Learning and Evaluation/Inspire Campaigns/Gender gap/What's next

From Meta, a Wikimedia project coordination wiki

Inspire campaign for gender diversity

Achievement against campaign objectives Community Resources identified three objectives[1] to measure progress against our goal to “proactively source and support new projects aimed at increasing gender diversity [on] Wikimedia projects”:

  1. Demonstrated significant increase in percent of IEG/PEG supported projects led by women (increased by over 50% from current IEG/PEG baseline of 30%)
  2. Demonstrate significant increase in female contributors: 1500 female participants, 1000 new female users, 200 active female users.
  3. Demonstrate increase in non-male gender-classified content, as measured by the increase in biographies about women. The target is to see a:
    • 2% increase in female biographies in Wikipedia projects with over 1 million articles
    • 4% increase in Wikipedia projects with article counts between 100 thousand and 1 million articles
    • 6% increase in Wikipedia projects with article counts less than 100 thousand articles.

Based on the Gender Gap Inspire grants completed by August 2016, we know the following:

  1. 86% of projects supported by Gender Gap Inspire grants were led by women.
  2. A total of 626 individuals were involved across the grants
    • At least 144 were women. We have minimal information about the total number of new or active female users.
    • 259 were new users.
    • 269 were users who remained active for the duration of the program (making at least 5 edits per month)
  3. 12,717 articles were created or improved across Wikimedia projects, of which at least 157 biographies were created, 126 about women.
    • On average, between July 2015 to July 2016, the number of biographies about women increased by 0.39% across all Wikipedia projects, which amounts to 126,194 new biographies being written about women across all Wikipedia projects.[2]
  • Based on the same data, we know the following breakdown:
Size of Wikipedia Project July 2015 baseline:
Female biographies, as % of total biographies (Average)
July 2015 to July 2016:
Average increase in female biographies
Large projects

(Over 1 milllion articles)

16.56% +0.27% or 1100 new bios per project
Medium projects

(100K to 1M articles)

15.65% +0.33% or 167 new bios per project
Small projects

(less than 100K articles)

16.11% +0.44% or 21 new bios per project

What we learned from this campaign

As an initial experiment in proactive grantmaking, we consider the first Inspire Campaign a success. It engaged around 600 people in addressing the gender gap, resulting in both quantitative and qualitative shifts: approximately 13,000 articles were created or improved, the number of women leading grant-funded projects increased, and our thinking about gender diversity expanded in both expected and unexpected ways.

That said, we did not meet our targets around participation and content. Further investigation shows that we chose targets that were both too high to be feasible and too narrow to reflect the diversity of ideas proposed during the campaign. Consequently, our targets inadvertently set our campaign participants up for perceived failure and undermined the significance of what they have achieved.

Reflecting on this in retrospect, we see a need for broader engagement in the design stage of the campaign, including input from the Wikimedia communities and other Foundation staff. This would have helped us to better predict how our communities might respond to a call to action around gender diversity. The first campaign originated from the forward-looking ideas of some of the most inspiring thought leaders within the Foundation’s staff, who created this initiative by repurposing existing grants workflows to support innovation. Consequently, there were limited resources to support a substantial feedback process around campaign design, including the feasibility of our targets. Going forward, we believe that budgeting for such a process is crucial to campaign success.

The goals and targets we set for campaigns inevitably shape the types of ideas we receive. In this case, focusing on participation and content drew projects skewed toward increasing raw article numbers and participation: we funded large numbers of editathons, or other offline outreach, focused on training women how to edit and on increasing content about women. Though there are exceptions, we saw relatively few projects focused on improving article quality or building awareness about gender diversity. We also saw few projects focused on non-binary approaches to gender diversity that explicitly encompass all gender identities, rather than women in particular.

In the future, we would like to consider opportunities to broaden the array of proposals prompted by an issue-focused campaign. We believe that a more intensive feedback in the design phase - where we actively solicit, interpret and incorporate feedback from across our movement - would help us improve the language, definitions and objectives in the framing of the campaign, potentially widening the field of ideas that we attract.

In regards to future campaigns, there are also two important lessons we’ve learned around setting targets:

We need to set objectives and targets that are aligned with our underlying values.

Setting a target around the number of female participants seemed like a relevant and straightforward idea at the beginning of this campaign. However, as they focused on attracting, motivating and retaining women, many of our grantees identified concerns about collecting this type of information.

