Consider the Redirect

In wikis, redirects are special pages that silently take readers from the page they are visiting to another page. Although their presence is noted in tiny gray text (see the image below) most people use them all the time and never know they exist. Redirects exist to make linking between pages easier, they populate Wikipedia’s search autocomplete list, and are generally helpful in organizing information. In the English Wikipedia, redirects make up more than half of all article pages.

seattle_redirectOver the years, I’ve spent some time contributing to to Redirects for Discussion (RfD). I think of RfD as like an ultra-low stakes version of Articles for Deletion where Wikipedians decide whether to delete or keep articles. If a redirect is deleted, viewers are taken to a search results page and almost nobody notices. That said, because redirects are almost never viewed directly, almost nobody notices if a redirect is kept either!

I’ve told people that if they want to understand the soul of a Wikipedian, they should spend time participating in RfD. When you understand why arguing about and working hard to come to consensus solutions for how Wikipedia should handle individual redirects is an enjoyable way to spend your spare time — where any outcome is invisible — you understand what it means to be a Wikipedian.

That said, wiki researchers rarely take redirects into account. For years, I’ve suspected that accounting for redirects was important for Wikipedia research and that several classes of findings were noisy or misleading because most people haven’t done so. As a result, I worked with my colleague Aaron Shaw at Northwestern earlier this year to build a longitudinal dataset of redirects that can capture the dynamic nature of redirects. Our work was published as a short paper at OpenSym several months ago.

It turns out, taking redirects into account correctly (especially if you are looking at activity over time) is tricky because redirects are stored as normal pages by MediaWiki except that they happen to start with special redirect text. Like other pages, redirects can be updated and changed over time are frequently are. As a result, taking redirects into account for any study that looks at activity over time requires looking at the text of every revision of every page.

Using our dataset, Aaron and I showed that the distribution of edits across pages in English Wikipedia (a relationships that is used in many research projects) looks pretty close to log normal when we remove redirects and very different when you don’t. After all, half of articles are really just redirects and, and because they are just redirects, these “articles” are almost never edited.

edits_over_pagesAnother puzzling finding that’s been reported in a few places — and that I repeated myself several times — is that edits and views are surprisingly uncorrelated. I’ll write more about this later but the short version is that we found that a big chunk of this can, in fact, be explained by considering redirects.

We’ve published our code and data and the article itself is online because we paid the ACM’s open access fee to ransom the article.

Another Round of Community Data Science Workshops in Seattle

Pictures from the CDSW sessions in Spring 2014
Pictures from the CDSW sessions in Spring 2014

I am helping coordinate three and a half day-long workshops in November for anyone interested in learning how to use programming and data science tools to ask and answer questions about online communities like Wikipedia, free and open source software, Twitter, civic media, etc. This will be a new and improved version of the workshops run successfully earlier this year.

The workshops are for people with no previous programming experience and will be free of charge and open to anyone.

Our goal is that, after the three workshops, participants will be able to use data to produce numbers, hypothesis tests, tables, and graphical visualizations to answer questions like:

  • Are new contributors to an article in Wikipedia sticking around longer or contributing more than people who joined last year?
  • Who are the most active or influential users of a particular Twitter hashtag?
  • Are people who participated in a Wikipedia outreach event staying involved? How do they compare to people that joined the project outside of the event?

If you are interested in participating, fill out our registration form here before October 30th. We were heavily oversubscribed last time so registering may help.

If you already know how to program in Python, it would be really awesome if you would volunteer as a mentor! Being a mentor will involve working with participants and talking them through the challenges they encounter in programming. No special preparation is required. If you’re interested, send me an email.

Community Data Science Workshops Post-Mortem

Earlier this year, I helped plan and run the Community Data Science Workshops: a series of three (and a half) day-long workshops designed to help people learn basic programming and tools for data science tools in order to ask and answer questions about online communities like Wikipedia and Twitter. You can read our initial announcement for more about the vision.

The workshops were organized by myself, Jonathan Morgan from the Wikimedia Foundation, long-time Software Carpentry teacher Tommy Guy, and a group of 15 volunteer “mentors” who taught project-based afternoon sessions and worked one-on-one with more than 50 participants. With overwhelming interest, we were ultimately constrained by the number of mentors who volunteered. Unfortunately, this meant that we had to turn away most of the people who applied. Although it was not emphasized in recruiting or used as a selection criteria, a majority of the participants were women.

The workshops were all free of charge and sponsored by the UW Department of Communication, who provided space, and the eScience Institute, who provided food.

cdsw_combo_images-1The curriculum for all four session session is online:

The workshops were designed for people with no previous programming experience. Although most our participants were from the University of Washington, we had non-UW participants from as far away as Vancouver, BC.

Feedback we collected suggests that the sessions were a huge success, that participants learned enormously, and that the workshops filled a real need in the Seattle community. Between workshops, participants organized meet-ups to practice their programming skills.

Most excitingly, just as we based our curriculum for the first session on the Boston Python Workshop’s, others have been building off our curriculum. Elana Hashman, who was a mentor at the CDSW, is coordinating a set of Python Workshops for Beginners with a group at the University of Waterloo and with sponsorship from the Python Software Foundation using curriculum based on ours. I also know of two university classes that are tentatively being planned around the curriculum.

