Good morning everyone,
Due to the pandemic, I conducted several online surveys. It is quite agile, not at all expensive, but there are several problems: the sample's representativeness and its bias, the difficulty of accessing social profiles that do not have social networks and a certain distrust. What other problems would you highlight? What solutions would you recommend? Are the results - fundamentally the search for associations between variables - reliable given a certain level (n> 50) of observations for each variable? Anyway, what are your opinions about this tool and how can we improve the research design to incorporate online surveys in an appropriate way? Thank you very much
Sample "coverage" is probably the main issue in online surveys, and journals such as Public Opinion Research have published articles on this topic for the past decade or more. One prominent option is "mixed mode" surveys where you collect data both online and by phone.
As for the association between variables, this kind of statistical issue is the same regardless of how the data are collected. Rather than informal criteria such as 50 observations for each variable, you should use a program such as G*Power to determine the necessary "power" of your test.
- Badges
- Science topic