Monday, March 16, 2009
This is one of those "seems like it should be obvious" questions, but...
A patron has asked me what percent of women in different age cohorts
(specifically, in their 40s, 50s, 60s, 70s at some specific point) over
time have given birth to at least one child. Percent who have never given
birth would also be acceptable, of course. I've been having a surprising
amount of difficulty with this. The SLID has a question on number of
children ever, and I think the 1993 person file PUMF includes variables
that will let me compute an estimate, but I was wondering if anyone had
any better or more recent suggestions. There are plenty of sources that
give averages, but not breakdowns.
1) The 1971 and 1981 individual censuses(PUSTs) asked number of children ever
born of (sometime-married) women; the 1991 census (PUMS, Basic Summary
Tabulation B9105) expanded it (in variable Fertility) to all women age 15+;
and then it was dropped from the Census, as far as I can tell.
2) The data described above is your best bet. I checked with both Census Division and Health Statistics Division (for vital statistics) and neither one of them can offer anything further.
The relevant products for "number of children ever born" are:
- CANSIM Table 075-0018: Historical statistics, number of children ever born per 1,000 ever-married women aged 15 years and over, every 10 years (persons) *Terminated* (for 1961 and 1971)
- 1981 Census print publication called "Population: Nuptiality and fertility" (Catalogue no. 92-906, Volume 1 - National series) contains a table called "Women Ever Married by Age Groups"
- 1991 Census print publication called "Fertility" (Data Products: Nation Series: catalogue no. 93-321-XPB)
3) I'm not sure if the desire is for a time series or just the distribution of women who
have given birth by age cohort. If its the latter, the data that immediately come to
mind are the family cycles in the GSS. For example, Cycles 20 (2006) and 15 (2001) both
contain questions about the number of children given birth to (COR_Q042). While you
won't be able to obtain estimates for smaller areas of geography, you can get an estimate
for Canada and the more populous provinces.
Would it be possible to have someone from Census Division offer their explanation of why they consider it no longer important to ask this question? I was under the impression that this is a rather fundamental element of demographic information.
1) There is a continuous battle of which questions will be on the Census because there is always great demand for new content and new questions. This question was asked every ten year and was dropped in 1981. The idea is that it may not necessary to ask that question on the Census but other survey may be as appropriate to get that information. But I don't think it was picked up in other surveys.
I would suggest that the timing is right if you want to voice this information as a need:
There is a forum for consultation for the 2011 Census. This is a discussion forum specifically for that type of questions and identifying needs for 2011.
2) Don't forget too the Comparison of the census and alternative sources of
data (Source: 2001 Census consultation guide./ Statistics Canada
The 2006 version doesn't include fertility, so isn't particularly useful,
but the 1996 one did:
Friday, March 13, 2009
A file called Community Profiles was released along with the pumf for the
1991 Aboriginal Peoples Survey. A graduate student would like a similar
aggregated file from the later APS surveys.
She is aware of the 2001 Census Aboriginal Population Profile and the 2006
Census Aboriginal Population Profile. It would be convenient for her to be
able to look at all the communities in one file instead of having to work
her way through them one community at a time. All the topic-based
tabulations with aboriginal identity, etc., lack census subdivision level
coverage as far as I can see.
So the question is, are there amalgamated Community Profiles for 2001 and
I found some aboriginal tables at the CSD level that may offer what your
For 2006, there is a topic-based tabulation in the Aboriginal Peoples of
Canada CD-ROM that is at the Census Subdivision level:
The file number and name is 92-598-XCB2006001 (Aboriginal Population Profile
for Canada, Provinces and Territories, Census Divisions and Census
Subdivisions, 2006 Census).
For 2001, there is a CSD-level table as well:
The file number and name is 94f0043xcb01001.ivt (Aboriginal Population
Profile for Canada, Provinces, Territories, Census Divisions and Census
Subdivisions, 2001 Census).
I am wondering if there is an easy answer to this question of “average age of schooling for Canadians”. Although not conveyed in the email from the Grad student, the key to this request (communicated during a phone chat) is the desire to have annual data on this topic, back to the 1970’s. Also, if possible, to have the data by province. The student adds via email:
Pertaining to our discussion yesterday, the data that I'm looking for is this:
What is the average years of schooling for Canadians?
By schooling, I mean elementary, high school (secondary), and university or an accredited/recognized post-secondary educational institution.
