Research talk:Formerly active editors survey
- Loaded questions (questions which suggest an answer) should be avoided. For the reasons to leave, there were some options in past General User Survey which could be reused to compare.
- Per  some demographic question is needed for comparison (maybe some ISTAT or Nielsen study?).
- On motives etc., it's important not to make this into yet another study on motivation; perhaps the survey should avoid to answer questions on motivation at all, or just keep them as comparison (copying those used by some relevant past study).
Comparable questions are useful to correct selection bias (imagine if only males answer, or only banned users, or only people who edited for work). --Nemo 17:11, 21 April 2015 (UTC)
Può essere utile: http://www.pewresearch.org/2016/05/02/evaluating-online-nonprobability-surveys/ Nemo 15:08, 3 May 2016 (UTC)