People and feedback are hard to do right
|
(English) This is an essay. It expresses the opinions and ideas of some Wikimedians but may not have wide support. This is not policy on Meta, but it may be a policy or guideline on other Wikimedia projects. Feel free to update this page as needed, or use the discussion page to propose major changes. |
Surveys are pieces of writing which reach a large amount of people, such as
- SurveyMonkey surveys (details on their usage here, for instance)
- Queries for feedback with a big audience (such as this one, advertised in global banners)
- Petitions, pledges, and letters spread widely, persuading readers of one specific thing to suppoer (see i.e. this one or that one or the other one or here)
Where people are allowed to read past discussions (this means #2 and #3), bias is likely to be introduced by definition. See herding instinct, bandwagon effect, crowd psychology.
Filling a survey is a obnoxiously hard job. Textbox. Tick box. Textbox. People end up filling it loosely without thinking just to get rid of it all. All feedback left should be voluntary and in free-form (one big textbox).
A large number of supporters may be suggestive of something, but I'd perhaps say it isn't even indicative that a problem even exists! When it does, they may be approaching one of its consequence without touching the root issue...
It is, in my view, important to not try to guess the response in advance. Otherwise it is very easy to misinterpret the result.
To get proper understanding, you need to write all facts briefly, and give people one big textbox for writing response. Structure it later. (For instance, you introduced new software. Don't try to guess that people will say "I like it", "I don't like it", "I want to improve it using ...". This limits peoples' thoughts. Maybe they want to say "this reminds me of ..." and this thought is useful for your future development?)
Where you would like people to participate and see past discussions, your ‘post-processing’ of feedback may also introduce bias.