- Email: [email protected]
The image-based alcohol-action implicit association test
J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
Contents lists available at ScienceDirect
Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep
Short report
The image-based alcohol-action implicit association test Tibor P. Palfai*, Carl K. Kantner, Kelli D. Tahaney Boston University, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 August 2014 Received in revised form 18 June 2015 Accepted 1 July 2015 Available online 6 July 2015
Background and Objectives: Previous work has shown that automatic alcohol-action associations, assessed by the Implicit Association Test (IAT), may play a role in hazardous drinking patterns. The majority of alcohol-related IATs have been constructed using verbal stimuli, and even those who have used pictorial stimuli have only represented beverage categories with pictures. To assess implicit appetitive responses among a broader population of alcohol users, such as those who experience limitations reading and understanding English, there may be utility in the development of an IAT that utilizes only non-verbal stimuli. Methods: The current study presents an initial effort to develop such a task and examine its association with drinking. One hundred and fifty-three university students participated individually in a laboratory study in which they first completed a pictorial alcohol-specific approach/avoid IAT, followed by selfreport measures of drinking. Results: As hypothesized, negative binomial regression analyses showed that IAT scores predicted the number of heavy drinking episodes and typical number of drinks per occasion. Limitations: The use of a university student sample for this initial study represents an important limitation of this work, which should be addressed in future research. Conclusions: These findings provide initial evidence for the potential use of non-verbal IATs to assess alcohol-related implicit cognition among adults. Implications for the assessment of hazardous drinking behavior across populations are discussed. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Alcohol Implicit Assessment Motivation Image
1. Introduction There is now considerable evidence to suggest that excessive drinking may be associated with stronger automatic appetitive responses to alcohol cues (Stacy & Wiers, 2010). These responses have been characterized as a function of automatic associations in memory between alcohol cues and appetitive responses. Although assessment of these automatic associations has been demonstrated using a variety of tasks (e.g., Field, Caren, Fernie, & De Houwer, 2011; Ostafin, Palfai, & Wechsler, 2003), the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998) has been the most commonly used index of automatic alcohol associations. Investigators have shown that individuals who engage in heavier drinking show stronger associations between alcohol cues in memory and positive affect (e.g., de Jong, Wiers, van de Braak, & Huijding, 2007), positive alcohol expectancies (e.g., Jajodia &
* Corresponding author. Boston University, Department of Psychology, 648 Beacon Street, 4th Floor, Boston, MA 02215, USA. E-mail address: [email protected] (T.P. Palfai). http://dx.doi.org/10.1016/j.jbtep.2015.07.002 0005-7916/© 2015 Elsevier Ltd. All rights reserved.
Earleywine, 2003), and approach motivation (e.g., Palfai & Ostafin, 2003). Consistent with the broader literature on the IAT, alcoholspecific IAT studies have traditionally used words to represent stimulus (e.g., alcohol) and attribute (e.g., approach) categories (e.g., Palfai & Ostafin, 2003; Wiers, Van Woerden, Smulders, & DeJong, 2002). However, to facilitate efforts to characterize individual differences in automatic associations to specific cues, there may be advantages to using pictorial representations of stimuli to represent categories (Ostafin & Palfai, 2006; Teachman, Gregg, & Woody, 2001). There is evidence to suggest that priming of automatic associations may be facilitated with pictorial cues (Fazio, Jackson, Dunton, & Williams, 1995) as more relevant schema may be activated by cues that better represent the stimuli encountered in environments. Pictorial cues of appetitive stimuli such as alcohol may be particularly salient to heavier drinkers, and as such may provide a stronger index of automatic, motivationally relevant responses to alcohol. A number of studies have shown that alcoholrelated IATs that make use of picture stimuli to represent target categories (Cohn et al., 2012; Lindgren et al., 2012; Ostafin & Palfai,
136
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
2006) show adequate internal reliability and predict alcoholrelated outcomes. Although there has been increased interest in the use of pictorial stimuli to represent target categories, the attribute categories (e.g., good/bad, approach/avoid) in these studies have typically been represented as verbal stimuli. The use of pictorial stimuli to represent attribute categories may be particularly useful for those who have limited verbal skills or limited verbal skills in the language in which the task is to be completed. Moreover, the use of non-verbal stimuli to represent attribute categories may provide a method of assessing motivationally relevant associations that are not verbally mediated (Slabbinck, De Houwer, & Van Kenhove, 2011). The few studies that have used pictures as stimuli for attribute categories have used them to assess implicit associations among children (e.g., Pieters, Haske van der Vorst, Engels, Wiers, 2010; Thomas, Smith, Ball, & Tasmania, 2007). Pieters, van der Vorst, Engels, and Wiers (2010), for example, found that 11-12 year-olds who showed stronger associations between angry adult facial expressions and alcohol pictures were more likely to use alcohol. The use of nonverbal stimuli to represent all categories in the IAT has not yet been examined among adult populations of drinkers. Moreover, no study to date has explored whether alcohol-action associations as measured by the IAT may be represented non-verbally. The current study presents an initial effort to develop this type of alcohol non-verbal IAT task and examine its association with drinking. In this initial phase, the task was developed with university undergraduates, the population on which a good deal of research on the alcohol IAT has been conducted. University student drinkers completed a non-verbal IAT prior to a series of self-report questionnaires. Based on previous work with the approach-avoid IAT (e.g., Ostafin & Palfai, 2006; Palfai & Ostafin, 2003), the main hypothesis was that response times on the image-based alcohol-action IAT would be associated with more frequent heavy episodic drinking in the past 30 days. Associations with typical drinking amounts per occasion and frequency of drinking were also examined. 