Participation biases in OSM: Survey now LIVE!

Posted by Geospa_gal on 3 August 2017 in English (English)

A couple of weeks ago I posted a diary entry introducing myself, my interests in OSM and showcasing an upcoming survey for OSM editors. I am very pleased to announce that the survey is now live and I would like to invite as many OSM editors as possible, to participate. The survey is anonymous and only takes a couple of minutes to complete.

The project is motivated by recent research that has found a strong male bias in OSM participation. This has led to assertions that various geospatial knowledge could be under represented or poorly recorded on the map. However, the actual consequences of this bias remain little explored or reported. By collecting information about contributors to OSM, which can then be analyzed along with their editing patterns, the impacts of this bias might begin to be measured and therefore better understood.

If you are interested in or would like to participate in the study, please click on the link below, which will take you to the Bristol Online Survey website where you will find more information and an opportunity to participate in the survey. As a small incentive, at the close of the survey in a few weeks’ time, 60 respondents will be drawn at random to receive a £15 Amazon voucher.

To participate in the survey, click on the link below:

Please do think about participating. It is hoped that knowledge about the way participation biases impact on crowdsourced maps will enable new strategies to be developed to address any resulting voids in the geospatial information provided by amateur mappers. In turn this could strengthen the role played by platforms such as OSM in urban planning and sustainability and raise the profile of the important mapping work that you all do.

In the meantime, if you would like to know more about me, my research activities or the project, please visit my University webpage (link below) and do not hesitate to get in touch directly or via the OSM messaging service.

Thank You

Comment from SimonPoole on 3 August 2017 at 16:03

Could you link to the " recent research that has found a strong male bias in OSM participation" It is not directly a secret that at least at the visible level there is little to no non-male contributors, but your wording seems to be a bit ambiguous and might be referring to something else.

Comment from SimonPoole on 3 August 2017 at 16:15

The other comment would be, after completing the survey, that it seems to be rather biased towards assuming an UK/US education system (for example by ignoring the dual-training system popular in German-speaking countries) even though DACH and the rest of Europe form the largest part of our contributor base (not to mention the language bias by only offering the survey in English).

Comment from imagico on 3 August 2017 at 16:23

Your survey description says:

However, once you have submitted your responses, your username will be replaced by your OSM personal ID number (openly available in OSM) to protect any inferred identity from your username.

You should be aware that resolving OSM user IDs to user names is trivial, you can for example do this using Pascal's tools:

If participation is not anonymous anyway i wonder why you do not require an OSM login via OAuth to make sure every participant participates with her/his user account and not that of someone else.

Comment from Geospa_gal on 7 August 2017 at 16:10

Hi SimonPoole and thanks for your feedback. With regard to the survey bias towards certain educational systems, as adding options relevant to each country could make the survey incredibly lengthy I have attempted to include alternative systems by including the wording 'or equivalent'. Knowledge of respondents' exact educational qualifications is uneccessary so in the absence of this option for non-UK/US educated respondents, an approximation is perfectly acceptable. The same rationale has been applied to offering the survey only in English (at this stage) and of course it will be made clear in any published results to which countries those results are relevant. Surveys such as these have their limitations and it is important to be aware of these.

With regard to the research demonstrating a strong male bias; below is a link to a presentation by Muki Haklay and Nama Budhathoki which presented a 96% male participation bias in OSM in 2010. This bias has been supported by subsequent work on the gender bias in VGI in the last 7-8 years by for example Monica Stephens, but also work by Manuella Schmidt, Sylvia Klettner and Renate Steinmann. I hope this is helpful.

Comment from Rovastar on 7 August 2017 at 23:18

This was a smaller than I expected survey. Basically for those who have not clicked it just asks you your username, gender, age range, nationality, etc.

I am confused about what you are going to do with this basic info. Is it just to get a new updated ratio of male to female? or are you going to looking at my and others editing patterns and determine if we have biases/any bias exists?

Comment from rorym 🏳️‍🌈 on 14 August 2017 at 08:45

Minor one: for gender you have "Prefer not to say" twice, weird that.

