). E P Q - R and suicide attempt: The relevance o f psychoticism. Personality and Individual Differences, 12, 199-902. Loxton, N . J., Dawe, S. (2001). Alcohol abuse and dysfunctional eating i n adolescent girls: the influence o f individual differences in sensitivity to reward and punishment.  Alcohol  Abuse and Eating Disorders, 29, 455-462. MacAndrew, C , & Steele, T. (1991). Gray's behavioral inhibition system: a psychometric evaluation. Personality and Individual Differences, 212, 157-171.  166 M a c K i n n o n , D . P., Warsi, G . , & Dwyer, J. H . (1995). A simulation study o f mediated effect measures. Multivariate Behavioral Research, 30, 41-62. MacLeod, A . K . , & Byrne, A . (1996). Anxiety, depression, and the anticipation o f future positive and negative experiences. Journal of Abnormal Psychology,  105, 286-289.  MacLeod, A . K . , & Salaminiou, E . (2001). Reduced positive future-thinking in depression: Cognitive and affective factors. Cognition and Emotion, 15, 99-107. MacLeod, A . K . , Pankhania, B . , Lee, M . , & Mitchell, D . (1997). Parasuicide, depression and the anticipation o f positive and negative future expectancies. Psychological  Medicine, 27,  973-977. MacLeod, A . K . , Tata, P., Kentish, J., & Jacobson, H (1997). Retrospective and prospective cognitions i n anxiety and depression. Cognition and Emotion, 11, 467-479. MacLeod, C , & Mathews, S. (1991). Cognitive-experimental approaches to the emotional disorders. In P. Martin (Ed.), Handbook of behaviour therapy and psychological  science:  An integrative approach (pp. 116-150). N e w Y o r k : Pergamon Press. Mathews, S., & M a c L e o d , C . (2002). Induced processing biases have causal effects on anxiety. Cognition and Emotion, 16, 331-354. McCreary, D . R., & Sadava, S. W . (2000). Stress, alcohol use and alcohol-related problems: the influence o f negative and positive affect in two cohorts o f young adults. Journal of Studies on Alcohol, 61, 466-474. M c K a y , D . , & Schare, M . L . (1999). The effects o f alcohol and alcohol expectancies on subjective reports and physiological reactivity: A meta-analysis. Addictive  Behaviours,  24, 633-647. M c K e e , S. A . , Hinson, R . E . , W a l l , A . M . , & Spriel, P. (1998). A l c o h o l outcome expectancies and coping styles as predictors o f alcohol use i n young adults. Addictive Behaviors, 23, 17-22.  167 M c N a l l y , R. J. (2002). Anxiety sensitivity and panic disorder. Biological Psychiatry, 52, 938946. Mehrabian, A . (2001). General relations among drug use, alcohol use, and major indexes o f psychopathology. Journal of Psychology,  135,71-86.  Mehrabian, A . , & Bernath, M . S. (1991). Factorial composition o f commonly used self-report depression inventories: Relationships with basic dimensions o f temperament. Journal of Research in Personality, 25, 262-275. Merikagas, K . R., Mehta, R . L . , Molnar, B . E . , Walters, E . E . , Swendsen, J. D . , Aguilar-Gaziola, S. B i j l , R., Borges, G . , Caraveo-Anduaga, J. J., Dewit, D . J., Kolody, B . , Vega, W . A . , Wittchen, H . , & Kessler, R. C . (1998). Comorbidity o f substance use disorders with mood and anxiety disorders: Results o f the international consortium in psychiatric epidemiology. Addictive Behaviors, 23, 893-907. Meyer, B . , Johnson, S. L . , & Carver, C . S. (1999). Exploring behavioral activation and inhibition sensitivities among college students at risk for bipolar spectrum symptomatology. Journal of Psychopathology  and Behavioral Assessment, 21, 275-292.  Miller, A . , & Tomarken, A . J. (2001). Task-dependent changes in frontal brain asymmetry: effects o f incentive cues, outcome expectancies, and motor response.  Psychophysiology,  38, 500-511. Morely, S., Williams, A . C . de C , & Black, S. (2002). A confirmatory factor analysis o f the Back Depression Inventory in chronic pain. Pain, 99, 289-298. Muris, P., Merckelbach, H . , Ollendick, T., K i n g , N . , & Bogie, N . (2002). Three traditional and three new childhood anxiety questionnaires: Their reliability and validity in a normal adolescent sample. Behaviour Research and Therapy, 40, 753-772.  