2. Methods 2.1. Participants Subjects consisted of undergraduate students over the age of 18 from a large private university in the northeastern United States. Students participated in exchange for course credit in introductory psychology. Recruitment and study procedures were approved by the university's IRB. A total of 153 subjects completed study procedures. The mean age of the sample was 19.24 (SD ¼ 1.18) years, and 75.2% of participants were female, which did not differ by group. The majority of the sample (56%) was Caucasian (25% Asian, 4% Black, 1% Hawaiian/Pacific Islander, 14% other), 8.5% identified as Hispanic and 77% identified English as their first language. Students had a mean of 1.79 (SD ¼ 2.86) heavy drinking episodes, consumed a mean of 2.18 (SD ¼ 2.49) drinks per occasion, and consumed alcohol on a mean of 2.9 (SD ¼ 3.60) occasions in the past month. Male students were more likely to engage in drinking more frequently than female students [3.95 (SD ¼ 4.27) vs. 2.55 (SD ¼ 3.33)] and to consume more drinks on a typical drinking occasion [3.16 (SD ¼ 3.45) vs. 1.97 (SD ¼ 2.15)] but not more likely to engage in heavy episodic drinking (using gender specific definitions) [2.18 (SD ¼ 3.44) vs. 1.66 (SD ¼ 2.65)].1
1
Univariate ANOVAs showed that males were more likely to engage in drinking more frequently [F(1, 151) ¼ 4.35, p < .05] and to consume more drinks on a typical drinking occasion [F(1, 151) ¼ 6.33, p < .05] but not more likely to engage in heavy episodic drinking (using gender specific definitions) [F(1, 151) ¼ .96, p ¼ .33].
2.2. Measures and procedures All procedures took place during a 1-h laboratory session. Subjects first provided informed consent and then were randomized to the IAT order condition. The IAT was administered, followed by selfreport questionnaires on individual differences and alcohol use behaviors. Alcohol use. The Daily Drinking Questionnaire- Modified (Dimeff, Baer, Kivlahan, & Marlatt, 1999) was used to collect information on past month alcohol use. Subjects indicated the frequency and typical quantity of drinking per occasion over the past 30 days, as well as the number of heavy drinking episodes (5 or more drinks on one occasion for males and 4 or more for females) in the past 30 days.
2.3. Non-verbal alcohol action IAT Automatic action associations to alcohol cues were examined with a non-verbal alcohol-specific IAT developed for this study. The task required that participants categorize a series of stimuli presented one at a time according to one of four categories: two target categories (i.e., alcohol, water) and two attribute categories (i.e., approach avoid). The stimulus set for the attribute and target categories were pictures. The alcohol and water pictures consisted of single or multiple bottle/can pictures of established brands (e.g., Budweiser, Coors, Corona, Jack Daniel's, Miller Light, Perrier, Evian, Dasani, FIJI, Aquafina) as well as unbranded water bottles. The approach and avoid pictures consisted of either a man or woman engaging in different actions. The approach pictures included, “man gesturing for embrace”, “woman and child running toward one another”, “woman gesturing come here”, “hand reaching for trophy”, “arm and hand extended out to shake”. The avoid pictures included actions such as, “woman running away”, woman gesturing stop”, man blocking his head with hands and turning away”, “man gesturing stop”, “woman shielding face.”.2 The approach set consisted of The IAT consisted of 7 blocksd(a) a 20trial alcohol-water picture discrimination task (e.g., right ¼ alcohol, left ¼ water), (b) a 20 trial approach-avoidance picture discrimination task (e.g., right ¼ approach, left ¼ avoid), (c) a 20-trial combination block (e.g., right ¼ alcohol or approach, left ¼ water or avoid), (d) a 40-trial combination block of the same combination, (e) a 20-trial single category discrimination block where the target categories are reversed (right ¼ water, left ¼ alcohol), (f) a 20 trial combined incongruent block (e.g., right ¼ water or approach, left ¼ alcohol or avoid), and (g) a 40-trial incongruent block. The IAT was presented in two orders that were counterbalanced across studentsdone with “alcohol” and “approach” in the first combined block and one with “alcohol” and “avoid” in the first combined block. Reaction time data were prepared for analysis by first removing incorrect responses and outliers that Exceeded 10,000 ms (Greenwald, Nosek, & Banaji, 2003). Participants who did not complete at least 80% of the trials correctly (n ¼ 7) were removed from subsequent analyses and one participant was removed because of inconsistencies in responding (e.g., reporting no drinking in the past month but 3 heavy drinking episodes). The implicit appetitive response index was based on the D2SD score (D3), which was calculated by examining the difference between response
2 As the approach-avoid picture stimuli were novel, each picture was rated by a pilot sample of 20 students on 7 point Likert-scale items ranging from 3 (very clearly avoid) to 0 (unclear/I don't know) to þ3 (very clearly approach). Mean ratings for the approach pictures were 2.60 (SD ¼ .31) and mean ratings for the avoid pictures were 2.4 (SD ¼ .35).