In many cases, maintaining anonymity about gender identity is a matter of safety and security. Consequently, some participants expressed concern about disclosing gender-identifying information to project organizers, or having such information linked to their usernames. Even when gender-identifying information is not specifically collected, protecting anonymity can still be problematic at events that are primarily designed to attract women. Simply adding one’s username to the sign-in list at an event predominantly attended by women is too great a risk of exposure for some users.

As coordinators of the Inspire Campaign, we found ourselves in a quandary. On the one hand, we wanted to collect this information to evaluate our success against our Campaign targets. On the other hand, we wanted to ensure user privacy to support safe conditions for editing. When these two goals were in conflict, we erred on the side of protecting privacy and safety - allowing participants to volunteer information as they felt comfortable - prioritizing the ultimate goals of the campaign over our ability to measure success in achieving them. In light of this, we should have anticipated the under-reporting of female participants, and especially female editors.

Given that we will continue to use the Inspire Campaigns to focus on sensitive topics - like the Harassment campaign held earlier this year - we anticipate always seeing some level of under-reporting of sensitive information. As such, going forward we aim to focus objectives on more qualitative outcomes, such as participants' feelings of safety or inclusion with our projects.

We need to set SMART-er objectives and targets, and refine them using available data to ensure they are (and remain) realistic. We believe we can do this better now, thanks to the baselines set in this campaign and the tools created by our grantees.

Just like our grantees, we struggled to accurately predict the outcomes of this campaign and set realistic targets based on those predictions. This was due in large part to:

  1. Having limited baseline information about female participation or biographies about women. For instance, using data from Wikipedia Human Gender Indicators (WHGI), we now know that on average the number of biographies about women increase ~0.5% per year, and that a 1% increase is a significant increase (regardless of the size of the Wikipedia). Our goals of increasing this by 2-6% were, in hindsight, highly unrealistic.
  2. Setting targets before knowing the kind of ideas that would be proposed, developed or funded. Ultimately, proposals were not selected solely on the basis of meeting our targets, instead focusing on meeting the overarching goal of the campaign. Grants for research and tools (such as WHGI) - which minimally contribute to the targets of people or articles - have been extremely valuable in improving our understanding of the gender gap and how or why it manifests.

However, this campaign leaves us with better baselines and tools, that we can use to set more realistic targets in the future.

What's next for our work on the gender gap?

We recognise that knowledge is differently understood, constructed, used and expanded in different contexts. Our most significant lens of diversity - particularly of geography, language and gender - implies that the motivations, means and methods of expanding knowledge on our projects are different for different communities. Our team’s work at times challenges the very core of the encyclopaedic work of Wikipedia: questioning what ‘notability’, ‘reliability’ and ‘citation’ means in order to progress towards the ‘sum of all human knowledge’, whether oral or written, offline or online, by multiple voices or just a few. Only 1 in 10 contributors worldwide is projected to be female: this implies a significant set of knowledge gaps that our team is trying to address, in partnership with our communities around the world.
— Anasuya Sengupta, Chief Grantmaking Officer under whom the Inspire Campaign for gender diversity originated.

Improving the diversity of our content and contributors is a core goal of Community Resources. We were surprised and inspired by the people who participated in this campaign, the discussions they ignited in their communities, and the results of their projects. We will continue our work in proactive grantmaking, learning from this experience in the ways discussed above.

In particular, we are interested in understanding and experimenting with different models for proactive engagement and grantmaking. In terms of future grant campaigns, this could include adapting the current Inspire model - adjusting the budget, timeframe, and specificity of the topic - or exploring new models such as seeding the discussion with new ideas from thought leaders and issue experts both within and outside the movement. However, beyond grant campaigns, this could include developing a gender gap strategy for the Community Resources team to running thematic quarters or years.

Right now, while we are committed to keeping a focus on diversity, including gender diversity, our future next steps depend heavily on the movement-wide strategy that is currently in development. As this movement-wide strategy will have implications on the organizational strategy of the Wikimedia Foundation, our team’s short term goal is to ensure that improving diversity continues to be a major theme in these strategy conversations. This will include ensuring that the various communities who work on diversity are also a part of these conversations.

In the long term, once the movement strategy is complete in August 2017, our team will revisit the topic of gender diversity under the new lens of the movement’s strategic priorities. By designing new or improved models for proactive grantmaking, our hope is that Inspire campaigns will continue to ignite innovative thinking how to address key issues like the gender diversity within our movement.

Notes & References