Talk:Wikimedian Demographics

I'm not sure that the intended purpose of this project is really "demographics". One goal might be to provide a way to get Wikimedians from multiple projects to talk about important issues, which might be better served by using a different name for the project. If the goal is really collection of demographic data, I think this approach has a serious problem with sampling.

I'm not sure that a yes/no format for answers is a good way to start. I suspect that if we asked 100 different Wikimedians to explain what "Creative Commons is the right license" means, we would get about 50 different answers, most of them indicating that people really do not understand the topic. At the very least, I think there needs to be an effort made to educate people about the topic before trying to get useful answers to yes/no questions about a topic. --JWS 13:49, 12 October 2007 (UTC)

I'm not particularly attached to yes/no options, but they do allow categorization. I think a lot of issues we'll be hitting are those that have had long debates on mailing lists, community discussion pages, etc., and we can link to those when we set up questions. The GFDL vs. CC debate is a handy start in some ways though, because it's been debated endlessly on a number of fora (particularly on Wikibooks), and it's really just intended as an example anyway.

Perhaps the best way to make new questions is to talk about them a bit first, and come up with list of options that are finite but as close to exhaustive as possible. Narrowly defined questions will be more adaptable than broad ones.

It is really demographics, if you think about it. Demos as a prefix generally points to the population of a particular region (or in our case virtual community), and collecting statistics about a population is demography. The goal is not to elect, legislate, execute, or pass official judgement: they're just polls. --SB_Johnny | ^talk 15:01, 12 October 2007 (UTC)

women in Wikimedia projects

Possible topic: Women in Wikimedia projects from Milos Rancic

Sound good... what kinds of questions?--SB_Johnny | ^talk 15:01, 12 October 2007 (UTC)

Are you a woman? Do you know of an active contributer that is a woman? Are wikimedians sexist? etc..--Rayc 07:04, 27 January 2008 (UTC)

I had a hard time loading the pages on that site, but the questions look good. Would you mind sending him an email to see if we can use them? --SB_Johnny | ^talk 13:29, 28 January 2008 (UTC)

Might be worth coordinating with Wikichix. Cormaggio ^talk 12:53, 29 January 2008 (UTC)

I've forwarded an email SB_Johnny sent me to the WikiChix mailing list. Feel free to keep us up to date on this by emailing wiki@ the domain above. Angela 14:07, 29 January 2008 (UTC)

Thanks very much Angela :-), and yes, I'll mail the list when we get a first draft done. Btw (all), Milos' questions have been updated slightly to this version. I've just emailed Milos about using them for this purpose, and asked him for feedback. Cormaggio ^talk 17:30, 29 January 2008 (UTC)

Update: Milos says his questions are under GFDL (he was just "too lazy" to say so :-)), so we're free to use them. I'll have a go at turnign them into questions for a questionnaire - was thinking of having some answers "on a scale from 0 to 5", so might need some guidance there... Cormaggio ^talk 19:46, 29 January 2008 (UTC)

See also Wikipedia vs. Women. BTW, 93% of the people who have answered our survey, so far, are male.--mikeu ^talk 01:44, 30 January 2008 (UTC)

There's also a lot to be gleaned from the conversation which ensued on foundation-l after Wikichix was set up. As I noted to Milos yesterday, the problem was often that people didn't recognise there was a problem. However, that is more in depth than what we're doing here. I've hacked together a few questions at Template:Wmdgs-women-and-wikimedia-survey1 - and I'll ask the Wikichix list for feedback on how the survey could be improved, and for their participation in the survey (which should go some way to addressing the gender imbalance of respondents. Cormaggio ^talk 09:44, 30 January 2008 (UTC)

I added license data on the blog (box at the right side; GFDL, of course) :) And I'll try to find some time for contribution to Wikimedian Demographics :) --Millosh 10:18, 30 January 2008 (UTC)

Please ignore the link I gave above and go instead to Wikimedian_Demographics/Women_and_Wikimedia (partly created as a result of Mikeu's comments in the #Concern section below). Please add ideas and comments on questions and methodology. I've also sent an email to the wikichix list to solicit feedback. Cormaggio ^talk 13:07, 30 January 2008 (UTC)

Changing the basic survey

Hopefully a few folks are watching :). I'd like to propose changing the questions on the basic survey a bit, and have the answers do something more than just categorize, and maybe move the surveys to the main user page.

