Skip to main content

Rape more common than smoking in the US

By January 6, 2012No Comments
The United States is experiencing an epidemic of sexual violence. New findings from the National Intimate Partner and Sexual Violence Survey (NISVS), a study launched by the Center for Disease Control and Prevention in 2010, report that nearly 1 in 5 women are estimated to have been the victims of rape, defined as unwanted completed or attempted sexual penetration, including victims who did not have the capacity to give consent (owing to intoxication, for example). In almost all cases, the perpetrator was someone the victim knew (91.9%) and more than half of the time was their own partner. Young adulthood was the period of highest risk for first sexual victimization. For 80% of female victims, first rape occurred before age 25; for 42%, before age 18.
Although the lifetime prevalence of stalking was surprisingly less than the prevalence of non-consensual sexual contact, incidence estimates for the past year suggest that this trend could be changing as new communication technologies introduce new opportunities for sexual harassment. In 2010, the study estimates that 1.27 million American woman were raped–equivalent to one woman every 29 seconds–and 5.1 million were stalked–equivalent to one woman every 7 seconds.
These are a few of the key findings from the first annual report of what will be an ongoing, nationally representative survey of sexual violence in the US. The NISVS will be a great resource to public health researchers. Using the data collected by the NISVS, investigators will be able to monitor national and state-specific trends in the prevalence of sexual violence and stalking for the first time, to characterize the type of individuals who are at the highest risk of being a victim or a perpetrator of a sexual crime (whether physical or psychological), and to investigate the health consequences of sexual victimization.
The first NISVS report has already made national headlines because of the high prevalence of forced sex that was found. To give some perspective, the study's figures, if true, indicate that number of American women who have been raped is greater than the number who are current smokers. Skeptics might question whether such statistics could be for real. After all, some surveys have been made famous for getting it wrong, like the 1948 polls that were made the laughingstock of newly elected President Harry S. Truman after they predicted that Thomas E. Dewey would win the election. A survey that makes predictions risks being ridiculed (often very publicly) but it benefits from the opportunity of self-correction. Surveys that only intend to provide a snapshot of a population never get such feedback. As a consequence, the study designers can never know whether they have managed to avoid distortion.
For survey research, coverage and response are the most important ingredients to getting an accurate picture of the population of interest. Surveys with good coverage give every member of the target population a chance to participate. Convenience samples, like polls conducted at shopping malls, are notoriously poor at coverage. When random digit dialing was introduced in the 1970s, it was a breakthrough for survey research because it largely solved the coverage problem.  At that time, 90% of the US population had a landline telephone. So, with RDD and phone interviews, the question was no longer how to get a sample with good coverage for almost any region of the States but how to do it in a cost-effective way.
With the increasing popularity of new communication technologies in the 21st centruy, there is a growing number of Americans who will never be reached through a landline. The challenge that cell phone use in particular poses for survey research has been a focus of the work of Paul J. Lavrakas, former Chief Research Methodologist for Nielsen and contributing author to Advances in Telephone Survey Methodology. In a Public Opinion Quarterly article summarizing the outcomes of several gatherings of expert panels between 2003 and 2008, Larakas and colleagues conclude that "…surveying persons reached on cell phone numbers in the United States currently is a very complex undertaking if one wants to do it 'right,' i.e. to do it legally, ethically, and in ways that optimally allocate one's finite resources to gather the highest quality data, and to analyze and interpret those data accurately.'' Though daunting, to avoid non-coverage bias, it is an undertaking that has to be faced. The Pew Research Center estimates that for 25% of current households the only phone service used is a mobile phone.
The 2010 NISVS was a telephone survey. A strength of the design is that both landline and cell phones were included. Of the 16,507 completed interviews, 9,046 of the participants were on a cell phone and 7,461 were on a landline.
Response is the more concerning piece of the NISVS methodology. Using the American Association for Public Opinion Research Response Rate 4 formula, response rate is defined as the number of fully and partially completed interviews over all eligible cases, which is the sum total of cases with completed interviews, refusals, non-contacts, other spoken language than that used for the survey, and a correction factor for cases of unknown eligibility. The NISVS response rate was an underwhelming 27.5%. The majority of non-response was due to non-contacts. Although this rate is actually better than average for national telephone surveys, it still raises that question of whether the experience of sexual violence, stalking, and intimate partner violence is different among those Americans in the 72.5% of the sample who were either away from their phone or who looked at the unknown number coming in and chose not to answer.
If non-participation were a random event (a doubtful but much hoped for circumstance for survey researchers), then the NISVS participants would still provide a nationally representative sample of Americans. However, there is evidence that chance was not entirely indiscriminate when selecting the 27.5% of responders when one compares the NISVS sample and the US population on several socio-economic characteristics. In contrast to the US population, participants in the NISVS were more likely to have a college or advanced degree (36.4% versus 29.6%), more likely to be divorced (12.9% versus 10.3%), more likely to never have been married (30.2% and 26.1%), and more likely to have an annual household income that was below the federal poverty level (19.6% versus 12.1%).
The NISVS has an important public health message but it is unclear to which public it applies.

Leave a Reply

Significance Magazine