Supporting Statement

State and Local Area Integrated Telephone Survey

OMB # 0920-0406

Three-year generic clearance granted 04/25/11

Expires 04/30/14

GenIC

2013 National Survey of Children in Nonparental Care (NSCNC):

Revised Questionnaire

Prepared by:

Matthew D. Bramlett, Ph.D.

Statistician - Health

CDC/NCHS

3311 Toledo Road, Room 2111

Hyattsville, MD 20782

301-458-4070 (voice)

301-458-4035 (fax)

[email protected]

April 18, 2013

The State and Local Area Integrated Telephone Survey (SLAITS)

OMB clearance number 0920-0406

Expiration 04/30/14

GenIC: 2013 National Survey of Children in Nonparental Care (NSCNC): Revised questionnaire

In January 2013, we submitted a GenIC request to add an additional topic to the SLAITS generic clearance: the National Survey of Children in Nonparental Care (NSCNC), as a followback survey to collect additional information from selected respondents to the previous 2011-2012 National Survey of Children’s Health (NSCH). We received OMB approval to administer the NSCNC pretest questionnaire as a Computer-Assisted Telephone Interview (CATI) dress rehearsal to a sample of 50 NSCNC-eligible NSCH respondents. If the rehearsal indicated no need for questionnaire content changes, approval was also given at that time to administer that questionnaire to the full sample of remaining NSCNC eligible cases. If changes to the questionnaire content were needed, a revised questionnaire would be submitted to OMB.

The CATI dress rehearsal did indeed indicate a need for a few questionnaire content changes, to address nonvariability in some response distributions, to shorten the questionnaire by deleting some items, and to improve questionnaire flow. Since the eligible sample, estimated burden, and data collection protocols have already been approved by OMB and have not changed, this package amendment includes only the revised questionnaire and a description and justification of changes made to the questionnaire content, for brevity of review. Although the burden is unchanged, one (1) hour will be added to the burden to facilitate the submission of this request.

What follows is a brief overview of the NSCNC, and a description of questionnaire deletions and other minor changes. The revised questionnaire is included in Attachment A, the content changes are described in Attachment B, and the original questionnaire is included in Attachment C.

NSCNC OVERVIEW

The State and Local Area Integrated Telephone Survey (SLAITS) mechanism is conducted by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC), under OMB generic clearance number 0920-0406 (expires 04/30/14). This three-year clearance allows NCHS to collect health and well-being data on children, families, and communities. NCHS received approval on 1/24/2013 to add a topic module genIC to the National Survey of Children’s Health (NSCH), which fits under SLAITS. This module is a longitudinal follow up on a subsample of participants from the 2011-2012 NSCH. This 2013 follow up is referred to in this information collection request package as the National Survey of Children in Nonparental Care (NSCNC). The NSCNC module is primarily sponsored by the Office of the Assistant Secretary for Planning and Evaluation (ASPE) of the Department of Health and Human Services (HHS), with assistance from the Annie E. Casey Foundation.

The very rare subgroup of vulnerable children that the NSCNC will follow back in this module are those who were living apart from their biological or adoptive parents. All children eligible for the NSCNC module were the subjects of completed interviews in the 2011-2012 NSCH. The NSCNC data is intended to supplement the data collected during the NSCH, for these children. The NSCNC data will characterize this subpopulation and provide critical needs assessment data for this population of children who are not well served by current programs and services. The particular issues faced by households in which children are living apart from their families arise within a number of HHS programs, and indeed those of other federal agencies as well, but do not clearly fall within any single agency’s program authority.

NSCNC data collection began on April 2, 2013 with a CATI dress rehearsal that lasted 3 days. It quickly became apparent that the questionnaire was too long, necessitating some question deletion. Because the questionnaire needed to be shortened, other minor changes to improve questionnaire flow were also implemented.

JUSTIFICATION FOR QUESTIONNAIRE CHANGES

Original timing estimates for the NSCNC interview length were derived from CATI programmer testing, and indicated that the timing was roughly comparable to the desired interview length of 30 minutes. Thus, only minor deletions were anticipated to reach that target. However, the pretest indicated a considerably longer average interview length than anticipated: an average of 43 minutes, 10 seconds. We believe this is due to the particularities of this sample and the salience of the survey topic to the respondents: most of the respondents are grandparents who are caring for their grandchildren and are very interested in sharing their stories about that caregiving. Thus, many responses tend to the garrulous, greatly increasing average interview length.

Because the available funding for data collection will not permit long interviews and response rates suffer when potential respondents are informed of long interview length when consent is requested, it is critical that the interview length be shortened to an average of 30 minutes. Given that the ratio of observed interview length to desired interview length is 1.4, applying that ratio to the number of questions in the survey indicated a need to delete about 63 questions.

We first deleted questions that seemed to cause significant delays in interviews because they were cognitively difficult for respondents to answer or for interviewers to process, and questions that we had concerns about regarding their validity (based on suspicions of socially desirable answers or interviewer perceptions obtained via interviewer debriefing). We next deleted questions that showed limited variability in response options or limited analytic utility for other reasons. Third, we eliminated questions that were redundant with other items and in some cases, we combined separate questions that measured different aspects of the same topic into single questions. In one case, we replaced a previously-validated 6-item measure of depression with an alternate previously-validated 2-item measure of depression. Finally, we had to make a few deletions based on prioritizing ASPE’s data needs. In making such priority-based decisions, items that were similar to items measured in the original NSCH data collection (with the exception of living arrangements, caregiver health and insurance, and household income, which we specifically wanted to track over time) were given lower priority for retention.

In addition to these questionnaire deletions, we made a number of small changes here and there to improve questionnaire flow or reduce respondent or interviewer confusion.

The revised questionnaire is included in Attachment A. Attachment B shows the question deletions, combinations, and replacements, along with the reason for each, and also lists the other small questionnaire content changes made. In our estimation, the changes made were minor enough that the pretest cases can be retained in the final sample. For the most part, this will involve blanking the data values for questions that have been deleted, with a few other minor edits to the data records to make them consistent with the balance of the sample.

For ease of review, attachment B shows the deleted or changed questions in questionnaire order. Most changes are deletions, so the indicated questions appeared in the original questionnaire (attachment C) but not in the revised questionnaire (attachment A). In a few cases, two questions were combined into one. Attachment B indicates which of the paired questions was deleted and which was retained, although with altered question wording to account for both previous questions (for example, the first listed item in Attachment B shows that L8 was deleted and combined with L7 – new question L7 has question text that combines the text from old questions L7 and L8). Items listed in Attachment B are shown in the order they appear in the questionnaire, and the letters in the variable names correspond with separate sections of the questionnaire (for example, “L” in L7 and L8 refers to the section on Living arrangements). In Attachment B, SC means “Sample Child” and R means “Respondent.”

Page 4 of 4