ECS 2019 - Sampling and weighting
In the European Company Survey (ECS) 2019, a total of 21,869 management interviews were completed, ranging from 122 in Cyprus to 1,498 in Italy. Data were collected from 3,073 employee representatives in total, ranging from 3 in Cyprus to 467 in Finland. (For more details, see Table A2 on p. 150 of the ECS 2019 overview report)
Sampling of establishments followed a multistage random sampling approach stratified by establishment or company size (10–49 employees, 50–249 employees, 250+ employees) and the broad sector of activity (production, construction, services). The approach aimed for a balance between representativeness at the level of the number of establishments and representativeness at the level of the number of employees working in these establishments. In countries with a company-level sampling frame, a screening procedure was used to randomly select an establishment within the company.
Contact strategy and respondent selection
Establishments were contacted via telephone. Interviewers asked to be put in contact with the most senior person in charge of personnel (in large establishments, the target was ‘a senior person in charge of personnel’). Once in touch with the target management respondent, the interviewers proceeded with the identification and selection of respondents for the employee representative interviews, based on a series of questions designed to match the institutional structure in each country.
The national-level experts from the Network of Eurofound Correspondents were asked to indicate which employee representation bodies could potentially be involved in workplace-level discussions on work organisation practices, and what the appropriate terminology was in the language(s) of their country. (For more details, see Table A3 on p. 152 of the ECS 2019 overview report)
The identification and selection questions were put to the management respondent as part of the screening interview. A further attempt to identify a respondent for the employee representation questionnaire was made at the end of the management questionnaire, at which point the manager was prompted once again to provide the contact details of an employee representative.
Weighting needs to be applied to ensure that the findings of this survey are representative of the number of establishments in terms of distribution across sectors, size classes and countries. It corrects for any disproportionalities in representation due to sampling design and non-response.
The weighting procedure for the ECS 2019 was carried out in four steps.
- Design weighting to correct for the fact that the sample was designed to overrepresent larger establishments.
- Non-response weighting to correct for differences between establishments that did and did not complete the screener interview, and then between respondents that did and did not complete the questionnaire online.
- Calibration against distributions found in the population statistics.
- Adjustment to reflect the relative size of the population in each of the countries in the survey.
An additional stage was required if the sampling frame was at the company level to correct for the fact that some companies consist of multiple establishments. After weighting, the deviations between population targets and the management respondent data by stratification cell remained below five percentage points for all cells.
Detailed information on sampling and weighting can be found in the Sampling and weighting report.
Add new comment