WWC2023/Program/Strategy/Research

• Home
• Research
• Hubs
• Engagement Plan

Introduction to Research

Session led by Nazia Akthar, Netha Hussain, Miriam Redi (online), and Rosie Stephenson-Goodknight

The pre-read pack containing past research related to gender in the Wikimedia movement was sent to all participants.

The Wikimedia Foundation identified that 81% of Wikipedia articles across languages are about men, 80% of active editors self-identify as men and 72% of pageviews come from people who self-identify as men.

Extended content
Knowledge gaps taxonomy: The knowledge gap taxonomy allows for finding and classifying evidence of inequalities in the Wikimedia projects. Readers: On Wikipedia, readers tend to read articles about people of similar demographics (gender and age) to them, which makes it important that there is adequate representation of women of all ages in all languages. In every part of the world, men tend to read Wikipedia more often than women. Men also read more articles when they visit Wikipedia than women. Participants: Time and again, researchers have shown that there is an alarmingly low proportion of women contributing to Wikimedia projects. However, there is some room for hope. Newcomers are more likely than earlier to identify as women and gender diverse than seasoned editors on Wikipedia, indicating that there is a steady increase in diversity among editors. Unfortunately, they are also more likely to indicate as having felt unsafe on wiki. Despite the challenges, they contributed to Wikipedia because they strongly identified with Wikipedia’s mission for free knowledge, wanted to share what they know with the world, or wanted to hone their writing and researching skills. Content: While Wikipedias in nearly every language have less number of articles about women compared to men, Wikipedia is likely to have more content and quantity of women’s biographies than traditional encyclopaedias. This gender gap spills over to Wikidata, where a larger gender gap exists in large language communities, with the gap narrowing over time. Although the majority of biographies are about men, the quality of biographies of women and individuals of non-binary gender are slightly higher than that of men. Articles with more interwiki links are likely to be about men, and there is a social bias on Wikipedia to assume male as the standard gender of the subject. Words related to family life, such as ‘marriage’ and ‘children’ are more likely to be associated with articles related to women, while career related words are more linked with men’s biographies. There is also a gender gap in terms of images on Wikipedia, with articles related to men having more images compared to that of women. Similarly, publications by women are less cited on Wikipedia than expected. Policies and practices: Women’s gender plays an important role in their perceived significance on Wikipedia, aggravating the long chain of historical inequalities and marginalizations in which women’s qualifications and achievements are undervalued. Activist editors and trainers of gender gap campaign have often noted harassment existing on Wikipedia and the difficulty in navigating through the large corpus of Wikipedia’s policies, which are not newcomer-friendly. The reliability guidelines on Wikipedia have also been criticised for excluding the complex, lived experiences of several communities. As a result, only a small number of privileged editors get to decide what content is reliable for Wikipedia.

Research Discovery & Centralization

Session led by Nazia Akthar, Netha Hussain, Rosie Stephenson-Goodknight

Links

• A pre-conference survey was sent to all attendees. Survey notes link (forthcoming). Survey analysis link (forthcoming).
• A pre-read pack of 66 research items was compiled by the Research cohort. It included item title, url, year of publication, year research was conducted, year/type of data dump, type of research, methods/notes, and key findings.

