Attachment K_State Selection Memo

ATTACHMENT K

STATE SELECTION MEMORANDUM

State Selection Memorandum

Final Submission

February 1, 2017

Introduction

We recommend ten States¹ be invited to participate in the Evaluation of Alternatives to Improve Elderly Access to Supplemental Nutrition Assistance Program (SNAP). The United States Department of Agriculture’s Food and Nutrition Service (FNS) commissioned this multi-year study to better understand how to maximize elder (age 60+) access to SNAP. The study has four key components: (1) an exploratory study to lay the groundwork for and guide the rest of the project; (2) a study of State interventions, based primarily on interviews with State SNAP administrators and staff from the selected study States; (3) a study of elder participant perspectives, based on interviews and focus groups with elder SNAP applicants, recipients, and eligible non-participants in the same States; and (4) a quantitative analysis of the impact of various interventions designed to increase elder SNAP access and reduce churn. This memorandum explains the criteria that were used to select potential study States and presents a list of these States.

Criteria for State Selection

The choice of study States is an important decision for the study because all the data collection activities for three of the four study components outlined above will take place within the sampled States. In the research proposal, the study team recommended a selection of States that captures both variation in the types and number of interventions implemented and variation in State characteristics.

The study team began by identifying a set of criteria to be used in selecting States for the study. We divided the criteria into three major groups: adoption of interventions aimed at increasing SNAP enrollment for the elderly; program participation rates; and program characteristics.

Adoption of Interventions

In this first category, we identified eight strategies that are aimed specifically at increasing the enrollment of elders in SNAP, or have the potential to significantly affect elder participation:

• Having an Elderly Simplified Application Project (ESAP)

• Setting a standard medical deduction (SMD) amount

• Participating in a Community Partnership Interview Demonstration (CPID)

• Participating in one of the data matching projects under the coordination of the Benefits Data Trust

• Participating in a Combined Application Project (CAP)

• Participating in a 36-Month Certification Demonstration

• Having a 36-month certification or recertification interview waiver.

Appendix A, attached to this memo, summarizes the current knowledge on adoption of interventions across States. SMD and CAP are the most popular interventions in this group, with 19 and 17 States, respectively, having adopted them. ESAP and recertification interview waiver (with eight States each) are the next most popular interventions. All other interventions have been adopted by five or fewer States. The Appendix shows that 13 States did not adopt any interventions, 20 States adopted one intervention, and 18 States adopted at least two interventions. In the selection process, we prioritized for inclusion States that have adopted at least two interventions because one of the main goals of the overall study is to assess how multiple interventions work in tandem. However, we wanted to select a few States that adopted only one intervention so that we could assess that intervention’s individual effect. It was necessary, as well, to ensure that each of the interventions had been adopted by at least one State in the sample.

To be sure, the types of interventions outlined above and in Appendix A are not the only ones in use among the States. A wide range of other interventions—such as allowing SNAP applicants or recipients to conduct certification or recertification interviews by phone instead of in person, offering the option to apply to the program online, and the availability of a call center to answer questions and facilitate program application—exist to facilitate increased enrollment and/or to decrease churning. However, the research team wanted to ensure that the States selected for the study primarily reflect variation in the set of eight interventions listed above because, as argued above, these interventions are deemed particularly relevant for the participation of elders.

Participation Rates

The second group of selection criteria consisted of variation in participation rates. We distinguished two criteria in this category: the level of participation rate (high or low) and the trend in participation rate (increase over time as opposed to remaining constant or getting lower). States vary considerably on both dimensions. As shown in Appendix A, the most recently available data (FY 2014) show participation rates ranging from a high of 72 percent in New York to a low of 21 percent in Wyoming. Similarly, participation rate trends between FY 2010 and FY 2014 vary from an increase of 18 percentage points in Maryland to a 1 percentage point decrease in Arkansas. The study team felt that it was important to choose States with all four possible combinations of participation characteristics (high participation trending higher, high participation staying constant, low participation trending higher, and low participation staying constant) in order to ensure that the outcome to be explained (SNAP enrollment) has a high degree of internal variation that supports meaningful research conclusions.

