Hierarchical Linear Modeling of Drinking to Cope with Anxiety among College Students

Previous research has assumed that drinking motives (reasons for drinking) remain fairly stable over time (traits). The current study examines how drinking motives of college students may vary over time (states). Weekly measures were completed by 125 undergraduates at Old Dominion University, which examined how average levels of anxiety interacted with anxiety coping motives to influence drinking. Using Hierarchical Linear Modeling, it was found that anxiety coping motives were less predictive for participants with elevated average anxiety levels, as opposed to participants with normal to low anxiety levels. Implications and possible explanations are discussed. D rinking motives, or self-reported reasons for drinking, are often examined as proximal precursors to alcohol consumption. Typically, these motives are treated as stable individual differences, in that they are assumed to vary from person to person (high between-persons variability), yet remain stable within each person over time (low within-persons variability). Research has suggested, however, that drinking motives do not appear to remain stable within-persons, especially with regard to motives related to coping with anxiety, or the self-reported negative mood state characterized by apprehension regarding the future (Grant, Stewart, O'Connor, Blackwell, & Conrod, 2007). Cox and Klinger (1988) examined the idea of motivation as the central pathway toward the decision to drink, by attempting to map the motivational influences that eventually lead to the decision. Their model begins with "historical factors" (i.e., personality characteristics, social/ environmental influences, and biological predisposition to alcohol's effects), from which the decision proceeds to "current factors" (i.e., current emotional state along with situational variables)(p. 171). According to Cox and Klinger, the final decision to drink is filtered through the "expected effects" from the alcohol, which the researchers divide into "direct, chemical effects" and "indirect, instrumental effects" (p. 171). The net result of Cox and Klinger's model is that the decision to consume alcohol can be characterized by the motivation to either (a) directly alter one's internal emotional state via its pharmacology, or (b) to influence externally generated rewards, such as social acceptance. Cooper (1994) suggested that these characterizations could be organized within a matrix which reflects their valence (i.e., positive or negative, or whether the individual hopes to gain something positive, versus avoid something negative), and their source (i.e., internal or external, or whether the individual is drinking in response to internal, versus external cues). As illustrated in Table 1, this organization results in four distinct motivations for drinking, which Cooper called enhancement (i.e., drinking to enhance the current emotional state), social (i.e., drinking to positively affect social interaction), coping (i.e., drinking to avoid a negative emotional state) and conformity (drinking to avoid negative social outcomes). From these dimensions, a four factor scale of drinking motives was developed and subsequent testing revealed that this model better accounted for alcohol consumption patterns than previous models, with the enhancement and coping motives being the strongest predictors of alcohol use, heavy drinking and problem drinking. The model was further supported by the fact that each of the motives revealed a unique set of drinking patterns and outcomes that were consistent with expectation. For example, social motives were significantly related to consumption of alcohol at parties or with friends, but were not related to problem drinking; while coping motives were correlated with solitary consumption in patterns consistent with problem drinking. In an effort to further examine the coping dimension of drinking motives, it was later proposed that it be subdivided into coping-depression and coping-anxiety (Grant et al., 2007). In the first half of their study, Grant et al. expanded Cooper's scale to include items representative of the two constructs, and found this new five-factor scale to exceed the previous scale in predicting patterns of consumption. They were further able to demonstrate that the two new dimensions were distinct from each other, in that each predicted different patterns of drinking and outcomes. However, inconsistent with Cooper's findings, the researchers found that neither of the coping motives were related to frequency of drinking, and only coping-depression was mildly (and inconsistently) related to number of drinks per occasion. Next, Grant et al. conducted a longitudinal analysis in which they found that high coping-depression motives prospectively predicted increased numbers of drinks, while high coping-anxiety motives were found to prospectively predict decreased numbers of drinks. Interestingly, however, high coping-anxiety motives prospectively predicted more alcohol-related problems (when frequency of drinking at time-point 2 was controlled for). Though this seems to contradict logic (high coping-anxiety predicts less drinking but more problems from drinking), it is consistent with other studies (Dawson, Grant, Stinson, & Chou, 2005; Morris, Stewart, & Ham, 2005), and may be a clue to the existence of within-persons motivational variation, which isn't typically analyzed because it is assumed to be stable. Therefore, the purpose of the current research was to investigate the possibility that coping-anxiety drinking motives may vary within-persons over time, allowing for differential consumption patterns in response to anxiety. Drinking motives are defined as the self-reported reasons for drinking. Anxiety is defined as the self-reported negative affective state characterized by apprehension regarding the future. Weekly measures were completed by 125 undergraduate students at Old Dominion University which sought to ascertain how average anxiety levels interacted with anxiety coping motives to influence drinking. It was believed that the relationship between anxiety and drinking would change over time in response to variation in anxiety coping motives.