Extended content
Part 1, pre-conference survey The survey analysis was conducted by Miriam, Tanja, and Netha before WikiWomenCamp convened. The results were presented by Miriam and Netha. Key findings based on survey responses are below. Involvement in gender research: Aha slide (forthcoming) Main areas of interest: Measuring and improving gender participation and contribution: analyzing differences in size and way of editing by gender; creating initiatives for women as active contributors; analyzing the relation between youth and women contributions; studying leadership roles and gender gap Studying the origin of gender bias and removing barriers to gender equality: analyzing gender bias in society, editors, and sources; barriers: digital divide; technical tools; perception of gender Measuring and addressing content gaps: gathering data about deleted bios and underrepresentation of women; studying the gender gap by language or in specific languages or occupations; representing the struggle of women in our society. Understanding the impact of support efforts for empowering women contributors: mentorship, on-wiki governance, capacity building, partnerships Studying and improving the safety, health and retention of gender diverse contributors Studying women’s participation beyond Wikipedia: e.g. WikiSource Generic interest: taking part in research; learning about the state of research (inc. affiliate-led research), the tools and the gaps; learning about state of gender equality and diversity. Part 2, other examples of finding gender-related content on wiki-meta Soukaina described the example of finding information about high school education through the Wiki4Education project based on Wikidata Ciell described the use of space banners to attract editors to wikiwomen events posted on the Celebrate Women wiki-meta page during Women’s History Month Vanj described social media methods to attract people to wikiwomen events Masana described the work that her team is doing on campaign-building Part 3, finding gender-related research We described where the research cohort found the research items that were included in the pre-read pack: Wiki-Meta (Celebrate Women) Wiki-Meta (Gender gap/Research) Wiki-Meta (Research:Index) Wiki-Meta (other?) Various language Wikipedias https://en.wikipedia.org/wiki/Wikipedia:WikiProject_Women_in_Red/Research Wikidata Social media Other Using Aha slides, we asked the attendees where they look for wiki gender-related research (forthcoming). Part 4, workshop The session attendees held discussions at four tables regarding the challenges and opportunities of various approaches to forming a centralized place (on Wiki-Meta, Wikipedia, Wikidata, or other) for links to gender-based wiki research. Aha slides responses (forthcoming). Aftermath Netha, Nazia, and Rosie met with Soukaina after the final Research cohort session to discuss the approaches for adding all wiki gender-related research into Wikidata, and then generating a Listeria list on wiki-Meta at Gender gap/Research on wiki-Meta. Rosie met with Soukaina again 2 weeks later during Wiki Indaba in Agadir to discuss the project further. Soukaina will begin work on the project in December 2023 or January 2024.

Research Community

Session led by Nazia Akhtar and Miriam Redi (online)

Objective

The overall goal of this session was to facilitate awareness, gather knowledge and information about the Wikimedia research community, as well as enable strong research links, networks, and connections between the participants, among whom a large number had expressed interest in conducting research on the gender gap in the Wikimedia Movement through the pre-camp survey.

Design

The session consisted of laying the groundwork for a discussion among the Strategy Cohort participants on what a research community in the Wikimedia Movement looks like/could look like and then conducting a group activity on the same. Its purpose was to collect and collate whatever information we already have and learn more about the research networks associated with the Wikimedia movement that we are not aware of and what more can be done to build on these to make gender-gap research smoother and more fruitful.

To this end, there were two overall questions underpinning the session:

1. What are the different streams of research that exist in the Wikimedia movement?
2. What are the intersections and collaborations between these streams?