• For the questions, change to let people rate their participation in the various projects. For example, 3=primary project, 2=regular contributor, 1=occaisional contributor, 0=rarely if ever contribute.
• Set it so that the answers automatically create userboxes (but with an option in the template that allows that to be turned off).
• Have all templates combined into one big template, to which you just add fields as you participate in more polls.

Some fields could also just categorize someone as having taken a "non-dichotomous" survey (long answers, essays, etc.).

I think the project could be an excellent research tool for resources that study Wikimedia projects, such as Wikipedia arbitration committee, etc. --SB_Johnny | ^talk 11:24, 1 January 2008 (UTC)

I just took the survey (although I've known about it for a while) and I think those suggestions sound good. I was also thinking it might be nice to have a dynamic content summary page where it lists the ongoing results. BTW, I just added a userbox for Beta. I've never been big on userboxes though I think they can be usefull for somethings.--mikeu 12:32, 1 January 2008 (UTC)

Well, I've made a new startup template: {{Wmdgs-survey 1}}. It needs an explainer, but I'll add that today. It uses modified userboxes to display results, and has more subtlety in the answers available (using numeric values rather than yes or no for most questions). --SB_Johnny | ^talk 10:40, 4 January 2008 (UTC)

We should include betawikiversity: in the survey. See also Template:User Wikiversity Beta --mikeu 13:51, 7 January 2008 (UTC)

For the basic survey probably not, since we're not asking about which language projects either (the survey would be way too long if we did). The Wikiversity survey could certainly ask this though. (I spoke with Mike on irc about this a while back, then forgot to also reply here.) --SB_Johnny | ^talk 10:30, 25 January 2008 (UTC)

Disability status of Wikimedians

I think it would be good to have a survey asking about the disability status of wikimedians. I'm not entirely sure how this would work; there are way too many disabling conditions to make the questions condition-specific. I guess we could have them in categories, such as:

• Physical disability (such as paralysis, missing limb, blindness)
• Neurological disability (such as autism, epilepsy, cerebral palsy)
• Chronic ilness (such as lupus, chronic fatigue syndrome)
• Psychological disorder (such as depression, PTSD, phobia)

Then we could also have interesting questions like:

• Does your disability affect your job?
• Do you receive adequate support for problems arising from your disability?
• Do you believe that your disability has some positive effects on your life?
• Would you take a cure if one was available?
• Is the wikipedia article on the Social model of disability informative?

What do you think? --Luai lashire 02:55, 20 January 2008 (UTC)

Sounds good (and sorry for the delay... my watchlist runneth amok lately). Should these all be yes/no questions? --SB_Johnny | ^talk 10:26, 25 January 2008 (UTC)

I think this is a good idea. We might be able to add a field where the person specified the 'condition' within the category - something like "Neurological|Asperger's". However, as for the second set of questions, I think we should align them more towards Wikimedia - like whether they find relevant information here (a la the last one), and whether the condition affects their ability to contribute. I think some of the questions you have above might be better as a specific research project (perhaps by developing a more detailed survey of the people who answer these questions). Cormaggio ^talk 10:36, 25 January 2008 (UTC)

Cormaggio- do you mean like, "Does your disability affect your participation in Wikimedia projects?" instead of "Does your disability affect your job?" We could always have both, or neither- these were only suggestions. As for making a separate research project, yes, that could work, but I thought it might be best to have all or most surveys be a part of this learning resource- and then any other learning resource could use the information gathered here. Also, members of other learning resources could create surveys and submit them here as part of the project. --Luai lashire 18:08, 25 January 2008 (UTC)