Table 1 Partial correlations among IAT and alcohol use variables, controlling for block order.
Heavy drinking episodes Quantity Frequency a b c
IAT
Hvy.
.184a .228b .134
.672c .766c
Qty.
.593c
p < .05. p < .01. p < .001.
times on the combined alcohol-approach, water-avoid blocks and the combined reverse blocks divided by the pooled standard deviation (Greenwald et al., 2003). D scores were computed such that larger scores represented stronger alcohol-approach motivation. The measure had good internal consistencydthe D3 split-half reliability was .691. 3. Results 3.1. Partial correlation alcohol use outcomes To examine the association among drinking variables and the IAT, partial correlations were conducted controlling for the order of IAT (i.e., whether approach or avoid was first paired with alcohol). As shown in Table 1, IAT scores were significantly associated with heavy episodic drinking and typical quantity per occasion. 3.2. Predictive value of IAT scores on alcohol use To examine the association between the non-verbal IAT and drinking variables, a series of negative binomial regression analyses were conducted with number of heavy drinking episodes, typical quantity per occasion and frequency of drinking as the main outcome variables. Negative binomial regression was chosen for these analyses due to the fact that these were over dispersed count data that were positively skewed. For each analysis, block order and gender were entered first as covariates followed by the D-score. The primary analysis showed a significant effect of implicit appetitive motivation on number of heavy drinking episodes, aIRR ¼ 2.25 (CI: 1.11, 4.58), p < .01. As shown in Fig. 1, higher levels of IAT approach scores were associated with higher predicted mean levels of heavy episodic drinking. IAT scores also predicted typical quantity of use, aIRR ¼ 1.74 (CI: 1.09, 2.80), p < .03 but did not significantly predict frequency of drinking over the past 30 days aIRR ¼ 1.55 (CI: .91, 2.62), p ¼ ns.3 4. Discussion The current study attempted to examine whether an implicit association test (IAT) that used non-verbal stimuli to represent both stimulus (e.g., alcohol) and attribute (e.g., approach) categories would predict drinking behavior among students. Results showed that participants readily categorized the stimuli in this task and that the measure showed adequate internal reliability. Moreover, automatic alcohol approach motivation as measured by the IAT significantly predicted frequency of heavy drinking and typical use
3 Secondary analyses were conducted on D3 scores that removed stimulus items that showed categorization error rates of greater than 8%. Two avoidance pictures and one water picture was removed. Associations between IAT scores and alcohol use did not change with these modified scores.
Frequency of heavy drinking episodes
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
137
3.5 3 2.5 2 1.5 1 0.5 0 -1SD
Mean
1SD
IAT Score Fig. 1. Predicted heavy drinking episodes by IAT level.
per occasion, however the association with frequency of use was not significant. These results are consistent with previous work (e.g., Ostafin & Palfai, 2006) that suggests that implicit alcoholapproach associations may be more predictive of quantity of use patterns. The study replicates and extends previous work on the use of the IAT to predict alcohol use patterns among college student populations. This type of non-verbal IAT may have utility for the assessment of implicit appetitive responses among a broader population of alcohol users, such as those who may have difficulty reading and understanding English, or to assess outcomes among alcohol dependent patients who have undergone approachavoidance training (e.g., Wiers, Eberl, Rinck, Becker, & Lindenmeyer, 2011). It is important, however, to emphasize the limitations of this initial study. First, the magnitude of these effects should be compared to “standard” approach-avoid IATs. It is unclear from the current work whether this sort of measure demonstrates the same level of association with drinking as well-established alcohol related IATs that have represented alcohol-approach associations with words. Analyses of a recent study conducted in our lab with a similar sample of 99 students (mean of heavy drinking episodes 1.85; SD ¼ 2.71) found that the standard approach-avoid IAT showed slightly higher predictive value for heavy drinking days [IRR ¼ 2.41 (CI: 1.37, 4.23), p < .01]. To clearly address this question, however, requires a direct comparison of these versions within the same experiment. Second the current study used the common practice of assigning specific keys to approach and avoid responses. Given the work on hemispheric specialization in approach and avoid motivation (e.g., Harmon-Jones, 2003), it would be important to address this potential concern through counterbalancing of response keys. Third, the use of a college student population to test the validity of this measure is a limitation that should be addressed in future research. Findings from this study provide promise for the development of a non-verbal approach that may be used to assess automatic appetitive responses among individuals who may have difficulty with written language, however these results with a college students is simply the first step in establishing the IAT in this regard. Finally, it will be important to provide prospective data on the utility of this task. The current work provides a promising first step for developing and IAT with non-verbal stimuli to assess appetitive responses.