So for example, I have:
elementary through high school so that's 12 years
my bachelor's degree so that's 5 years
one master's degree so that 2 years
my second master's degrees (in progress) 2 years.
total years of schooling is 12 + 5 + 2 + 2 = 21 years.
The data can be aggregated to all Canadians.
Thanks for any help. We have looked at the Labour Force Survey (educational attainment) via SDA, but there appears to be no simple (single) answer for the question asked, in terms of annual data for “average years” of schooling for Canadians.
Here is a summary of what the author divisions have told me about the years of schooling variable:
- census data for this variable is not available for 2006 due to changes in the questionnaire. This matches some information I found on the following page of the Statcan site (scroll down to the heading "Modified Questions - Education)": http://www12.statcan.ca/english/census06/reference/dictionary/ovpop4a.cfm#04. This means they cannot provide a custom tabulation for 2006 on years of schooling
- total years of schooling is available for census years between 1971 and 2001 as a custom tabulation. Note that the population for 1976 was 1/3 sample and 1971 1/3 sample. Here is the link to the 2001 Census Dictionary stating the comparability years for total years of schooling:
- LFS has only annual data on "educational attainment" starting in 1990. They suggested that you check the CANSIM table 282-0004 to see if it would correspond to your needs.
Will this product be made available via DLI as data product ?
The PUMF for the APS 2006 is expected to be released sometime this month, and it will be made available through the DLI.
Here's the silliest of silly questions - I'm looking at the SPSS syntax
provided to us for the 2005 and the 2006 SHS PUMF file. Why is it in 2005
all the variables are coded as string and in 2006 they're coded as numeric?
I missed this the first time - but my faculty didn't
FYI - the SPSS syntax for SHS 2006 has been changed so that it is consistent
with the coding for earlier years. I will be sending out an announcement to
this effect soon.
Two researchers here at York, using the pumfs from the 2006 SLID, have
reported problems with the syntax files in SPSS and SAS. Both are
receiving error messages for missing value specifications. I am not
very familiar with these applications and don't know how to help. Any
suggestions would be very welcome.
The SLID 2006 SPSS files have been updated. I will be sending out an
announcement regarding these and other updated SLID files soon.
Will there be a PUMF for this one time survey?
In January, the author division told me that there would be no PUMF for this survey. However, the microdata is available through the Research Data Centres, and custom tabulations are also available. I found the answer in the listserv archives:
It seems there's a problem with accessing E-Stat lately. I am getting error messages when clicking on links from the menu at right.
The url needs to be updated with the .gc portion
I thought we were assured that the URLs would be redirected for a
number of months. Has that not happened?
I don't think redirection is happening to a great extent.
I just found out that some weeks ago we had to change the urls for over 500
Stat Can publications that were linked from our catalogue. The Collections staff
just went ahead and did it without letting me know there was a problem.
Apparently a Find and Replace technique worked for many of the urls.
But I think it would be useful to be told ahead of time that something
like this was about to occur so we could plan for it.
We been working with Dissemination Division to find solutions to the redirection issues you have found with E-STAT and the Statistics Canada web site in general. Here are the results of their discussions:
1. Redirection for E-STAT addresses has been problematic and the solutions are not straightforward. Therefore, to ensure the E-STAT works correctly you must access it through the following URLs:
2. Redirection should be working for the rest of the Statistics Canada site. However, it is clear from your feedback that there are URLs that are not redirecting. If you find a link that is not redirecting properly, please let us know so we can have Dissemination Division fix it. The intention is still that redirects (outside of E-STAT) will work for at least the next 2 years so it is important that we correct any that are not working at the moment.
Thank you all for your feedback and patience on this matter. We will be sure to inform Dissemination Division of any problems you report to us regarding redirection.
Would this be a good time to update all the www.statcan.ca links to www.statcan.gc.ca that are scattered across our websites?
If I am asking for E-Stat to be updated everywhere I may as well ask for other links to be done as well
Or are there some areas where this would cause problems?
There are some areas of the site that still have issues with their URLs. I would recommend holding off on updating all the URLs for the time being until most or all of them are resolved.
In the meantime, please be sure to report any URLs that you find that are not redirecting properly.