Yes, I was thinking about keeping the survey relevant to Wikimedia projects, and having anything that's not as a separate survey to "Wikimedian Demographics" (though it could of course use/modify the same structure/templates/etc.). Cormaggio ^talk 12:15, 27 January 2008 (UTC)

Well, the problem with that is that it would make data harder to compare (at least not unless we get some rather beefy tools that can connect categories over 2 or more pages. For example, we might want to compare relative proportions of people who are "primarily wikipedians" vs. "primarily wikibookians" and different disabilities, sexual preferences, nationalities, political affiliations, etc. Keeping to one page as much as reasonable works a lot better for interpreting the data using the category systems. I do agree that the pages will eventually become ungainly if someone were to take all of the available surveys, but I suspect what we're going to find is that people won't take every survey that comes up anyway (for example, I would be unlikely to answer a survey about the politics of New Zealand (I know nothing about it), or my experience in the military (I have none)). --SB_Johnny | ^talk 16:52, 27 January 2008 (UTC)

Ok, what I take from that is that we're saying that these surveys can be about anything, so long as they are answered by Wikimedians. (I was wondering if my definition of "Wikimedian demographics" was too narrow.) In this case, we will probably need those beefy tools to deal with people taking many surveys. :-) Cormaggio ^talk 18:02, 27 January 2008 (UTC)

That was why I had named it "Wikimedian Demographics", as opposed to "Wikimedia something-or-other" :-). It's more about the contributors and users rather than just a set of political polls on Wikimedia-related subjects (though of course we should include those too). We should however perhaps do some bot shopping before we hit the limits on the DPL (it's throttled to only allow up to 100 results on a category comparison). --SB_Johnny | ^talk 13:36, 28 January 2008 (UTC)

Right, so a trip down to the bot-ique is in order. (Sorry :-)) But seriously, does such a thing exist? Wouldn't it be great to be able to find out - in one place - what bots exist, how to use them, and how to request a new one for a specific need? Cormaggio ^talk 18:39, 28 January 2008 (UTC)

Best place to start would most likely be #wikimedia-toolserver. My keyboard time has been pretty busy lately though (not to mention lacking in abundance). If you have time, please ping them there... what we need is a tool that can

1. count the number of pages in a category, and then log it on the category page.
2. compare the number of pages in linked categories (for example, of all pages in category x, n pages are also in category y)

I think that should do it for now :). --SB_Johnny | ^talk 12:06, 29 January 2008 (UTC)

I'm having problems with IRC of late - might need to try other clients/options - but if anyone's reading this and feels like bringing it up, please feel free to do so. Meantime, I've added a note to the toolserver wiki. Cormaggio ^talk 12:32, 29 January 2008 (UTC)

Discussion seems to have ceased here. I can't actually make a finished survey myself, so what can I do to help move this along? Should I try to make a revised list of questions and then post it here for discussion? --Luai lashire 00:28, 6 February 2008 (UTC)

I (or someone) needs to make better documentation, but {{Wmdgs-simplesurveyquestion-2}} is actually very easy to use, and quite suitable for these questions since they're all yes/no. I'll make a start of it on {{wmdgs-disabilities}} and maybe you can follow along (just copy the template there over and over for each question, then remove the < pre > and < / pre > tags when you're done). --SB_Johnny | ^talk 10:02, 6 February 2008 (UTC)

On second thought, don't remove the preview tags. Just fill in the questions and what the responses should look like, and then save it in the preview mode (if you don't, it's much harder to fix mistakes later). I'll add the table formatting and instructions when you're done. --SB_Johnny | ^talk 10:08, 6 February 2008 (UTC)

I tried adding some more of the questions. I realized that, while this will work for now, in order to do some more complicated questions (such as allowing people to specify what their disability is), we are going to need to alter it more. I will finish the questions first, though, and then we can figure out how to do the complicated stuff. --Luai lashire 22:54, 6 February 2008 (UTC)