Conflict of interest None to declare.
138
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
References Cohn, A. M., Cameron, A. Y., Udo, T., Hagman, B. T., Mitchell, J., Bramm, S., et al. (2012). Delineating potential mechanisms of implicit alcohol cognitions: drinking restraint, negative affect, and their relationship with approach alcohol associations. Psychology of Addictive Behaviors, 26(2), 318. Dimeff, L. A., Baer, J. S., Kivlahan, D. R., & Marlatt, G. A. (1999). Brief Alcohol Screening and Intervention for College Students (BASICS): A harm reduction approach (p. 200). New York, NY: Guilford Press. Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobtrusive measure of racial attitudes: a bona fide pipeline? Journal of Personality and Social Psychology, 69(6), 1013. Field, M., Caren, R., Fernie, G., & De Houwer, J. (2011). Alcohol approach tendencies in heavy drinkers: comparison of effects in a relevant stimulus-response compatibility task and an approach/avoidance Simon task. Psychology of Addictive Behaviors, 25, 697e701. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology, 74, 1464e1480. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197e216. Harmon-Jones, E. (2003). Clarifying the emotive functions of asymmetrical frontal cortical activity. Psychophysiology, 45, 275e278. Jajodia, A., & Earleywine, M. (2003). Measuring alcohol expectancies with the implicit association test. Psychology of Addictive Behaviors, 17, 126e133. de Jong, P. J., Wiers, R. W., van de Braak, M., & Huijding, J. (2007). Using the extrinsic affective Simon test as a measure of implicit attitudes towards alcohol: relationship with drinking behavior and alcohol problems. Addictive Behaviors, 32, 881e887.
Lindgren, K. P., Neighbors, C., Teachman, B. A., Wiers, R. W., Westgate, E., & Greenwald, A. G. (2012). I drink therefore I am: Validating alcohol-related implicit association tests. Psychology of Addictive Behaviors, 27(1), 1. Ostafin, B. D., & Palfai, T. P. (2006). Compelled to consume: the implicit association test and automatic alcohol motivation. Psychology of Addictive Behaviors, 20, 322e327. Ostafin, B. D., Palfai, T. P., & Wechsler, C. E. (2003). The accessibility of motivational tendencies toward alcohol: approach, avoidance, and disinhibited drinking. Experimental and Clinical Psychopharmacology, 11, 294e301. Palfai, T. P., & Ostafin, B. D. (2003). Alcohol-related motivational tendencies in hazardous drinkers: assessing implicit response tendencies using the modified IAT. Behaviour Research and Therapy, 41, 1149e1162. Pieters, S., van der Vorst, H., Engels, R. C., & Wiers, R. W. (2010). Implicit and explicit cognitions related to alcohol use in children. Addictive Behaviors, 35(5), 471e478. Slabbinck, H., De Houwer, J., & Van Kenhove, P. (2011). A pictorial attitude IAT as a measure of implicit motives. European Journal of Personality, 25(1), 76e86. Stacy, A. W., & Wiers, R. W. (2010). Implicit cognition and addiction: a tool for explaining paradoxical behavior. Annual Review of Clinical Psychology, 6, 551. Teachman, B. A., Gregg, A. P., & Woody, S. R. (2001). Implicit associations for fearrelevant stimuli among individuals with snake and spider fears. Journal of Abnormal Psychology, 110(2), 226. Thomas, S. R., Burton Smith, R., & Ball, P. J. (2007). Implicit attitudes in very young children: an adaptation of the IAT. Current Research in Social Psychology, 13(7), 75e85. Wiers, R. W., Eberl, C., Rinck, M., Becker, E. S., & Lindenmeyer, J. (2011). Retraining automatic action-tendencies to approach alcohol in hazardous drinkers. Addiction, 105, 279e287. Wiers, R. W., Van Woerden, N., Smulders, F. T., & De Jong, P. J. (2002). Implicit and explicit alcohol-related cognitions in heavy and light drinkers. Journal of Abnormal Psychology, 111(4), 648.
