PALS 2001

Question

A student who wants to use the PALS 2001 survey has asked me where or if there is a designation for province of residence. She could not find a variable relating to it. I have looked through the
questionnaire, where I saw that province information was collected at the beginning and the end of it. Using our access to IDLS, however, I do not see either a direct or derived variable relating to province. Have we overlooked something?

Looking at the user's guide, there is no province variable listed in the data dictionary.

Section 6.5, which talks about non-disclosure, mentions that "the level of detail of the pumf is not as fine as that of the master file ... Actions were also taken to make the microdata file more secure ... These actions concern the geography included in the file, ..." Is this the first PUMF where geography is completely suppressed?

Answer

It was decided that the richness of this survey was within the variables rather than the geography. A conscious decision was made to put as much detail as possible in the file
therefore sacrificing geography.

Unfortunately variable details that were released in previous versions of surveys do not always get approved for release with the current version. There could be a number of reasons for this including sample size, response rate, complexity and level of detail of questions asked etc. Most of all there are increased confidentiality concerns in all of our surveys. In fact I am of the opinion that previous surveys would not contain the same level of detail now if they were brought before the Microdata Release Committee.

Each time a survey manager makes changes to a public use file and brings it before the Release Committee requires the expenditure of a lot of time and resources. At this point in time it would be beyond the means of the budget to try and add even one additional variable to a released file.
Even one variable requires the manager to go through the entire release process - which could cost close to $50,000 and take months to do. Once a file is released, unless it is to fix an errors in the data, the actual data file is rarely altered.