I'm not sure what you could do about education system. Apparently there's a European Qualifications Framework which might help you rank different qualifications. Maybe link to it? (But people would probably have to look it up)

Comment from escada on 23 August 2017 at 11:34

Was there any research so far on "This has led to assertions that various geospatial knowledge could be under represented or poorly recorded on the map' ? Do we know what female mappers would add or what female users of a map would like to see on a map ?

Are any of those features mapped less by male mappers ? Are the specific regions where this make (or does not make) a difference ?

I honestly have no idea. So far I managed to interview 3 female mappers for my Mapper of the Month series and I see no difference in what they map and what I map.

I once read that there are a lot of pubs in OSM and that this is because there are a lot of male mappers. In Belgium we use "pub" as well for places where families go on an afternoon because there is a playground and one can drink coffee/tea as well as beer and eat pancakes/ice cream. In the evening those places are used for dinners. Not really family-unfriendly places. But we also use the same tag for what we call "Café" in Dutch and French, which are close to the stereotype pub of drinking and smoking.

So counting male vs. female contributors is one thing, and rather easy to accomplish. But is the mapping behaviour different ?

Maybe we have only people that love to bike and hike (where the needs of female/male contributors are more or less the I think).

Lot's of questions, unfortunately I have no answers

Comment from SeleneYang on 25 August 2017 at 12:22

I believe this survey has much more to do with the sex of contributors rather than gender, as gender is understood as a social construction. I just had a talk about gender gap in our community in the State of the Map Japan, I'm part of #Geochicas, a latin american women group, so just to answer Escada, we're starting to know what women do need to be mapped in order to have a accurate and complete map. We're not talking about how pubs are unrelated to women, because we also go drinking and not everything for us is related to family friendly spaces, we do not considere it a male oriented tag, but there's a difference for example on the tags vending=condom (used 1393 times) and vending=femine_hygiene (used 0 times).

If you want to see my presentation about gender in SOTM Japan, here are the slides:

Comment from SimonPoole on 25 August 2017 at 13:53

Gender is mainly a grammatical concept that got started used as a social concept only recently, much to the confusion of people that actually speak languages that are grammatically gendered.

So lets stick with the sex of contributors to try and avoid that confusion. I think the "pub" example is particularly bad anecdotal evidence, not only because we don't actually have an extraordinary number of pubs mapped compared say to toilets and other objects, associating them with just one sex is invoking a stereotype as @SelineYang points out.

The good thing about OSM is that the data is open, and so it is easy to test claims against the real numbers, and I find it very difficult to see any pattern at all with respect to objects mapped, if at all and you want to invoke sexual stereotypes things that are typically associated as being male tend to have very low counts, and such that are higher counts associated with female stereotypes.

In the case of vending=condom vs. vending=femine_hygiene it is comparing a near zero number to zero and I'm not sure that there is any takeaway from that, outside of that universally people don't seem to map stuff in their toilets.

Comment from SeleneYang on 25 August 2017 at 14:02

Only recently you mean the past 50 years? About the example of vending=condom vs vending=feminie_hygine it cannot be reduce to say that people do not seem to map stuff in their toilets, it is a clear example that the number and gender of OSM contributors do matter, and that things that are priority on women's life are not well represented in the map.

Comment from imagico on 25 August 2017 at 14:30

Only recently you mean the past 50 years? About the example of vending=condom vs vending=feminie_hygine it cannot be reduce to say that people do not seem to map stuff in their toilets, it is a clear example that the number and gender of OSM contributors do matter, and that things that are priority on women's life are not well represented in the map.

Actually no, these numbers do not give you a clear indication of a bias within OSM in that regard. Unless you can proof that vending machines for feminie hygine articles are universally as frequent as vending machines for condoms the observed bias could be fully explained by bias outside of OSM. This does not mean there is no such bias in OSM but these numbers do not proof such a bias. It is a well known phenomenon that in OSM (or in any geodata collection for that matter) there is preferential treatment of common features over less common features. That is a problem of bias and discrimination in itself but it is not the same as gender/sex discrimination.

What would indeed be a very valuable thing to study is if there are systematically different preferences in what things people consider worthy or important to map between men and women. But this is extremely hard to properly study in a neutral, unbiased way - because if you take men and women who already map in OSM and study their mapping preferences this is obviously not representative for men and women in general.