Contents lists available at ScienceDirect
Journal of Behavior Therapy and Experimental Psychiatry journal homepage: www.elsevier.com/locate/jbtep
Short report
The image-based alcohol-action implicit association test Tibor P. Palfai*, Carl K. Kantner, Kelli D. Tahaney Boston University, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 4 August 2014 Received in revised form 18 June 2015 Accepted 1 July 2015 Available online 6 July 2015
Background and Objectives: Previous work has shown that automatic alcohol-action associations, assessed by the Implicit Association Test (IAT), may play a role in hazardous drinking patterns. The majority of alcohol-related IATs have been constructed using verbal stimuli, and even those who have used pictorial stimuli have only represented beverage categories with pictures. To assess implicit appetitive responses among a broader population of alcohol users, such as those who experience limitations reading and understanding English, there may be utility in the development of an IAT that utilizes only non-verbal stimuli. Methods: The current study presents an initial effort to develop such a task and examine its association with drinking. One hundred and fifty-three university students participated individually in a laboratory study in which they first completed a pictorial alcohol-specific approach/avoid IAT, followed by selfreport measures of drinking. Results: As hypothesized, negative binomial regression analyses showed that IAT scores predicted the number of heavy drinking episodes and typical number of drinks per occasion. Limitations: The use of a university student sample for this initial study represents an important limitation of this work, which should be addressed in future research. Conclusions: These findings provide initial evidence for the potential use of non-verbal IATs to assess alcohol-related implicit cognition among adults. Implications for the assessment of hazardous drinking behavior across populations are discussed. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Alcohol Implicit Assessment Motivation Image
1. Introduction There is now considerable evidence to suggest that excessive drinking may be associated with stronger automatic appetitive responses to alcohol cues (Stacy & Wiers, 2010). These responses have been characterized as a function of automatic associations in memory between alcohol cues and appetitive responses. Although assessment of these automatic associations has been demonstrated using a variety of tasks (e.g., Field, Caren, Fernie, & De Houwer, 2011; Ostafin, Palfai, & Wechsler, 2003), the Implicit Association Test (IAT; Greenwald, McGhee, & Schwartz, 1998) has been the most commonly used index of automatic alcohol associations. Investigators have shown that individuals who engage in heavier drinking show stronger associations between alcohol cues in memory and positive affect (e.g., de Jong, Wiers, van de Braak, & Huijding, 2007), positive alcohol expectancies (e.g., Jajodia &
* Corresponding author. Boston University, Department of Psychology, 648 Beacon Street, 4th Floor, Boston, MA 02215, USA. E-mail address: [email protected] (T.P. Palfai). http://dx.doi.org/10.1016/j.jbtep.2015.07.002 0005-7916/© 2015 Elsevier Ltd. All rights reserved.
Earleywine, 2003), and approach motivation (e.g., Palfai & Ostafin, 2003). Consistent with the broader literature on the IAT, alcoholspecific IAT studies have traditionally used words to represent stimulus (e.g., alcohol) and attribute (e.g., approach) categories (e.g., Palfai & Ostafin, 2003; Wiers, Van Woerden, Smulders, & DeJong, 2002). However, to facilitate efforts to characterize individual differences in automatic associations to specific cues, there may be advantages to using pictorial representations of stimuli to represent categories (Ostafin & Palfai, 2006; Teachman, Gregg, & Woody, 2001). There is evidence to suggest that priming of automatic associations may be facilitated with pictorial cues (Fazio, Jackson, Dunton, & Williams, 1995) as more relevant schema may be activated by cues that better represent the stimuli encountered in environments. Pictorial cues of appetitive stimuli such as alcohol may be particularly salient to heavier drinkers, and as such may provide a stronger index of automatic, motivationally relevant responses to alcohol. A number of studies have shown that alcoholrelated IATs that make use of picture stimuli to represent target categories (Cohn et al., 2012; Lindgren et al., 2012; Ostafin & Palfai,
136
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
2006) show adequate internal reliability and predict alcoholrelated outcomes. Although there has been increased interest in the use of pictorial stimuli to represent target categories, the attribute categories (e.g., good/bad, approach/avoid) in these studies have typically been represented as verbal stimuli. The use of pictorial stimuli to represent attribute categories may be particularly useful for those who have limited verbal skills or limited verbal skills in the language in which the task is to be completed. Moreover, the use of non-verbal stimuli to represent attribute categories may provide a method of assessing motivationally relevant associations that are not verbally mediated (Slabbinck, De Houwer, & Van Kenhove, 2011). The few studies that have used pictures as stimuli for attribute categories have used them to assess implicit associations among children (e.g., Pieters, Haske van der Vorst, Engels, Wiers, 2010; Thomas, Smith, Ball, & Tasmania, 2007). Pieters, van der Vorst, Engels, and Wiers (2010), for example, found that 11-12 year-olds who showed stronger associations between angry adult facial expressions and alcohol pictures were more likely to use alcohol. The use of nonverbal stimuli to represent all categories in the IAT has not yet been examined among adult populations of drinkers. Moreover, no study to date has explored whether alcohol-action associations as measured by the IAT may be represented non-verbally. The current study presents an initial effort to develop this type of alcohol non-verbal IAT task and examine its association with drinking. In this initial phase, the task was developed with university undergraduates, the population on which a good deal of research on the alcohol IAT has been conducted. University student drinkers completed a non-verbal IAT prior to a series of self-report questionnaires. Based on previous work with the approach-avoid IAT (e.g., Ostafin & Palfai, 2006; Palfai & Ostafin, 2003), the main hypothesis was that response times on the image-based alcohol-action IAT would be associated with more frequent heavy episodic drinking in the past 30 days. Associations with typical drinking amounts per occasion and frequency of drinking were also examined. 