Survey of Employers on Workers' Skills (SEWS)

Detailed information for 2021

Status:

Active

Frequency:

Occasional

Record number:

5317

The survey is designed to collect information from businesses on issues of skills requirements and skills development, as well as on the strategies deployed to mitigate the consequences of skill gaps within their businesses.

Data release - October 3, 2022

Description

Statistics Canada, in partnership with Employment and Social Development Canada (ESDC), is conducting this business survey to capture information from employers on issues related to the demand for skills in the economy. The survey will primarily collect information on employers' skills needs and skills gaps as well as their human resources management practices, work organization, training programs, and talent recruitment and retention programs.

The data collected will be used to support ESDC's policies and programs pertaining to skills and training needs of Canadian employers. As well, the collected information and in particular the prevalence of skill gaps across industry and firm size classes, is important in order to stimulate policy actions to address pressures and imbalances in the labour market.

Data sources and methodology

Target population

All business with 1 employee or more, with a yearly revenue higher than $30,000 and do not operate in the following industry class from the North American Industry Classification System (NAICS): 118131, 814 and 91.

Instrument design

The collection instrument for this survey is an electronic questionnaire. The questionnaire is the result of collective input from stakeholders both internal and external to Statistics Canada.

Sampling

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

A stratified random sample of business locations classified to the North American Industry Classification System (NAICS) Canada 2017

Sampling unit
Establishment

Data sources

Data collection for this reference period: 2022-03-08 to 2022-05-04

Responding to this survey is voluntary.

Data are collected directly from survey respondents.

Collection method: Electronic questionnaire
Method of initial contact: Pre-contact (phone)
Follow-up method: Email remainders and phone follow-up
Language(s): French and English
Average time: 1 hour

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

Error detection

Error detection is an integral part of data processing activities. Prior to imputation, a series of edits are applied to the collected data to identify errors and inconsistencies. Errors and inconsistencies in the data are reviewed and resolved by referring to data for similar units in the survey and information from external sources. If a record cannot be resolved, it is flagged for imputation. Finally, edit rules are incorporated into the imputation system to detect and resolve any remaining errors, as well as to ensure that the imputed data are consistent.

Imputation

After microdata verification, a variable was created for each of the survey variables to identify those that had either failed the verification rules or had missing values. Imputation was performed to reduce the amount of missing, inconsistent or incomplete data. The missing data were imputed using a randomly selected donor inside the imputation class. These imputation classes were formed based on statistical analysis performed with frame information or previous variables on the questionnaire.

A minimum number of units was required within each imputation class. When imputation classes were too small, larger classes were created by combining several classes together.

Imputation of survey variables was performed in an automated way using BANFF, a generalized system designed by Statistics Canada.

Estimation

Estimation is a process by which Statistics Canada obtains values for the population of interest so that it can draw conclusions about that population based on information gathered from only a sample of the population. For this survey, the sample used for estimation comes from a single-phase sampling process.

An initial sampling weight (the design weight) is calculated for each unit of the survey and is simply the inverse of the probability of selection. The weight calculated for each sampling unit indicates how many other units it represents.

However, since some of the selected units did not answer the survey, reweighting is performed on the responding units so that their final weights still represent the whole target population. While imputation is used to correct for item non-response, complete non-response is treated through reweighting. The weights of respondent records were adjusted to compensate for the non-respondents.

After the reweighting is performed, a calibration process is performed so that the weighted totals per calibration groups equal the population totals.

Estimation of proportions and other statistics is done using the final calibrated weights to calculate the population parameters for the domains of interest. The response values for sampled units were multiplied by the final calibrated weight in order to provide an estimate for the target population.

Quality evaluation

Estimates were reviewed to ensure that the findings are logical and quality checks were carried out to ensure that estimates are consistent. Atypical results were flagged for investigation and were corrected as necessary.

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

There are two types of errors which can impact the data: sampling errors and non-sampling errors.

Estimates are subject to sampling error. This error can be expressed as a standard error. For example, the proportion of businesses in the target population that would respond YES to a given question is estimated to be 50%, with a standard error of 4%. In repeated sampling, the estimate would be expected to fall between 46% and 54%, nineteen times out of twenty. The following rules based on the standard error (SE) are used to assign a measure of quality to all of the estimates of percentages.

A = Excellent (0.00% to less than 2.50%)
B = Very good (2.50% to less than 5.00%)
C = Good (5.00% to less than 7.50%)
D = Acceptable (7.50% to less than 10.00%)
E = Use with caution (10.00% to less than 15.00%)
F = Too unreliable to be published (Greater than or equal to 15%, data are suppressed)

Non-sampling errors may occur for various reasons during the collection and processing of the data. For example, non-response is an important source of non-sampling error. Under or over-coverage of the population, differences in the interpretations of questions and mistakes in recording and processing data are other examples of non-sampling errors. To the maximum extent possible, these errors are minimized through careful design of the survey questionnaire and verification of the survey data.

Report a problem on this page

Is something not working? Is there information outdated? Can't find what you're looking for?

Please contact us and let us know how we can help you.

Privacy notice

Date modified: