Canadian Survey on the Provision of Child Care Services (CSPCCS)

Detailed information for April 2022

The purpose of this survey is to collect information on the provision of child care services in Canada for children ages 12 and under at the national, provincial and territorial level. Data is collected from licensed and unlicensed home-based and centre-based child care providers. Questions will be asked about staff, services provided, enrollment and daily fees as well as the extent of challenges related to COVID-19. The data will be used by Employment and Social Development Canada (ESDC) for policy research and development.

Data release - March 22, 2023

• Questionnaire(s) and reporting guide(s)
• Description
• Data sources and methodology
• Data accuracy

Description

Statistics Canada is conducting the Canadian Survey on the Provision of Child Care Services (CSPCCS), in collaboration with Employment and Social Development Canada (ESDC). Data will be gathered on:
- status and type of licensing of child care providers,
- type of child care services offered,
- number and age group of enrolled children and related daily fees,
- type of building where child care services are provided,
- number, type of tenure, and relevant training of child care employees,
- sources of operating funds, and
- existence and extent of challenges related to COVID-19;

This is intended to gather national, provincial and territorial level information on child care provision in centres, licensed and unlicensed homes.

Subjects

• Child care
• Children and youth

Data sources and methodology

Target population

The target population for this survey includes all child care locations in Canada with an annual revenue of at least $2,500 per year. To identify these units, the sampling frame was built by taking locations coded to North American Industry Classification System (NAICS) 624410 on the Business Register in addition to centres matched from licensed provincial lists, whose primary activity does not fall under NAICS 624410 code. All units with revenue less than $2,500 were excluded from the frame.

Instrument design

The questionnaire was developed by Statistics Canada in partnership with Employment and Social Development Canada (ESDC). Questionnaire Design Resource Centre (QDRC) conducted cognitive testing with selected respondents and made recommendations that were implemented to finalize the questionnaire. Data will be collected electronically. An HTML format of the electronic questionnaire is uploaded on the IMDB page.

Sampling

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

The sampling frame is extracted from Statistics Canada's Business Register (BR). The sampling unit is the location, as defined in the Business Register.

Sampling unit
The sampling units are all home-based licensed, home-based unlicensed and centre-based child care locations in Canada.

Stratification method
Units were stratified into three groups: Licensed Centre Predominance (LCP), Licensed Home Predominance (LHP) and Unlicensed Home Predominance (UHP). These three strata were further stratified into eleven regions, consisting of the ten provinces and the territories combined into one region.

Centres on the frame that were identified as having a particularly large number of employees on the Business Register formed their own stratum where all locations within it were selected in the sample with certainty.

Sampling and sub-sampling
The total sample size was set at 20,000 units, and units were allocated to ensure similar quality across regions and balance among the estimation domains, which are centres, licensed home-based child care and unlicensed home-based child care in each of the eleven regions. The allocation to strata was made so that the coefficient of variation for estimated proportions of 50% within estimation domains is expected to be around 2.9% for centres as well as unlicensed home and 4.3% for licensed homes.

Data sources

Data collection for this reference period: 2022-04-07 to 2022-07-06

Responding to this survey is voluntary.

Data are collected directly from survey respondents.

Data will be collected via electronic questionnaire (EQ).

Respondents with email address in the Business Register System will receive email invitation while the remaining receive invitation letter with Security Access Code (SAC) to the EQ in the mail.
Computer assisted telephone interview (CATI) will be done for failed edit (FEFU) and non-response (NRFU) cases.

The average time required to complete the interview / survey is about 35 minutes.

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

Error detection

Data editing is the application of rules to detect missing, invalid or inconsistent entries or to identify data records that are potentially in error. In the survey process, data editing is done at two different time periods.

First, editing is done during electronic questionnaire collection. Edits during data collection generally consist of validity and some simple consistency edits. Follow-up with respondents occurs to validate potential erroneous data following any failed preliminary edit check of the data. Once validated, the collected data is regularly transmitted to the head office in Ottawa.