Muris, P., Schmidt, H., Merckelbach, FL, Schouten, E. (2001). Anxiety sensitivity in adolescents: factor structure and relationships to trait anxiety and symptoms of anxiety disorders and depression. Behaviour Research and Therapy, 39, 89-100.  Nelson, D. V., & Novy, D. M . (1997). Self-report differentiation of anxiety and depression in chronic pain. Journal of Personality Assessment, 69, 392-407. Nishith, P., Resick, P. A., & Mueser, K. T. (2001) areas sleek difficulties and alcohol use motives in female rape victims with postretirement stress disorder. Journal of Traumatic Stress, 14, 469-479. Noelen-Hoeksema, 1990, Sex differences in depression. Stanford, C A : Stanford University Press. Nunes, E. V., McGrath, P. J., Quitkin, F. M . , Ocepek-Wlikson, K., Stewart, J. W., Koenig, T., Wager, S.,'& Klein, D. F. (1995). Imipramine treatment of cocaine abuse: possible boundaries of efficacy. Drug and Alcohol Dependence, 39, 185-195. Oliver, J. M . , & Burkham, R. (1979). Depression in university students: duration, relation calendar time, prevalence, and demographic correlates. Journal of Abnormal Psychology, 88, 667-770. O'Malley, P. M . , Johnston, L. D., & Bachman, J. G. (1998). Alcohol use among adolescents. Alcohol Health and Research World, 22, 85-93.  Overholser, J. C , Freiheit, S. R., & DiFilippo, J. M . (1997). Emotional distress and substance abuse as risk factors for suicide attempts. Canadian Journal of Psychiatry, 42, 402-408. Papineau, K. L., Roehrs, T. A., Petrucelli, N . R., Leon, D., & Roth, T. (1998). Electrophysiological assessment (The Multiple Sleep Latency Test) of the biphasic effects of ethanol in humans. Alcoholism: Clinical and Experimental  235.  Research, 22, 231 -  Park, J. Y . , Huang, Y . H . , Nagoshi, C . T., Yuen, S., Johnson, R . C , Ching, C . A . , & Bowman, K . S. (1984). The flushing response to alcohol use among Koreans and Taiwanese. Journal of Studies on Alcohol, 45, 481-485. Peirce, R. S., Frone, M . R., Russell, M . , & Cooper (1994). Relationship o f financial strain and psychosocial resources to alcohol use and abuse: The mediating role o f negative affect and drinking motives. Journal of Health and Social Behavior, 35, 291-308. Peterson, R. A . , & Reiss, S. (1992). The Anxiety Sensitivity Index manual (2nd ed). Worthington, O H : International Diagnostic Systems. Petrakis, I., Carroll, K . JVL, N i c h , C , Gordon, L . , Kosten, T., & Rounsaville, D . (1998). Fluoxetine treatment o f depressive disorders in methadone-maintained opioid addicts. Drug and Alcohol Dependence, 50, 221-226. Pickering, A . D , Corr, P. J., & Gray, J. A . (1997). Interactions and reinforcement sensitivity theory: A theoretical analysis o f Rusting and Larsen. Personality and Individual Differences, 26, 357-365. Pickering, A . D . , & Gray, J. A . (1999). The neuroscience o f personality.-In L . A . Pervin & O.P. John (Eds.), Handbook of personality:  Theory and research (2nd ed., pp. 277-299). N e w  York: Guilford Press. Pickering, A . D . , Corr, P. J., Powell, J. H . , Kumari, V . , Thornton, J. C , & Gray, J. A . (1997). Individual differences i n reactions to reinforcing stimuli are neither black nor white: To what extent are they Gray? In H . Nyborg (Ed.), The scientific study of human nature: Tribute to Hans J. Eysenck at eighty (pp. 36-67). N e w Y o r k : Elsevier Science. Pickering, A . D . , Diaz, A . , & Gray, J. A . (1995). Personality and reinforcement: A n exploration using a maze-learning task. Personality and Individual Differences,  18, 541-558.  Pihl, R. O., & Peterson, J. B . (1995). Alcoholism: The role o f difference motivational systems. Journal of Psychiatry and Neuroscience,  20, 372-396.  