Strategy/Wikimedia movement/2018-20/Working Groups/Community Health/Survey results
Summary of CHWG survey responses
Wikipedia is an enormous success, launched in 2001 to create a global repository of knowledge accessible by all. There are nearly 6 million articles in English and an average of 571 new articles in English each day. Wikipedia is now available in 285 languages and there are over 48 million pages globally; over 20,000 new articles are created monthly. Wikipedia continues to grow, with greater reach.
As Wikipedia has grown, so has its community. However, this growth and success have been accompanied by problems in the community. Problems include targeted incivility toward others on the site, and an unwelcoming environment for newcomers to the community and others. These issues are so widely known that they have been the focus of academic papers about Wikipedia. In light of these widely recognized issues, Wikimedia has determined to take action to enhance the overall health of its community. In this context, community health refers not to physical health or the health of an individual, but rather to the social cohesion of the community of Wikipedians, and the feeling of belonging and co-ownership of Wikipedia. Wikimedia’s Movement Strategy acknowledges that improvement of Wikipedia’s community health is necessary, and that there are issues within the Wikimedia community that adversely affect community health.
In support of the 2030 Wikimedia Movement Strategy Direction, the Community Health Working Group (CHWG), the CHWG interviewed people during Wikimedia Summit 2019, and by email, to ask what barriers to participation are and to solicit suggestions about how to resolve these barriers and other issues. These interviews were used to draft a preliminary set of Recommendations for the Foundation’s Board of Trustees regarding community health within the Wikimedia movement. This was done in compliance with the Focus and Rationale described on Meta-Wiki. The draft recommendations are titled:
- A joint set of rules we all agree to live by (a.k.a. Code of Conduct)
- Redefining power structures to better serve the communities
- Building the leadership of the future
- Structure for handling conflicts- before, during and after
- Investing in building an inclusive global community
- Newcomers are a key to the success of the movement
- “Democratizing” participation (making Wikipedia/Wikimedia everyone’s responsibility) ->Advocacy WG (Hexagon) AND Reducing barriers for participation
- Privacy and security for everyone - All-terrain readiness
- Opening the Circle (diversity, with resources and support)
- Network structure to continually support community health
- Aligning resource allocation with community health goals
- Investing in equity-centered technologies
The Community Health Working Group, part of the Strategy Movement, launched a survey about community health to gage the level of community support for the draft recommendations. Data was collected until August 18 using the Wikimedia survey in English (N=80), French (N=12), Indonesian (N=5), and Spanish (N=9). The survey was translated into Portuguese but no responses were received in this language. Of the 106 responses, 8 were entirely empty. Excluding empty forms leaves 98 responses, most of which were nearly complete. Many answers were very thoughtful and went into paragraphs of text. The extensive responses indicate that the people who responded to the survey are very invested in Wikipedia. Below is a summary of the survey responses received and how they relate to the CWHG preliminary recommendations, drafted after consultation with community members through interviews (see the synthesis of interview responses).
Survey responses reflect the need to address issues of community health
Here is a sample of quotations from survey responses that reflect the need to address long-standing issues previously identified in the interviews:
- Treating topics related to anything outside the post-internet world and the white male 30-something ‘brotopia’ with hostility and as irrelevant, not-notable, and to be removed or suppressed. (timestamp: 8/7/2019 21:34:00)
- Among established editors, probably those most likely to have a bad experience on the project are members of underrepresented groups who are highly visible within the community or who focus on highly controversial content areas. People in these categories are particularly likely to receive harassment and abuse, and this can be difficult to control because it often involves prolific sockpuppeting and/or off-site harassment by email, on other websites, on social media, etc. Much like the dynamic in social media, prominent women are often subject to specifically sexual harassment. This kind of behavior is usually easy to ignore if it happens once, but when it happens in high volume or in response to an upsetting incident, it can be very distressing for the victim. I feel that anti-harassment efforts to date have tended to focus in the wrong areas, targeting friction between established editors rather than this kind of behavior by long-term abusers or off-site groups, which are exactly the areas the community is least effective at addressing. (timestamp 30-7-2019 7:55:35)
- Don't be afraid to ban hundreds or thousands of people who persistently show bad behavior. The community will recover and flourish. (timestamp: 17-6-2019 0:21:58)
- decline will continue based on organizational culture, until an investment is made to shift the culture (timestamp: 17-7-2019 13:12:21)
- [Identify the key influencers] on "culture" and make a dedicated effort to make them very civil (timestamp: 17-7-2019 21:31:09) [spelling corrected for ease of comprehension]
- Make incivility warning more formal and clear, with at least some token consequences for repeatedly failing, even when individual mistakes are not actionable. (timestamp 19-7-2019 1:14:25)
- I know some women get more negative attention when they reveal they are women. That shouldn't happen. (timestamp: 30-7-2019 20:37:24)
What are things within your community or project that create barriers to involvement for you?
