Sunday, February 25, 2007
Wikipedia burnout: an analysis
It's a frequent assertion that there's an amount of stress in volunteering to help Wikipedia: for no other reason than adding content can often be taking part in an endless argument with a constantly-changing parade of people with unclear qualifications. It has become a truism that after enough time this leads to burnout, and Wikipedia loses another valuable member.
As this impression is based on anecdotal evidence (some of which was presented in this recent thread on WikiEN-l), I thought I'd take a stab at evaluating some evidence. I began with a simple assumption: I have always thought the WikiEN-l list was the forum where the serious Wikipedians gathered and talked policy. (I have since unsubscribed because I found it too much of a distraction from my primary interest in Wikipedia -- writing articles.) I looked at my collection of emails from January, 2003 when I first subscribed to this list, and found that 42 unique individuals had contributed at least one email in that month. So what has happened to them since?
In performing my analysis, I discarded two persons immediately: the first was Jimbo Wales, for obvious reasons; the second was an alias of a notorious troublemaker, who has since been banned from Wikipedia and would provide no help to answering my question concerning burnout. Then I worked on pairing each person with an account on the English-language Wikipedia, and was able to succeed with 33 of the remaining 40 names.
Next, I did a search to see which ones were granted Admin powers (a complete listing can be found here), and found that 27 of these 33 people did receive Admin rights, most within a year. I believe that this confirms my impression of the WikiEN-l list -- at least during that time frame. I also believe that this shows that anyone at the time with the minimum qualifications who wanted to be an Admin received the rights; there was nothing special about having those powers. The six exceptions only prove the rule. A search with Interiot's Tool shows that three who never became Admins stopped contributing to Wikipedia soon after. One (Eclecticology) has publically stated that he does not want to be an Admin on Wikipedia; a second is well-known for being contentious, and was notorious for making claims to WikiEN-l that he was the victim of anti-Semitic attacks -- only to have a quick investigation reveal that his opinion or edit was simply being challenged by another Wikipedian. This leaves one person (Ericd) who has continued to make edits and yet has never been grated Admin powers; I guess it is because he has never wanted them.
I then turned to these 27 Admins, and looked at their editting history (from Interiot's Tool), studied their user page (both current and older versions), and drew upon what I knew of these individuals (admittedly, some I know better than others, but I would not claim that I know any of them very well). I found that they could be placed into four groups.
1. Those who explicitly stated that they were quitting Wikipedia, or had quietly stopped contributing. I counted eight in this group, although one (JeLuF) could have moved over to the German Wikipedia. Three of them (jtdirl, Tannin, and Tarquin) had explicit comments on their user pages saying that they had quit Wikipedia. Jtdirl's was very specific about one of his reasons for leaving:
I put Koyaanis Qatsi to this group because he had asked to be de-Adminnned. A few months after the time his request was granted, his edit totals had fallen off, so it is likely he knew he might be leaving, and so felt guilty about keeping the Admin powers.
As for those who just quietly stopped contributing, it's difficult to be certain why they left, but Camembert, who had been appointed in the first round to the Arbitration Committee (or ArbCom) states he has left Wikipedia on his user page. In contrast, DanKeshet's user page alludes to several lengthy moves across the United States, and these disruptions may have been the cause for his departure. Clearly, not all departures are due to stress.
2. Those who had taken one or more breaks from Wikipedia, but as of January, 2007 were still contributing. Taking breaks (also known as "Wikibreaks") is not uncommon; a quick glance at the number of links to the Wikibreak template shows that about 548 people are currently on declared Wikibreaks, and not all who are on Wikibreaks use that template -- or even announce the fact.
I found six fell into this group, one of whom (The Cunctator) still participates in WikiEN-l. I don't know whether one should join this group with the first, because it is clear that they felt stress (Ortolan88 explicitly mentions it as one reason for his breaks from Wikipedia), or with the fourth group below because the break allowed them to cope with the stress and return to Wikipedia.
3. Those who had moved on to other projects related to Wikipedia. This is the largest group, with nine members. One could consider this group will not offer any useful insight about burnout, since their direct participation in Wikipedia suffered due to other activities, but one could also consider this group very useful, since many of them (at least six: Anthere, Danny, Eloquence, Maveric149, Sannse and Brion Vibber) went on to assume more responsibility for running Wikipedia-related activities : two are members of the Wikimedia Foundation board, one is an employee of the Foundation, two more have volunteered their time to help it, while one is a community support person for Wikia.
4. All others. This is a catch-all group for four individuals (including myself) who do not fall into one of the other three groups, or could be considered prime candidates for burnout -- yet continue to edit as frequently now as we ever have.
