Potential effects of sample size expansion in CPS: March 2000 to March 2001


From 2000 to 2001, the number of respondents in the March Supplement increases. The increase is due to the use of the SCHIP file rather than the original March supplement.

I am looking at educational attainment by citizenship status (by year). Accordingly, I have divided 18-60 year old individuals into the following educational attainment categories: High school diploma or less, some college but without a baccalaureate degree, and Baccalaureate degree or more.

I’d like to know if and how the expansion of the sample affects the proportions on educational attainment for either group. Are the samples before and after 2001 really comparable (with respect to educational attainment)? I’d appreciate very much your answers and references on published papers or forum discussions relevant to my question.


Shoumi Mustafa

I don’t know of any published papers specifically looking at the effect of the sample expansion on educational attainment. However, the Census Bureau investigated this and the effect of changes in population controls on estimates of educational attainment. Their findings can be found here.