I am helping an economics professor with the Survey of Household Spending and he has a question about the age of the respondent variable. He would like to know if there is any way to obtain the exact age as opposed to just an age group. I checked the questionnaire and there is indeed a question about the age so the data must be available somewhere… As a neophyte, I would also be curious to know the rationale for providing age groups rather than the exact age figure.
Anyway, the prof in question is used to the Survey of Consumer Finance (stopped in 1998) which apparently provided the exact age. So he would like to be able to obtain the same information with the SHS. Is there a master file available from a RDC for that survey?
The author division confirmed that the exact age is available on the RDC internal file for the SHS:
"Any Research Data centre should have a copy of the Internal files for the SHS for all years, or if they don't have one should get one on request. The exact ages are on the internal file, The Pumf uses age groups for confidentiality."
They also confirmed that custom tabulations are available, provided that the sample size is sufficient and there are no confidentiality issues with the particular variables being crossed. However, the Division encourages people to use the Research Data Centres to perform their own custom tabulations, so the professor can submit a proposal to use the RDC file if he wants to avoid the cost of a custom tabulation.
Do we have a date when the PUMF for the CTUMS 2008 Cycle 1 will be released to DLI? (http://www.statcan.gc.ca/daily-quotidien/090119/dq090119b-eng.htm)
Also, could you please confirm that the CCHS 2007 is still due to be released this summer (and perhaps a more specific date)?
The PUMF for CTUMS 2008 Cycle 1 was released to the DLI some time ago, but the author division subsequently found there were problems with the bootstrap weight files so we were asked not to load it. I asked for an update just now, and was told that they are still correcting the bootstrap weight files. However, they said that we could go ahead and load all the other files for now and they will send us the bootstrap weight files separately once they are corrected. Jackie is loading the files as we speak and I will announce them as soon as they are available on the sites.
As for the CCHS 2007, the author division confirmed that "end of summer 2009" is still the anticipated release date (they cannot provide a more accurate date at this point).
I know this has just been released, but any possibility of getting this or previous years through the DLI? I have an SME researcher that I know will be asking.
This is another survey for which no standard data products were released. I have asked the author division if they are willing to provide the DLI with any cross-tabulations and will keep you posted.
The contact for this survey explained that aggregate data is available upon request, free of charge, from the division. However, it is also being made available through Industry Canada's website. I checked their site and found the following under the "SME Financing Data Initiative" section of the site:
Anyone can request the data (free of charge, in PDF or Excel formats) for 2000, 2001, 2004 and 2007 through this site, which also contains other information on SME Financing that may be of interest to your researcher.
We're currently working on the Survey of Family Expenditures. It doesn't look like you have the surveys for the following years: 1969, 1974, 1978, 1982, 1984, 1986, 1990, and 1992. Are we able to obtain them?
The surveys you mentioned are now available on our site. I'm going to send a bilingual announcement soon.
I have cool question from an undergraduate student this afternoon - can anyone direct me to a source that could tell me the number of dual income earners in a household before and after 1960. The student has looked at Census - and pulling out the # of earners in a household is not as straightforward... I'm stuck on the 1960 bit.
Any suggestions would be greatly appreciated.
I looked in the 1951 Census print publications and found one table that may be relevant. It can be found in Volume III: Housing and Families:
Table 128. Families by size, types, numbers of members in the labour force...1951
This table provides the number of families with either 0, 1, or 2+ members in the labour force for Canada, provinces, counties and census divisions.
There are 2 similar tables for other levels of geography:
Table 129 = incorporated cities, towns, and villages of 1,000 population and over
Table 130 = Census Metropolitan Areas
I'm searching for the number of doctors in Quebec and Ontario for the years 1993 to 2000. I found in "Aperçu des statistiques sur la santé 1999" data from 1977 to 1992 and 2000 to 2007 on the CIHI (Canadian Institute for Health Information) site. Does a source exist for 1993 to 2000?
1) The last of the reports should answer your question. It is in the paper version: Number, distribution and migration of the Canadian doctors - Years chosen, 1961-1995.
2) CIHI published numerous publications for 1993-2000, but they are not always available online. I suggest you consult the pertinent versions of the following titles (and to order printed versions via inter-library loans):
1. Médecins équivalents à temps plein, Canada : http://secure.cihi.ca/cihiweb/dispPage.jsp?cw_page=AR_17_F
2. Nombre, répartition et migration des médecins canadiens : http://secure.cihi.ca/cihiweb/dispPage.jsp?cw_page=AR_14_F&cw_topic=14
3. Chapitre 21 du rapport Tendances relatives au personnel de la santé au Canada, de 1995 à 2004 :
Moreover, you always have the option of ordering data from CIHI if the aforementioned publications are not enough. Also keep in mind that the Health Stats Division (here at Statcan) told me that CIHI is the best resource for the statistics that you want.
