Wednesday, August 17, 2016

Applied Research in Post-Secondary Institutions


I have a researcher looking for information of applied research divisions of post-secondary institutions in Canada. Specifically, he would like to know the overall impact these departments have on their greater community including:

- Number of students/faculty/staff hired in these divisions

- Number of community partnerships/commercial business involved

- Operating costs

- Profits, if any


Statistics Canada (2009), 'Estimates of Research and Development Expenditures in the Higher Education Sector, 2007/2008', in Science Statistics, Catalogue no. 88-001-X, vol. 33, no. 5, Ottawa: Statistics Canada.

CANSIM Tables of potential relevance,3580026,3580001

External links
Council of Ministers of Education, Canada
Canadian Education Association

· Canadian Association of University Business Officers (CAUBO)

· The Applied Research and Commercialization Initiative- FedDev Ontario

Some Statcan Resources


I'm working with Census data from 1986-2006, and with the NHS 2011. I am trying to calculate mortality rates using the death counts by CSD in the Vital Statistics, and with the population counts by CSD in the short form 100% population census. In order to compute these rates, I need to match the CSDs that appear in the Vital Stats and in the Censuses, and I've been having a lot of trouble with this, as it appears there are more CSDs in the Vital Stats than in the Census. The only reasons I can think of as to why this might be occurring are: 1. that the codes have changed over the years, and 2. that some CSDs have been suppressed in the Census (whether they were incompletely enumerated reserves, or of poor data quality). I am using the concordance tables on Stat Can's website to track the changes in codes over the years, but this only accounts for a few hundred of the few thousand unmatched CSDs over the period of time I'm interested in. From this I have two questions:

What I really need for each census year (for short form & long form questions) is a list of the suppressed CSDs, showing the reason for the suppression, similar to what is available for 2011 NHS (


Unfortunately, we did not publish online any suppression lists prior to 2006. For 2006, we only published the names (no reasons for the suppression). In 2011, for both the Census and the NHS both the names and the reasons were published. Here are the links to these lists.

NHS Census subdivisions (CSD) not released

2011 Census subdivision (CSD) suppression list

2006 Census subdivision suppression list with names - 100% data

2006 Census subdivision suppression list with names - 20% sample data

Wednesday, August 3, 2016

PALS 2006


I have a question regarding the Participation and Activity Limitation Survey, 2006. I downloaded the PALS data, looking for the variable "ADFT_Q01 " (regarding autism developmental disability) , but it is not in the data set. Do you know why?


To collect information on adults whose everyday activities are limited because of a condition or health problem, the PALS questionnaire identifies ten types of disabilities. However, for reasons of confidentiality related to the PUMF, four types of disabilities were grouped as “Other.” As a result, the typology provided to PUMF users contains seven types of disabilities, namely hearing, seeing, speech, mobility, agility, pain and “other” disabilities. The PALS 2006 Technical and Methodological report contains a description of this typology for adults.

PALS 2006 Technical and Methodological report link:

Variable in PUMP file: OTH_LIM (Other Limitation Derived Variable (Learning, Memory, Development, Psychological, Unknown))

A note of caution in the use of the developmental variable as a measure of autism, as autism is only of many developmental disabilities represented by this variable.

Included are links to more recent information that may be useful.

Developmental disabilities among Canadians aged 15 years and older, 2012 Released February 29th 2016

115-0003: Adults with disabilities, by type, sex and age group, for Canada, provinces and territories.

115-0004: Adults with disabilities, by severity, sex and age group, Canada, provinces and territories.

Monday, July 11, 2016

Maternal Deaths


I am helping a researcher who is looking for statistics on maternal death from abortions in the 1930's and 1940's. She has found some information for B.C. (from B.C. Vital Statistics), but she is interested in similar statistics for Ontario, or for Canada as a whole.

While searching the StatsCan site, I found the following notice. StatsCan has data from 1970 to 2006, and CIHI now provides access to this data:

I guess the real question is which agency (or agencies) would have data for the 1930s and 1940s. Any advice would be appreciated.

There are some digital collections of the Ontario, Office of the Registrar General – Vital Statistics made available online at the Internet Archive:

The chances of obtaining reliable figures would be somewhere in the vicinity of nil owing to the fact that abortion was prohibited until the Criminal Law Amendment Act of 1968-69. You might actually be better off exploring crime statistics to see how many doctors and others were charged (one famous case being

There is a very good chapter (Chapter 2, “Abortion as Birth Control”, pp. 32-53) in the following book which outlines many of the issues influencing the availability and reliability of statistics on maternal deaths due to abortion in Canadian history. 
McLaren, Angus, and Arlene Tigar McLaren. The Bedroom and the State: The Changing Practices of Contraception and Abortion in Canada, 1880-1997. (2nd ed.) Toronto: Oxford University Press, 1997.
Here are two snippets of info from the book that might be useful:
“In a 1934 study of 334 maternal deaths in one year in Ontario, an important finding was unearthed. Researchers found that fifty-nine, or 17 percent of all maternal deaths, were due to abortion.” (Phair & Sellers, 1934, as cited in McLaren & McLaren, 1997, p.45)
Original research article cited:
J.T. Phair and A.H. Sellers, “A Study of Maternal Deaths in the Province of Ontario.”Canadian Public Health Journal, 25 (1934), pp. 563-79.
“It is not easy … to pin down the number of abortion deaths. How a maternal death was classified depended ultimately on the differing judgements and conflicting concerns of doctors, coroners, and magistrates. There are indications that only two of every three abortion deaths were reported by vital statistics and presumably even a lower ratio of deaths known to medical authorities may have come to the attention of legal authorities.” (p. 45)
“We have established that in the three decades between 1920 and 1950 probably close to 500 women in British Columbia died as a result of abortion-related deaths.” (p. 51)
Citing their own research article:
McLaren, Angus, and Arlene Tigar McLaren. “Discoveries and Dissimulations: The Impact of Abortion Deaths on Maternal Mortality in British Columbia.” BC Studies, 64 (1984/85), pp. 3-26. Available online from:

Labour Market Forecast Data


We’ve got a request for labour market forecast data for curator positions in Canada. The researcher basically wants the Canadian equivalent of these US data. Does this exist anywhere?


Employment and Social Development Canada (formerly HRSDC) maintains the Canadian Occupational Projection System, which I believe is what your user is looking for in this case:

Monday, July 4, 2016

CCHS 2012 - Mental Health variable, Household Income


My researcher initially asked if a 15-year-old could have been responsible for answering questions about household income, but I'm wondering, more broadly, if someone who is under 18 or even someone who is not considered one of the heads of the household could be responsible for answering household questions.

It started with this question, "I am using the Canadian Community Health Survey - Mental Health (2012; CCHS-MH, 2012) data set, and I have a question. The age of respondents begins at 15, and goes to 90 and above. I note in the user guide, it says no proxy interviews were permitted. I am wondering, for the income variables ( ex. total income in the household), were 15 year olds responsible for providing this?"

After some back-and-forth, it turns out that she's interested in people aged 15-29, and wondered why she had so many of them seem to be in a higher-than-expected income group (using household income). I replied that although she's able to pull out the age group in question doesn't mean they were living alone (i.e., every member of the household influences household income) and I have suggested she might want to consider the individual income variable. I'm waiting to hear from her to see if there is any follow-up to that suggestion.

So barring anything unexpected, I think I've actually found a solution for her research (not that hard, I admit), but a general question remains unanswered in my mind: is there an age-cutoff for answering the household questions? I looked at the questionnaire and the data dictionary, and see that when someone starts answering the questions they are rejected if they are under 15. I also see that in section 5.4 of the User's Guide, "The interview," there is a statement indicating that "a knowledgeable household member was asked to supply basic demographic information on all residents of the dwelling. One member of the household aged 15 years or older was then selected for a more in-depth interview, which is referred to as the C2 interview." In 5.4 e) Proxy interviews, the UG says that no proxy interview were permitted. This, I believe, is where my researcher's confusion stems.

Can someone confirm for me that the income questions would have been asked of the person most knowledgeable of the household? I feel like a dork for even asking, since it seems pretty obvious (I've read the Income Imputation section, btw), but I'd rather be safe than sorry.


This pertains to CCHS Mental Health, so this response does not apply to the Annual.

As stated in the User Guide, the beginning of the survey is done by a knowledgeable person in the household. The interviewer is instructed to ask for a knowledgeable person when they first make contact with the household. This person completes the household roster so we get the most accurate information on name, age, sex, & marital status (the “basic demographic information”) of all household members. Then we do the selection of the survey respondent. The respondent does the rest of the survey themselves and income is part of that. Therefore, income is asked to 15-year-old respondents. If they don’t feel qualified to answer the question, they do have the option to say “Don’t know” or “Refuse”, as with any other question on the survey. Income was imputed when there was a non-response.

Monday, June 27, 2016

Households and Environment (2013) residential heating fuels


I have a researcher looking at the Households and the Environments survey (2013). Is it possible to get a breakdown of the approximately 3100 "non-CMA" responses for question EH_Q02 A through G by Census Division or Dissemination Area?


Analysis at the Dissemination Area (DA) level is not possible. We do not have DAs attached to records on our master file so there would have to be some cost-recoverable work undertaken to achieve this. However, we would recommend against this as DAs are generally small geographic units, which means it is extremely unlikely that there would be many, if any, DAs with a sufficient number of records to be releasable, assuming that the CVs would meet our release criteria. Given that the area of interest is non-CMA Ontario, I would be surprised if any met this criteria.

Analysis at the Census Division (CD) level is, in theory, possible. Constraining the analysis to only non-CMA households in Ontario will result in many CDs having variables that will be flagged as “use with caution” or unreleasable due to our data quality guidelines. However, this is work that we could undertake on a cost-recoverable basis with the understanding that the costs would be payable by the client regardless of the number of releasable values in the results. Thus, I would not advise this as it could be quite costly and not produce satisfactory results. Still, if there is interest in pursuing this, please let me know and I will prepare an estimate.

If the researcher is eligible to use the RDC, they could undertake the analysis at the CD level themselves. The 2013 HES microdata in the RDCs contains the necessary geographic variables and we could provide some advice on how to work with the non-CMA households in Ontario.

Finally, the 2013 release ( talks about how households heat their homes, stating that “those from Ontario and in the West mostly opt for forced air furnaces burning natural gas”. In particular, it notes that the forced air furnaces were the dominant type of heating system in Ontario (73% of households) and that, “[i]n general, natural gas is available to most households from Ontario to British Columbia”. I know this is not as specific as the researcher is interested in, but he (or she) may find it helpful.