Powell, J. H . , A l - A d a w i , S., Morgan, J., & Greenwood, R. J. (1996). Motivational deficits after brain injury: Effects o f bromocriptine i n 11 patients. Journal of Neurology, Neurosurgery  and Psychiatry, 60, 416-421.  Powell, J., Dawkins, L . , & Davis, K . E . (2002). Smoking, reward-responsiveness, and response inhibition: Tests o f an incentive motivation model. Biological Psychiatry, 51, 151-163. Rapee, R. M . , Brown, T. A . , Antony, M . M . , & Barlow, D . H . (1992). Response to hyperventilation and inhalation o f 5.5% carbon dioxide-enriched air across the DSM-IIIR anxiety disorders. Journal of Abnormal Psychology, 101, 538-552. Reid, L . D . (1996). Endogenous opioids and alcohol dependence: Opioids alkaloids and the propensity to drink alcoholic beverages. Alcohol, 13, 5-11. Reiss, S., Peterson, R . A . , Gursky, D . M . , & M c N a l l y , R . J. (1986). Anxiety sensitivity, anxiety frequency, and the prediction o f fearfulness. Behaviour Research and Therapy, 24, 1-8. Roth, P. (1994). M i s s i n g data: A conceptual review for applied psychologists. Personnel Psychology, 47, 537-560. Rusting, C . L . , & Larsen, R. J. (1997). Extraversion, neuroticism, and susceptibility to positive and negative affect: A test o f two theoretical models. Personality and Individual Differences, 22, 607-612. Saklofske, D . FL, K e l l y , I. W . , & Janzen B . L . (1995). Neuroticism, depression, and depression promise. Personality and Individual Differences, 75, 27-31. Sanderson, W . C , DiNardo, P. A . , Rapee, R. M . , & Barlow, D . H . (1990). Syndrome comorbidity in patients diagnosed with DSM-III-R anxiety disorder. Journal of Abnormal Psychology, 99, 308-312. Schafer, J . C , & Leigh, B . C . A comparison o f factor structures o f adolescent and adult alcohol effect expectancies. Addictive Behaviors, 1996, 21, 403-408.  Schmidt, N . B . , Lerew D . R., & Jackson R. J. (1997): The role o f anxiety sensitivity i n the pathogenesis o f panic: Prospective evaluation o f spontaneous panic attacks during acute stress. Journal of Abnormal Psychology, 106, 355-364. Schmidt, N . B . , Lerew D . R . , & Jackson R. J. (1999): Prospective evaluation o f anxiety sensitivity i n the pathogenesis o f panic: Replication and extension. Journal of Abnormal Psychology, 108, 532-537. Schmidt, N . B . , Lerew D . R., & Trakowskim, J. H . (1997). B o d y vigilance i n panic disorder: Evaluating attention to bodily perturbations. Journal of Consulting and  Clinical  Psychology, 65, 214-220. Schmitz, J. M . , A v e r i l l , P. M . , Stotts, A . L . , Moeller, F. G . , Rhoades, H . M . , & Grabowski, J. (2001). Fluoxetine treatment o f cocaine-dependent patients with major depressive disorder. Drug and Alcohol Dependence, 63, 207-214 Schuckit, M . A . (1998). Biological, psychological, and environmental predictors o f alcoholism risk: a longitudinal study. Journal of Studies on Alcohol, 36, 117-126. Schuckit, M . A . , Tipp, J . E . , Bergman, M . , & Reich, W . (1997).Comparison o f induced and independent major depressive disorders in 2,945 alcoholics. American Journal of Psychiatry, 154, 948-957. Schuckit, M . A . , Tipp, J. E . , Bucholz, K . K . , Nurnberger, J. I, Hesselbrock, V M , Crowe, R. R., & Kramer, J. (1997). The life time rates o f three major mood disorders and four major anxiety disorders i n alcoholics and controls. Addiction, 92, 1289-1304. Schwartz, F., Carr, A . C , M u n i c h , R. L . , Glauber, S., Lesser, P., & Murray, J. (1989) reaction time impairment schizophrenia and effective illness: The role o f attention. Biological  Journal  Psychiatry, 25, 540-548.  Segal, Z . V . , & , Ingram, R. E . (1994). M o o d priming and construct activation i n tests of cognitive vulnerability to unipolar depression. Clinical Psychology Review, 14, 663-695.  Silverman, W . K . , Fleisig, W . , Rabian, B . , & Peterson, R. A . (1991). Childhood Anxiety Sensitivity Index. Journal of Clinical Child Psychology, 20, 162-168. Smith, G.T., & Goldman, M . S . (1994). Alcohol expectancy theory and the identification o f highrisk adolescents. Journal of Research on Adolescence,  4, 229-248.  