MAT 141 (Statistics)

This page was updated on: January 31, 2014.

The United States and Canada each conduct a census every ten years.  In addition to the nationwide census, more detailed surveys are sent to a smaller number of households in each country.  Scroll down to see the forms used in each survey and read how the processes used in the US and Canada differ.

Link to the website of the US Census Bureau:  www.census.gov 

The census occurs every ten years, as mandated by the US Constitution.  The most recent census took place in 2010.

The Census Bureau attempts to collect information on every individual living in the US using a "short form" mailed to each household.  The form consists of 10 questions, most of which concern the age, gender and racial/ethnic background of every individual living in the household.  Participation in the census is required by law.  A representative of the Census Bureau will follow up with households that have not returned a form.

The cover letter that accompanied the form stated "It is improtant that each household respond only once."  Interestingly, two forms were mailed to my parents' household, although they only completed one form.

The Census Bureau also collects more detailed information from a much smaller sample of households.  In the past, this was done by mailing a "long form" of the census to one-sixth of the households at the time the census was taken.  However, in 2005 the long form was replaced by the American Community Survey (ACS). ACS data are collected continuously throughout the year and throughout the decade from a relatively small sample of the population (250,000 households every month, or 3 million addresses annually). Since the ACS is conducted on an ongoing basis, rather than once every 10 years, it provides more current data than the long form. Participation in the ACS is also required by law.

The ACS form consists of 48 questions.  Many of the questions are very straightforward, easy to answer, and likely to result in accurate responses (e.g., Question 12, which asks how many vehicles are kept at the household).  One has to wonder about some of the other questions.  For example, Question 14 asks for the amount paid for utilities during the most recent month.  Will every household take the time to find their utility bill to answer this question accurately?  Question 19 asks homeowners to estimate the current sale price of their home?  An accurate estimate would require an appraisal by a real estate professional, which is unlikely to happen unless the homeowner is planning to sell the home.

(Includes information from the following source: US Census Bureau)

Link to the census website of Statistics Canada (government agency responsible for the census): http://www12.statcan.gc.ca/census-recensement/index-eng.cfm

A national census in Canada is conducted every five years, as required by the Constitution Act of 1867.  The most recent census took place in 2011.

As in the US, a short form consisting of 10 questions is mailed to every household.  Participation is required by law.  To collect more detailed information, a long form of the census had been mailed to one out of five households and these households were required by law to complete the forms. 

However, in 2011, the mandatory long form census was replaced by a voluntary 35-page National Household Survey (NHS).  Anticipating a lower response rate due to the voluntary nature of the survey, the NHS was mailed to a much larger sample (one in three households, instead of one in five).

The response rate for the 2011 NHS was 68.6%, compared to a 94% response rate for the 2006 long form.  Even though the NHS response rate was higher than anticipated (Statistics Canada expected a 50% response rate) and generated more data than the 2006 mandatory long form (because many more forms were mailed out), response rates varied considerably from one community to another.  In some places the response rate was as low as 25%.  In a few communities, the response rate was zero.   According to David Bellhouse, a statistics professor at the University of Western Ontario, "when you get a response rate that is low, especially from a government survey, there could be quite a difference between the people that have responded and those that haven't." 

As in the US, the government of Canada uses survey data to plan social services.  Some experts are concerned that the people who are most in need of these services - poor people, immigrants and aboriginal people, for example - are less likely to respond and are therefore underrepresented in the results.

Additional concerns about potential bias in the NHS are discussed in the following article on the Huffington Post website: "National Household Survey That Replaced Long-Form Census is Unreliable: Experts." 

(Includes information from the following source: Statistics Canada, Wikipedia, Huffington Post.)

Canada Census Short Form

Link to 2011 Canada Census Short Form

Sampling Methods

Statistics Canada conducts a number of other surveys to collect data on labor, business and agriculture.  Sampling is used for these surveys and participation is required by law.  The following statement on the Statistics Canada website  describes how samples are selected:

"How was I selected for this survey?  Participants for a sample survey are selected randomly in order to avoid any biases.  Once you have been selected, we cannot ask anyone else to complete your questionnaire because the sample would no longer be random."