[purple_turtle]

> internet questionnaire posted to social media sites

This has got to have a strong selection effect.

[autoexec]

I think so, but that hasn't stopped it from being a popular way to gather data, or from often being the only source of data used in a paper. I get that internet surveys are dirt cheap and it's easy to get large number of responses, but you have to take the results with a such a massive boulder of salt that it seems more like a convenient way to churn out papers (or even an easy way to get a desired result) than a way to conduct actual worthwhile research.

[RobotToaster]

> an easy way to get a desired result

Reminds me of that bit from yes prime minister https://www.youtube.com/watch?v=6GSKwf4AIlI

[JackeJR]

Well a good surveyor would take into account order effects, i.e. the order in which questions were asked for example using Latin square designs.

[HPsquared]

An uncontrolled selection bias that has changed over time as the social media landscape has changed.

[philipwhiuk]

Yes, which is why the parent points out that it wasn't this.