Sex education in schools has little or no effect on when or whether teenagers have sex, use contraceptives or get pregnant, an analysis by two University of Washington physicians has found.

The analysis - a critical review of five previous studies - does not condemn sex education in the schools. It simply says there is little evidence it has any effect on behavior."The existing data suggest that a classroom course alone cannot be expected to change sexual behavior . . . in opposition to the adolescent's sexual world as molded by the television, motion picture, music and advertising industries, as well as peer group and adult role models," says the report by Dr. James Stout and Dr. Frederick Rivara.

Stout, a third-year resident in pediatrics at Children's Hospital and Medical Center in Seattle, said in an interview that studies show sex education generally does increase knowledge of the subject. But behavior is another matter.

"There is no inherent difference between sex education and other forms of education or health education," said Stout, the lead author of the study.

"Most people who smoke know that cigarettes are bad for them. They have that knowledge, but it doesn't translate into changed behavior."

Stout and Rivara, a physician and UW associate professor of pediatrics, analyzed five previous studies of sex-education programs around the country. Their conclusions were reported Friday in Pediatrics, the journal of the American Academy of Pediatrics.

One of the five studies found a slight increase in sexual activity among teenagers who had had sex education, and the other found a decrease. No effect was found in the other studies.

Two studies found very small increases in contraceptive use among some youngsters who had sex education; the others found no effect. In only one study was teen pregnancy slightly decreased among girls who had had a sex-education course.

Stout and Rivara say all five of the studies have flaws. For example, there was no control over who received the sex-education classes; students' memories of their activities may not have been accurate; and content, length and quality varied among the programs.

Better-structured, long-term studies are needed to better measure sex education's effects, the physicians stressed.

One reason for the lack of a measurable effect, they said, is that "we may be asking these programs - and our schools - to accomplish too much. . . .To place the burden of counteracting the prevailing forces in our society toward premarital sex on our schools alone is both naive and inappropriate."

Bud Turner, health curriculum specialist with the Seattle School District, agreed that schools cannot accomplish all the needed sex education.

"It's got to come from many different areas . . . from home, church, many areas, to be effective," he said.

One problem, he said, is that many parents are afraid to talk about sex with their children and say it is the schools' responsibility. And there are teachers who are uncomfortable with the subject and say it is the parents' responsibility, he said.

Stout said they examined one very effective program in a South Carolina town. Parents, teachers, ministers and community leaders all were given courses in subjects such as anatomy, contraception, ethics, self-esteem and parenting skills.

Along with the media, they were asked to supplement schools' sex-education efforts with the goal of delaying youths' first sexual intercourse and promoting contraception. The results three years later: a 35 percent decrease in teen pregnancies, compared with a 14 percent increase in communities without such a program.