Internet Skills and Wikipedia’s Gender Inequality

with Eszter Hargittai and Aaron Shaw
January 21, 2014 at 12:30PM
Berkman Center for Internet & Society, Harvard
Link to event page

The luncheon at the Berkman Center.

The luncheon at the Berkman Center.

I was excited to find out in a tweet from Stine yesterday that this event would be taking place in Cambridge, just as I happen to be here a few more days. I registered hastily, emailed Eszter Hargittai to let her know I’d be there and what our project is about, and today I made my way shyly to 23 Everett St, a cute typical Cambridge yellow house with carpet-covered stairs leading up to the Berkman Center, home to quirky door signs and funky looking internet researchers. The room where the luncheon was to take place filled up very quickly with people who all seemed to know each other if not from current then past co-operation, and I wondered, clutching my dozen sheets of WikidGrrls propaganda in one hand and my Twitter-ready phone in the other, what will I say if I have the chance to speak for 20 seconds about WikidGRRLs? Here’s what I came up with:

Last January a few colleagues and I were awarded a 15k seed grant from the Future of Information Alliance at the University of Maryland. We were given the opportunity to put the scholarship about Wikipedia’s gender gap to practical use in the real world: i.e., a chance to see where this gap is formed and if this gap could be fixed by an intervention in middle school education. We were dispatched to four local schools (mostly in Prince George’s County, Maryland) were each one of us taught a group of about ten girls for ten weeks. We introduced them to the Internet as a tool for creating and sharing knowledge, specifically using a wiki site. We found that although teens are quite comfortable with the Internet as a place for seeking out information and connecting with friends, they barely knew how to “use” it to produce, transform, edit and share information, knowledge and/or experience for the benefit of others. One thing that especially struck me when I asked the girls why they think girls contribute so disproportionately little to Wikipedia compared to boys, one of them answered: because we have better things to do, like taking care of our families.

Well, I never got to say precisely that, but I did speak up twice to ask questions and to introduce our work, and I did circulate our propaganda around the table. But enough of this subjective banter – what was the talk about? I would direct you first and foremost to Nate Mathias‘ fantastic summary of the whole thing, which includes a very neat Prezi made by Willow Brugh, and all the links you need to find out more. Here are a few notes of my own to top it off:

Wikipedia is obviously important as an everyday source of information to millions of people: it is one of the five most visited websites on the whole web. More interestingly, it is an amazing mine of “volunteer labor” it is estimated that 41 million hours of free work have been put into its creation. However, the disheartening aspect of this creation is that women make up only 16% of global editors, and 23% of US adult editors, a statistic that the Wikimedia foundation has been struggling with and trying to counter for a number of years.

Eszter Hargittai from Northwestern University explained that her longitudinal data of hundreds of college students is beneficial because it lets her show that internet skills do not improve vastly over time but stay relatively the same. In each of the three waves of data collection, she and her students used pencil and paper surveys, which I found very interesting: it was to avoid biasing the responses towards those who are internet savvy and active. The data were collected between 2009 and 2012 and the population was diverse, with around 40% Whites, 20% Hispanics and 20% Asian Americans. 61% of the respondents were female. Eszter noted that the sample was pretty much representative of this age group in the US, as 48% of her respondents graduated from college, which is close to the national average.

The survey measured basic variables pertaining to internet use, such as number of use years, number of access locations (a.k.a. autonomy of use), weekly hours spent online, and evaluated respondents using an internet skill index (scale developed by Eszter). It also contained Wikipedia specific measures: confidence in editing WP, school assignment to edit WP, etc. The results were rather disappointing in terms of the prevalence of contribution to Wikipedia: of all the respondents,

Read an entry: 99%
Edited a mistake: 25%
Edited by adding: 18%
Started a new entry: 9%
Added an image: 10%
Contributed in any way: 27.6%

When it came to the talk’s topic, ie. the gender gap, it turned out that men are much more likely (about 2-3 times) to report having contributed. However, the researchers noted that having the assignment in school does level the playing field both between races and between levels of parents’ education. Hopeful in terms of possible interventions…such as WikidGRRLs! This is confirmed by the following graphic:


The gender difference in contributions is slightly reduced when assignments are given in school to edit Wikipedia.

Aaron Shaw, who also works at Northwestern and like Hargittai was a fellow at the Berkman Center, said that the only things that are statistically significant in explaining the differences between those who edit and those who don’t are gender, internet skills, and whether or not you were assigned in school to edit Wikipedia.

The most discouraging conclusion drawn from their model was that as skills were higher, the disparity between men and women in terms of probability to edit was ever greater. Indeed, look at this graphic:


The “Male” curve is the top one.

As you move out along the skills variable, the women in the sample remain relatively unlikely to edit. Men on the contrary become increasingly more likely to be editors than women (50% vs 30%). In conclusion, women are less likely to have high skills and even at high skills they are less likely to edit Wikipedia than men. Interestingly, even controlling for socio-economic status does not reduce the effect of gender on disparity of internet skills: men’s internet skills are more spread out than women’s, but their mean is higher.

Takeaways from this talk:

  • The gender gap really matters among the higher skilled users.
  • Skills really matter. People with low skills just don’t contribute.
  • Skills have long term effects in terms of the behavior that people engage in.
  • It seems that “confidence in editing” doesn’t predict who edits.

Questions for future research:

  • Why aren’t skilled women more likely to contribute?
  • How can Wikipedia (the world’s largest free knowledge resource) address these barriers to entry for low skills internet users in general?

YouTube video from this talk (yours truly tries to sell WikidGRRLs at 54:51):



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s