This is an archived article that was published on sltrib.com in 2012, and information in the article may be outdated. It is provided only for personal research purposes and may not be reprinted.

Washington • In declaring that women's bodies can somehow prevent pregnancy in the event of a rape, Missouri Republican Rep. Todd Akin sparked a national conversation about a scientifically unsubstantiated theory, sending reporters scurrying in search of the data that would prove him right or wrong.

Surprisingly few hard facts and figures were available about the prevalence of rape-related pregnancies. Many news outlets, including this one, cited a 1996 study published in the American Journal of Obstetrics and Gynecology, which estimated that more than 32,000 women experience a rape-related pregnancy each year. The report also concluded that 5 percent of rape victims become pregnant, which would mean that 640,000 rapes occur each year.

But that figure doesn't jibe with other reputable sources. For example, FBI data show that 95,769 forcible rapes were reported in 1996. The 2005 National Crime Victimization Survey calculates that 64,080 rapes occurred in 2004 and 2005.

Because pregnancy is believed to occur with about the same frequency whether a woman was raped or engaged in consensual sex - about 5 percent of the time - the wildly different estimates in occurrences of rape have produced wildly varying estimates of the number of pregnancies that result.

Scott Berkowitz, founder of the Rape, Abuse and Incest National Network, said he relied on the National Crime Victimization Survey, an annual review conducted by the Justice Department, which he said offered "the best data" available.

"The sample size it uses is vastly bigger than any other study," Berkowitz said. He estimated, based on the data, that about 3,200 pregnancies resulted from rape each year.

But Dean G. Kilpatrick, a professor and doctor at the Medical University of South Carolina and author of the 1996 study that assumes a much higher number of rapes, said government figures vastly understated the incidences of rape, and therefore, rape-related pregnancies.

"People tend to pay a lot of attention to government statistics, but there are serious problems with the major government statistics about rape," Kilpatrick said.

There are two main sources of government data about rape: the FBI and the Justice Department.

Figures provided by the FBI only count rapes that were reported to police - Kilpatrick says his research shows that at least 80 percent of all rapes go unreported - and they don't take into account rapes in which the victim was intoxicated or otherwise unable to give consent. Until earlier this year, the FBI defined forcible rape as "the carnal knowledge of a female forcibly and against her will."

The Justice Department's National Crime Victimization Survey is a bit more complete, Kilpatrick said. The survey goes beyond crime data and interviews people by telephone or face to face. But there is so little confidence in the survey's methodology that the government has convened a Committee on National Statistics panel to compare different approaches and "develop an optimal design" for a survey to gather rape and sexual assault data.

"They wouldn't be doing the study if they were satisfied with the way they are measuring things right now," Kilpatrick said.

Kilpatrick's study from the 1990s sampled 4,008 women in a three-year survey. He said rapes were even more common today, making it possible that as many as 50,000 rape-related pregnancies occurred each year.