|
Unit of Analysis
| individual |
|
Universe
| Population of the Czech Republic aged 15 and older. |
|
Time Method
| panel survey |
|
Data Collector
| Public Opinion Research Centre (CVVM), Institute of Sociology of the Czech Academy of Sciences |
|
Frequency
| The unweighted dataset is biased towards a more educated population. |
|
Sampling Procedure
| Panel respondents are recruited using a multi-stage stratified cluster sampling with repetition. The sampling frame is based on the Register of Census Districts and Buildings (RSO). The strata consist of regions crossed with municipality size (66 strata). Within each stratum, the primary sampling unit is the municipality. The selection of a municipality takes into account its population size (a larger municipality has a higher probability of being selected) and the selection allows for a repeated selection of a municipality. Within a municipality, 6 building addresses are always randomly selected, again with probability based on the population of the building (with repetition). Within a building, a household is randomly selected (without repetition) and within a household, a member is randomly selected. |
|
Major Deviations for Sample Design
| The unweighted dataset is biased towards a more educated population |
|
Collection Mode
| Computer-Assisted Web Interviewing (1720 respondents). Computer-Assisted Personal Interviewing (30 respondents). |
|
Type of Research Instrument
| Standardized questionnaire. |
|
Actions to Minimize Losses
| Primarily, respondents are contacted by interviewers who either take their contact details and pass them on to CVVM staff or complete a registration questionnaire directly with them. In the former case, CVVM staff contact respondents by email or telephone to remind them to complete the registration questionnaire. Each respondent is provided with materials (a letter describing the panel, contact details and a unique login code, and a leaflet about the presence of the interviewer) that s/he receives when s/he is invited to the panel. At the same time, these materials are distributed to the selected households, even if they are not convinced to participate or no one was present at the time of the interviewer's visit. The interviewer contacts the household repeatedly (at least 3 times on different days and at different times of the day, at least one visit must be after 5 pm or on a weekend). Respondents receive monetary incentives for completing the registration questionnaire (not the contact questionnaire) and for completing the questionnaires (the CVVM is part of a public research institution, neither donations nor some other types of incentives can be used as incentives due to the laws governing the funding of public research institutions). CVVM commonly publishes the results of its surveys on its website https://cvvm.soc.cas.cz/cz/), so respondents will be regularly informed of the press releases issued. We regularly assess open-ended questions for those who refuse to participate in the panel and train interviewers to address these objections. A common response is that respondents do not feel competent enough to participate in the project and answer questions. Interviewers are tasked with explaining that this is not a matter of expertise, but rather the respondent's subjective opinion. |
|
Control Operations
| The distribution of basic sociodemographic characteristics (age, sex, education, region, size of municipality) was checked. |
|
Weighting
| Three types of weights are used: the design weight, the post-stratification weight and the election results weight. The design weight (variable name: w8_design_weight) takes into account the number of household members from which the respondent is selected and is the inverse of the number of household members. The aim is to adjust for the within household probability of selection. The post-stratification weight (variable name: w8_poststrat_weight_trim) uses population information from external sources and its aim is to reduce the potential sampling error and non-response bias. The weight was constructed using information on gender, age group, education, region and the size of the municipality as follows: 1) Gender (1 = male, 2 = female); 2) Age (1 = 15-19 years, 2 = 20-29 years, 3 = 30-39 years, 4 = 40-54 years, 5 = 55-64 years, 6 = 65+ years); 3) Education (1 = primary school + without matriculation exam = lower than secondary education; 2 = with matriculation exam = completed secondary education; 3 = university and other degrees = tertiary education); 4) Cross classification of region (NUTS 2 level (Eurostat NUTS) and municipality population (1 =up to 1 999 inhabitants, 2 = 2 000 – 4 999 inhabitants, 3 = 5 000 – 19 999 inhabitants, 4 = 20 000 – 49 999 inhabitants, 5 = 50 000+ inhabitants). For descriptive analyses, a weight was created in the dataset based on the actual results of the parliamentary elections (variable w8_elec_results_weight). This weighting variable automatically excludes from the analysis respondents who did not have the right to vote or refused to answer or did not know the answer for the party choice (i.e. these respondents who have no alternative in the election results have a value of 0 for the w8_elec_results_weight). |
|
Cleaning Operations
| Attention checks are included in the questionnaire and the length of the questionnaire is controlled. The combination of too short a time to complete the questionnaire and failure to complete the attention check leads to the questionnaire being dropped. |
|
Response Rate
| Recruitment rate (RECR): 31.3%, Profile rate (PROP): 70.5%, Actual survey completion rate (COMR): 92.8%, Cumulative response rate (CUMRR): 20.5%, Response rates are calculated according to AAPOR guidelines: AAPOR. „Standard Definitions - Final Dispositions of Case Codes and Outcome Rates for Surveys", 2016. |
