Submission of burden changes to the Rape and Sexual Assault Pilot Study

MEMORANDUM TO: Rochelle W. Martinez

Desk Officer

Office of Statistical Science and Policy

Office of Management and Budget

THROUGH: Lynn Murray

Clearance Officer

Justice Management Division

FROM: Allen J. Beck, Ph.D.

Senior Statistical Advisor

Bureau of Justice Statistics

SUBJECT: Submission of burden changes to the Rape and Sexual Assault Pilot Study

DATE: August 5, 2014

The purpose of this memorandum is to request an increase of 29.5 hours in the burden authorized for the Methodological Research to Support the NCVS: Self-Report Data on Rape and Sexual Assault: Pilot Study (OMB #1121-0343). The increase in burden results from the need to conduct further testing of initial assumptions related to recruiting respondents through random digit dialing (RDD).

The testing will finalize the parameters for the cell-phone component for the Rape and Sexual Assault (RSA) Pilot study. The additional testing is necessary because the initial feasibility study indicated eligibility rates that were significantly different from our initial assumptions related to recruiting RDD respondents for the cellphone sample. The proportion of respondents that completed a screening interview that were female was 30%. Our initial estimate was that this should be closer to 50%. In addition, we had estimated that approximately 95% of the cell phone sample would reside within one of the five metro areas. This was closer to 80% in the feasibility.

The differences in expected versus actual rates are large enough to necessitate conducting a small, second, feasibility test to better approximate the rates that will be expected on the larger Pilot test. We intend to test four changes to the procedures:

1. Increase the number of callbacks and to conduct refusal conversion. The first feasibility test was designed to test the operational procedures for the two interview modes (in-person; telephone) and the three different sample types of General Population, Volunteer and Service Provider. The protocols were not set up to maximize response rates, so there was very little follow-up of telephone numbers that did not respond to initial attempts to contact the household.

The second feasibility study will call back all cases, multiple times, using the same protocol to be used on the Pilot survey. The analysis of refusals suggests that females were less likely than males to answer the telephone when called. By calling back multiple times, the second feasibility test will ascertain whether it is possible to increase the number of female respondents who answer the screening items. If successful, the test will provide an estimate of how much of an increase can be expected during the Pilot test.

In addition, respondents that refuse to complete the screener will be contacted with a second request to participate. This second request will be made approximately 2 weeks after the initial refusal and will be administered by a specially trained refusal conversion interviewer. Follow-up contacts will not be made if the respondent exhibits hostility at the first call.

2. Shorten the initial introduction. In response to comments by the interviewers, the new introduction now first tells the respondent the agency that is sponsoring the call and then describes the study:

Hello, my name is {DISPLAY D1}. I'm calling on behalf of the U.S. Department of Justice. Please let me know if you’re currently driving a car or doing any activity that requires your full attention, so I can call you back at a later time.

[confirm speaking to an adult and personal use of cell phone]

We are doing a scientific study on health and safety. I am calling to see if you are eligible to take part in this study. If you are selected, you will be asked questions about health and safety and will be sent $20 in appreciation for your time after you complete the study. It will take about 2 minutes to see if you qualify.

3. Change what is displayed on the caller-id on the cell phone. In the first feasibility test, a local Maryland number was displayed (area code 301). We will switch to displaying an 800 number based on discussions with CDC which uses a similar protocol in administering the NISVS survey.

4. Attempt to code the gender of the voice on voice mail messages. Interviewers will be trained to code whether the name provided on the voice mail message is clearly female, clearly male, or cannot be determined. They will also attempt to code whether the voice used to record the voice mail message is clearly female, clearly male, or cannot be determined. This information will serve two purposes. One, it will allow us to understand whether females send the call to voice mail at a different rate than do males, and two, it will allow us to be more targeted with refusal conversion efforts, such that known males will not need to be exposed to refusal conversions.

The respondent universe of the second test comprise individuals in the non-institutionalized population that use cell phones. The sample will consist of 2,000 randomly generated cell phone numbers, which will result in approximately 1,400 numbers after purging the sample of inactive numbers. We anticipate this will result in 309 completed screening interviews where eligibility is determined (Table 1).

This should provide enough data to set the parameters for the pilot. The percent of screened respondents will have a confidence interval of +/- 5%. This will provide the study with enough accuracy to set the sample size for the cell phone sample.

Table 1. Sample size and projected number of interviews for the second feasibility test

Total Numbers	2,000
Flagged as Not Working before calling	600	30%
Calling	1400
Not Working or Non-residential	70	5%
Residential Status Undetermined	448	32%
Determined Residential	882	63%
Completing a screener	309	35%
Age eligible (18+)	247	80%
Gender eligible	124	50%
Location eligible (within 5 SMSA’s)	117	95%
Complete Main Interview	70	60%

The test will administer a reduced version of the NSHS interview. (See Attachment T.) Since the goal of this test is to estimate the eligibility rate, there isn’t a need to administer the main portion of the NSHS interview. The proposed interview will include the initial demographic section and skip directly to the two vignettes. We have modified the debriefing section to ask about the respondent’s use of caller ID. It will not ask any of the victimization questions. By deleting the main part of the NSHS interview, burden, costs and the time needed to complete the feasibility test are significantly reduced.

The consent statement will contain two significant changes. First is a change to the description of burden, estimating 12 minutes for completion rather than 20. The second change is that we have deleted the statement about breaking confidentiality because of committing suicide or harming someone else. Because we do not ask any personal experiences related to unwanted sexual activity, this statement should not be necessary.

The estimated burden and cost to respondents associated with the second feasibility study is shown in Table 2. The estimate of time to complete each component is based on the first feasibility test. A total of 29.5 hours of respondent burden is estimated for the test at a cost of $707.

Table 2. Estimates of Respondent Burden and Costs to Respondents for Second Feasibility Study

	Burden estimate	# of Completes	Total Burden	Cost to respondents+
Determine eligibility	3 minutes	309	15.5 hours	$371.69
Main Interview	12 minutes	70	14 hours	$335.72
Total			29.5 hours	$707.41

+ Based on the average hourly earnings of $23.98 hour¹ for private nonfarm payrolls (Source: http://www.bls.gov/web/empsit/ceseesummary.htm).

If there are any questions concerning this request, please contact Allen J. Beck, Senior Statistical Advisor, Bureau of Justice Statistics, at (202) 616-3277 or by email at [email protected].

1