Sobell, L . C , & Sobell, M . B . (1986). Can we do without alcohol abusers' self reports? Behavior Therapist, 9, 141-146. Sobell, L . C , & Sobell, M . B . (1990). Self report issues in alcohol abuse: State o f the art and future directions. Behavioral Assessment, 12, 77-90. Spielberger, C . D . , Edwards, C , Montouri, J., & Lushene, R. (1973). State-Trait Anxiety Inventory for Children—Trait  Subscale. Palo Alto, C A : M i n d Garden.  Spielberger, C . D . , Gorsuch, R . L . , Lushene, R. E . , Vagg, P . R., & Jacobs, G . A . (1983). Manual for the State-Trait. Anxiety Inventory. Palo A l t o , C A : Consulting Psychologists Press. Steed, L . (2001). Further Validity and Reliability Evidence for Beck Hopelessness Scale Scores in a Nonclinical Sample. Educational and Psychological  Measurement,  61, 303-316.  Stewart, S. H . , Comeau, M . N . , Loba, P., Conrod, P. J., Maclean, A . M . , & Watt, M . C . (2002,  i September). Efficacy ofpersonality-matched  preventative  interventions for at-risk teen  drinkers: a pilot study. Paper presented at The Addictions Conference, Eindhoven, Netherlands. Stewart, S. H . , & Devine, H . (2000). Relations between personality and drinking motives in young adults. Personality and Individual Difference, 29, 495-511. Stewart, S. H . , Finn, P. R., & Pihl, R. O. (1992). The effects o f alcohol on the cardiovascular stress response i n men at high risk for alcoholism: A dose response study. Journal of Studies on Alcohol, 53, 499-506. Stewart, S. H . , H a l l , E . , W i l k i e , H . , & Birch, C . D . (2002). Affective priming o f alcohol schema in coping and enhancement motivated drinkers. Cognitive Behaviour Therapy, 31, 68-80.  Stewart, S. H . , Loughlin, H . L . , & Rhyno, E . (2001). Internal drinking motives mediate personality domain-drinking relations i n young adults. Personality and Individual Differences, 30, 271-286. Stewart, S. H . , Taylor, S., & Baker, J. M . (1997). Gender differences in dimensions o f anxiety sensitivity. Journal of Anxiety Disorders, 11, 179-200. Stewart, S. H . , Peterson, J., & Pihl, R. (1995). Anxiety sensitivity and self-reported alcohol consumption rates in university women. Journal of Anxiety Disorders, 9, 283-302. Stewart, S. H . , Zvolensky, M . J., & Eifert, G . H . (2001). Negative-reinforcement drinking motives mediate the relation between anxiety sensitivity and increased drinking behaviour. Personality and Individual Differences, 31, 157-171. Strober, J. (2000). Prospective cognitions in anxiety and depression: Replication and methodological extension. Cognition and emotion, 14, 725-729. Suominen, K . , Isometsa, E . , Henriksson, M D . , Ostamo, A . , & Loennqvist, J. (1997). Hopelessness, impulsiveness and intent among suicide attempters with major depression, alcohol dependence, or both. Acta Psychiatrica Scandinavia, 96, 142-149. Tanaka, E . , Sakamoto, S., Ono, Y . , Fujihara, S., Kitamura, T. (1990). Hopelessness in a community population: Factorial structure and psychosocial correlates. Journal of Social Psychology, 138, 581-590. Taylor, S. (1995). Anxiety sensitivity: Theoretical perspectives in recent findings. Behaviour Research and Therapy, 33, 243-248. Taylor, S., K o c h , W . J., & M c N a l l y , R. J. (1992). H o w does anxiety sensitivity vary across the anxiety disorders? Journal of Anxiety Disorders, 6, 249-259. Teichman, M . , Barnea, Z . , & Ravav, G . (1989). Personality and substance use among adolescents: A longitudinal study. British Journal of Addiction,  84, 181-190.  