Second, editing known as statistical editing is also done after data collection. Units with large deviations from a predictive frame variable are not used at the imputation stage. Furthermore, small outliers indicate a misselection of the unit of measure and are converted into the most plausible unit of measure. Large and small outliers cannot be used at the imputation stage.

Imputation

In the case of partial non-response, imputation is used to fill in information not provided by the respondent. Imputation makes it possible to have a complete set of data if one cannot collect it during the collection period.

The imputation was primarily done using donor imputation, where missing values in units are replaced with the value from a donor unit belonging to the same imputation class. Imputation classes are defined so that units within each one are similar in terms of characteristics of interest for the survey. Relationships between survey variables and variables used to define the imputation class are tested to ensure the existence of significant relationships. A small number of variables are imputed deterministically, based on pre-specified rules.

Estimation

Estimating the characteristics of a population from a survey is based on the assumption that each sampled unit represents a certain number of non-sampled units in the population. An initial weight is assigned to each unit to indicate the number of units in the population represented by that unit in the sample. Very important or otherwise unique units are assigned a weight of one to ensure that they only represent themselves.

Adjustments are made to the initial weights to improve the representativity of the sample and the reliability of the estimates. The weights are adjusted to compensate for total non-response, and then calibrated to benchmark counts.

Quality evaluation

Prior to publication, survey results are analyzed for comparability. This includes a detailed review of the estimates, and a comparison to other industry sources, historical trends and general economic conditions.

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 does not apply to this statistical program.

Data accuracy

The estimates obtained from sample surveys are subject to both sampling and non-sampling errors.

Non-sampling errors may occur throughout a survey for many reasons, such as non-response, coverage and classification errors, differences in the interpretation of the question, incorrect information from respondents, as well as mistakes during data capture, coding, and processing. Efforts to reduce non-sampling errors include careful design of questionnaires, editing of data, follow-up, imputation for non-responding units, and thorough control of processing operations.

The use of sampling frames results in coverage errors, notably undercoverage. Undercoverage occurs when the information on a location is incomplete in the Business Register. This normally happens in the case of new locations that have not yet filed tax forms with the Canada Revenue Agency.

Sampling errors occur because observations are obtained from a sample rather than from the entire population. Estimates based on a sample can differ from statistics that would have been obtained if a complete census had been taken using the same instructions, interviewers and processing techniques. This difference is called the sampling error of the estimate.

The true sampling error is unknown. However, it can be estimated from the sample itself by using a statistical measure called the standard error. The standard error can be used to build a confidence interval for the estimate. When the standard error is expressed as a percentage of the estimate, it is known as the relative standard error or the coefficient of variation (CV).

Estimates from this survey are assigned a quality indicator in the form of a letter to indicate their quality level (A-best, F-worst). The indicators take into account various factors that affect the quality of the data, notably the CV, the non-response errors, and the imputation errors. These indicators are updated each cycle to reflect the quality of the current estimates.

Non-sampling error
Non-sampling errors may occur throughout a survey for many reasons, such as non-response, coverage and classification errors, differences in the interpretation of the question, incorrect information from respondents, as well as mistakes during data capture, coding, and processing. Efforts to reduce non-sampling errors include careful design of questionnaires, editing of data, follow-up, imputation for partial non-response, and thorough control of processing operations.

Non-response Bias
The units in the sample are assigned initial weights at the sampling step, to make them represent a certain number of units in the population. The weights of the responding units are further increased to make these responding units represent non-responding units, as well as the population units they were meant to represent initially.

In order to reduce possible non-response bias, this weight adjustment is performed on groups of units sharing the same characteristics as much as possible ("homogeneous response groups") defined using variables from the survey frame or paradata obtained from the collection step.

Coverage error
The use of sampling frames results in coverage errors, notably undercoverage. Undercoverage occurs when the information on a location is incomplete in the Business Register. This normally happens in the case of new locations that have not yet filed tax forms with the Canada Revenue Agency.

The Business Register is kept up to date continuously.