I have to say that I find this the most fascinating of the four groups because, despite all else, the members of this group continue to contributes at least as many edits despite all other reasons not to. In alphabetical order:
To return to the subject of my investigation, if there is qualifiable evidence that stress leads to burnout, I'd have to say that the evidence could be seen that way: for those who did not move away from the English Wikipedia (18 of 27), one could lump the "Quitters" with the "Wikibreakers", and show that 14 out of 18 left Wikipedia within four years, in contrast to half (3 of 6) of the non-Admin group drifting away. However, one could also argue that taking a break from Wikipedia is an effective tool, and lump the "Wikibreakers" with the "Others" (who appear to have found a way to deal with the stress, however successfully) to show that 10 out of 18 (or about half) continue to contribute to Wikipedia, the same share as with non-Admins.
Then again, losing as few as half of a group over four years seems to me a turnover rate that is worrisome. Most students in a four-year institution (whether high school or college) actually manage to finish the program, and most high school students have less motivation than the sample group I looked at. I can only hope that I've enticed someone, with a better grasp of statistics and analysis, to gather a larger sample, subject it to a better analysis and provide a better answer.
Geoff
Technorati tags: wikipedia
As this impression is based on anecdotal evidence (some of which was presented in this recent thread on WikiEN-l), I thought I'd take a stab at evaluating some evidence. I began with a simple assumption: I have always thought the WikiEN-l list was the forum where the serious Wikipedians gathered and talked policy. (I have since unsubscribed because I found it too much of a distraction from my primary interest in Wikipedia -- writing articles.) I looked at my collection of emails from January, 2003 when I first subscribed to this list, and found that 42 unique individuals had contributed at least one email in that month. So what has happened to them since?
In performing my analysis, I discarded two persons immediately: the first was Jimbo Wales, for obvious reasons; the second was an alias of a notorious troublemaker, who has since been banned from Wikipedia and would provide no help to answering my question concerning burnout. Then I worked on pairing each person with an account on the English-language Wikipedia, and was able to succeed with 33 of the remaining 40 names.
Next, I did a search to see which ones were granted Admin powers (a complete listing can be found here), and found that 27 of these 33 people did receive Admin rights, most within a year. I believe that this confirms my impression of the WikiEN-l list -- at least during that time frame. I also believe that this shows that anyone at the time with the minimum qualifications who wanted to be an Admin received the rights; there was nothing special about having those powers. The six exceptions only prove the rule. A search with Interiot's Tool shows that three who never became Admins stopped contributing to Wikipedia soon after. One (Eclecticology) has publically stated that he does not want to be an Admin on Wikipedia; a second is well-known for being contentious, and was notorious for making claims to WikiEN-l that he was the victim of anti-Semitic attacks -- only to have a quick investigation reveal that his opinion or edit was simply being challenged by another Wikipedian. This leaves one person (Ericd) who has continued to make edits and yet has never been grated Admin powers; I guess it is because he has never wanted them.
I then turned to these 27 Admins, and looked at their editting history (from Interiot's Tool), studied their user page (both current and older versions), and drew upon what I knew of these individuals (admittedly, some I know better than others, but I would not claim that I know any of them very well). I found that they could be placed into four groups.
1. Those who explicitly stated that they were quitting Wikipedia, or had quietly stopped contributing. I counted eight in this group, although one (JeLuF) could have moved over to the German Wikipedia. Three of them (jtdirl, Tannin, and Tarquin) had explicit comments on their user pages saying that they had quit Wikipedia. Jtdirl's was very specific about one of his reasons for leaving:
One of the saddest experiences I had was to see a very very senior figure, a world-renowned expert who could command fees in the tens of thousands for a ten minute speech, hounded off here by a group of ignorant fools who knew nothing about the topic but made that gentleman's life a misery. Encyclopaedias like Brittanica would have offered that person a blank cheque to write for them. A small bunch of idiots drove him away.
I put Koyaanis Qatsi to this group because he had asked to be de-Adminnned. A few months after the time his request was granted, his edit totals had fallen off, so it is likely he knew he might be leaving, and so felt guilty about keeping the Admin powers.
As for those who just quietly stopped contributing, it's difficult to be certain why they left, but Camembert, who had been appointed in the first round to the Arbitration Committee (or ArbCom) states he has left Wikipedia on his user page. In contrast, DanKeshet's user page alludes to several lengthy moves across the United States, and these disruptions may have been the cause for his departure. Clearly, not all departures are due to stress.
2. Those who had taken one or more breaks from Wikipedia, but as of January, 2007 were still contributing. Taking breaks (also known as "Wikibreaks") is not uncommon; a quick glance at the number of links to the Wikibreak template shows that about 548 people are currently on declared Wikibreaks, and not all who are on Wikibreaks use that template -- or even announce the fact.