Tellegen, A . (1985). Structures o f mood and personality and their relevance to assessing anxiety, with an emphasis on self-report. In A . H . Tuma & J. D . Maser (Eds.), Anxiety and the anxiety disorders (pp. 681-706). Hillsdale, N J : Erlbaum. Thombs, D . L . , & Beck, K . H . (1994). The social context o f four adolescent drinking patterns. Health Education Research, 9, 13-22. Thorpe, L . , Whitney, D . K , Kutcher, S. P, Kennedy, S. H , & C A N M A T Depression Work Group. (2001). Special Populations. Canadian Journal of Psychiatry, Supplement 1, 46, 747-768. van Balkom, A . J. L . M . , Beekman, A . T. F., de Beurs, E . , Deeg, D . J. H . , van Dyck, R., van Tilburg, W . (2000). Comorbidity o f the anxiety disorders in a community-based older population i n The Netherlands. Acta Psychiatrica Scandinavia,  101, 37-45.  V a n Tilburg, M . A . L . , & Vingerhoets, A . J. J. M . (2002). The effects o f alcohol on mood induced by an emotional film: A study among women. Journal of  Psychosomatic  Research, 53, 805-809. Varlinskaya, E . I., Spear, L . P., & Spear, N . E . (2001). Acute effects o f ethanol on behavior o f adolescent rats: Role o f social context. Alcoholism: Clinical and Experimental  Research,  25, 377-385. Velting, D . M . , (1999). Personality and negative expectations: Trait structure o f the Beck Hopelessness Scale. Personality and Individual Differences, 26, 913-921. Watkins, P . C . (2002). Implicit memory bias in depression. Cognition and Emotion, 16, 381-402. Watson, D . , & Clark, L . A . (1992). O n traits and temperament: General and specific factors o f emotional experience and their relation to the five-factor model. Journal of Personality, 60, 441-476.  Wechsler H , Lee J, K u o M , and Lee H . (2000). College binge drinking in the 1990s: A continuing problem: Results o f the Harvard School o f Public Health 1999 College Alcohol Study. Journal of American College Health, 48, 199-210. Weissman, M . M . , Bland, R . , Joyce, P.R., Newman, S.,Wells, J.E., Wittchen, H . - U . , 1993. Sex differences in rates o f depression: cross-national perspectives. Journal o f Affective. Disorders, 29, 77-84. West, S. G . , Finch, J. N . , & Curran, P. J. (1995). Structural equation models with nonnormal variables: problems and remedies. In R. H . Hoyle (Ed.), Structural equation modeling: concepts, issues, and applications,  (pp. 56-75). Thousand Oaks, C A : Sage.  Whisman, M . A . , & Pinto, A . (1997). Hopelessness depression in depressed inpatient adolescents. Cognitive Therapy and Research, 21, 345-358. White, H . R., & Labouvie, E . W . (1989). Towards the assessment o f adolescent problem drinking. Journal of Studies on Alcohol, 50, 30-37. Williams A . , & Clark D . (1998). Alcohol consumption in university students: the role o f reasons for drinking, coping strategies, expectancies, and personality traits. Addictive  Behaviors,  25,371-378. W i l l s , T. A . , Sandy, J. M . , Shinar, O., & Yaeger, A . (1999). Contributions o f positive and negative affect to adolescent substance use: Test o f a bidimensional model in a longitudinal study. Psychology of Addictive Behaviors, 13, 327-338. Woicik, P. B . , Conrod, P. J., Stewart, S. H . , Pihl, R. O., & Dongier, M . (2002). The substance use risk profile scale (SURPS): an instrument for identifying motivational profiles for substance use. Unpublished manuscript, University o f British Columbia. Yamamoto, J., Lee, C . K . , Lynn, K . M . , & Cho, K . H . (1987). A l c o h o l abuse in Koreans. The American Journal of Social Psychiatry, 7, 210-214.  Zack, M . , Poulos, C . X . , Fragopoulos, F., & MacLeod, C . M . (2003). Effects o f negative and positive mood phrases on priming o f alcohol words in young drinkers with high and low anxiety sensitivity. Experimental  and Clinical Psyche-pharmacology,  11, 176-185s.  Zhang, J., & Jin, S (1996). Determinants o f suicide ideation: a comparison o f Chinese and American college students. Adolescence,  31, 451-467.  Zinbarg, R. E . , & M o h l m a n , J. (1998). Individual differences i n the acquisition o f effectively valence associations. Journal of Personality and Social Psychology,  74, 1024-1040.  Zinbarg, R. E . , & Revelle, W . (1989). Personality and conditioning: A test o f four models. Journal of Personality and Social Psychology, 57, 301-314. Zuckerman, M . , & Kuhlman, D . M . (2000). Personality and risk-taking: Common biosocial factors. Journal of Personality,  (55,999-1029.  