Students' perception of how much their peers are drinking is one of the most important predictors of whether their alcohol use will increase in college -- even more so than the actual drinking habits of their randomly assigned roommate, according to a new study by researchers at Virginia Commonwealth University.
The study, "Genes,Roommates, and Residence Halls: A Multidimensional Study of the Role of PeerDrinking on College Students' Alcohol Use," relied on a genetically informed sample of 755 randomly assigned college roommates to investigate the degree to which several factors would predict if a student would be more likely to increase their alcohol use during college.
It found that students were likely to drink more if they had a genetic predisposition to alcohol use, if their roommate drank heavily and if they perceived friends around them to be drinking frequently. Of these factors, the students' perception of peer drinking habits was the most influential.
"Of the things that we looked at in this study -- genetic predisposition, your roommate's drinking, and how much you think the people around you are drinking -- all three of those are important. They all independently predict future substance use," said Danielle M. Dick, Ph.D., a VCU professor in the Department of Psychology in the College of Humanities and Sciences and in the Department of Human and Molecular Genetics in the School of Medicine. "But the most important thing is how much you think people are drinking. It's not your predisposition, it's not how much your roommate is actually drinking, it's how much you think people are drinking."
The study will be published in a forthcoming issue of the journal Alcoholism: Clinical and Experimental Research and was led by Rebecca Smith, a doctoral student in the Department of Psychology.
"The big thing we found was that both how much you think your friends are drinking and how much your roommate reports drinking, both influence if you increase your drinking over time. So both of them matter," Smith said. "Interestingly, we found that perceptions -- how much you think your friends are drinking -- matters most. That was more important than how much your roommate is actually drinking."
The study is based on data collected through VCU's Spit for Science project, which aims to understand why some people are more likely to develop problems associated with substance use and emotional health by examining how individual genetic predispositions come together with environmental factors to contribute to these outcomes.
As part of the Spit for Science project, VCU student volunteers fill out surveys examining a long list of topics, including living situation; personality; family history and childhood upbringing; experience of stressful life events; social support; involvement in extracurricular activities; alcohol and other drug use; mental health related problems, such as depression, anxiety and eating disorders; caffeine use; friends' behavior; nicotine use; and religiosity.
The student participants also have an opportunity to provide a saliva sample, which gives researchers the ability to analyze students' survey results to explore how genetic and environmental factors contribute to the development of problems associated with the use of alcohol and other substances, and emotional health.
For this study, the researchers used anonymized Spit for Science data -- no student's identity was accessible to them -- to explore the survey results and genetics of randomly assigned roommates living in campus residence halls.
"This study addresses the really interesting question of what happens to students' drinking when they go to college. Our idea was: Let's use the natural experiment that exists when you get assigned to a random person that you are going to live with, which happens to a good number of college students," Dick said. "And, because we have the Spit for Science data that so many of our students have completed, we are able to match those survey responses to the people who were randomly assigned to each other. So that gives us both each individual's report of their substance use -- so we can look at what the effect is of a randomly assigned roommate on subsequent substance use in that individual."
As part of the study, the researchers also investigated whether the residence hall in which a student lives is a predictor of whether they will drink more.
"This was done on the back end -- we didn't see any of the identifying information -- but we match survey responses for students who were assigned to the same room, and who were in the same residence hall on campus," Smith said. "We then calculated an average score [showing] how much, on average, the people in [each] residence hall were drinking."
With that average score, they were able to compare that to individual residents' self-reported drinking habits.
"If you live in a residence hall where a lot of people drank, does that influence your own drinking?" Smith said. "We actually did not find that to be the case. So even if you lived in a residence hall with people who drank more alcohol, that did not predict whether or not your alcohol [consumption] would increase or decrease."
The researchers believe residence halls were not a major predictor of student alcohol use because of the design of VCU residence halls, which can be many stories tall and lead to fewer interactions among residents on different floors. Residence halls at other colleges, where the dorms have more students per floor, might see a different effect, Smith said.
The study's key finding -- that students' perception of peer drinking is an important predictor of their own drinking behavior -- could be useful for colleges' efforts to curtail student drinking.
At VCU, for example, the university's Wellness Resource Center places flyers, called the "Stall Seat Journal," in public restrooms across campus to convey survey results about how much VCU students actually are drinking. The idea is that by correcting students' over-estimates of how much others are drinking, the students' alcohol consumption will be lowered and harm will be reduced.
"We know that, for example, the vast majority of people are not ending up drunk in the bushes on Friday night and having no idea how they got home. But who do you talk about and remember and chat about in Starbucks on Sunday morning? It's not, 'Hey, I was studying at the library last night and I got a ton done and I'm really set for classes on Monday morning.' It's 'Oh my gosh. Did you see Jim passed out in the bushes?'" Dick said. "What that does is it creates this illusion that everybody's out there drinking, but we know that that's actually not the reality."
Smith, a former social worker who earned a master's degree from the VCU School of Social Work and who worked in psychiatric hospitals, has long been interested in substance use and mental health, and this study reflects the important influences of genetics and environmental factors.
"It's the person and it's the social context," she said. "The biological, psychological and social factors around you all influence what behaviors you engage in and the kind of outcomes you experience. By having Spit for Science, this huge data set gives us lots of great information. And so, with this study, [it shows] how peers ... can play such key roles in the lives of adolescents and college students."
