Justification for Change

Request for Non-substantive change to the

National Intimate Partner and Sexual Violence Surveillance System Survey

(NISVSS)

OMB #0920-0822

The National Center for Injury Prevention and Control requests to change the NISVSS race question to include the category “other, specify” to question RC4 (page 12) and R8 (page 52). This Non-Substantive change request does not include changes to the currently approved burden and/or costs.

The reasons for the request to include the “other, specify” race category include:

1. NISVSS will employ race in the post-stratification of the sample weights. Population totals for race categories are obtained from the census estimates. A mismatch error, potentially leading to bias and increased variance due to weighting could occur if the race questions are different in NISVSS and in the 2010 U.S. Census. It is imperative that the questions used in post-stratification are the same in the survey and in the benchmark (the 2010 U.S. census).

2. Omitting the “Other, specify” response option in the NISVSS race question would lead to more respondents not providing any response (coded as “Don’t Know” or “Refused”) or be forced into a less applicable race category. Those who do not provide a response would need to be imputed, so that someone who would respond “Other, specify” to the census question would be categorized into a race category such as “Black or African-American” in the survey. Those who select a race category may tend to select some options more than others, altering the racial distribution. Providing an “Other, specify” option as in the census would alleviate these sources of bias and variance.

3. While some respondents do not select any of the race categories that are provided to them, we have found in NISVSS that they often provide open-ended responses to “other, specify” that fall into one or more of the race categories. Thus, another use of this open-ended category is to code correct race categories for some respondents, further improving the classification of respondents in weighting adjustments.

4. In addition to the 2010 U.S. Census, numerous other large national surveys employ an “Other, specify” response option in the race question. Examples include , such as the National Survey on Drug Use and Health, the Behavior Risk Factor Surveillance System, and the National Health Interview Survey.

5. Without the inclusion of “other, specify” the current NISVSS will produce flawed race/ethnicity data that seriously undermines the ability to weight the data properly and the ability to produce valid national estimates and cross talk with other national surveys.