Hope this helps you.
Monday, March 9, 2009
I'm looking for the Saskatchewan birth rates from 2005 to 2008. I found the previous years (1994-2004) in the publication #91-213, but for the more recent years they don't give the number of births (in all the publications I consulted including Cansim and the Monthly Statistical Review of Saskatchewan). Is this data available?
I found the Saskatchewan birth rates for 2005 to 2008 on the following Statistics Canada web page:
By default, this table presents the number of births, but they can choose birth rates from the menu on the left (under the title "Dans cette série"). Please note the figures for 2007-2008 are provisional.
Thank you! The same research asks if such data exists for different regions in the interior of the provinces (or even another province besides Quebec)?
Sorry for the delay in answering your question. I needed to hear back from a number of Statistics Canada people to be certain.
An analyst at the demography division says the birth rates are not easily available for the other geographic provincial levels. It should be calculated with the population estimates and number of births (a personalized totalization). Nevertheless, they recommend using fertility rate totals, which is available at the CMA level for 2005 at the following publication:
I hope this helps.
Friday, March 6, 2009
Right near the end of the questionnaire for the Canadian Community Health Survey, cycle 2.1, there is a Nurses' Supplement. The data for this supplement does not appear to show up in either the PUMF or the documentation for the PUMF. However, it does show up in the data dictionary for the Master File,
http://www.statcan.gc.ca/imdb-bmdi/document/3226_D3_T9_V2-eng.pdf pages 1031 through 1036 of 1044. There doesn't appear to be any information about this previously on the dlilist. I would assume that the data has been removed from the PUMF for confidentiality purposes, but thought that it might be appropriate to get a confirmation of this? Thanks
in advance for any information you might shed on this matter.
I have a patron looking for birth rate statistics for Canadian rural areas. I've looked through ESTAT and the Census webpages with no joy. Am I missing something?
1) Figure 2.6 in the publication, "Demographic Changes in Canada from 1971
to 2001 across an Urban-to-Rural Gradient" shows the
total fertility rate for rural and urban areas from 1971 to 2001. The
publication is available at:
These rates were based on Census data.
The online publication, "Births" (84F0210X) contains a section on fertility but doesn't
break out the differences between urban and rural.
The data source for this publication is the Births Database in Vital
The General Social Survey's Family Cycle (c20 is the latest) provides
some fertility data. The public use file has a variable (LUC_RSTC) indicating
rural/urban residence, but only for Quebec, Ontario and B.C.
2) For Quebec, you can have birth rate statistics by regions through the Institut de la statistique du Québec Web Site.
If you follow this link, you will get a map and then will be able to click on each region, click on "demographie" and then on "taux de natalité".
I would be surprised these data not being available for other provinces.
3) There may be a specific reason why the individual is looking for birth rate, and, if that is the case, as indicated, births will be reported in the provinces' vital statistics. Here is one link that should take you to the vital stats for each province.
However, I wonder how reliably "birth rate" for a specific provincial geopolitical area[i.e. county] in a province can be translated into "rural". I think there may be a problem if provincial registrar's haven't categorized birth data as "rural/nonrural" now and through time. I don't think they have. As well, there would be a problem with definition.
Bollman's and Clemenson's working paper published late last year may be useful for definitions and data:
For subprovincial areas, the researcher may wish to look at census figures over a given period, ages 0-4, for CDs, for example.
4) I haven't paid close attention to this so I hope what I have to say is relevant; if I'm off-topic, I apologize for jumping in. Anyhow, I think it's important to realize that many of the births in the urban areas would have occurred in the rural area, if facilities and personnel were in place to provide this service.
My colleague was working with these files and made the following observation:
On the just released Survey on Financing of Co-operatives, 2007, the
data are in these two files, English and French.
efco2007dat.xls H1, H2. Caractéristiques des coopératives,
selon les valeurs moyennes de l’actif et du
passif de la coopérative, Canada, 2008
sfco2007dat.xls H1, H2. Characteristics of co-operatives, by
average assets and liabilities of
If one of these files said 2008, I would assume it was a typo. However,
both say 2008, so either the title of the survey is wrong or the file
descriptions are wrong.