2. Methods 2.1. Participants Subjects consisted of undergraduate students over the age of 18 from a large private university in the northeastern United States. Students participated in exchange for course credit in introductory psychology. Recruitment and study procedures were approved by the university's IRB. A total of 153 subjects completed study procedures. The mean age of the sample was 19.24 (SD ¼ 1.18) years, and 75.2% of participants were female, which did not differ by group. The majority of the sample (56%) was Caucasian (25% Asian, 4% Black, 1% Hawaiian/Pacific Islander, 14% other), 8.5% identified as Hispanic and 77% identified English as their first language. Students had a mean of 1.79 (SD ¼ 2.86) heavy drinking episodes, consumed a mean of 2.18 (SD ¼ 2.49) drinks per occasion, and consumed alcohol on a mean of 2.9 (SD ¼ 3.60) occasions in the past month. Male students were more likely to engage in drinking more frequently than female students [3.95 (SD ¼ 4.27) vs. 2.55 (SD ¼ 3.33)] and to consume more drinks on a typical drinking occasion [3.16 (SD ¼ 3.45) vs. 1.97 (SD ¼ 2.15)] but not more likely to engage in heavy episodic drinking (using gender specific definitions) [2.18 (SD ¼ 3.44) vs. 1.66 (SD ¼ 2.65)].1
1
Univariate ANOVAs showed that males were more likely to engage in drinking more frequently [F(1, 151) ¼ 4.35, p < .05] and to consume more drinks on a typical drinking occasion [F(1, 151) ¼ 6.33, p < .05] but not more likely to engage in heavy episodic drinking (using gender specific definitions) [F(1, 151) ¼ .96, p ¼ .33].
2.2. Measures and procedures All procedures took place during a 1-h laboratory session. Subjects first provided informed consent and then were randomized to the IAT order condition. The IAT was administered, followed by selfreport questionnaires on individual differences and alcohol use behaviors. Alcohol use. The Daily Drinking Questionnaire- Modified (Dimeff, Baer, Kivlahan, & Marlatt, 1999) was used to collect information on past month alcohol use. Subjects indicated the frequency and typical quantity of drinking per occasion over the past 30 days, as well as the number of heavy drinking episodes (5 or more drinks on one occasion for males and 4 or more for females) in the past 30 days.
2.3. Non-verbal alcohol action IAT Automatic action associations to alcohol cues were examined with a non-verbal alcohol-specific IAT developed for this study. The task required that participants categorize a series of stimuli presented one at a time according to one of four categories: two target categories (i.e., alcohol, water) and two attribute categories (i.e., approach avoid). The stimulus set for the attribute and target categories were pictures. The alcohol and water pictures consisted of single or multiple bottle/can pictures of established brands (e.g., Budweiser, Coors, Corona, Jack Daniel's, Miller Light, Perrier, Evian, Dasani, FIJI, Aquafina) as well as unbranded water bottles. The approach and avoid pictures consisted of either a man or woman engaging in different actions. The approach pictures included, “man gesturing for embrace”, “woman and child running toward one another”, “woman gesturing come here”, “hand reaching for trophy”, “arm and hand extended out to shake”. The avoid pictures included actions such as, “woman running away”, woman gesturing stop”, man blocking his head with hands and turning away”, “man gesturing stop”, “woman shielding face.”.2 The approach set consisted of The IAT consisted of 7 blocksd(a) a 20trial alcohol-water picture discrimination task (e.g., right ¼ alcohol, left ¼ water), (b) a 20 trial approach-avoidance picture discrimination task (e.g., right ¼ approach, left ¼ avoid), (c) a 20-trial combination block (e.g., right ¼ alcohol or approach, left ¼ water or avoid), (d) a 40-trial combination block of the same combination, (e) a 20-trial single category discrimination block where the target categories are reversed (right ¼ water, left ¼ alcohol), (f) a 20 trial combined incongruent block (e.g., right ¼ water or approach, left ¼ alcohol or avoid), and (g) a 40-trial incongruent block. The IAT was presented in two orders that were counterbalanced across studentsdone with “alcohol” and “approach” in the first combined block and one with “alcohol” and “avoid” in the first combined block. Reaction time data were prepared for analysis by first removing incorrect responses and outliers that Exceeded 10,000 ms (Greenwald, Nosek, & Banaji, 2003). Participants who did not complete at least 80% of the trials correctly (n ¼ 7) were removed from subsequent analyses and one participant was removed because of inconsistencies in responding (e.g., reporting no drinking in the past month but 3 heavy drinking episodes). The implicit appetitive response index was based on the D2SD score (D3), which was calculated by examining the difference between response
2 As the approach-avoid picture stimuli were novel, each picture was rated by a pilot sample of 20 students on 7 point Likert-scale items ranging from 3 (very clearly avoid) to 0 (unclear/I don't know) to þ3 (very clearly approach). Mean ratings for the approach pictures were 2.60 (SD ¼ .31) and mean ratings for the avoid pictures were 2.4 (SD ¼ .35).
