Because we depend upon survey statistics to describe attributes of the current caseload, it is important that the sample we use is a good representation of the population. Using data from our administrative database, we compared survey respondents and non-respondents on selected demographic characteristics and used Chi-square tests to test the significance level of the differences.
|Education of Case Head|
|Ethnicity of Case Head|
|Age of Case Head|
|20 or Under||12.6||11.5||2.9||0.413|
|Number of Children|
|Months on AFDC/TANF in the Past 10 Years|
|Have UI Wages April-June 2002|
We found that the respondents and non-respondents are similar in the type of geographical area in which they resided, their years of education, age, number of children in the case, past use of AFDC/TANF, and UI wages earnings during the quarter which included the sample month (June 2002).
The non-respondents were more likely to be white (or non-black minorities including Hispanics, American Indians and Asians) and less likely to be blacks. Statistically, the difference in race composition of respondents and non-respondents is not significant. Therefore, we conclude that nonresponse bias is not a problem in this study, and that the sampled cases that completed the survey adequately represent the study population.