3
$\begingroup$

I'm working with IPUMS ACS and Census data in Stata. I'm interested in learning about income distributions and variability for specific subpopulations defined by education level, occupation, race and sex.

The survey data has probability weights, but I want to use commands that don't accept probability weights. They will accept frequency weights, though.

I want to estimate: Kernel density distribution Mean Coefficient of Variation Standard Deviation Percentiles

Can I just use frequency weights instead of probability weights? These seem to give me accurate point estimates for the mean, but standard errors that are too small.

Will that throw off the kernel density, standard deviation, and coefficient of variation?

$\endgroup$
1
$\begingroup$

mean command with pweight gives you mean and sd estimates, which in turn gives you estimate of the coefficient of variation.

pctile also takes pweight. It will generate percentiles.

kdensity only gives point estimates, not confidence intervals of the density estimates, so I think using fweight instead of pweight is fine. But I'm not certain about this.

Alternatively, svy: commands may make your life easier when working with survey data. See here for a quick intro.

Also, think through whether you want to describe your sample, or to make inference about the population. See here for an example. An relevant quote from the link:

It was intentional that summarize does not allow pweights. summarize’s purpose, as I see it, is to provide descriptive statistics for the sample, not to provide inferential statistics for the population. By this criterion, I argue that pweights do not belong here since pweights are used to provide estimates of the population parameter mu.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.