A limited study of five NCAA Division I-A schools suggests that more than a quarter of the nation's football players may have used steroids at least once.

A professor in Penn State's college of health and human development said Monday his study sets an "upper limit" of actual anabolic steroid use among college athletes, whereas an NCAA study done last year set a lower limit."The mean overall projected rate of use of anabolic steroids across all sports surveyed was 14.7 percent for male athletes and 5.9 percent of females," Professor Charles Yesalis said.

Yesalis, who worked with researchers from Penn State and five other schools, said the study shows that steroid use is more widespread than earlier thought. Last year's NCAA study, based on self-reporting, estimated steroid use at less than 5 percent.

"For years, a number of researchers have believed that with self-reporting there's an underreporting of use," Yesalis said. "This represents an upper bound, and for the first time there are boundaries and somewhere in between is reality," Yesalis said.

Athletes at five schools that agreed to participate in the study were asked to estimate steroid use among their opponents. Even so, Yesalis said the athletes said a lot about their own programs.

Previous studies have shown that a significant amount of steroid use is tied to athletes' perception that their opponents are using them - in other words - "I only use them because he uses them."

"It's like going to expert informants on what's going on," Yesalis said. "It lets you project your behavior on others without endangering yourself. The literature does say that people who use drugs see more use around them."

The director of the NCAA's drug testing program said the self-reporting study probably did not detect all use and that college sports' largest governing body gave credence to the Yesalis report.

"Perceptions are very important in this area," said Frank Uryasz, the NCAA's director of sports sciences. "Athletes tell us if they were sure the guy across the line wasn't using steroids, they wouldn't.

"They still believe they're competing against dirty athletes," Uryasz said.

The more recent study was conducted at five institutions across the country. The schools were guaranteed anonymity. The researchers were from Michigan State, Ohio State, Penn State, UCLA, Vanderbilt and West Virginia.

The study projected steroid use at 29.3 percent for football teams, 20.6 percent among men's track and field teams and 16.3 percent for track and field.

More athletes also perceived steroid use to be increasing in college athletic programs, Yesalis said.

Rick Albrecht, a research assistant at Michigan State's office of medical education, cautioned that Yesalis' questions differed from those in the 1989 NCAA steroid study conducted by Michigan State.

The NCAA study asked about steroid use in the last year. Yesalis' asked about steroid use at any time.

"I think he's correct in assuming that we have a bottom boundary here and that actual use is somewhat higher - but how much higher? That's the $64,000 question," Albrecht said.

For Uryasz, though, it doesn't matter what the specific numbers are.

"I'm not sure it's important to get at the number," Uryasz said. "If there's any use, there's a problem. I don't care if its 1 percent, 5 percent or 50 percent."