Yeah, there would need to be a large number of options for that. The first 2008 US prez elections template allowed using candidate names rather than numbers from a list, so could probably be done that way. Assuming that you find that substituted template easier to work with than raw-scripting, would it be helpful to create a "massively multi-answer questionmaker"? Also, I'm planning to do a religion survey at some point which will have an associated "add religion" template similar to {{Wmdgs-geography-addcountry}}. Would it be better to take that approach, or just have all disabilities accounted for from the start? --SB_Johnny | ^talk 12:07, 7 February 2008 (UTC)

Well, there's no way we can actually come up with every possible disability that exists.... I think it would probably be best to go with a system like the one used for the election questions. That way, if no one has previously responded with that disability, it will just make a redlink to a new category, right? One problem though, is that this system will not link together related conditions and responses could be variable, so we may end up with, for example, people who answer "on the autism spectrum" and people who answer "ASD" and people who answer "Asperger's", and none of these will be connected. I'm not sure what we could do to prevent that from happening. --Luai lashire 00:02, 8 February 2008 (UTC)

Yup, unfortunately templates are pretty much if-then statements, so they can only do what's specifically scripted. The "show instructions" trick is the only idea I've come up with, but I've been lurking on irc to find someone to do a better job (someone mentioned that he might be able to do a javascript fix, but not sure). We could have an answer that just displays whatever someone enters, but it would be useless for catgegories. --SB_Johnny | ^talk 10:27, 9 February 2008 (UTC)

Survey of Wikipedians

Wikimedia Foundation and UNU-MERIT announce First Survey of Wikipedians

Heh, scooped again :). --SB_Johnny | ^talk 10:27, 25 January 2008 (UTC)

No matter. :-) I think there's a place for both of these (and quite a few other surveys)... Cormaggio ^talk 10:39, 25 January 2008 (UTC)

Concern

I am concerned that this survey does not appear to enable anonymizing one's responses. This could deter editors from responding honestly to some questions. 81.98.153.136 18:50, 28 January 2008 (UTC)

Agreed. I would like to answer all the questions, but am a little uncomfortable with my level of sexual activity appearing on a public page with my name next to it. I suspect that the more we delve into such sensitive subjects, the more people will be dissuaded from answering them, due to the public nature of the answers. --Luai lashire 23:33, 28 January 2008 (UTC)

I suspect we'll find over time that some questions just won't get as many responses as others, but not every question needs to be replied to. Wiki software doesn't really permit a way to make it anonymous, unfortunately, except if you create a single purpose account for surveys (which isn't ideal either, but perfectly acceptable if someone really wants to answer but also really wants to be anonymous). Would adding a field for "prefer not to answer" be a good idea? --SB_Johnny | ^talk 11:19, 29 January 2008 (UTC)

Yes, you always need to give an option not to answer, particularly with questions such as sexual activity (I didn't know there were any). On anonymity, it seems like an inherent feature of doing a survey on a wiki (except for a solution like what User:Surveytaker has done) - large-scale anonymised surveys would have to be carried out in other ways. I've been offered to use SurveyShare, which has been used in a study on Wikibookians, if it's of any use... Cormaggio ^talk 12:43, 29 January 2008 (UTC)

Personally, I started skipping questions when we got to politics. I am actually rather opionated on the topic, but prefer to keep that off-wiki. For me, seeing a userbox on one of my user pages about condom use falls into the Category: Way too much information. More generally I think we do need to consider how survey takers opting out influences the results. Perhaps this learning project could include a lesson on w:Questionnaire construction and w:Selection bias? Or maybe a page where we critique and try to improve a set of questions before asking for participation? In general I think that exploring a wiki-based questionnaire is a wonderfull idea and I am impressed with how the project is going and the potential. But if we don't give some thought to methodology now we might find ourselves analyzing inconsistent data. (half the people take an old version of the survey with one set of questions, and half take a version with additional choices - can the results be directly compared?) Having a "prefer not say" choice will help in that we can at least count the nubmer of people who did not want to answer. --mikeu ^talk 23:33, 29 January 2008 (UTC)

Good points. Do you think it best if we start a general learning topic Research methodology (with links off to different methodologies based on different epistemological stances), or have a methodology subpage for every existing research project? Cormaggio ^talk 11:42, 30 January 2008 (UTC)

I guess the opt-outs should be for the whole survey then? If someone doesn't want to answer one question in particular, they can just skip it, but if they are uncomfortable with an entire survey the easiest thing might just be to have an option for "No thanks". I myself might be prone to take that option on Luai's survey, for example.