Summary: The most frequent barriers to respondent's participation named were along the themes of hostility (N=27), not inclusive (16), abuse (6) and abuse of admin power (3).
While respondents did not name the recommendations, 40 responses to this question supported draft recommendations such as creating a CoC, opening the circle, building an inclusive global community, training and building leadership, ongoing training of admins, democratization of participation, changes to policy regarding conflict prevention and resolution, and creating a structure to monitor and evaluate CH.
Two responses conflicted with recommendations, one about privacy, and the other said that "mentoring people is demanding" which could be interpreted to mean that more people should help mentor newcomers or that this individual is against mentoring newcomers. (I feel like I spend way too much time trying to mentor editors who are entrenched in their behaviors and aren't willing to change in order to follow policies and procedures of the WP community. timestamp 10/7/19 18:37) This respondent overall describes trying to do a great deal to mentor editors, but that editors are sometimes resistant to adapt or change, and her answers across the 12 substantive questions give the impression that she is more burnt out than against mentoring.
No answer: 25/17 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms).
What are the things within your community or project that create barriers to involvement for some people due to their personal identity?
Summary: The most frequent things within respondent's community or project that create barriers to involvement for some people due to their personal identity were sexism (N=18), hostility (15), identity (14), race (10), homo- and/or transphobia (4), abuse (3), religion (3), and reporting (2). Reporting was linked to backlash and abuse. Religion was mentioned in tandem with the anti-semitism and the Israelli-Palestinian conflict and bias against conservative viewpoints.
While respondents did not name the recommendations, 53 responses to this question supported draft recommendations such as building an inclusive global community, increasing diversity, opening the circle, democratizing Wikimedia, and low tech reporting to reduce barriers, privacy for all and other technical recommendations on Privacy and Security, training for editors and admins, a CoC, creating a structure to monitor CH, mentoring and welcoming newcomers, and conflict prevention and resolution.
No responses seemed counter to draft recommendations.
Other interesting quotes: bias against offline sources and its effects on the global south [16-7-2019 21:19:26]; strongly held right-wing political beliefs with an example from wikiTalk about not allowing other races on police forces [16-7-2019 23:45:43]; and As WP is so based around 'consensus' (however vague that concept is), a single individual is relatively powerless, which also limits the disruptive abilities of a single person. It's when they assemble into groups that the trouble really starts. [19-7-2019 13:25:24]
The geographic demographics of people who indicated race are impossible to generalize by geography; they came from Australia, Belgium, Germany, Japan, and Europe and the US.
No answer: 16/8 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Responses to questions 1 and 2 reflect some of the interview responses. These responses were similar to those identified in the interviews. When asked “Are there practices, policies and regulations within your community or project that create barriers to involvement?” interviewees identified problems related to the projects, structures, community issues, new editors, and more. For example, within Wikipedia / the projects themselves
- the complexity of editing was mentioned as a barrier to involvement. This relates to wikisyntax, the layout of articles, the many rules and conventions etc. It is difficult to get started, and this is related to the treatment of newcomers.
- especially criteria concerning notability make it difficult for underrepresented groups to contribute in a way that is meaningful for them, as relevant quality sources can be scarce, and this has implications for diversity, particularly in places where there are fewer published resources and people may not be notable to outsiders (perhaps the majority of the Wikipedia community).
Regarding structure and processes
- the absence of clear processes and rules can also be a barrier to involvement. It makes it very difficult - especially for newcomers, but not just for them - ‘to do the right thing’.
- the rules that do exist are sometimes ‘selectively implemented’, echoing the survey answers about abuse of admin power and long-term abusers.