Program Characteristics

The third category consisted of a single criterion: whether the program is run the State level or at the county level. We wanted to include States from both categories of this criterion.

Other Considerations

The research team felt that after all the criteria set out above were considered, the resulting choices had to be balanced with respect to geographic distribution and state size. Additionally, we considered the proposals for State selections we received from the FNS National Office during the interviews that were conducted in late 2016 for the purpose of drafting the Exploratory Memo. These suggestions largely coincided with the ones received from representatives of regional offices and national intermediaries during interviews held in late 2016.

State Selections

Based on prior experience in working on similar projects, we anticipate the possibility that some States may opt not to participate in the study. For this reason, we have selected six alternate States in addition to 10 primary study States. The six alternates will be considered in the case of a higher-than-anticipated number of refusals from the primary States. We will select alternates that best match the State(s) that opted out with respect to the key selection criteria.

Exhibit 1. Proposed Study States

Primary Selections		Alternate Selections
Alabama	Arkansas	Arizona	Colorado
Florida	Idaho	Maryland	North Dakota
Massachusetts	Minnesota	South Carolina	Texas
Nebraska	New York
Pennsylvania	Washington

Appendix A: State Characteristics and State Selection

		Interventions									Participation		Administration
	State	ESAP	SMD	CPID	Benefits Data Trust	Increased Benefit Amount	CAP	36-Month Certification	Recert. Interview Waiver	Total Interventions	Participation Rates	Trend FY 2010-2014	State-run	County-run
	Alabama									2	30.4	3.7	
	Alaska									0	37.5	10.1	
	Arizona									1	24.7	-1.0	
	Arkansas									2	23.5	-1.0	
	California									0	21.3	12.3		
	Colorado									2	37.6	11.3		
	Connecticut									1	49.9	12.2	
	Delaware									0	31.0	8.3	
	District of Columbia									1	43.0	8.9	
	Florida									3	55.7	12.9	
	Georgia									2	36.1	4.8	
	Hawaii									0	54.7	17.6	
	Idaho									1	34.9	8.7	
	Illinois									0	48.2	16.5	
	Indiana									0	33.1	4.5	
	Iowa									1	36.6	9.3	
	Kansas									2	32.4	6.7	
	Kentucky									1	35.0	-1.1	
	Louisiana									2	29.8	1.5	
	Maine									0	52.0	5.5	
	Maryland									4	43.0	18.1	
	Massachusetts									3	62.9	16.5	
	Michigan									1	45.1	6.5	
	Minnesota									1	41.1	13.8		
	Mississippi									2	32.9	5.3	
	Missouri									2	39.0	3.7	
	Montana									0	35.5	9.3	
	Nebraska									1	32.4	7.4	
	Nevada									1	34.7	8.1	
	New Hampshire									1	41.5	13.7	
	New Jersey									1	49.3	15.7		
	New Mexico									1	36.8	8.0	
	New York									2	71.6	17.0		
	North Carolina									1	34.4	8.4		
	North Dakota									1	33.3	0.2		
	Ohio									0	39.3	8.8		
	Oklahoma									1	28.0	-1.2	
	Oregon									1	62.9	15.8	
	Pennsylvania									3	41.7	5.7	
	Rhode Island									1	58.7	18.2	
	South Carolina									4	31.2	2.6	
	South Dakota									2	37.2	4.0	
	Tennessee									0	46.8	7.6	
	Texas									4	37.3	5.7	
	Utah									0	25.6	4.6	
	Vermont									1	67.8	14.0	
	Virginia									2	30.9	1.4		
	Washington				(pending)					2	55.0	14.5	
	West Virginia									0	34.2	0.3	
	Wisconsin									0	45.8	15.6		
	Wyoming									1	20.6	1.5	

Key		Primary Selection
		Alternate Selection

1 The project proposal stated that nine to twelve states would be invited to participate in the study. The decision to invite ten states was judged as optimal in balancing the requirement to capture as much State variation as possible while ensuring that each study State would receive the required level of attention and resources.