Families on TANF in South Carolina: Employment Assets and Liabilities. Appendix A. Sample Weighting

10/01/2004

The sample for the study consisted of 1,493 families who were on TANF in South Carolina in June 2002 and who were “mandatory” for work participation. The sample was stratified to include three groups of families, as follows:
 
  • clients who had not been given a temporary work exemption or an extension of their 24 monthtime limit and had fewer than 24 months of welfare history (Stratum 1);
  • clients who were mandatory for work but who had been granted a temporary exemption from work requirements (Stratum 2); and
  • clients who were mandatory for work but who had been granted extensions of the state’s 24-month time limit (Stratum 3).
The sample of 1,493 cases was selected from a statewide caseload of 11,002 cases in which clients were mandatory for work participation in May 2002. When the sample was selected, there were 290 cases (among the caseload of 11,002 cases) for which it was not possible to classify into a stratum, based upon admin istrative records information. After the survey was completed, however, the status of all but 74 of the 290 cases was determined.
 
Exhibit 1 shows the sample sizes and sample weights. For cases in Stratum 1, we initially selected 579 cases from the known caseload of 8,108 cases in this stratum. For cases in Stratum 2, we initially selected 639 cases from the known caseload of 2,407 cases. For cases in Stratum 3, we initially selected all 197 cases in the caseload. From the 290 cases where we could not initially determine the case’s stratum, we selected 78 for the sample.
 
After the surveys were completed, we found that 185 of the 290 cases whose status was initially unknown were, in fact, Stratum 1 cases. Of these 185 cases, 43 had been selected for the sample. In addition, we found that 31 of the 290 cases whose status was initially unknown were, in fact, Stratum 2 cases. Of these 31 cases, we had selected 6 for the sample.
 
None of the 290 cases whose status was initially known were later found to be Stratum 3 cases. Finally, in 74 of the 290 cases whose status was initially unknown, we still could not determine the stratum to which they should be assigned. Of these 74 cases, 29 had been selected for the sample.
 
For analyses involving the whole sample, it was necessary to develop six sample weights, as shown in Section A of Exhibit 1. These weights were used to adjust for the fact that the six types of cases had a different probability of being selected for the sample. In addition, we had to develop special sets of weights for analyses in which we looked only at individual strata, as illustrated in Sections B, C, and D of Exhibit 1.
 
Exhibit 1
Sample Weights Used in the Data Analysis
 
Sample Strata
Stratum Known When Sample Selected?
Population Size
Included in Sample
Weight Applied
A. Analyses of the Whole 
Stratum 1 Yes 8,108 579 1.9003
No 185 43 0.5838
Stratum 2 Yes 2,407 639 0.5112
No 31 6 0.7011
Stratum 3 Yes 197 197 0.1357
No 0 0 N/a
Total   11,002 1,493  
 
B. Analyses of Stratum 1 Cases Only Yes 8,108 579 1.0503
No 185 43 0.3227
 
C. Analyses of Stratum 2 Cases Only Yes 2,407 639 0.9966
No 31 6 1.3669
 
D. Analyses of Stratum 3 Cases Only Yes 197 197 1.000
No 0 0 -

 

View full report

Preview
Download

"emp-assets-liab-223E.pdf" (pdf, 977.84Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®