- harassment was frequently mentioned, often related to the treatment of women but basically of anyone who upsets the community status quo
- frequently, there is a failure of the community to act against harassment
- a bias against women-editors/edits by women is not unusual
- many communities have a well established ‘old guard’, who are resistant to change and are able to impose their way of doing things. Survey respondents linked this to mob behavior.
- lack of patience with, tolerance for, and appreciation of newcomers is a common phenomenon and discourages them to continue
- coaching/mentoring programmes aimed at retention of new editors are lacking
- a lack of resources (money, time) prevents communities dealing with issues they recognise, and while Wikimedia cannot influence people’s time constraints, WMF could direct some resources to specific communities.
- emerging communities should be allowed to find their own way of doing things, and not be forced to copy the models of Europe and the USA,
- there is sometimes a discrepancy between local culture and the basic concept of Wikipedia, and these last two points have implications for building inclusivity, opening the circle and increasing diversity.
What do you think would improve participation in your community?
While respondents did not name the recommendations, 47 responses to this question supported draft recommendations such as developing a CoC, ongoing training for admins and editors, democratization, inclusivity, privacy for all, removing and reducing barriers especially to reporting, welcoming and mentoring newcomers, redefining power structures, conflict prevention and resolution, and a structure to monitor CH. One person suggested a very detailed code of conduct which not only specifies what behaviors are and are not acceptable, but also how one obtains relief from victimization, protection from false accusation, and outlines the steps involved in filing a complaint, reviewing evidence and potential outcomes. (timestamp: 19-6-2019 19:23:22)
Two responses seemed to advocate against growing the number of Wikipedians/Wikimedians.
Other ideas to consider included term limits for admins, streamlining policies and guidelines, creating tools to maintain a list of entries/pages needing improvement, identifying uncovered areas of knowledge so editors and newcomers could see what expertise and information would be helpful; better recognition for contributors, which is related to another suggestion about information about how users use the content such as counts on visits to pages for those who ask If I make an edit, does anyone even see it? (timestamp: 9/7/19 0:38) One person suggested professional editors to resolve content disputes. A better interface was recommended (see sheet 'mobiles.')
No answer: 13/5 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Unanticipated finding: Mobile interface improvement
Summary: Mobile devices are deemed a way to access WP but 6 respondents indicated that the mobile interface is not up to editing. At least one highlighted that this will have implications for recruiting new editors as younger people and many people in countries with limited access to computers read WP on their phones. An unfriendly or difficult or even simply ugly mobile interface could be an obstacle to building a diverse and inclusive global community. These 6 responses came from the US (2), the UK (2) and Italy, including 2 women, 2 men and 2 people who declined to indicate their gender.
What prevents contributors from taking action against bad behaviour?
Summary: A Code of Conduct with very clear definition of what is acceptable and what is not is the overwhelming suggestion. In response to what prevents users from taking action against bad behavior, 29 respondents referred to non-enforcement of current standards, 19 referred to reprisals (including mob reprisals), 9 referred to people perceived to be able to get away with harassing others (some called them 'unblockables'). Doxxing and long-term harassment were each named once, but may be factors referred to by people who listed reprisals and mob behavior among deterrents to action.
While respondents did not name the recommendations, 34 respondents supported draft recommendations, included anonymous reporting tools (lower technical barriers), welcoming and mentoring newcomers, ongoing training for admins, redefining power structures.
One response ran counter to draft recommendations, with the idea that incivility is part of everyday life off-wiki. Some responses questioned the premise of the question, and referred to defining what is unacceptable behavior, and that making someone uncomfortable is sometimes what happens when people receive valid criticism of their edits. This could have implications for mentoring, but rather seemed to have implications for burnout.
No answer: 11/3 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Interviewees were asked a very similar question, “What prevents good faith contributors and functionaries from taking effective action against bad behaviour?” and their replies also emphasized the need for lower barriers to reporting such as private or anonymous reporting, mob behavior when people gather as a group to act against a particular contributor or newcomer, that bad behavior by reliable contributors is tolerated, and that people who address bad behavior face hostility and burnout, and those who mentor newcomers also develop burnout. The next question for interviewees was what are the main barriers that prevent taking effective action against bad behavior, and answers akin to those from the survey included:
- Bad behavior and certain biases are accepted and normalized
- Gaps between different communities or abuses of power
- Complexity to report or identify bad behavior
- Lack of training
- Complex procedures to set and implement policies
- Lack of responsibility
What would you change in your community to address these problems?
