On any given weekend, at least 10 percent of students at a single college could be hosting a party, and on average, party hosts who live off campus are drinking more and engaging in more alcohol-related problem behaviors than are the students attending their bashes, research suggests.

In contrast, hosts of parties held on campus tend to drink less than do the students attending their gatherings, according to the study.

The research also suggests that college party hosts are more likely than the students attending parties to be male, living off campus, members of a Greek organization and in their second year or higher of college, and have more money to spend each month than other students.

The results come from an online survey of 3,796 students over the course of two academic years.

The findings could guide efforts by university personnel to curb excessive drinking at college parties, researchers say.

"Party hosts set the context for the attendees. They decide what kind of drinks are going to be there and how many people are going to attend," said Cynthia Buettner, assistant professor of human development and family science at Ohio State University and lead author of the study. "So if you could get people to think about hosting a party in a particular way, you could reduce the risks for the people who attend."

The study is published in a recent issue of the journal Addictive Behaviors.

Buettner used data from a larger study of campus drinking behavior to zero in on the activities of party hosts. To her knowledge, hers is the first study to examine how the behavior of party hosts differs from that of the students who attend college gatherings.

"It's all in the name of intervention. The more information you have, the better able you are to target prevention efforts," Buettner said.

The researchers contacted a random sample of registered students by e-mail, asking them to report on their alcohol use during eight different weekends from 2005 to 2007. If they reported in the online survey that they had attended or hosted a party, they were eligible for this study.

Of the 3,796 participants, 433 - or more than 12 percent - had hosted weekend parties. "It's not a small group," Buettner said. "That finding alone surprised us."

About 80 percent of the parties reported in the survey had been held at off-campus locations. The average number of guests at parties attended by survey respondents ranged from 25 to 60.

Off-campus party hosts consumed an average of almost nine drinks, compared to the 7 ½ drinks consumed by party guests. On campus, the trend was reversed: Party hosts reported drinking an average of about 4 ½ drinks, compared to the 7 ½ drinks consumed by attendees. The total range of drinks consumed spanned from zero to 30, according to the survey.

Off-campus party hosts were more likely to participate in problem behaviors associated with drinking than were attendees at any party and on-campus party hosts. These included verbal arguments, public urination, flashing or mooning, vandalism to the party location or to nearby property, rioting, physically fighting, driving after drinking and riding with someone who had been drinking.

On the other hand, hosts of parties held on campus were less likely than party attendees at either type of location to observe risky drinking and related consequences. These behaviors included heavy drinking, underage drinking, unwanted sexual advances, verbal arguments, physical assault, public urination, flashing or mooning, vandalism or spontaneous rioting.

Though the questionnaires weren't designed to pursue more details about these outcomes, the researchers said the finding that hosts of on-campus parties drink less than their guests is probably associated with the risks of getting in trouble with the university.

"It's logical to think that off-campus party hosts would be more likely to drink a lot. They know they're not going to drive, they're home and they probably started before everyone arrived. Our theory is that on-campus party hosts may be worried about potential sanctions," Buettner said.

The findings could be used to influence intervention efforts on college campuses, the researchers say. For example, beyond advising students to "party smart," potential off-campus party hosts could be informed of their increased risk for heavy drinking.

"I'd be willing to bet, though we wouldn't know until we did the research, that there is a group of students who tend to be the host over and over again. This gives you a group of students for whom a very particular type of intervention would be helpful," Buettner said.