Canadian Survey on Disability (CSD)

Detailed information for 2017

Status:

Active

Frequency:

Every 5 years

Record number:

3251

The purpose of the Canadian Survey on Disability (CSD) is to provide information about Canadian youth and adults whose everyday activities are limited due to a long-term condition or health-related problem. This information may be used to plan and evaluate services, programs and policies for Canadians living with disabilities to help enable their full participation in society. The survey is sponsored by Employment and Social Development Canada (ESDC).

Data release - November 8, 2019

Description

The CSD gathers information about Canadians aged 15 and over whose daily activities are limited due to a long-term condition or health-related problem.

Information from the CSD may be used by all levels of government, as well as associations for persons with disabilities and researchers working in the field of disability. Data may be used to plan and evaluate policies and programs for Canadians with disabilities to help enable their full participation in society. In particular, information on adults with disabilities is essential for the effective development and operation of the Employment Equity Program. Data on disability are also used to fulfill Canada's international agreement relating to the United Nations Convention on the Rights of Persons with Disabilities.

The survey collects information on: the type and severity of disability, use of aids and assistive devices, daily help received or required, use of various therapies and social service supports, educational attainment, labour force participation details, requirements and unmet needs for accommodations at school or work, the experience of being housebound, veterans of the Canadian Armed Forces with disabilities, Internet use, methods used to access government services and sources of income.

Subjects

  • Disability
  • Equity and inclusion
  • Health
  • Society and community
  • Work, income and spending

Data sources and methodology

Target population

The population covered by the CSD was composed of all persons aged 15 and over (as of May 10th 2016, Census Day), and who reported having difficulty "Sometimes", "Often" or "Always" to one of the Activities of Daily Living questions on the 2016 Census of Population long form. This included persons living in private dwellings in the 10 provinces and three territories. Persons living on a First Nations reserve were not included, nor were those living in collective dwellings, such as institutional residences and Canadian Armed Forces bases.

The target population for the CSD corresponded to a subset of the covered population, namely persons who reported on the CSD that they were limited in their daily activities due to a disability.

Instrument design

The questionnaire was developed in collaboration with Employment and Social Development Canada (ESDC). A content review of the 2012 CSD was conducted with subject matter experts and stakeholders to identify any potential data gaps. Input was obtained from ESDC's Technical Advisory Group (TAG) on disability which consisted of representatives from various community associations across Canada. Specialized consultation was also provided by Veterans Affairs Canada and Service Canada.

New questions were added to the 2017 CSD, including seven new modules in the areas of episodic disabilities, general health, use of various therapies and social services supports, the experience of being housebound, veterans of the Canadian Armed Forces, Internet use and accessibility of government services. Two age-related questions were introduced for each of the 10 disability types. The first asked respondents at what age they began having difficulty with a health-related problem or condition. This question identified the age at which their difficulty or condition first started (the onset). The second question determined at what age the difficulty or condition began to limit their daily activities.

The labour force activities component of the questionnaire also underwent modifications to better reflect standard employment indicators found on labour surveys by Statistics Canada (known as harmonized content). This approach was taken to be consistent with other Statistics Canada surveys.

The 2017 CSD underwent the transformation from a Computer Assisted Telephone Interview (CATI) mode of collection to an Internet-based Electronic Questionnaire (EQ). This required some adaptations in order to craft a user friendly EQ model. Modifications included presentation and wording of some questions and answer categories as well as the addition of help text.
The 2017 questionnaire was tested in both official languages throughout the development. Qualitative content testing was conducted by Questionnaire Design Resources Centre (QDRC) at Statistics Canada and in several off site locations across Canada. The EQ application underwent qualitative testing by QDRC in conjunction with ESDC.

One final step was taken with respect to content development for the 2017 CSD: an in-depth review of 2016 Census variables for their potential record linkage with the CSD dataset. This resulted in the addition of approximately 300 Census variables covering 15 distinct subject matter areas linked to the final CSD data files for 2017.

Sampling

This is a sample survey with a cross-sectional design.

The sample design was a stratified two-phase design based on the 2016 Census. The first phase was the Census itself, and corresponded to the sample of households selected to receive the long form, so about one household out of four, systematically selected across Canada. Phase 2 corresponded to the sample of persons who reported having difficulty on the Activities of Daily Living questions on the long form Census.