Summary: 48 responses to "What would you change in your community to address these problems?" echoed specific recommendations of the CHWG, without referring to the draft recommendations. These included a clear CoC that described not only what is acceptable and unacceptable but also procedures for reporting and response, as well as opportunities to defend against allegations. Ongoing training for admins, and removing barriers by supporting private or anonymous reporting were popular. Responses supported welcoming and mentoring newcomers. Creating a structure to monitor CH was supported, as was redefining power structures (especially regarding taking input at local levels and changing 'unblockable' status.)
Multiple responses suggested changing the definition of bad behavior to address actions rather than feelings, because feelings are too subjective. Multiple responses advised addressing bad behavior and doing so promptly, and doing so using the same rules for everyone regardless of their edit count or status in the community. As with previous questions,'unblockable' people's bad behavior, especially among admins, was deeply resented and discouraging.
5 responses were counter to recommendations. 3 of these 5 responses specifically calling for more transparency and against anonymous or private reporting; these three responses typically supported other recommendations, like, for example, instating a CoC. 2 responses seemed to be against the Strategy Movement and anti-WMF.
Other things to consider: ‘Professional editors’ (undefined) to resolve conflicts over content was recommended by one person, based on quality of content. Ways to flag divergent issues were suggested, with a drop-down menu to flag issues with either content or behavior, and for these to have different processes perhaps with different administrators. One person suggested not building a hierarchy based on the number of edits, which reflects redefining power structures. At least two people referred to positive reinforcement.
No answer: 19/11 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Who is responsible for taking action against bad behaviour now?
Summary: On this question, there is a lot of acknowledgment that everyone and admins should handle bad behavior, but responses reflect a lot of frustration that this is no one and everyone's job, and so addressing bad behavior sometimes is not done. Responses reflect a lot of understanding of the process as intended. One person suggested encouraging newcomers to delete insults, which could affect levels of empowerment or rely on broader empowerment of volunteers. French responses overwhelmingly focused on admins, but also recognition of inaction and that admins (and others) should act even before reporting.
The most common answers were everyone and the admins, but it was acknowledged that no one is really held accountable. The recommendation “Redefining power structures” could address this, perhaps with a new role.
Answers to this question were generally short, just a few words (e.g., everyone, admins) but 10 longer answers corresponded to recommendations for ongoing training for admins, building an inclusive community, diversity, welcoming and mentoring newcomers, and redefining power structures. The most commonly reflected recommendation was for ongoing training, reflecting a perception that admins are supposed to do this work but may be unresponsive.
No answer: 16/8 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Who should be responsible for taking action against bad behaviour?
Summary: As with the previous question, when asked "Who should be responsible for taking action against bad behaviour?" responses were short, but 37 responses said this should be everyone's responsibility (Note: If "checkuser" means editors, then one more for everyone), while 34 said that this should be an admin and/or arbcom responsibility (these are not mutually exclusive.. Some responses said that users should report to admins. Bystander training could be useful.
Some responses specify new staff in this role like a helpdesk, others say elected admins (not admins for life) and a quorum of admins (which is more work for admins), others want bystander interventions (which is everyone). Some specify not WMF, and some refer to T&S. This echoes the interview data about dis/satisfaction with AN/I.
17 longer responses reflected specific CHWG draft recommendations, most importantly the recommendation to redefine power structures, ongoing training for admins, mentoring new editors, and adopting a CoC. 6 responses included WMF, 1 included T&S, and 2 said 'not WMF.' 4 responses reflected a desire to maintain the status quo.
No answer: 14/6 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
What do you think would improve civility in your community?
Summary: 50 responses - more than half of filled-in questionnaires! - to "What do you think would improve civility in your community?" reflected the CHWG draft recommendations. These focused on a clear CoC, with admins included, as reflected in draft recommendation for non-automatic admin renewal and ongoing training for admins. Welcoming and mentoring newcomers was also supported.
30 responses suggested enforcement; these included references to admins and unblockables and penalties. 10 responses suggested clear guidelines about what is acceptable and what is not acceptable; these suggested less emphasis on bad words and politeness going forward.