Since the survey was conducted in February and March of 2008, so the
2007 must refer to the reference year, and one way or another, both
years are correct.
However, the real problem with the file description is that the file
contains not only table H1, H2, but a total of about 40 tables, of which
H1, H2 is only the last sheet. I have a list of the titles of all the
English language tables (sorry, haven't checked the French file) at:
under the Notes section, should anyone find it useful.
I have a graduate student interested in studying Aboriginal women's experience of violence. We found this report --
Addressing violence against Aboriginal Women, the Elderly and Children from July 2008, which makes some claims about incidence of violence against Aboriginal women, which it footnotes to:
Measuring Violence Against Women: Statistical Trends.
This contains the reference:
"For the first time, residents of the three territories (Yukon, Northwest Territories and Nunavut) were interviewed for the 2004 General Social Survey (GSS) as part of a pilot test and results show that women who live in the territories also experience higher levels of violence."
However, when I search variables in the 2004 GSS, there is no indicator with "aboriginal" as a category, nor is there any "territorial data" at all.
Is there some way of acquiring the survey data that has been collected?
An analyst in the author division explained that the aboriginal variable was suppressed on the public use microdata file in order to guard against disclosure. The analyst went on to explain that the data is not even available in the RDCs:
"The data for the territories is not available in the RDCs. Due to concerns about data quality (which are explained in the attached report), the microdata are not available to researchers outside of the Canadian Centre for Justice Statistics who published the report. We are trying again in the territories for the current GSS (2009) on Victimization, hoping that the improved methods we are using will yield better quality results that could be made available to researchers. They will not be available until summer 2010 though."
A researcher is seeking EI benefits by gender. She found CANSIM Table 276-0005, but this table doesn’t have a gender variable.
1)What about the Employment Insurance Coverage Survey (EICS)?
2) That is correct - the codebook for the EICS does indicate that there are variables for sex (SEX), benefits amount (BENAMNT) and benefits type (BENTYP). However, the PUMF is for 2007, so I have also forwarded your question to the author division in the hope that there is a way to get more current data. I will keep you posted.
3)It turns out that the EICS 2007 PUMF is your best bet. There are two separate surveys related to EI:
1. Employment Insurance Coverage Survey (EICS) 2007: this is the annual survey Sandra mentioned. The DLI has the PUMF and it contains the variables you need, but only for 2007.
2. Employment Insurance Statistics - Monthly (EIS): this is a monthly survey for which PUMFs are not produced. The CANSIM table you found comes from this survey.
Unfortunately, the author division of the latter (monthly) survey confirmed that they cannot provide the data presented in CANSIM table 276-0005 broken down by sex, not even for a fee.
Monday, March 2, 2009
An anxious professor here is wondering if you could give a better idea when the the Aboriginal Peoples Survey 2006 Public Use Microdata File (PUMF) might be released? I checked Products Release page and see it's supposed to be released this month.
My contact in the Division won't provide me with a specific date at this time, but he did reconfirm their intention to release the PUMF by the end of this month. To be safe, I would expect it to be released towards the end of March.
Sorry I couldn't be of any further assistance.
A researcher here has inquired about the Canadian Survey of Experiences
with Primary Health Care 2008. The Daily from Thursday February 5, 2009,
states that "The microdata file ... is now available." I checked the
dlilist and found the appended message from 2008, indicating that there
would be no PUMF, but that the master file would be available via the RDCs.
The researcher has already checked with the RDC and was told that it is not
presently available there, which is what I also found from their web pages.
Would it be possible to check again with the author division to confirm if
and when this survey might be made available to researchers? Thanks, Gary
The author division provided me with answers to your questions as well as an update on the status of the PUMF for the Canadian Survey of Experiences with Primary Health Care 2008:
1. A PUMF will be released and is planned for the fall of 2009.
2. The February 5, 2009 release is announcing that the division itself has microdata from which it can produce custom tabulations. It did not mean that a microdata file was available to researchers.
3. While the master file is not presently available at the RDCs, it can be made available through the RDCs if a researcher submits a proposal. Your researcher should go back to the RDC and enquire about that if they want access to the master file.