Wednesday, August 31, 2016

CCHS - weighting of optional questions


I am looking at the CCHS survey questions. I have a good understanding about the sampling technique followed and the weighting process undertaken for the responses achieved off of the core content questions.

But I am not sure about the optional content questions. These questions are optional and by optional it means that they have been selectively chosen by some provinces interested. So for example, the question about insurance covering part or all prescription medication, was an optional content question, that was included only in the surveys for the Ontario, New Brunswick and Yukon in cycles 2013 and 2014.

I understand how the respondents were chosen in these provinces. But I don’t know how I can generalize these responses i.e weigh them Canada wide or at lease provincial wide.


The CCHS 2013-2014 User Guide, on p. 8, it states:
It should be noted that, unlike the modules included in the common content, the resulting data from the optional content modules is not easily generalized across Canada.
Any analysis on optional content should be done using the master files. Optional content is not included on sub-sample or rapid response files (with the exception of when it is used as theme). Even if it were, the records would be a subset from the master anyways, so it would limit analysis to look at anything but a master.

Optional content is not the best name. We should call it sub-national content, meaning that it is not available for producing estimates at the Canada level. For example, HUI was only selected in Manitoba in 2012. So with 2012 data, you can only produce estimates for Manitoba. In 2013-2014, HUI was used as theme content, meaning that it is collected nationally and thus can be used to produce Canada level estimates. For 2012 though, there is nothing that can be done to make inference about HUI estimates in Canada based solely on the data from Manitoba. In some years you can get close with some of the content – food security, for example, is an optional module that gets selected by most of the provinces in the years it is not theme. For those years, you can be close to having national estimates, but it would still be underrepresented by the provinces that did not select the content.

If the client wants to send us the content they are interested in, we can write up a list of which years and which provinces they can create estimates for. A good number of the optional modules have been used as theme at some point, so there is likely that there will be a year for which national estimates can be produced.

Police Administration Survey


I have a researcher would like data for personnel and expenditure by municipality for the years 2013-2015 from the Police Administration Survey. Is this available? Would it be a custom tabulation?


The data is available for police service, not by municipality. While many municipalities do have standalone police services they are not necessarily mapped directly to municipal boundaries.

Data at the national level: Police resources in Canada, 2015


Police resources in Canada, 20154


Police resources in Canada, 2013


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.