Table 1 Partial correlations among IAT and alcohol use variables, controlling for block order.
Heavy drinking episodes Quantity Frequency a b c
IAT
Hvy.
.184a .228b .134
.672c .766c
Qty.
.593c
p < .05. p < .01. p < .001.
times on the combined alcohol-approach, water-avoid blocks and the combined reverse blocks divided by the pooled standard deviation (Greenwald et al., 2003). D scores were computed such that larger scores represented stronger alcohol-approach motivation. The measure had good internal consistencydthe D3 split-half reliability was .691. 3. Results 3.1. Partial correlation alcohol use outcomes To examine the association among drinking variables and the IAT, partial correlations were conducted controlling for the order of IAT (i.e., whether approach or avoid was first paired with alcohol). As shown in Table 1, IAT scores were significantly associated with heavy episodic drinking and typical quantity per occasion. 3.2. Predictive value of IAT scores on alcohol use To examine the association between the non-verbal IAT and drinking variables, a series of negative binomial regression analyses were conducted with number of heavy drinking episodes, typical quantity per occasion and frequency of drinking as the main outcome variables. Negative binomial regression was chosen for these analyses due to the fact that these were over dispersed count data that were positively skewed. For each analysis, block order and gender were entered first as covariates followed by the D-score. The primary analysis showed a significant effect of implicit appetitive motivation on number of heavy drinking episodes, aIRR ¼ 2.25 (CI: 1.11, 4.58), p < .01. As shown in Fig. 1, higher levels of IAT approach scores were associated with higher predicted mean levels of heavy episodic drinking. IAT scores also predicted typical quantity of use, aIRR ¼ 1.74 (CI: 1.09, 2.80), p < .03 but did not significantly predict frequency of drinking over the past 30 days aIRR ¼ 1.55 (CI: .91, 2.62), p ¼ ns.3 4. Discussion The current study attempted to examine whether an implicit association test (IAT) that used non-verbal stimuli to represent both stimulus (e.g., alcohol) and attribute (e.g., approach) categories would predict drinking behavior among students. Results showed that participants readily categorized the stimuli in this task and that the measure showed adequate internal reliability. Moreover, automatic alcohol approach motivation as measured by the IAT significantly predicted frequency of heavy drinking and typical use
3 Secondary analyses were conducted on D3 scores that removed stimulus items that showed categorization error rates of greater than 8%. Two avoidance pictures and one water picture was removed. Associations between IAT scores and alcohol use did not change with these modified scores.
Frequency of heavy drinking episodes
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
137
3.5 3 2.5 2 1.5 1 0.5 0 -1SD
Mean
1SD
IAT Score Fig. 1. Predicted heavy drinking episodes by IAT level.
per occasion, however the association with frequency of use was not significant. These results are consistent with previous work (e.g., Ostafin & Palfai, 2006) that suggests that implicit alcoholapproach associations may be more predictive of quantity of use patterns. The study replicates and extends previous work on the use of the IAT to predict alcohol use patterns among college student populations. This type of non-verbal IAT may have utility for the assessment of implicit appetitive responses among a broader population of alcohol users, such as those who may have difficulty reading and understanding English, or to assess outcomes among alcohol dependent patients who have undergone approachavoidance training (e.g., Wiers, Eberl, Rinck, Becker, & Lindenmeyer, 2011). It is important, however, to emphasize the limitations of this initial study. First, the magnitude of these effects should be compared to “standard” approach-avoid IATs. It is unclear from the current work whether this sort of measure demonstrates the same level of association with drinking as well-established alcohol related IATs that have represented alcohol-approach associations with words. Analyses of a recent study conducted in our lab with a similar sample of 99 students (mean of heavy drinking episodes 1.85; SD ¼ 2.71) found that the standard approach-avoid IAT showed slightly higher predictive value for heavy drinking days [IRR ¼ 2.41 (CI: 1.37, 4.23), p < .01]. To clearly address this question, however, requires a direct comparison of these versions within the same experiment. Second the current study used the common practice of assigning specific keys to approach and avoid responses. Given the work on hemispheric specialization in approach and avoid motivation (e.g., Harmon-Jones, 2003), it would be important to address this potential concern through counterbalancing of response keys. Third, the use of a college student population to test the validity of this measure is a limitation that should be addressed in future research. Findings from this study provide promise for the development of a non-verbal approach that may be used to assess automatic appetitive responses among individuals who may have difficulty with written language, however these results with a college students is simply the first step in establishing the IAT in this regard. Finally, it will be important to provide prospective data on the utility of this task. The current work provides a promising first step for developing and IAT with non-verbal stimuli to assess appetitive responses.
