First preliminary results from UNU-Merit Survey of Wikipedia Readers and Contributors available
April 16th, 2009

From late October to early November 2008, the Wikimedia Foundation and UNU-Merit conducted the first multilingual survey of Wikipedia readers and contributors in 20 languages. In total, more than 130,000 Wikipedia readers and contributors completed the extensive survey questionnaire (out of more than 300,000 people total who took at least part of the questionnaire).* This level of response far exceeded our expectations, and the data that was collected provides a wealth of information about the Wikipedia community. English, German and Spanish were the most responsive Wikipedia editions and together make up two thirds of the responses.

The UNU-Merit team has spent the previous months cleaning and preparing the data, and is now making available first results for some of our priority questions. Key outcomes of this first analysis include:

* 65% of respondents self-described as readers, and 35% as (mostly occasional) contributors. Former contributors are analysed separately.
* Respondents came from over 200 countries, ranging from 10 to 85 years completed the survey; their average age is 26 years, and 25% of the respondents are younger than 18 years. Female respondents are a bit younger than the average (24 years)
* Among these, readers and contributors are on average in their mid-twenties, and predominantly male (75%)
* Women, with a share of 25% in all respondents, are more strongly represented among readers (32%) and less strongly represented among contributors (13%).
* Both educational levels and age are slightly higher among contributors than among readers.
* Regarding their motivations to contribute, respondents mentioned as their top two reasons that (1) they liked the idea of sharing knowledge, and (2) that they had come across an error and wanted to fix it.
* The concern that they might not have enough information to contribute is the main reason holding back potential contributors, mentioned by 51% of this group. Fourty-eight percent mentioned they were happy readers of Wikipedia, and saw no reason to get involved as contributors.
* The most common reason why respondents have not donated money to the Wikimedia Foundation, mentioned by more than 42% of respondents, is that they don’t know how. (If you happen to be one of them, we suggest you go to donate.wikipedia.org ;-) )

Ruediger Glott and Philipp Schmidt from UNU-Merit have made available additional data in the online workbook of their analysis (PDF file), and we’re planning to give you regular updates with new data every couple of weeks from now on. The survey team also maintains its own website at wikipediastudy.org.

This is a landmark moment in the history of Wikipedia and the Wikimedia movement. These and future findings that will result from this data will help to shape our efforts to reach new contributors and new readers. The Wikimedia Foundation wishes to thank everyone who has made this survey possible, especially the UNU-Merit Team and the community of translators.

Erik Moeller
Deputy Director, Wikimedia Foundation

* In addition to the 130,000 responses overall, we’ve received 40,000 responses from the Russian Wikipedia, which very significantly overrepresents this group in the total response set. The survey team has excluded this group from the data until the possible causes for this overrepresentation can be fully understood.

Finally, female Wikipedians are significantly younger (24.2 years) than male community members (26.4 years).

Hmm, the part about 13% of the contributors being women caught my eye...

Off of the WMF blog

My completely uneducated stab-in-the-dark guess is that the median age is probably about 20.

First preliminary results from UNU-Merit Survey of Wikipedia Readers and Contributors
* 65% of respondents self-described as readers, and 35% as (mostly occasional) contributors. Former contributors are analysed separately.

The status of the respondents regarding partnerships also corresponds to the age structure of the Wikipedia community, as only 30% of the respondents say they have a partner. Correspondingly, the share of Wikipedians with children is only 14%. The difference between readers and contributors in this regard is negligible.

Pretty close. Survey says "Half of the respondents are are [sic] younger than 22 years".

But the real question is how many of those were born as women.

... So I assume consensus will never get all the inappropriate pics off of Wikipedia, if we define "inappropriate" as "what most parents would want to keep off of their children's computer screens."

Off of the WMF blog



That's the part I'm eager to see.

The sentence was apparently written ad spellchecked by this half.

I'm rather surprised that by their mid-20s most of Wikipedia's contributors had worked out which sex they were. Its clearly not my impression.

I am surprised that there is no discussion of selection or that we would even entertain the results without any information about selection. That the survey indicates a mere 2:1 ratio between "reader" and "contributor" indicates that something is seriously amiss. Looking things up on Wikipedia is very, very common. Editing Wikipedia is rare.

Self-selection bias. "Contributors" are probably 10x to 100x more likely to participate in a survey on the site than "Readers" would be. As for completing a survey once started, the differential is likely even higher again by a factor of 2 or 3 to 1.

This is what I do for a living -- remind stakeholders of the inherent biases in survey data collection.

Greg