2011 Census Long-Form Policy
Issue: Should the Government of Canada conduct the 2011 Census using a compulsory long form, or a voluntary long form?
Statistics Canada, reporting to the Minister of Industry, carries out a Census of Population every five years.
According to Statistics Canada, the Census of Population produces:
- Detailed information on population subgroups and for small geographical areas, which cannot be generated through surveys, required to assess the affects of specifically targeted policy initiatives and serves as a foundation for other statistical surveys.
- Population estimates, projections and in-depth information on special populations, such as operators of agriculture holdings and Aboriginal peoples.
- Population counts and estimates used in determining electoral boundaries, distribution of federal transfer payments, and the transfer and allocation of funds among regional and municipal governments, school boards and other locally based agencies within provinces.
- Statistical information for use in per capita measures in fiscal policies and arrangements and other economic analysis, and in program and service planning
- Statistical information ranging from demographic, social and economic conditions of the population, annual and quarterly estimates of the population, households and families and their projections to number and types of farms and farm operators, and the socio-economic conditions and well-being of Aboriginal peoples.
Prior to 2011, Statistics Canada obtained data for the Census of Population by requiring 80 percent of Canadians to respond to a short form, and 20 percent of Canadians to respond to a long form. The short form sought basic information, while the long form sought information on a range of subjects including income, occupation, employment, involvement in the labor force, ethnic background, religion, languages, housing conditions, Aboriginal status, country of birth, etc.
Statistics Canada uses statistical sampling theory to make observations about the Canadian population, and subgroups within the population, from the information generated through the long form. In a large population subgroup (such as inhabitants of a Census subdivision in a large city), a relatively small percentage of the subgroup's population needs to complete the long form. In a small population subgroup (such as inhabitants of a census subdivision in a small, remote community), a relatively large percentage of the population needs to complete the long form. Statistics Canada selects those individuals required to complete the long form in a way that ensures that there are a sufficient number of respondents to enable Statistics Canada to confidently make statistical observations about the subgroup.
The Census of Population is carried out under the budget heading "Census, Demography and Aboriginal Statistics", within which the Census of Population is the major activity. The cost of the activities within this budget vary over a five year period, with much of the work carried out around Census years. The total cost of the budget label over a five year period 2006/2007 to 2010/2011 is estimated at $675.1 million, based on actual data where available and budgets where actual data is not available. See BUDGETS AND EXPENDITURES.
For the 2011 Census, the Government Minister responsible for Statistics Canada and the Census introduced a plan to replace the compulsory long form with a voluntary long form. The Minister indicated that Statistics Canada agreed with the plan. Apparently, Statistics Canada did not agree with the proposal, and because the Minister had misrepresented Statistics Canada's position, the Chief Statistician resigned.
While some commentators have questioned whether the information in the long-form is necessary, the parties most knowledgeable on the subject, including interest groups, researchers, federal departments, provincial and local governments, businesses and their organizations, have indicated the information acquired through the long form is essential to their operations.
Commentators have noted that some countries have abolished or do not conduct Censuses of Population, that the Government of Canada already has much of the information sought in the Census of Population in its administrative records, and that some foreign governments have accessed administrative records to generate information that might otherwise come through a Census of Population. For 2011, the Government of Canada is not in a position to use administrative records to replace the Census of Population, although this could be an option for 2016.
Compulsory Long Form : This option is the status quo. It entails the distribution of the long form to 20 percent of Canadians, carefully selected to ensure adequate number of responses to enable Statistics Canada to make statistical observations, particularly for small population sub-groups and small geographic areas. The option is backed by the potential for legal action against individuals that do not respond to the long form questionnaire.
Voluntary Long Form: This option involves the distribution of the long form to a percentage of the population much larger than 20 percent. Recipients will be asked to voluntarily complete the long form, and there will be no potential for legal action should they not complete the form. The percentage of the population receiving the long form will be much larger than 20 percent Because some individuals will likely not respond to the long form questionnaire, and to ensure adequate response rates for small population subgroups and small geographic areas.
Issues: These refer to the subject areas against which the options will be compared.
Effectiveness: The Census of Population is a tool for the development of Canada. Businesses (particularly larger businesses) use it to make investment decisions, and to make the case to others for financing. Governments at all levels - municipal, provincial, federal - use it to make policy decisions that will affect the future, as well as a tool for assessing the effectiveness of existing policies. Non-governmental organizations use it to plan their work, justify their activities to others, and measure the success of their initiatives.
An unreliable Census of Population will undermine economic progress in Canada. It will lead to bad investments by government, business and other organizations, and the continuation of ineffective activities when they should be terminated.
There are several elements to the effectiveness of the Census: the quality of the data from the 2011 Census for subjects covered by the Census; the coverage of subject areas; and the comparability of the data from the 2011 Census with the data from previous and future Censuses.
Data Quality from the 2011 Census for Subjects Covered: This refers to absence of bias.