Because a growing number of groups have been contacting us about running their own events based on the CDSW — and because we are currently making plans to run another round of workshops in Seattle late this fall — I coordinated with a number of other mentors to go over participant feedback and to put together a long write-up of our reflections in the form of a post-mortem. Although our emphasis is on things we might do differently, we provide a broad range of information that might be useful to people running a CDSW (e.g., our budget). Please let me know if you are planning to run an event so we can coordinate going forward.

Community Data Science Workshops in Seattle

Photo from the Boston Python Workshop – a similar workshop run in Boston that has inspired and provided a template for the CDSW.
Photo from the Boston Python Workshop – a similar workshop run in Boston that has inspired and provided a template for the CDSW.

On three Saturdays in April and May, I will be helping run three day-long project-based workshops at the University of Washington in Seattle. The workshops are for anyone interested in learning how to use programming and data science tools to ask and answer questions about online communities like Wikipedia, Twitter, free  and open source software, and civic media.

The workshops are for people with no previous programming experience and the goal is to bring together researchers as well as participants and leaders in online communities.  The workshops will all be free of charge and open to the public given availability of space.

Our goal is that, after the three workshops, participants will be able to use data to produce numbers, hypothesis tests, tables, and graphical visualizations to answer questions like:

  • Are new contributors to an article in Wikipedia sticking around longer or contributing more than people who joined last year?
  • Who are the most active or influential users of a particular Twitter hashtag?
  • Are people who participated in a Wikipedia outreach event staying involved? How do they compare to people that joined the project outside of the event?

If you are interested in participating, fill out our registration form here. The deadline to register is Wednesday March 26th.  We will let participants know if we have room for them by Saturday March 29th. Space is limited and will depend on how many mentors we can recruit for the sessions.

If you already have experience with Python, please consider helping out at the sessions as a mentor. Being a mentor will involve working with participants and talking them through the challenges they encounter in programming. No special preparation is required.  If you’re interested,  send me an email.

The Wikipedia Gender Gap Revisited

In a new paper, recently published in the open access journal PLOSONE, Aaron Shaw and I build on new research in survey methodology to describe a method for estimating bias in opt-in surveys of contributors to online communities. We use the technique to reevaluate the most widely cited estimate of the gender gap in Wikipedia.

A series of studies have shown that Wikipedia’s editor-base is overwhelmingly male. This extreme gender imbalance threatens to undermine Wikipedia’s capacity to produce high quality information from a full range of perspectives. For example, many articles on topics of particular interest to women tend to be under-produced or of poor quality.

Given the open and often anonymous nature of online communities, measuring contributor demographics is a challenge. Most demographic data on Wikipedia editors come from “opt-in” surveys where people respond to open, public invitations. Unfortunately, very few people answer these invitations. Results from opt-in surveys are unreliable because respondents are rarely representative of the community as a whole. The most widely-cited estimate from a large 2008 survey by the Wikimedia Foundation (WMF) and UN University in Maastrict (UNU-MERIT) suggested that only 13% of contributors were female. However, the very same survey suggested that less than 40% of Wikipedia’s readers were female. We know, from several reliable sources, that Wikipedia’s readership is evenly split by gender — a sign of bias in the WMF/UNU-MERIT survey.

In our paper, we combine data from a nationally representative survey of the US by the Pew Internet and American Life Project with the opt-in data from the 2008 WMF/UNU-MERIT survey to come up with revised estimates of the Wikipedia gender gap. The details of the estimation technique are in the paper, but the core steps are:

  1. We use the Pew dataset to provide baseline information about Wikipedia readers.
  2. We apply a statistical technique called “propensity scoring” to estimate the likelihood that a US adult Wikipedia reader would have volunteered to participate in the WMF/UNU-MERIT survey.
  3. We follow a process originally developed by Valliant and Dever to weight the WMF/UNU-MERIT survey to “correct” for estimated bias.
  4. We extend this weighting technique to Wikipedia editors in the WMF/UNU data to produce adjusted estimates of the demographics of their sample.

Using this method, we estimate that the proportion of female US adult editors was 27.5% higher than the original study reported (22.7%, versus 17.8%), and that the total proportion of female editors was 26.8% higher (16.1%, versus 12.7%). These findings are consistent with other work showing that opt-in surveys tend to undercount women.

Overall, these results reinforce the basic substantive finding that women are vastly under-represented among Wikipedia editors.

Beyond Wikipedia, our paper describes a method online communities can adopt to estimate contributor demographics using opt-in surveys, but that is more credible than relying entirely on opt-in data. Advertising-intelligence firms like ComScore and Quantcast provide demographic data on the readership of an enormous proportion of websites. With these sources, almost any community can use our method (and source code) to replicate a similar analysis by: (1) surveying a community’s readers (or a random subset) with the same instrument used to survey contributors; (2) combining results for readers with reliable demographic data about the readership population from a credible source; (3) reweighting survey results using the method we describe.

Although our new estimates will not help us us close the gender gap in Wikipedia or address its troubling implications, they give us a better picture of the problem. Additionally, our method offers an improved tool to build a clearer demographic picture of other online communities in general.