2 responses were counter to recommendations, with 1 response being against welcoming newcomers and 1 responses being against creating any new structures to address projects or sections that this respondent works on. An additional respondent expressed sentiments against WMF involvement.
No answer: 16/8 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Who is responsible for the individual health of contributors in your community?
Summary: This is another question with primarily short answers, which can be summarized as 30 indicating everyone, and 12 indicating higher roles such as admin, Arbcom, and even WMF and T&S. These answers are not mutually exclusive. The 'thank feature' and bystander interventions were each mentioned once in separate responses.
13 responses echoed CHWG draft recommendations, mostly CoC, democratization and inclusion, welcoming and mentoring newcomers, and ongoing trainings. It was noted that a CoC needs enforcement, and that T&S is perceived as accessible only to WMF insiders.
No responses seemed to run counter to draft recommendations.
Note: More than 10 people thought this question referred to physical health.
Interviewees also said that everyone in the community is responsible for contributors’ wellbeing, However, the responsibility for taking actions against bad behaviour is given to the affiliates, community leaders, admins, and WMF. Interviewees acknowledged that there is a lack of process and policy in reporting bad behaviour as well as the need to be more inclusive and welcoming. Some advised training and capacity building as important aspects in dealing with conflict resolution.
What do you think would improve the individual health of contributors in your community?
Summary: 28 responses reflected CHWG draft recommendations, with a real emphasis on conflict prevention and resolution. Other recommendations reflected included developing and enforcing a CoC, welcoming and mentoring newcomers, ongoing training (part of redefining power structures),
Responses emphasized enforcement, positive feedback and reinforcement, and the need for community buy-in of recommendations. One person described having been themselves been blocked when they were dealing with external stressors, unfortunately without insight about acting out; this is a good reminder that external issues affect people's behavior on-wiki.
One response was counter to draft recommendations, expressing sentiments against welcoming newcomers and for 'faster blocking.' Another response was not counter to recommendations but suggested less interference from WMF.
One response included a well phrased warning about resistance to forthcoming changes to WP: "There are the wikipedians who think that this Strategy is nothing to do with them, who believe that Wikimedia is a reflection of society and the community that makes it up and that the bias in it is inherited from our culture and therefore nothing can be done. ‘We’re not going to change society, are we?’ they say when asked about the strategic process. And so in their daily work, they hinder any progress towards the changes people want to see by 2030."
Note: 12 responded as though this were about physical health.
No answer: 22/14 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
What else do you suggest to improve community health in the Wikimedia community?
Summary: Answers to the question "What else do you suggest to improve community health in the Wikimedia community?" were diverse and echoed differing opinions seen previously; for example, 10 answers could be summarized as 'ban harassers' with still more recommending taking action against bad behavior. 3 replies were 'pro-Fram' and some used community-buy-in as an argument for that. 6 responses referred to community buy-in. 4 recommended more meet-ups or meetings about policy. 3 replies seemed to imply that WMF staff should participate on-wiki before making decisions. 3 recommended more positive feedback, with 1 of these suggesting rewards for long-term editors.
21 responses echoed CHWG draft recommendations, including implementing a CoC, mentoring and welcoming newcomers, conflict prevention and resolution, creating a structure to monitor CH, privacy for all, building a diverse global community, and democratization. One person suggested a system:
- Short, low-stake topic blocks, ibans for WP:CIV, WP:BITE violations.
- Admin accountability: the need to justify admin actions with diffs - 4+ diffs for short 1/3-day sanctions, 7+ diffs for 1-week, 10+ diffs for a month, 20+ diffs for a year
- WP:Administrators/Problems, similar to Probleme
- Non-voluntary admin re-election
- Right to defend oneself: editing of WP:ANI/<UserName> when blocked. Editing right for WP:ANI, WP:AN, WP:ARB, if not abused.
