Churning - Responses to OMB Comments 2 7 13

Responses to OMB Comments

February 7, 2013

Understanding the Rates, Causes, and Costs of Churning in the

Supplemental Nutrition Assistance Program (SNAP)

This document was prepared in response to OMB comments following a briefing held on February 4, 2013.

Responses by FNS are provided in red.

1. We understand that the Project Team has chosen 6 states from which to draw respondents to the qualitative aspect of their study. From these 6 states, it seems that 6 sites overall will be selected, and from those sites, 10 focus group members will be recruited (allowing for some nonresponse in the latter).

1. In the proposal, it seems that a site may be selected in each state, or they may be concentrated in fewer states.

There will be one site selected within each of the 6 States recruited for the study.

2. Please clarify the criteria that will be used to select sites and, from each site, focus group members. This can be a general discussion if some of the selection criteria are still under discussion with the contractor and the states, but it would be helpful to include as much detail as possible at this point.

Site Selection: Sites will be selected based on the same criteria used in the selection of States and will best reflect the variations of policy and operation environments from the State selection process. Selection criteria includes technological capabilities, waivers, involvement of CBOs, rural/urban status, and length of recertification period. Site selection will be done through a collaborative process between the contractor and the States.

Focus Group Selection: Focus group participants will be selected from State administrative caseload data of identified recent churners. All focus group participants must be at least 18 years of age and identified as head of households. Other specific details of the selection criteria are still under discussion between the contractor and State. In sites where there are significant Spanish-speaking populations, one focus group will be conducted in English and one will be in Spanish. In others sites, groups may be divided based on household characteristics related to churn (e.g., sources of income, previous length of time in SNAP).

3. Please state more explicitly the criteria by which CBO representatives, administrators, and caseworkers will be selected among participating states. Please indicate how the selection process may affect your interpretation of the resulting data.

Administrator Selection: Administrators will be selected by virtue of their role in the administration of the SNAP program. Ideally, the administrators would hold positions typically titled State SNAP Director, Local Site SNAP Director, Local Site Caseworker Supervisor for SNAP. Although these titles may vary by State and local office, the contractor will work to identify the person performing the functions typically performed by administrators in those roles.

CBO Representatives: The contractor will attempt to document the CBOs working in the local site and their role in SNAP. Through discussions with the local SNAP administrator, the contractor will select the CBOs and representatives from these CBOs based on level of direct engagement with SNAP and knowledge of the content areas addressed by the interview questions. The contractor will document reasons provided by the SNAP administrator for their recommendations.

Caseworkers: Caseworkers will be selected on the advisement of the local office Director of SNAP based on time availability and organizational structure of SNAP operations.

These interviews will be conducted as part of the qualitative component of the study. Due to timing and resource constraints, these interpretations are understood to be suggestive and will be reflected as such in the language in the reports.

2. Please clarify how and when focus group participants are informed of the $30 incentive. (NASS’ comment.) Generally, we would expect this information to be conveyed at the time of recruitment and at the initiation of the focus group activity.

Focus group participants will be informed of the incentive at the time of recruitment. This is provided for in the scripts of both the introductory recruitment phone call as well as the reminder phone call.

3. Please describe in greater detail the selection of the one state for linked unemployment earnings. If the selection and availability of these data are not clear, please state that.

On the advisement of USDA ERS researchers who conducted a similar study into linked data analysis, consideration was given to States that participated in the ERS study since they had existing data sharing arrangements in place between Universities and States. Of the 14 States identified as potential candidates during the initial stages of this study, 5 were also participants on the ERS study. Of these 5 ERS states, Florida was the only State to be recruited to participate in this study. Florida has a data sharing arrangement with the University of Missouri, which has a contract pending with the contractor for this study. FNS does not expect to accept delivery of any actual data from this linked data analysis, and will only receive tabulated results to be incorporated in our reporting and analysis.

4. Generally, impact analyses would involve the use of control groups and randomized controlled trials. Observational analyses with exposure-outcome focus would minimally involve longitudinal data. When discussing analyses of multi-year administrative data without a control group, it may be more appropriate to describe the analysis as examining factors associated with the outcomes of interest, rather than impact. When discussing analyses of qualitative data collected at one time point, interpretations of impact would be suggestive only.

We agree and use of language in all reports will carefully adhere to the suggestions above.

5. Where confidentiality is referenced, please reference either the Privacy Act (if invoked) “responses will be kept private under the Privacy Act” or replace “confidential” with “secure to the fullest extent permitted by law.”

All references to “confidential” were revised as recommended.

6. Should the Project Team have the opportunity to analyze administrative records prior to launch of the full qualitative component, we would welcome seeing a non-substantive change to the instruments reflecting what they have learned from the initial quantitative analysis. From our discussion yesterday it sounds like changes to instruments based on admin data are not anticipated given timing and resource constraints, but wanted to reiterate this suggestion in case circumstances change.

Due to timing and resource constraints, the qualitative and quantitative components of the study will be conducted largely concurrently. However, if time and resource constraints change, we will consider the option as recommended.