|
Abstract:
|
The Assisted Housing Income (Re)certification Study calls for a national sample of housing projects and tenants from three assisted housing programs. Unfortunately, different programs predominate in different localities. In the past a composite measure of size was used to select clusters, and allocations were made for each PSU so as to insure sampling the same total number of tenants for each program. For the current cycle, a two-phase sample was used to select projects from each program so that probabilities within programs remain proportional to size. While this method leads to a lower design effect due to weighting, the work load is slightly uneven across PSUs. Options, such as a two-phase sample for two of the programs and a two-stage sample equalizing the total samples by PSU, are examined through simulations.
|
