STATS ARTICLES 2012
Sexual Assault and Statistical Confusion in the US Military
S. Robert Lichter, Ph.D., July 30, 2013
Is there a crisis of sexual assault in America’s armed forces? You would certainly think so from a Department of Defense survey that generated strong media buzz ever since it was released in May. The report gave scientific credibility and political heft to concerns raised by a series of heavily covered cases of sexual assault and the military’s apparent inability to either prevent such attacks or punish the offenders.
For example, CNN reported that “the number of service members anonymously reporting a sexual assault grew by more than 30% in the past two years. The Defense Department estimated that more than 26,000 troops experienced an episode of ‘unwanted sexual contact,’ a huge jump from 19,300 in the 2010 report.”
And the Baltimore Sun declared, “A crisis is the best way to describe it. According to a survey released earlier this year, the number of sexual assaults taking place in the U.S. military has risen sharply, from about 19,000 in 2010 to about 26,000 last year…. What's shocking is not simply the volume of crime but the lack of reporting by the victims, most of whom are women.”
These alarming and widely quoted figures come from the Department of Defense’s “2012 Workplace and Gender Relations Survey of Active Duty Members,” which is part of its “Annual Report on Sexual Assault in the Military.” The report contains data on surveys administered in 2006 and 2010, as well as 2012.
The survey has not been without critics. Most prominently, Marine Corps Captain Lindsay Rodman authored a May 20 Wall Street Journal op-ed, which argued “the 26,000 figure is such bad math – derived from an unscientific sample – that no conclusions can be drawn from it.” Captain Rodman also criticized the survey for defining “sexual assault” as any “unwanted sexual contact,” including “unwanted touching” of a woman’s breast or buttocks. And the report shows that unwanted touching accounts for one out of three reported assaults.
There is “definition inflation” here, which is common in surveys that attempt to call public attention to a social problem. The larger the number of respondents who are classified as victims, the more attention the study gains. This is not to suggest that unwanted sexual touching is trivial or excusable. However, the survey could have avoided confusion by asking respondents directly if they had been sexually assaulted.
Captain Rodman’s further claim that the survey is unrepresentative and “unscientific” is harder to gauge. Some unlikely extrapolations from the survey that she mentions are discussed in detail below. According to the report’s methodology section, over 108,000 individuals were surveyed, and about one in five returned completed questionnaires. The report does not directly state that the sample was randomly based, which would obviously be necessary to generalize from the survey to all active duty personnel. Some weighting of the results was clearly done to bring the sample into line with the actual distribution of males and females in the military.
If we take the data at face value, though, the often-cited finding that sexual assaults increased from 19,300 in 2010 to 26,000 in 2012 raises troubling questions. These numbers are extrapolated from the survey’s finding that 6.1 percent of women and 1.2 percent of men surveyed reported unwanted sexual contact, up from 4.4 percent of women and 0.9 percent of men in 2010 – an overall increase of 35 percent.
However, the report also includes findings from the 2006 survey, ignored by the media, which found that 6.8 percent of women and 1.8 percent of men reported unwanted sexual contact. The “2013 DoD Sexual Assault Prevention and Reporting Strategic Plan” extrapolates from this percentage an estimated total of 34,200 service members – 32 percent higher than the much-quoted 2012 total, and an astonishing 77 percent higher than the 2010 total.
Thus, if we extrapolate from these surveys to the entire military population, sexual assaults first dropped sharply from 2006 to 2010 and then rose sharply from 2010 to 2012. So any crisis of sexual assault in the military, it seems to have appeared, abated, and then reappeared – all in the course of six years. And although assaults are more frequent now than they were in 2010, they are still 24 percent below 2006 levels.
Moreover, in light of these apparently dramatic changes, it is surprising to learn from the report that the numerical differences are rarely statistically significant. Thus, the overall increase in assaults from 2010 to 2012 was statistically significant for women, which means it was unlikely to be accounted for by sampling error.
However, none of the individual subcategories – which included unwanted touching, attempted or completed vaginal, oral, and anal sex, and unspecified contact – increased to a statistically significant degree. Further, there were no statistically significant differences in the number of sexual assaults, either overall or by subcategories, between 2012 and 2006. All this suggests that the sole significant finding of an increase in unwanted contact among women from 2010 to 2012 is not robust and should be interpreted with caution.
The report’s graphic illustration of these data also includes annual figures on the number of actual reports of sexual assault to the DoD. Of course these numbers are much smaller, as sexual assault is notoriously underreported. But notice the trendline, which is slowly and gradually moving upward, increasing 15 percent from 2006 to 2010 and another 11 percent by 2012.
We cannot know to what degree this increase reflects higher rates of reporting by victims, as the problem gained public visibility and the Pentagon was increasingly forced to respond to it. But such a gradual year-by-year increase stands in sharp contrast with the dramatic rise and fall shown by the survey results.
Finally, extrapolations from the survey findings lead to another unlikely conclusion. Virtually the entire discussion in the media has concerned female victims of sexual assault while ignoring males. But the survey paints a different picture.
According to Defense Dept. figures, in 2010 there were 1,429,995 active duty military personnel, of whom 210,485 (14.7 percent) were female and 1,219,510 were male. (These numbers are relatively stable over time. For example, in 2013 CNN reported that active duty personnel totaled 1,468,364, of whom 14.6 percent were female.) Based on these figures, the 6.1 percent of females surveyed who reported unwanted sexual activity would total 12,840 women. The 1.2 percent of males reporting such activity would total 14,634 men.
Thus, if the survey is correct, it seems to show that more males than females are victimized by sexual assaults than females. This would contradict the Baltimore Sun’s statement that “most of [the victims] are women,” along with (as she notes) Captain Rodman’s statement that “empirically males do not constitute anywhere near the majority of victims of unwanted sexual contact.” To be fair, she might regard this as validating her argument that “no conclusions” can be drawn from a survey based on “such bad math.”
S. Robert Lichter, Ph.D., July 30, 2013