Subject Coverage from the 2011 Census: This refers to the areas covered, particularly small geographic areas and specific target populations. Small area data is particularly important for local governments and business decision-making.
Comparability with Past Censuses: This refers to the ability to use data from the 2011 and previous Censuses to make observations about trends, progress, etc.
Census Public Cost: The cost to the public of the 2011 Census is the result of the production cost of the 2011 Census, less whatever revenues Statistics Canada can generate from the sale of the data.
Enforcement: This refers to the costs to external agencies for enforcing the legal requirements to complete the long form Census. These costs include the costs of preparing cases for prosecution, prosecution costs, court costs, and prison costs.
Privacy: In current public policy discussions, "privacy" means that if one provides personal information to one agency for a particular purpose, that agency should keep the information private and not pass the information on to third parties. In the context of the Census, information provided to Statistics Canada is not to be passed to third parties under the Statistics Act and privacy policies of Statistics Canada. The privacy risk as a consequence is from illegal activities and would be directly related to the number of completed forms in the hands of Statistics Canada. Based on past history, this risk is low.
It is worth noting that at the completion of the 2006 Census, Statistics Canada carried out an extensive feedback process involving 1,200 government agencies, municipalities, non-governmental organizations, academics, businesses and citizens. The results were incorporated into a 53 page report. Privacy was not mentioned.
The Privacy Commission reported only 3 privacy complaints related to the last two Censuses, and 50 complaints in the last twenty years. There were 33 complaints related to the 1991 Census, based primarily on the questions asked at that time.
Note that while privacy is largely a non-issue, the Industry Minister raised the issue as a defence for the decision for a voluntary long-term Census.
Public Response Burden: This refers to the time members of the public must spend to fill out the Census long form, and is the product of the number of people that complete the form, times the average time for completion.
One should note that a disproportionate burden of response to the long form falls on rural populations, where a larger portion of the population is required to complete the form. Ironically, the voluntary long form does not change this response burden, but merely shifts it from those that do not want to complete the form to others in the area.
Intrusiveness: This refers to the extent to which the government gathers information about individuals. The concern has two aspects. One relates to why the government needs the information. The second aspect relates to individual concerns about what the government might do with the information.
It is worth noting that Canadian governments (municipal, provincial, federal) already have much of the information requested in the Census through income tax records, property tax records, public health citizens, utilization of public programs (old age security, Canada Pension Plan, Employment Insurance), passport records, citizenship and immigration records, previous Census forms, etc. This reinforces the first point regarding intrusiveness. If governments already have the information, why are they asking for it again through the Census. It also undermines the second point. If governments already have much of the information, the net new intrusion into the private lives of citizens is significantly less.
The fact that citizens cannot get access to a comprehensive explanation of why questions are asked in the long form, what Statistics Canada will do with the information, who will be expected to use it, and how they will use it points to a weakness in Statistics Canada's communications programs.
Coercion: This involves government laws and regulations that force individuals to do or not do things they would do or not do. Coercion is always justified in terms of the public good. In Canada, coercion is always subject to the Charter of Rights and Freedoms. The power to coerce comes from Parliament. Through the Statistics Act, Parliament created penalties for not complying with Statistics Canada's requests for information. The rationale is that the public good is served by forcing individuals to provide information. While Statistics Canada could take action against individuals that do not complete the compulsory long form, historically it has not done so.
Political Considerations: These ultimately involve appeal to voters either nationally or in targeted ridings, sub-regions, provinces. This issue is relevant to the "efficiency constituency" and the "freedom constituency".
Efficiency Constituency: This is the group of Canadians looking for efficient, competent government. This constituency is national in scope, and includes most Canadians.
Freedom Constituency: This constituency values more individual freedom and less government regulation. While this constituency exists to some extent across Canada, it represents a larger portion of the population in Western Canada, where reliance on federal government programs has historically been less pronounced than in other regions, and in rural areas, where restraints on individuals are less pronounced and interaction among individuals tends to be less. Residents of urban areas live in closer proximity to other Canadians and see greater value in restraints on other individuals as a benefit to themselves and therefore in the public good.
