Publication details
Srovnání kvality dat z „pravděpodobnostního“ a nepravděpodobnostního vzorku : Příklad českého výzkumu „GGP-Současná česká rodina“
| Title in English | Comparing data quality in a probability and a non-probability sample : An investigation based on the GGP-contemporary Czech family project |
|---|---|
| Authors | |
| Year of publication | 2024 |
| Type | Appeared in Conference without Proceedings |
| MU Faculty or unit | |
| Citation | |
| Attached files | |
| Description | Traditional methods of collecting sociological data using questionnaire surveys from random samples of the population face increasing difficulties, including diminishing returns, (potentially) declining representativeness, and increasing cost. One way to address these difficulties is to use data from non-probability (opt-in) panels. In this paper, using data from the GGP-Contemporary Czech Family Survey (from 2020-2022), we compare the quality of data obtained in the traditional way (a probability sample obtained by randomly selecting address points and dialing random mobile phone numbers, N=5583) and data obtained from a pilot study of the same survey conducted on an opt-in panel of the SCaC agency (N=1313). Using descriptive statistics, we show that for a number of variables, the variables have the same distribution and the two samples are indistinguishable. Similarly, using multivariate methods, we show that (controlling for other variables) the effect of sample type on the dependent variable is often negligible. For many purposes, samples can be effectively pooled. However, before any analysis, the specific variables need to be carefully examined to rule out sample effects. For future research, it is always advisable to include the possibility of collecting data in opt-in non-probability panels. |
| Related projects: |