Zvolensky, M . J., Feldner, M . T., Eifert, G . H . , & Stewart, S. H . (2001). Evaluating differential predictions o f emotional reactivity during repeated 20% carbon dioxide enriched air challenge. Cognition and Emotion, 15, 767-786.  177  Appendix Substance Use Risk Profile Scale ( S U R P S ; W o i c i k et al., 2002) HO  1. I am content.  AS  2. In stressful situations, I often fear that no one will reach me in time.  IMP  3. I often don't think things through before I speak.  SS  4. I would like to skydive.  HO  5. I am happy.  AS  6. I get frightened and feel that I am losing my mind when I cannot concentrate on the things that I need to do.  IMP  7. I often involve myself in situations that I later regret.  SS  8. I enjoy new and exciting experiences even i f they are unconventional.  HO  9. I have faith that my future holds great promise.  AS  10. It's frightening to feel dizzy or faint.  IMP  11. The most interesting and exciting things are usually illegal or immoral.  SS  12. I like doing things that frighten me a little.  HO  13. Sometimes I think I am no good at all.  AS  14. It frightens me when I feel my heart beat change.  IMP  15. I usually act without stopping to think.  SS  16. I would like to learn how to drive a motorcycle.  HO  17. I feel proud of my accomplishments.  AS  18. I get scared when I'm too nervous.  IMP  19. Generally, I am an impulsive person.  SS  20. I am interested in experience for its own sake, even i f it is illegal.  HO  21. I feel that I'm a failure.  AS  22. I get scared when I experience unusual body sensations.  178  IMP  23. I'm stubborn and strong-minded and act upon my thoughts despite others' opinions.  SS  24. I would enjoy hiking long distances in wild and uninhabited territory.  HO  25. I feel pleasant.  AS  26. It scares me when I'm unable to focus on a task.  IMP  27. I feel I have to be manipulative to get what I want.  HO  28. I am very enthusiastic about my future.  Note. H O = hopelessness; A S = anxiety sensitivity; I M P = impulsivity; SS = sensation-seeking  179 Modified Drinking Motives Questionnaire ( D M Q - M ; Blackwell, 2003) Social  1. A s a way to celebrate.  Depression  2. T o relax.  Enhancement 3. Because I like the feeling. Social  4. Because it is what most o f my friends do when we get together.  Depression  5. To forget m y worries.  Enhancement 6. Because it is exciting. Social  7. To be sociable.  Anxiety  8. Because I feel more self-confident or sure o f myself.  Enhancement 9. T o get a high. Social  10. Because it is customary on special occasions.  Anxiety  11. Because it helps me when I am feeling nervous.  Enhancement 12. Because it's fun. Social  13. Because it makes a social gathering more enjoyable.  Depression  14. To cheer me up when I'm in a bad mood.  Conformity  15. To be liked.  Depression  16. To numb m y pain.  Depression  17. Because it helps me when I am feeling depressed.  Conformity  18. So that others won't k i d me about not using.  Anxiety  19. To reduce m y anxiety.  Depression  20. T o stop me from dwelling on things.  Depression  21. To turn off negative thoughts about myself.  Depression  22. To help me feel more positive about things in m y life.  180 Depression  23. To stop me from feeling so hopeless about m y future.  Conformity  24. Because m y friends pressure me to use.  Conformity  25. To fit in with a group I like.  Enhancement 26. Because it makes me feel good. Depression  27. To forget painful memories.  Conformity  28. So I won't feel left out.  Note. Conformity, Social, and Enhancement = conformity motives, social motives and enhancement motives from Cooper's (1994) Drinking Motives Questionnaire-Revised; Depression and Anxiety = Blackwell's (2003) depression motives and anxiety motives.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0091248/manifest

Comment

Related Items