Appendix AL.2. NASS Comments and Westat Responses

Appendix AL.2: NASS Comments and Westat Responses

#	Reviewer comment	Westat response
Supporting Statement A
1	There appears to be three separate places that the information for the study will be obtained from. The first is from the SNAP participants, the second are the FINI Grantees, and the third are the retail outlets. The universe for the areas in which the data are to be collected seems defined. However, it is unclear where all the data will be obtained. For the SNAP participants will the data be coming just from the administrative data files from the state agencies? For the grantees and the outlets where is all the data to be included in the sample going to be obtained from?	We will construct the sampling frame for the SNAP participants using the following four data sources: (1) SNAP administrative data from the State agencies, which includes SNAP household records with contact information, SNAP status and benefit, and basic demographic information; (2) ALERT data from FNS, which includes the SNAP participants’ transaction records that can be matched to the SNAP participant records by EBT card number, and to the outlet records by FNS number, (3) STARS data from FNS, which includes a list of SNAP authorized farmers markets and grocers with store type, status, and geographic information, etc., and can be matched to the outlet records by FNS number; and (4) outlet information from the FINI grantees for the FINI outlets.
2	The cluster division is talked about how it will apply to the SNAP participants but not with the grantees and outlet.	Characteristics of the outlets will drive the formation of the clusters. Using a combination of outlet type (i.e., farmers markets/farm stands and grocery stores) and incentive match rate, we will create four unique intervention outlet clusters. Comparison clusters will be defined by outlet type only 1) farmers markets and farm stands; and 2) grocery stores.
3	Will the same sample design be applied to all the separate areas in which the data will be collected? The specific sampling methods used to select the samples should be included in the documentation as well.	We will select a sample of SNAP participants from each frame using a same sampling method. That is, before the sample selection, the SNAP participant records on the frame will be sorted by State, urbanicity, and outlet ID, and then, from the sorted frame, a SNAP participant sample will be selected with equal probability systematic sampling method.
4	It is not discussed how the sampling frame for the SNAP participants will be placed into the 6 clusters.	By linking the data sources (described in the response to question 1) and by using geocoding, a list of SNAP participants living within an outlet’s catchment area can be established for each of the intervention and comparison clusters. Sampling frames for each of the 6 clusters--based on outlet type and incentive ratio match amount--will be determined as follows: Treatment clusters: we will geocode FINI outlets and SNAP households. Next, we will compute distance between SNAP households and the FINI outlets. The sampling frame will include all SNAP households within a 2-mile radius (8 miles for rural outlets) of the FINI outlet. Comparison clusters: Using data matching procedures (such as propensity scores), we will match similar SNAP households who live near similar non-FINI farmers markets/farm stands or grocery stores using data from the STARS database. We will geocode the SNAP households and compute distance between SNAP households and non-FINI outlets. The sampling frame will include all SNAP households within 2 miles in urban areas (and 8 miles in rural areas) from the nearest non-FINI retailer. For both treatment and comparison clusters, we will divide the sampling frame using EBT transaction data contained in the ALERT database. SNAP households with transactions in the FINI outlets and selected non-FINI outlets will be considered shoppers and those with no EBT transactions at the stores will be considered nonshoppers.
5	The clusters in which the study is divided into is based on the business while the participants will be people/households. In a specific area/neighborhood isn’t there the possibility of the different cluster businesses in the same area of the participants? In other words couldn’t a SNAP participant shop at a farmers market as well as a grocery store?	There is a possibility that some catchment areas may overlap with each other, especially in densely populated areas or States. In these instances, it is likely that a SNAP participant shops (i.e., has EBT transactions) at more than one FINI and/or selected non-FINI outlet. To address this potential issue, we will have a reviewing step in the frame construction process to see if there are any outlets that are close to each other and will result in an overlap. If such cases are identified, we will do a random selection among the overlapped catchment areas to keep only one in the frame.

#	Reviewer comment	Westat response
Supporting Statement B
1	Remove Westat Logo address and contact information. This data collection is sponsored by USDA/FNS	This change has been made.
2	[date of submission] – update	This change has been made.
3	follow recommendations from title page PART A	These changes have been made.
4	Delete appendice reference in PART B since these are denoted in PART A	This change has been made.
5	How were these households selected? What criteria used.	We have updated the text to include these details.
6	[sampling procedures for SNAP participants] - This needs to be reference in PART A question 2. How is this data obtain? Is this already collected? If so, reference form, OMB# and expiration date of already approved data collection that we plan to used as secondary data. Include the time in burden it will take for the agencies to gather already approved data and submit to FNS.	We plan to obtain administrative data from the SNAP eligibility system. Prior to OMB approval, we will initiate contact with SNAP agencies in selected States but will obtain the administrative files only after OMB approval.
7	Indicated whether or not these grantees were selected and awarded during the RFA process	This change has been made.
8	Where are the States in this section? Include which group contain State Agencies.	We will separate the State agencies from non-State agency grantees.
9	In the footnote, include the OMB# and expiration date when referencing a government sponsored study.	This change has been made.
10	Using what instrument? How is this information transmitted to us?	The SNAP eligibility system data already includes this information; we will obtain these data from State agencies that have FINI outlets, using a secure transfer method such as an FTP site or encrypted email.
11	Insert appendix references	This change has been made.
12	Again reference the appendices denoted in TOC PART A for these reminder notifications	This change has been made.