Question
For the 1986 Census PUMF - Family and Individual files AND 1981 Census PUMF - Individual file, there are no weight variables. Should these be here? Where can I go to get them?
Answer
In the 1986 documentation we found the following explanation. It turns out the '81 and '86 provides weights that are to be used for population estimates rather than to correct for sampling methodology. Unfortunately, the author division did not include a weight variable in the file so you must include it yourself.
For example, in SPSS you would include the following comand:
compute weightp=50.
In SAS:
weightp=50
Within each stratum, the sample was systematically selected with a random start and probability proportional to the first stage weight. Sample sizes within strata were determined such that each individual selected represented fifty individuals in the target population. Special procedures, described below, were required for temporary residents and overlap with other microdata files.
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.