The sampling unit for phase 1 (the Census) was the household, while that of phase 2 was the person.

Strata were defined so as to guarantee sufficient sample sizes in each domain of estimation and optimize sample allocation. The domains of estimation consist of the provinces cross-classified with the following age groups: 15 to 24, 25 to 44, 45 to 64, 65 to 74, and 75 and over. In Prince Edward Island, those aged 15 to 24 were grouped with those aged 25 to 44 into a single domain. In the territories, all age groups were combined (15 years and over).

Strata were formed within each domain by taking into account the Census sampling design so as to group together persons with similar preliminary sampling weights (i.e. remote areas vs non remote areas) and potential severity of the disability (i.e. mild, moderate or severe) which is approximated by considering the answers to the Activities of Daily Living questions (No, Sometimes, Often or Always) and the number of them which were positive. Hence six strata were created within each domain by crossing three severity levels by the fact of living in a remote area or not.

Allocation method
Sample allocation was done to be able to estimate a minimum proportion with a maximum coefficient of variation (CV) of 16.5% in each estimation domain. A CV of 16.5% corresponds to the upper limit for a CV so that the estimate is considered to be of an acceptable quality. The minimum proportion to estimate was set to 10% for those in the group aged 15 to 24 years old, to 8.5% for the groups aged 25 to 44 and 45 to 64, to 12% for those aged 65 to 74, and to 13% for those aged 75 years and over. For Prince Edward Island, the minimum proportion to estimate for those aged 15 to 44 years old was set to 8.5% while in the three territories for those aged 15 and over, it was set to 9%.

Sample size
Sample sizes computed with these parameters were further increased to compensate for two types of losses : the false positives (persons who reported difficulties to the Activities of Daily Living questions on the Census but who do not have a disability according to the CSD) and for non-response.

The total sample size for the CSD was 50,000 persons.

The sample was drawn using systematic sampling with the frame being sorted by collection unit to minimize the chance of selecting more than one person per household.
While the CSD did not cover persons who responded "No" difficulties or conditions on the Activities of Daily Living questions on the Census questionnaire, a sample of these individuals (called the NO sample) was nonetheless included in the final CSD data files. These people are all considered to be people without a disability. This sample allows the computation of disability rates, which require estimates for the entire population, not just persons with a disability.

An additional sample of approximately 5,000 persons was also drawn to be part of a methodological research project. This sample of persons was also drawn among persons who did not report any difficulties or conditions on the Activities of Daily Living questions on the Census. The Disability Screening Questions were asked of these persons to see if they had a disability or not. This allowed methodologists to determine the extent to which the new questions on the Census covered persons with a disability.

For more details on the sampling design, the domains of estimation, the stratification and the allocation, please consult the 2017 CSD: Concepts and Methods Guide.

Data sources

Data collection for this reference period: 2017-03-01 to 2017-08-31

Responding to this survey is voluntary.

Data were collected directly from survey respondents and also linked from the 2016 Census.

Data collection for the CSD was done using an Internet-based electronic questionnaire (EQ), and involved two types of collection methods: a self-reporting method with the questionnaire completed directly on-line by the respondent (rEQ) and an interviewer-led method conducted by telephone (iEQ). Initial determination of collection method for a CSD respondent was based on the method the respondent had used for the 2016 Census. Respondents all received an invitation to participate in the survey by the mail, with the rEQ respondents receiving a link to the EQ and a secure access code, while the iEQ respondents were informed that they would be contacted by telephone. Reminder letters were sent to rEQ respondents approximately two weeks apart for the duration of collection. Collection for rEQ and iEQ were done in parallel for the first part of collection, and all rEQ non-respondents were transferred to iEQ for follow-up. Overall, approximately 40% of respondents completed the self-reporting rEQ and 60% completed the telephone interview iEQ. All survey responses were kept highly secure through industry-standard encryption protocols, firewalls and encryption layers.

Data was obtained directly from the selected respondents, but proxy interviews were allowed under some conditions.

The EQ was available in English and French. Interviews lasted about 35 minutes on average. To reduce interview time, Statistics Canada has combined information from the CSD to selected data from the 2016 Census. It is possible that data from other surveys or administrative data sources may be added at a later date.

View the Questionnaire(s) and reporting guide(s) .