|Compulsory Long Form||Voluntary Long Form|
|Effectiveness||There are three components to effectiveness: data quality from the 2011 Census for the subjects covered; subject coverage from the 2011 Census; and comparability with past Censuses.|
|Data Quality from the 2011 Census for Subjects Covered||Because recipients of the compulsory long form are randomly selected and because all respondents to the Census long form are expected to answer the questions in the form, there is relatively little bias in the responses.||Because recipients of the voluntary long form can voluntarily respond to the questions in the form, it is expected that certain groups will tend to not respond, including the less educated, the poor, the socially disadvantaged, and those that feel the Census is "intrusive" or have a fundamental distrust of government or are disillusioned by it. These non-responses will lead to a bias in the responses, and significantly undermine the data quality for subjects covered.|
|Subject Coverage from the 2011 Census||Sufficient recipients of the compulsory long form are selected to ensure that the Census can generate reliable information for small geographic areas and for specific target populations. For small geographic areas and specific target populations, it is usually necessary to select a relatively large percentage of the population to generate reliable statistics. Because recipients of the compulsory long form are expected to respond, Statistics Canada can reliably produce information for all small geographic areas and all specific target populations.||With the voluntary long form, Statistics Canada cannot be confident that it will get sufficient responses for all small geographic areas and all specific target populations. The likelihood is that it will get sufficient responses for some small geographic areas and some target populations. Statistics Canada then faces the dilemma of providing information for some small geographic areas but not others, or some target populations by not all. As a national statistics agency, Statistics Canada will likely choose not to produce information unless it can cover the entire country. In short, the subject coverage in a voluntary long form is likely to be less.|
|Comparability with Past Censuses||Compulsory completion of the compulsory long form is compatible with past Censuses. Users of the data can reliably compare results from previous Censuses and the 2011 Census to identify trends, track progress, etc.||The voluntary long form will not be compatible with previous Census because of biases in the response, and because the subject areas covered are likely to be less.|
|Census Public Cost||
With the compulsory long form, Census forms would be distributed to about 20 percent of the population. The vast majority would complete the form, so data entry would be required for about 20 percent of the population. As the compulsory long form has historically not resorted to prosecutions for of those that did not complete the Census, there would be no change in enforcement costs to Statistics Canada. Normally, some people do not complete the Census on time, so follow-up is required, and could be expected to be at historic levels.
Revenue generation from the Census would be in line with historic levels, because the data quality and subject coverage would be in line with historic levels.
Overall, the public cost of the compulsory long form would be less.
With the voluntary long form, Statistics Canada would need to receive completed forms from more than 20 percent of the population, to cover the unevenness of voluntary responses. In order to get enough information in all areas of the country, Statistics Canada would get more than enough information in some. Call this the "data redundancy factor". In addition, additional forms would need to be distributed to cover those who choose not to complete the Census. Consider this the "voluntary non-response factor". The two factors would lead to the distribution of forms to substantially more than 20 percent of the population. Since there is uncertainty about both factors, the response to uncertainty would be in the first instance pre-Census surveys to estimate both. These surveys are a cost item. Ultimately, responses would likely be substantially more than 20 percent of the population because of the redundancy factor, leading to extra data input costs compared with the compulsory long form. With the voluntary long form, there would be no enforcement costs. However, extensive follow-up would be needed from initial non-respondents and would cover those who choose not to complete the voluntary long form and those who are tardy in completing the form. Follow-up calls would be more time consuming with the voluntary long form, and would typically involve more explanation and persuasion than would occur with the compulsory long form.
Revenue generation would be less than with the compulsory long form, because the data quality would be inferior and the subject coverage less.
Overall, the public cost of the voluntary long form would be more than with the compulsory long form.
|Enforcement||While the compulsory long form includes the possibility of legal action against those that do not complete the form, historically legal action is not taken. For example, a number of First Nations have refused to participate in past Censuses, and no legal action was taken against them.||With the voluntary long form, no enforcement is possible.|
|Privacy||Privacy risks would be less, because Statistics Canada would have completed forms from approximately 20 percent of the population.||Privacy risks would be greater, because Statistics Canada would have completed forms from more than 20 percent of the population, as a result of the "data redundancy factor".|
|Public Response Burden||The public response burden would be in line with historic levels.||The public response burden would be higher than historic levels, because of the "data redundancy factor".|
|Intrusiveness||The level of intrusiveness is the same with the compulsory or voluntary long form, because the long form is the same in both cases.||The level of intrusiveness is the same with the compulsory or voluntary long form, because the long form is the same in both cases.|
|Coercion||The compulsory long form is potentially more coercive than the voluntary long form, although the degree of coercion has historically not been high.||The voluntary long form does not involve coercion.|
|Political Considerations||There are two components of political considerations: political appeal to these who want efficient, competent government (the Efficiency Constitutency), and political appeal to those who do not want to be required to do things by governments (the Freedom Constituency).|
|Efficiency Constituency||The compulsory long form is the more efficient, producing better quality data covering historic subject areas for 2011, allowing better comparisons with past Censuses, all for a lower public cost, including a lower production cost and higher revenue generation potential.||This voluntary long form is the less efficient, producing poorer quality data covering fewer subject areas for 2011, limiting comparisons with past Censuses, all for higher production costs and lower revenue generation potential.|
|Freedom Constituency||The compulsory long form is the less appealing, because of its potential restrictions on freedom. The appeal is based to some extent on the erroneous perception that the coercion is applied in a rigorous and systematic way.||The voluntary long form is the more appealing, because there is no coercion.|
Budget and Expenditures
Statistics Canada, Departmental Performance Report, 2006/2007 to 2008/2009
Statistics Canada, Report on Plans and Priorities, 2009/2010 and 2010/2011