Extended content
Summary Miriam Redi of the Research Division (RD) at Wikimedia Foundation (WMF) discussed the role of the RD in developing and supporting research communities. As part of this, she stressed: the conducting of regular events for the communities, such as Wikimedia Research Showcase Wiki Workshop 2023; rewarding research that has critical potential through initiatives, such as the annual WMF Research Award, which in 2023 included two papers related to gender-based research; offering research funds through the Wikimedia Research Fund, which concentrates on funding projects that are both critical to the Wikimedia ecosystem as well as boost diversity within the movement; releasing a bi-annual research report every June and December; and offering technical support to the research community to support them throughout their research projects with mentoring as well as facilitating access to data and other resources. Miriam’s presentation also offered open research questions to the Strategy participants to encourage them to conduct research. These were the specific research directions she mentioned: Addressing knowledge gaps Preserving knowledge integrity Building the foundation When asked by a participant whether there was data available on Wikimedia projects other than Wikipedia, Miriam responded that the research team has: recently released image files for Wikimedia Commons; will be releasing quality scores for Wikidata; and is eager to do more to help foster computer vision research. The second half of the session was conducted by Nazia, who identified known streams of research and then posed two AHA slide survey questions. The first question posed asked what were some of the new and emerging challenges associated with women’s representation in the Wikimedia movement? The responses to this question stressed: Political, social, and cultural factors Right-wing interference in anything considered "woke" Media representation of women The dependence of the new generation on social media for information lack of recognition and compensation for women's work Topics that are tailored to women's general interests or needs Biased sources Shortcomings within the Wikimedia Movement Resources Limited funding and other resources, and The role of bureaucracy in securing access to resources Logistics and Infrastructure Global coordination for content campaigns Lack of communication An affiliate to lead conversations on hubs Marginalization Redlists Harassment and Targeting, Backlash from men Burnout Representation Shrinking of communities and pushing out of women, non-binary individuals, and other minorities from the movement; A decreasing number of women in administrative and organizing positions and a reduction in their sustainable growth; and A decline in the number of Trust and Safety staff. Design/Tech Lack of metrics for articles that are not biographies Lack of knowledge on the WikiWomen Movement from other regions In response to an AHA survey question, 6 participants out of 13 said that they did not know of other streams of research in the Wikimedia movement. Three said that they were not sure, but 4 also said that they had heard of such streams. The next question asked the participants to list other sources or streams of research that they had heard of. The responses that were relevant to this inquiry included: The United Nations Diff blog posts Individual contributors, senior citizens, and collective and oral memory/history Grant reports News sources Podcasts Affiliates The next aspect of this session was a table discussion (in pairs) on: What do volunteer Wikimedian Researchers, academic researchers and the WMF research division bring to the table? How can we connect, collaborate, and collectively produce better research on the gender gap? We asked a few pairs to present. Since we did not have enough time to have each and every pair of participants present their ideas before the room, we collected all the sticky-note observations and displayed them on an easel board at the front of the room. Notes were also taken by an official note-taker as well as by a volunteer on the easel board at the front of the room. What follows is a summary of the insights gleaned from all these sources: In response to the first question: Different researchers specialize in different areas and bring different kinds of expertise and perspectives. There should be a platform for all the streams to connect. There was consensus that synthesis is required between external research and WMF research because whatever research individual streams have done does not give one stream the full picture. All the voices are needed, and they must connect and collaborate. There was consensus that difference in movement perspectives enhances knowledge. Wikimedians have insider knowledge from the ground and local levels, cultural knowledge, legitimacy, “response-ability,” a greater sense of accountability, good assessment of the challenges and implementation plan, the ability to communicate better and consult with the communities, but no human capacity that is sustainable. On the other hand, academia is good at/knows more about machine learning, managing large datasets, and tends to have existing Wiki connections and skills in methodology. The Research Department at WMF brings resources, tools, expertise, coordination capacity, and technical capacity at a global scale. can also conduct Research conferences to foster knowledge exchange and peer learning. Researchers have specifications (?) and critical lenses. In response to the second question: There was widespread agreement in the room that the RD at WMF can conduct research conferences to foster knowledge and capacity exchange and peer learning. Related ideas included the following: Have research tracks in more conferences. Regular research events should be promoted well. There will be language diversity and differing perspectives during research conferences, so the challenge is how to bring them all together meaningfully during such an initiative. There are social media networks, academic networks (ex: Research Gate), and citation databases such as Mendeley operating. They make research and researchers visible. Collaboration can be built over these networks. Wiki-Meta can plan to have such networks to make their researchers and research visible. There were also other responses that do not appear to be concerned or connected to imagining the forms or modes in which different streams could come together and collaborate. But they did offer several ideas about themes that the different streams could come together and undertake as research topics: The WMF RD and academic researchers have data and this can be used not only to enhance capabilities to analyze article quality through the language used, but also isolate the dynamics and linguistics/words used (for example, for articles related to women). This may be facilitated by creating templates for article creation in different language WPs. There should be research about organizers, who often bring people into the movement, apart from the usual user data analysis. Better implementation of content tools on articles is required. Content analysis tools can be applied to WP to understand the qualitative gender bias on WP. This can also be extended to other languages and, thus, make research more inclusive, and allow us to see how different languages talk about different genders and related aspects/to help us better understand the various languages. We should also leverage the gender focus in a more qualitative way and look at existing research that is not primarily focused on women, but analyzes women’s needs. More gender-focused research should be conducted; this will show us the relationship patterns between language and gender and what this relationship looks like for each language community. One need of the hour is for a very factual and experiential body of knowledge, both local and global. research output in laypeople’s terms and discarding the use of jargon that is difficult to follow if one does not have a formal background in higher education. Instead, there should be easy information available in the form of capsules, podcasts that can be understood by ten-year-olds, and feedback mechanisms. Qualitative research on understanding relationships and patterns between communities (micro/macro) If we want to increase information about women/reduce the gender gap - how possible is it given the lack of secondary and tertiary resources about women? Can this be quantified? We must be mindful of resource drainage. Methodologies should be improved. Data language analyst → lack of categories; data templates → deletion discussions.

Building a Research Network

Summary is forthcoming.