Error detection

All responses to the 2017 CSD questions were captured directly in the EQ application, both for the interviewer-led (iEQ) component and the respondent self-reporting (rEQ) component. The EQ application, just like any other computerized questionnaire, reduces processing time and costs associated with data entry, transcription errors and data transmission.

For some CSD questions, data underwent a preliminary verification process when respondents were completing the survey. This was accomplished by means of a series of edits programmed into the EQ. That is, where a particular response appeared to be inconsistent with previous answers or outside of expected values, the interviewer or the self-reporting respondent was notified with an on-screen warning message, providing them with an opportunity to modify the response provided.

Once survey responses were transmitted to head office, more extensive data processing for the CSD began. This involved a series of steps to convert the questionnaire responses from their initial raw format to a high-quality, user-friendly database involving a comprehensive set of variables for analysis.

A series of data operations were executed to clean files of inadvertent errors, remove duplicate records, edit the data for consistency, code open-ended questions, create useful variables for data analysis, and finally to systematize and document the variables for ease of analytical usage.

Imputation

For the CSD, discrepancies, logical inconsistencies and missing information were resolved wherever possible either automatically using customized deterministic editing rules or through manual interventions.

Estimation

In a sample survey, each respondent represents not only himself or herself but also other people who have not been sampled. For that reason, each respondent is assigned a weight which indicates the number of people that he or she represents in the population. To maintain data coherence and ensure that the results accurately represent the target population and not just the individuals sampled, that weight must be used to compute all estimates.

There are several steps in calculating the weights for the CSD. The first step is to assign each selected unit an initial weight based on the sample design. The initial weight is the inverse of the probability of inclusion. For the CSD, the initial weight is the product of two factors: the Census long form questionnaire weight and the CSD subsampling weight (the inverse of the sampling fraction). Then a number of adjustments are made to the weights to control for exclusions during collection, for non-response and to avoid extreme weights in the estimation domains. The final step is to calibrate the survey weights on the Census estimated totals and make certain adjustments to account for units that were in scope during the May 2016 selection process but out of scope at the time of the survey in 2017.

Weights were also computed for the approximately 140,000 persons selected for the NO sample added to the analytical file to allow comparison of disability rates. Again, the "NO sample" were those who did not report any difficulties or conditions on the Activities of Daily Living questions on the Census. The initial weight for the NO sample was the product of the Census long form final weight and the inverse of the CSD sampling fraction of the NO sample. A calibration of these weights was done first on the Census estimated totals, followed by an adjustment to account for the estimated number of units who became out of scope between Census day and time of CSD collection.

The general bootstrap method for two-phase sampling was used again in 2017 to compute bootstrap weights. This special method was developed in 2006 for the Aboriginal Peoples Survey (see É. Langlet, J.-F. Beaumont, and P. Lavallée. (2008), "Bootstrap Methods for Two-Phase Sampling Applicable to Postcensal Surveys", paper presented to Statistics Canada's Advisory Committee on Statistical Methods, May 2008, Ottawa.) Slight modifications were brought to the 2012 CSD method to account for the return of the mandatory Census. For the purpose of calculating the variance only, the Census sampling design was viewed as a two-phase design, where the first phase consisted of sampling approximately one in four households, and the second phase consisted in the sample of respondents to the Census. Although the Census response rate is quite high, this second phase allows to account for this non-response in variance calculation. To be able to use the general bootstrap for two-phase sampling, the two census phases were combined into one phase, while the CSD sampling was considered to be the second phase.

Quality evaluation

Quality assurance measures were implemented at every collection and processing step. Measures included recruitment of qualified interviewers, training provided to interviewers for specific survey concepts and procedures, observations of training of interviewers as well as interviews to correct questionnaire design problems and instruction misinterpretations, procedures to ensure that coding errors were minimized, and edit quality checks to verify the processing logic. Data were verified to ensure internal consistency and were also compared to other sources when available.

For 2017, the CSD included the full implementation of the Disability Screening Questions (DSQ) used for identifying persons with disabilities. In 2016, the Activities of Daily Living question on the Census, which serves to create the sampling frame for the CSD, was replaced by new filter questions taken from the DSQ framework. Qualitative and quantitative testing have shown that the new filter questions allows for better coverage overall of persons with disabilities, and especially of persons with disability types that are less visible, such as disabilities related to pain, memory, learning, development and mental health. One important consequence of this full implementation is that the disability rates observed in the 2017 CSD are not comparable to those of the 2012 CSD, but are very much consistent with what was expected and observed during testing.