By the way there is a fairly obvious geographic bias in mapping of vending=condoms:

Comment from SimonPoole on 25 August 2017 at 15:37

Yes 50 years is only recently.for a social science concept, or put differently it hasn't really proved its value yet (lets discuss that in a 100 years or so).

To be clear this is not about if there should be more female participation in OSM or not; yes there should be, naturally. Different point of views are naturally welcome, OSM is from a number perspective dominated by white European males, but that doesn't mean we don't want to accommodate or support adding stuff that is not part of our culture, daily life etc but in general you need to take the initiative yourself instead of waiting for an OK or something like that (see below).

Back to the example:humans of the male sex are unlikely to be able to map vending=femine_hygine from a practical pov,

But even so the problem is the numbers are so low, the 1'400 odd condom vending machines were added just by 500 contributors, and a single person could probably easily skew the numbers one way or the other within a week. That is assuming that the vending machines in question actually exist in larger numbers, which again is difficult to determine for the already mentioned reason.

Are there other similar cases of probably sex-based physical access restriction biases to be found?

The other side of the issue is that there isn't actually a preset in any of the major editors for the specific value. Which makes it rather difficult to map such facilities to start with (spelling and so on). I've opened an issue for iD and vespucci to fix that, but see my point above: if you want to make something easy to mapped regardless of any sex related issues, you need to add presets for it.

Comment from escada on 27 August 2017 at 06:15

SeleneYang I will certainly look at your video presentation.

Thanks for suggesting a tag that is not mapped by male mappers (but apparently also not by female mappers), but I wonder whether you will find any vending machine for female hygiene on the outside of buildings, unlike vending machines for condoms. At least in the area where I live. According to taginfo, there are 2 vending tags that contain feminine_hygiene (together with some other values). Maybe there are other words used for the same item ?

But I see your point, areas that are forbidden for men (such as toilets) cannot be mapped in detail without women. Are there other problems, or missing or better underrepresented tags ? I find this question more interesting than how many female mappers we have (not enough that is for sure).

Comment from rorym 🏳️‍🌈 on 28 August 2017 at 11:55

Just a note (because no one else has mentioned it), many if us want to make OSM more diverse, gender is one big problem, but let's not forget trans/nb people. Some women have penises and might need condoms. Some men menstrate and need menstruation products. And so on. (I wonder, does that mean we shouldn't tag them as feminine_hygiene 🤔🤔)

Comment from Geospa_gal on 30 August 2017 at 11:23

Thank you all for your comments. The thread makes fascinating reading for me and I'm glad to have sparked some debate on this issue.

I agree with escada: the more interesting question (and which I'm hoping to address through my research and this survey) is the impact of the gender bias rather than the bias itself. By analysing mapping behaviour by gender, I hope to be able to identify differences in what or how men and women map. The example of supposedly gender specific vending machines highlighted by SeleneYang being a really good example of an impact of the gender bias (although, as imagico says, it is not direct proof of a gender bias in OSM, it could be an indicator). By identifying these impacts the assumed 'problem' of the gender bias can be qualified and hopefully addressed so that these voids can be filled and the efficacy of VGI elevated. This is the long view at least.

Comment from philippec on 4 September 2017 at 18:04

But there are many women making a living studying and speaking about the absence of women in OSM.

Comment from joost schouppe on 6 September 2017 at 06:33

There might be a smaller association between gender/sex and mapping habbits in OSM than expected, since women mapping in OSM are a small subset of women. The same of course goes for men. For example, there could be as much interest in mapping power lines among both male and female mappers, even if (say) in the general population there is much more general interest in power lines among men than women. The same goes with differences explained in the general population by educational attainment, age, income, whatever. Finding any effect at all would thus be quite significant.

The bias might not be in the mappers, but in missing the things that are of interest to the non-mapper. I would say a necessary step in a bias analysis is to define real world things as either male or female (good luck), find good spatial data about them (good luck) and compare them to what is mapped in OSM. But that process sounds a bit paternalistic (unless you can find things that are extremely gendered, like the feminine hygiene example) and possibly more likely to replicate the personality of the researcher than an actual bias. So again, good luck :)

Another approach could be to find some resource which measures the presence of people by gender (by old school counting people, or with some novel approach like sexing geolocalized twitter or flickr streams), and then "just" compare the completeness of the map between places with a more female or a more male presence.

Login to leave a comment