CT-level Census data for 1961, 1971, and 1981

Question

REQUEST to add CT-level Census data for 1961, 1971, and 1981 to E-STAT.

A Faculty Member in our History Department is looking for an effective, straight forward, non-time consuming, and uncomplicated way, for his upper year undergrad history students (Historical Research Methods), to have access to the historical census data in geographic contexts. Here is an excerpt from the Prof's email to me:

"I am a professor in the history department, and I would like to use historical census data for a methods class that I will be teaching and research next semester. I would like students to be able to select variables and download their own census tables for analysis. I was hoping to use data with geographic specificity, particularly the tract-level data that is available to our students via the CHASS server at the University of Toronto (Canadian Census Analyzer). Unfortunately, the [historical] data available through CHASS are not readily readable in SPSS, the software that I would like students to use. Instead, the command files are missing important information (likely they are quite old) and they generate files that need a good deal of processing before they can be used. Do you know whether we can access this same data through some other service? "

In subsequent, very helpful, discussions with Chuck Humphrey, this idea came forward: "... a request could be made of E-STAT to add CT-level Census data for 1961, 1971, and 1981. "

How best to submit this request? Would there be general support/interest, in the DLI community, regarding this E-STAT development request?

Answers and Responses

1. Thank you for the excellent suggestion. We are aware this data is missing from E-STAT and we are looking at this matter. We will let you know of our decision. Do not hesitate to contact the Learning Resources team (education@statcan.ca) should you have any questions.

2. First of all, I should clarify that the Census profile files that are available via the CHASS interface for the 1961 and 1971 Censuses are not Stats Can products at all.

One of our graduate students here at UT took and aggregated the 1961 from the EA-level basic summary tables, and produced the files that are available in the CHASS interface. For 1961, the file is not strictly speaking a profile file, but is all characteristics that were in the original 1961 EA-level files, with the exception of any averages (means, medians, etc) that we could not recreate from the original EA files, because we lacked either the numerator, the denominator, or both in those files. However, since these files for the most part are either uni- or at most bi-variate distributions, we decided that they were more profile-like than not. We have done some spot checking to make sure we were getting the same numbers as the print profiles, but have checked by no means all of the CTs in Canada.

For the 1971 profile file, we extracted from the many 1971 CT-level basic summary tapes those characteristics that most closely approximated the print profile product from that census. Like with 1971, we did some spot checking against the print publications, but nothing like a comprehensive check. One or two variables we would have liked to include we had to drop because we couldn’t recreate what was in the print publications for them (housing value was I believe one of them).

In other words, for those two profile files, Statistics Canada may be the author of the content, but UofT Data Library Service is responsible for the edition. Whether it is appropriate for STC to put these into E-stat, or instead recreate the work we did and produce their own editions of these profiles, is a question on which there might be several points of view.

The 1981 census profile was the first machine-readable census profile file that I am aware of Statistics Canada having produced. I see no problems therefore in STC putting those profiles in E-stat.