As far as sampling troubles, we could also recategorize respondent categories for particular questions within a survey. For example, anyone who answers any of the questions on {{Wmdgs-sexuality and relationships-survey1}} is automatically added to Category:WMGDS/Sexuality and relationships-1/Participant. If we were interested only in sexual orientation, an additional respondent category could be added when someone answers that question (e.g. Category:WMGDS/Sexuality and relationships-1/Orientation.

Another idea we should think about is having a survey on the survey. E.g.: should personal questions be asked? should political questions be asked? etc. We can also organize some way of voting or consensing on what surveys get included in preload/all or in major updates. --SB_Johnny | ^talk 12:59, 30 January 2008 (UTC)

If we leave it that people can opt out of individual questions within surveys by not answering them, would there be a way to see how many times a particular question is not being answered? It might tell us that a particular question is either too sensitive or phrased/constructed badly. I'll have to think about your categorisation proposal - perhaps I'd have to see it in practice to understand how it would be useful. And as for a survey on surveys, why not leave it to normal talk page processes and raise points for discussion? By forcing people to use a prepared survey in order to critique surveys, we'd remove the ability for someone to raise concerns that haven't been addressed by the survey-survey. :-) Cormaggio ^talk 13:19, 30 January 2008 (UTC)

Help with templates please

Please can someone add the following countries to the question about travel/residence in Europe: Sweden, Denmark, Norway, Czech Republic, Slovakia, Hungary, Romania, Bulgaria, Greece, Turkey,

And the following to the question relating to Asia: United Arab Emirates, Pakistan, India.

Thank you. Surveytaker 00:15, 29 January 2008 (UTC)

Sweden and Norway are actually on the template already, but the answer options were not added to preload/all. We should probably set up a DPL tracker for that,

Turkey is already in the Asia survey, actually, though I suppose there may be disagreements about which continent it's on (technically both, but mostly Asia I think) :).

For adding countries, you can use {{wmdgs-geography-addcountry}}, like this:

{{subst:Wmdgs-geography-addcountry|
| Country name = Bulgaria
| Flag image (do not include the "Image:" prefix) = Flag of Bulgaria.svg
}}

The addcountry template is very easy to use, but it does take a while to hunt down the current flag on commmons (the filenames are usually but not always in the form of "Flag of Country.svg"). I won't have much wiki time today, but if someone would care to hunt down the image files, I loaded template lists on Template talk:Wmdgs-nationality, residence, and travel (Asia) and Template talk:Wmdgs-nationality, residence, and travel (Europe). --SB_Johnny | ^talk 11:51, 29 January 2008 (UTC)

• Turkey should be in both Europe and Asia - I have visited Turkey-in-Europe, but not Turkey-in-Asia. Surveytaker 21:26, 29 January 2008 (UTC)
  • Fair enough. --SB_Johnny | ^talk 12:43, 30 January 2008 (UTC)

Editing templates?

If I were to create some images to go with the results templates for some of the answers (for example, an image for the response "_USERNAME_ is bisexual"), how would they be added to the answer box? I know I can rely on other users to do that part if I provide the image, but I'd like to learn to do it myself. --Luai lashire 23:43, 11 February 2008 (UTC)

Actually, I just went and had a look at Wikimedia Commons and found several images I think we could use for the sexuality/gender questions. Shall I link to them? --Luai lashire 00:04, 12 February 2008 (UTC)

General survey problem

The #3 choice comes out as the #0 choice: "Projects you regularly contribute to, though not as much as your primary" turns into "Projects you don't contribute to, but do read and use" in the template. Chantessy 17:56, 18 July 2008 (UTC)

Which question? --SB_Johnny | ^talk 10:19, 20 July 2008 (UTC)