Binge Drinking, Marijuana Use, and Friendships: The Relationship Between Similar and Dissimilar Usage and Friendship Quality

Abstract

While it is commonly understood that the substance use of peers influences an individual's substance use, much less is understood about the interplay between substance use and friendship quality. Using a sample of 2,148 emerging adults nested within 1,074 dyadic friendships, this study separately investigates how concordance and discordance in binge drinking and marijuana use between friends is related to each friend's perceptions of friendship quality. Because “friendship quality” is a complex construct, we employ a measure containing five sub-elements – companionship, a lack of conflict, willingness to help a friend, relationship security, and closeness. Results for both binge drinking and marijuana use reveal that individuals in friendship pairs who are concordant in their substance use perceive significantly higher perceptions of friendship quality than individuals in dyads who are dissimilar in substance use. Specifically, concordant binge drinkers estimate significantly higher levels of companionship, relationship security, and willingness to help their friend than concordant non-users, discordant users, and discordant non-users. However, the highest amount of conflict in friendships is found when both friends engage in binge drinking and marijuana use. Several interpretations of these findings are discussed. Overall, concordance between friends’ binge drinking and marijuana use appears to help some elements of friendship quality and harm others.

Keywords:

• binge drinking
• dyads
• friendship quality
• marijuana
• substance use

Notes

1. “Binge drinking” and its operationalizational methods represent an evolving understanding about the nature of alcohol use and, as such, have recently changed slightly. Within this study, binge drinking refers to heavy episodic drinking during one setting as opposed to heavy drinking over a period of weeks or months as Jellinek (1952) and others have conceptualized it. We utilize a “five or more drinks in one setting” measure for both genders since the survey utilized does not contain information that allows for the typically accepted 5/4 measure (see Wechsler & Nelson 2001), nor does it allow for a binge drinking measure accounting for body mass or drinking-episode duration (see Lange & Voas 2001). Recent work, however, suggests that BAC measurement and other techniques may not be as capable as traditional measures (such as ours) at identifying risks related to short-term alcohol use (Fillmore & Jude 2011).

2. Because individuals are nested within dyads, violations of the independence assumption can occur. We checked for this possibility using the F-star and the W-test corrections to ANOVA. These tests share in common their use of cluster (or robust) modifications to the standard errors. Results were identical to the regular ANOVAs, and thus we report the standard F-tests in the results section.