Conflict of interest None to declare.
138
T.P. Palfai et al. / J. Behav. Ther. & Exp. Psychiat. 50 (2016) 135e138
References Cohn, A. M., Cameron, A. Y., Udo, T., Hagman, B. T., Mitchell, J., Bramm, S., et al. (2012). Delineating potential mechanisms of implicit alcohol cognitions: drinking restraint, negative affect, and their relationship with approach alcohol associations. Psychology of Addictive Behaviors, 26(2), 318. Dimeff, L. A., Baer, J. S., Kivlahan, D. R., & Marlatt, G. A. (1999). Brief Alcohol Screening and Intervention for College Students (BASICS): A harm reduction approach (p. 200). New York, NY: Guilford Press. Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Variability in automatic activation as an unobtrusive measure of racial attitudes: a bona fide pipeline? Journal of Personality and Social Psychology, 69(6), 1013. Field, M., Caren, R., Fernie, G., & De Houwer, J. (2011). Alcohol approach tendencies in heavy drinkers: comparison of effects in a relevant stimulus-response compatibility task and an approach/avoidance Simon task. Psychology of Addictive Behaviors, 25, 697e701. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology, 74, 1464e1480. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85, 197e216. Harmon-Jones, E. (2003). Clarifying the emotive functions of asymmetrical frontal cortical activity. Psychophysiology, 45, 275e278. Jajodia, A., & Earleywine, M. (2003). Measuring alcohol expectancies with the implicit association test. Psychology of Addictive Behaviors, 17, 126e133. de Jong, P. J., Wiers, R. W., van de Braak, M., & Huijding, J. (2007). Using the extrinsic affective Simon test as a measure of implicit attitudes towards alcohol: relationship with drinking behavior and alcohol problems. Addictive Behaviors, 32, 881e887.
Lindgren, K. P., Neighbors, C., Teachman, B. A., Wiers, R. W., Westgate, E., & Greenwald, A. G. (2012). I drink therefore I am: Validating alcohol-related implicit association tests. Psychology of Addictive Behaviors, 27(1), 1. Ostafin, B. D., & Palfai, T. P. (2006). Compelled to consume: the implicit association test and automatic alcohol motivation. Psychology of Addictive Behaviors, 20, 322e327. Ostafin, B. D., Palfai, T. P., & Wechsler, C. E. (2003). The accessibility of motivational tendencies toward alcohol: approach, avoidance, and disinhibited drinking. Experimental and Clinical Psychopharmacology, 11, 294e301. Palfai, T. P., & Ostafin, B. D. (2003). Alcohol-related motivational tendencies in hazardous drinkers: assessing implicit response tendencies using the modified IAT. Behaviour Research and Therapy, 41, 1149e1162. Pieters, S., van der Vorst, H., Engels, R. C., & Wiers, R. W. (2010). Implicit and explicit cognitions related to alcohol use in children. Addictive Behaviors, 35(5), 471e478. Slabbinck, H., De Houwer, J., & Van Kenhove, P. (2011). A pictorial attitude IAT as a measure of implicit motives. European Journal of Personality, 25(1), 76e86. Stacy, A. W., & Wiers, R. W. (2010). Implicit cognition and addiction: a tool for explaining paradoxical behavior. Annual Review of Clinical Psychology, 6, 551. Teachman, B. A., Gregg, A. P., & Woody, S. R. (2001). Implicit associations for fearrelevant stimuli among individuals with snake and spider fears. Journal of Abnormal Psychology, 110(2), 226. Thomas, S. R., Burton Smith, R., & Ball, P. J. (2007). Implicit attitudes in very young children: an adaptation of the IAT. Current Research in Social Psychology, 13(7), 75e85. Wiers, R. W., Eberl, C., Rinck, M., Becker, E. S., & Lindenmeyer, J. (2011). Retraining automatic action-tendencies to approach alcohol in hazardous drinkers. Addiction, 105, 279e287. Wiers, R. W., Van Woerden, N., Smulders, F. T., & De Jong, P. J. (2002). Implicit and explicit alcohol-related cognitions in heavy and light drinkers. Journal of Abnormal Psychology, 111(4), 648.