- Right to question admin actions and long-time editors without the too common backfire of "personal attack". (WP:ADMINABUSE applied)
- The oversight and cooperation of WMF with ArbCom to handle (not ignore) cases of harassment and high-profile editors / admins. If ArbCom fails to act in a timely manner (week / month, depending on seriousness), then WMF should act, without letting ArbCom delay the action. (timestamp 18-7-2019 23:49:06)
5 responses run counter to recommendations; 3 say everything is fine, but 1 of these responses is also supportive of recommendations on conflict prevention and resolution. 1 response advocated for requiring user photos, counter to the recommendation of privacy for all. 1 response was against anonymity (privacy for all.) 1 response expressed hostility toward WMF and T&S; this individual's responses have consistently been counter to change except to block WMF involvement.
Note: 2 answered as if this were about physical health.
No answer: 37/29 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Survey questions 10 and 11 asked respondents for their recommendations to improve community health. The email to the Gender and Diversity Mapping asked for recommendations. They recommended
- Creating friendly space for new editors
- Improving gender balance and diversity
- Education and empowerment of how to deal with harassment and sexism
- Behavior policy and reporting mechanism for bad behavior
- Appreciation for good faith contributors.
These recommendations are similar to the survey responses, particularly about welcoming newcomers, implementing and enforcing a CoC, and increasing diversity and building inclusivity.
Summary: 11 responses echoed CHWG recommendations addressing CoC, conflict prevention and resolution, technical changes to reduce barriers (to reporting) and for privacy for all, diversity, and a structure to monitor CH. Urgency was also expressed, that WMF is late to this issue and must act to make changes that will be effective, and that going after bad words instead of actual harassment is unacceptable.
2 responses were counter to recommendations and for the status quo. 5 respondents cautioned alienating long-term contributors, among whom 1 specified the need to learn from their long experience.
Some respondents seemed to want to revisit discussions and arguments that had been resolved, with one asking for closed discussions to be reopened, and others warning against alienating long-term users, including a reference to a banned long-term user. One expressed that WMF, "a distant central power," can never have "enough visibility in a local situation."
No answer: 44/36 (lower number indicates missing data in filled-in surveys, larger number includes the 8 empty forms)
Interviewees were also asked for any other comments, most of which fell into categories of transparency and privacy/anonymity, taking action against harassment and bad behavior, having a process for conflict resolution, extending friendly spaces and welcoming newcomers, and one recommendation to demand resources for CH.
Two demographic questions were asked, about region and gender. No questions about other demographics, such as race and age, were included.
Comparing survey data with interview data, interviews seem to have had more representation of women and people from regions underrepresented in the survey, but there were few responses to regional data from the interviews.
Survey respondents were overwhelming from Europe and North America, and most responded in English. 76 responses to the question “Please share your region, if you feel comfortable sharing” were received in the 106 survey responses (98 of which were filled in). Answers that were not geographic (Spiritual 2, Christian 1, apatheist 1) were counted as declines.
Declined to answer: 26/18
Only a few answers go to city or sub-regional level, mostly from the UK and US, 1 from Belgium, 3 from Spain. Cities and states or provinces were included as their country. Basque country and Catalonia were included as Spain. These last two answers came from the Spanish language surveys.
Table 1. The breakdown by region:
Any respondents from other regions did not share this information.
Figure 1. Pie-chart of responses by region
61 survey responses indicated a country, while 15 more included only a region. The list of countries named by survey respondents reflects the survey languages. Indonesia and Japan were the only Asian nations named in the responses. Bahasa Indonesian was the only Asian language in which the survey was available, and this may have lead to the strong participation - nearly 5% of surveys were from the Indonesian language version of the survey. It may be possible that representation from other places would have been higher if the survey had been translated into local languages.
Table 2. Geographic responses by country
Interviews offered far fewer respondents to this question, however, regions underrepresented in the survey data were better represented in interviews, with 3 people from Africa and 2 from South America. These numbers remain very small.
Table 3. Region of respondents
Summary: 28 female, 46 male, 30 declined to respond
Table 4. Gender breakdown
|Decline to answer||30|
The response "male-ish" coded as male, and can be revised. No respondent identified as anything else.
Two answers that were not gendered ('Really comfortable about it" and decline) were coded as having declined to answer.
Figure 2. Pie-chart of gender responses
Interview data is different with regard to gender, with a smaller proportion of men and larger proportions of men and unknown genders.
- Ford & Wajcman. Anyone can edit, not everyone does: Wikipedia and the gender gap; Wulczyn, Thain & Dixon, Ex Machina: Personal Attacks Seen at Scale