How should we interpret the "preliminary" label on full-count data?


#1

Michelle Pratt’s helpful answer to my last question (Avoid occupation info in preliminary full-count data?) raises a more general question in my mind about what constitutes appropriate use of the preliminary full-count data sets.

“Preliminary” could mean “these data sets are missing some variables but the variables currently made available are ready for use.” Or it could mean “the available variables are significantly less trustworthy than those in data sets not marked ‘preliminary’.” Or some mix of those.

I’m just trying to figure out how comfortably I should lean on the current full-count data sets, while recognizing that all such data sets are to some extend perpetual works in progress. I’m using STATEFIP, COUNTY, SEA, PERWT, SLWT, SEX, AGE, RACE, BPL, SCHOOL, HIGRADE, OCC1950, CLASSWKR, INCWAGE, INCNONWG, OCCSCORE, SEI, and the derived OCCSCORE_MOM, OCCSCORE_POP, OCCSCORE_MOM2, OCCSCORE_POP2.


#2

In the context of the full count samples, “preliminary” means that there is only a limited set of variables available. All the variables currently available are ready for use. Acknowledging that there are always improvements being made to the data, the variables you’ve listed should all be trustworthy.