Characterizing the role of affect in substance use in daily life: a novel application of latent profile analysis for daily diary data
Restricted (Penn State Only)
- Author:
- Van Doren, Natalia
- Graduate Program:
- Psychology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- May 16, 2022
- Committee Members:
- Robert Roeser, Outside Unit & Field Member
Rina Eiden, Major Field Member
Michelle Newman, Major Field Member
Jose Soto, Co-Chair & Dissertation Advisor
Ashley Linden-Carmichael, Co-Chair & Dissertation Advisor
Bethany C. Bray, Special Member
Kristin Buss (she/her), Program Head/Chair - Keywords:
- emotion
substance use
latent class anaylsis
heavy episodic drinking - Abstract:
- Emotion plays a critical role in the development and maintenance of problematic substance use. Extant research on emotion and substance use in college students frequently focuses on mean levels of affect, neglecting to examine how more complex, daily patterns of affect can impact substance use. This may be important as emotions differ on other dimensions beyond mean level of positive and negative affect (i.e., valence), such as their level of arousal (i.e., how activated or deactivated one feels). In addition, emotions within each valence represent discrete emotional states (e.g., anger vs. sadness), which have been linked to distinct behavioral signatures. Each of these aspects of emotional experience may be important for substance use, but are often examined in isolation, precluding the examination of how different aspects of emotion cluster together at the level of occasions (e.g., days) to predict substance use. Such an approach may more accurately capture emotional experience at the daily level and has the potential to improve the prediction of substance use in daily life, thereby aiding in intervention efforts. The goal of the present study was to extend knowledge of the effects of emotion on substance use in daily life by applying a novel methodological approach that can identify how latent profiles of affect may differentially contribute to substance use. The specific aims were to: 1) identify patterns of affect at the day level; 2) examine the links between day-level patterns of affect and substance use (probability and level of alcohol and cannabis use, respectively); and 3) examine gender as a person-level predictor of daily affect profiles. Drawing from a sample of 154 college students with daily diary data for up to 14 days, I used a mixed-indicator latent profile analysis (LPA) to identify and describe affective patterns across all person-days, allowing for differences in affective patterns across days within-persons (Aim 1). Next, I used daily latent profile membership to predict daily substance use, taking into account the non-independence of observations across days (Aim 2). Finally, I fit a mixed-indicator LPA with covariates using a generalized estimating equations approach to examine gender as a person-level predictor of the day-level affective profiles identified in Aim 1 (Aim 3). Results revealed five distinct day-level latent profiles of emotion: undifferentiated negative emotion days, undifferentiated positive emotion days, high arousal positive days, mixed emotion days, and low reactivity days. Moreover, day-level profiles were, in turn, differentially linked to probability of alcohol use, such that undifferentiated positive emotion days, high arousal positive emotion days, and low reactivity days were linked to significantly greater odds of using alcohol compared to undifferentiated negative emotion days. In addition, high arousal positive days were linked to significantly heavier alcohol use compared to undifferentiated negative emotion days. Contrary to hypotheses, there were no effects of profile membership on probability of cannabis use or levels of cannabis use, and there were no gender differences in profile membership. Together, findings suggest that daily emotional patterns may be meaningfully associated with daily alcohol use and level of use, such that experiencing positive affect in the absence of negative affect may be particularly risky for alcohol use, while high arousal positive states may have stronger associations with greater levels of alcohol use. Strengths, limitation, future directions, and implications for intervention are discussed.