Stat Computing > Seminars > Stata10

What's New in Stata 10: Survey Statistics

• Nearly double the number of procedures that work with the svy: prefix – now there are 48, previously “only” 27
  • Survival analysis, including Cox regression and parametric survival models (streg)
  • Other new additions include cloglog, glm, ivregress, ivprobit, nl, treatreg, truncreg, and zero-inflated models.
• New options in the svyset command include several methods for handling singleton PSUs, and a new option for Fay’s adjustment to BRR weights.
  • The issue of singleton PSUs (e.g., one PSU per strata) was a problem in previous versions of Stata, because when this happened, usually because of missing data or specifying a subpopulation, you didn’t get standard errors in your output. 
  • Now you have some different choices about how you would like to handle this situation.  These options include:  1) missing, which results in missing standard errors and is the default, 2) certainty, meaning that the singleton PSUs be treated as certainty PSUs; certainty PSUs are PSUs that were selected into the sample with a probability of 1 (in other words, these PSUs were “certain” to be in the sample) and do not contribute to the standard error, 3) scaled, gives a scaled version of the certainty option.  The scaling factor comes from using the average of the variances from the strata with multiple sampling units for each stratum with one PSU, and 4) centered, meaning that strata with one sampling unit are centered at the grand mean instead of the stratum mean.
• The svydes command has been renamed svydescribe; svydes continues to work.  The difference between the two commands is that svydescribe will put missing values in a new variable that you create using svydescribe when those observations are outside the estimation sample.
• The estat command has two new subcommands:  estat sd used after svy: means will give the standard deviation, and estat strata will give the number of singleton and certainty PSUs in at each sampling stage.

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services