(a) Sampling designs and plans. A plan for the random cases selected for the monthly sample is generated by the Planning and Research Unit in the Office of Finance in compliance with federal guidelines, and selection is made by computer and reported to ARU by DSD. Methodologies are outlined in detail and submitted to USDHHS and USDA for approval. Sampling plans for each program present the arrangement of sampling frames, probabilities of selection, sample designs, and provisions for selection of the sample. Sampling designs and selections are provided for review and verification by federal agency staff when requested.
(b) Defined sampling frames. A sampling frame is specifically defined for each separate program.
- (1) The Medicaid active case universe includes all ABCD categorically eligible cases, and all M category cases active as of the sample month. Retroactively certified months are excluded from the sampling frame. The monthly Medicaid negative action case universe consists of all denial and closure actions occurring in, or effective for, a given month.
- (2) The Food Stamp active case universe, including A, B, C, D, and S category cases, includes all households receiving food stamps during the month via either direct coupon issuance, ATP issuance, or EBT issuance. Each participating household is represented exactly once per sampling frame. The sampling frame for negative case actions is represented by a monthly computerized listing of such actions which were either handled as the preceding month's business, considering the effective date of the action, or were taken in the sample month effective the sample month. The negative action frame is arranged by date of action.
- (3) The TANF active case regular roll sampling frame includes each case in which a benefit amount was issued for, or issued in and for, the month sampled, and is represented once. Retroactive benefits are disregarded. Cases representing reissued or canceled benefits are subject to sampling as they originally appear on the TANF register according to the date of benefit issuance. The TANF negative action case universe consists of all case closures effective the sample month, based on failure to meet TANF work requirements.
- (4) The Day Care sampling frame includes all active day care cases as of the first day of the sample month, excluding TANF, Adult, and Child Welfare involved day care cases. The Day Care negative action case universe consists of all denial actions taken in a given month plus all case closures effective that month.
(c) Sampling procedures. Systematic sampling, with random starts, is applied to all frames which are scrambled prior to random selection. Both USDHHS and USDA recognize systematic sampling as unbiased because of the arbitrary file arrangements and because no known cyclical effects exist. For all frames, each month's samples begin with random starts taken in advance of the report period from computer-generated random numbers and restricted by the estimated size of the monthly universe. DSD is notified by memo from the Planning and Research Unit at the start of each report period of random starts and intervals. Sampling designs provide for sufficient overpull, or margin, to satisfy USDHHS and USDA requirements for effective sample sizes necessary to meet the specified statistical precision and confidence levels. ARU may drop a sample case in accordance with federal policies. Reasons may include, but are not limited to, improper selection, inability to locate, or unwillingness to cooperate.
(d) Other uses for the random samples. The sample selections used for QC purposes are also used for the program needs assessment surveys of clients conducted by ARU and provided to the Planning and Research Unit for tabulation and reporting to FSSD and OFO.