2017 Canadian Survey on Disability (CSD) and disability data from other surveys:
Several other surveys use the Disability Screening Questions (DSQ) to identify persons with disabilities (PWD) to allow comparisons of survey characteristics between PWD and persons without disabilities (PWoD).

Given that each survey has its own target population, reference period, collection mode, content and question order, non-response pattern and rate, and survey methodology, estimates regarding the prevalence of disability and/or estimates of the number of PWD based on the questions from the disability module used in these surveys should not be published. Instead, this DSQ module only serves to produce estimates of proportions among PWD showing particular characteristics to compare with the proportions among PWoD.

However, one exception is made for another post-censal survey: the Aboriginal Peoples Survey (APS). The APS specifically targets First Nations people living off-reserve, Métis and Inuit in Canada. For the first time in 2017, the APS used the DSQ to identify persons with a disability. Since the APS has a much larger sample of Aboriginal persons with a disability than the CSD and the APS sample is considered more representative of the Aboriginal population, the APS is the official source of disability rates for Aboriginal persons.

To ensure that the APS rates are comparable to those of the CSD, the CSD methodology was applied to the APS: anyone in the APS sample who did not report having a difficulty or long-term condition on the Activities of Daily Living question on the 2016 Census of Population was considered not to have a disability, regardless of their answers to the DSQ on the APS. For more information on the CSD methodology, refer to the 2017 CSD Concepts and Methods Guide.

The official source for disability-specific data in Canada, such as prevalence and counts, is the Canadian Survey on Disability. However for disability counts and prevalence with regards to Aboriginal persons, the Aboriginal Peoples Survey (APS) is the official source.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects that could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

Revisions and seasonal adjustment

This methodology type does not apply to this statistical program.

Data accuracy

Two types of errors occur in surveys: sampling errors and non-sampling errors.

The sampling error measure used for the CSD is the coefficient of variation (CV) of the estimate, which is the standard error of the estimate divided by the estimate itself. In this survey, when the CV of an estimate is less than or equal to 16.6%, the estimate can be used without restriction. When the CV is greater than 16.6% but less than or equal to 33.3%, the estimate will be accompanied by the letter "E" to indicate that the data should be used with caution. When the CV of an estimate is greater than 33.3%, the cell estimate will be replaced by the letter "F" to indicate that the estimate was suppressed for reliability reasons.

Non-sampling errors arise primarily from the following sources: non-response, coverage, measurement and processing. Total non-response will produce a bias if non-respondents have different characteristics from respondents and if non-response is not corrected properly. Non-response adjustments, combined with a relatively high response rate, helped reduce this risk of bias substantially for the CSD. Non-response to specific questions is often due to difficulty understanding the questions. Thorough quality reviews and questionnaire testing were carried out before the survey, which reduced the extent of partial non-response.

Collection for the CSD ended with an overall response rate of 69.5%. This response rate is the number of complete respondents divided by the number of cases sent to collection minus the out-of-scope cases (cases which did not meet eligibility criteria for the survey). Hence, this rate reflects the percentage of cases that completed the interview relative to the number of cases that should have completed it.

Response rates for the provinces ranged from 66.5% for New Brunswick to 78.5% in Quebec. In the territories, the response rates ranged from 51.5% in Nunavut to 65.6% in the Yukon. By age group, the response rates ranged from 62.1% for those aged 15 to 24 years old, to 77.0% for those aged 65 to 74 years old.

Coverage errors occur when there are differences between the target population and the sampled population (or survey population). In particular, undercoverage can be problematic. Since the CSD sample is selected from the Census, Census non-respondents could not be sampled for the CSD. With the very high response rate of the 2016 Census long form (97.8%), the scope of this problem is minimal.

Measurement errors occur when the response provided differs from the real value. Such errors may be attributable to the respondent, the interviewer, the questionnaire or the collection method, for example. Every effort was made to develop questions that would be understandable, relevant and appropriate for respondents through extensive content development activities and qualitative testing. Other measures were also taken, including the use of skilled interviewers, extensive training of interviewers, and observation and monitoring of interviewers.

Processing errors may occur at various stages, including data capture, coding and editing. Quality control procedures were applied at every stage of data processing to minimize these types of errors.

Documentation

Date modified: