WASHINGTON, Jan. 4 (AP)--Just saying no to sex works, at least for a while. A study found that teenagers who publicly pledged to remain virgins until marriage delayed having sex about 18 months longer than other teens.

Among those who formally promised to avoid unmarried sex, about 50 percent remained virgins until about age 20, said Peter S. Bearman, a Columbia University sociologist and co-author of a study in the American Journal of Sociology. Among nonpledgers, he said, 50 percent were no longer virgins by age 17. "The average delay among pledgers is 18 months," Bearman said in an interview Wednesday. "That is significant. And that is a pure pledge effect."

Bearman and his co-author, Hannah Brueckner, a sociologist at Yale University, analyzed the effect of virginity pledges on the behavior of teenagers enrolled in the National Longitudinal Study of Adolescent Health, a federally funded survey of children in the seventh through 12th grades. Data from the study suggested that by 1995 a church-led, voluntary effort had prompted about 2.5 million teenage boys and girls to make spoken or written pledges to refrain from sex until marriage.

Bearman and Brueckner analyzed data from interviews of 20,000 teenage virgins in 1994 and 1995. A follow-up survey in 1997 included 14,000 of those in the original study. Bearman acknowledged that some of characteristics that would lead kids to make the abstinence pledge would have led them to avoid early sex anyway. But even when the researchers made statistical allowance for these factors, "there was an additional delay added by the pledge," he said. "The more religious kids pledge, as do kids who are more oriented toward school," said Bearman. "Those are protective effects that would delay their entry into sex anyway. But the pledge effect is in addition to that."

The virginity pledges seemed to have the greatest effect on those who took the oath at age 16 to 17, while there was little effect found for those pledging at 18 or older. The effect of pledging earlier, at age 14 or 15, depended on the students' social environments, the study found. Ironically, Bearman said the more popular taking the pledge becomes in a school, the less effective it becomes. The maximum effect, he said, comes when about 30 percent of students in a school pledge. As the total reaches 40 and 50 percent, the pledge effect is eroded.

In effect, he said, as soon as the virginity pledge becomes too common "and something that everybody does," the persuasive social impact is lost. For the whole study group, Bearman cited a "significant" difference between the percentage of pledgers and nonpledgers who remained virgin through their teen years.

For instance:

Among 15-year-olds, 82 percent of the nonpledgers were virgins, compared with 90 percent of the pledgers. At 16, 68 percent of nonpledgers had abstained, while 79 percent of the pledgers were still virgin.

At age 17, half the nonpledgers had initiated sex, while 65 percent of the pledgers were still virgin. At 18, 32 percent of nonpledgers were virgins, compared with 54 percent of the pledgers.

"This takes into account all the other characteristics of the pledgers," said Bearman. "This is a pure pledge effect."

Bearman said he doesn't know how many of the teens actually fulfilled their pledge to refrain until marriage. That may become clear in a follow-up study planned next year.

J. Richard Udry of the University of North Carolina, director of the national study from which the virginity data was extracted, said the results surprised him. "It doesn't seem like an intervention like that would have an effect on sexual behavior, and a rather big one," he said.

Another part of the study found that when pledgers break their promise and first engage in sex, they are more likely to do so without using a contraceptive. Bearman said this is not surprising, since it would be illogical for a student both to pledge abstinence and to carry a contraceptive.