'm looking at the 2014 ASEC SPMpoverty measure.

A colleague mentioned that the 2014 ASEC involved partial implementation of some new question protocols, and that as a result I might not be able to look at SPMpoverty measures for the full sample.

In terms of SPMPOV in particular, can I run the data for the full sample, or do I need to control for HFLAG=0?

For the runs I’m doing, the results do vary slightly (by a few percentage points), depending upon whether i use the full sample or the 5/8 sample.

I’d rather use the full sample to have bigger n, but just not sure if that is appropriate, given whatever questionnaire changes were implemented that year.

I should also say–I am *not* trending the data at all, so that is not a concern for me; I’m just trying to get a 2014 snapshot.

TIA for any help,

Gretchen Livingston

The Census Bureau fielded the CPS 2014 ASEC sample using an experimental redesign. 5/8ths of the sample were given income questions consistent with prior ASEC surveys, while the other 3/8ths were given redesigned income questions. The Appendix Table of “The Supplemental Poverty Measure: 2014” report shows a side-by-side comparison of traditional and redesigned SPM poverty rates, including confidence intervals. Ultimately, whether or not to combine the 2014 samples is left to the discretion of the researcher.

Hope this helps.