I found six fell into this group, one of whom (The Cunctator) still participates in WikiEN-l. I don't know whether one should join this group with the first, because it is clear that they felt stress (Ortolan88 explicitly mentions it as one reason for his breaks from Wikipedia), or with the fourth group below because the break allowed them to cope with the stress and return to Wikipedia.
3. Those who had moved on to other projects related to Wikipedia. This is the largest group, with nine members. One could consider this group will not offer any useful insight about burnout, since their direct participation in Wikipedia suffered due to other activities, but one could also consider this group very useful, since many of them (at least six: Anthere, Danny, Eloquence, Maveric149, Sannse and Brion Vibber) went on to assume more responsibility for running Wikipedia-related activities : two are members of the Wikimedia Foundation board, one is an employee of the Foundation, two more have volunteered their time to help it, while one is a community support person for Wikia.
4. All others. This is a catch-all group for four individuals (including myself) who do not fall into one of the other three groups, or could be considered prime candidates for burnout -- yet continue to edit as frequently now as we ever have.
I have to say that I find this the most fascinating of the four groups because, despite all else, the members of this group continue to contributes at least as many edits despite all other reasons not to. In alphabetical order:
- AxelBoldt. A college professor in Saint Paul, Minnesota; he has given at least one
presentation about Wikipedia. I know the least about him, but see below for one possible insight about him. - Ed Poor. At one time Ed was considered to be the model Wikipedian, which is remarkable given that he was a self-identified member of the Unification Church and given to very conservative political opinions; his patience and efforts to create tolerance on Wikipedia had won him many fans -- until the stress caught up with him and he suffered a meltdown. The ArbCom took away his Bureaucrat and Admin powers and he considered quitting Wikipedia -- however, Interiot's Tool shows that he is still making edits at his usual rate.
- Fred Bauder. Fred was one of the original members of the ArbCom, appointed by Jimbo Wales in late 2003, and from what I have heard from other ArbCom members is responsible for most of the work of their decisions; and his contributions to Wikipedia continue at the same level. Perhaps most of these edits are to pages related to the ArbCom -- yet he does not show the fall-off in edits that the members of the third group do.
- llywrch, or me. Probably the one person I know best in this list, the results of Interiot's Tool fails to reveal much about what I felt or what I was experiencing -- such as the ambivalent feelings I have described, nor my recent medical problems. About the only thing these statistics show is my increased number of edits, caused by upgrading to a broadband internet connection in February. On the other hand, a glance at the record of my use of Admin powers shows that I have used them rarely; only a little more often than AxelBoldt, who has no recorded blocks of other users. Perhaps like him and ericd, by avoiding the stress of enforcing policy and confronting troublemakers I have been able to continue to be a reasonably productive contributor to Wikipedia.
To return to the subject of my investigation, if there is qualifiable evidence that stress leads to burnout, I'd have to say that the evidence could be seen that way: for those who did not move away from the English Wikipedia (18 of 27), one could lump the "Quitters" with the "Wikibreakers", and show that 14 out of 18 left Wikipedia within four years, in contrast to half (3 of 6) of the non-Admin group drifting away. However, one could also argue that taking a break from Wikipedia is an effective tool, and lump the "Wikibreakers" with the "Others" (who appear to have found a way to deal with the stress, however successfully) to show that 10 out of 18 (or about half) continue to contribute to Wikipedia, the same share as with non-Admins.
Then again, losing as few as half of a group over four years seems to me a turnover rate that is worrisome. Most students in a four-year institution (whether high school or college) actually manage to finish the program, and most high school students have less motivation than the sample group I looked at. I can only hope that I've enticed someone, with a better grasp of statistics and analysis, to gather a larger sample, subject it to a better analysis and provide a better answer.
Geoff
Technorati tags: wikipedia
Labels: wikipedia
Comments:
<< Home
Geoff, did you include me among your tally? (Obviously, I'm not an admin.) But, considering Jimbo's recent lifting of the content I considered libelous on my User page, I just may be on a "reverse burnout" and coming back soon to Wikipedia. With good intentions, of course.
My study only looked at the people who posted to the WikiEN-l list in January, 2003. That means I omitted all consideration of many people who were otherwise actively contributing to Wikipedia not only during that month, but since then.
If you did not post then, you were not included in my analysis.
Geoff
If you did not post then, you were not included in my analysis.
Geoff
Whoops, I misinterpreted the date of data collection. I would most certainly NOT have been a part of that list at that time. Neat analysis, though, anyway Geoff!
Hi Geoff. Although Wikia takes up a lot of my time, I /could/ find time to edit if I wanted to. I've just found that other places (mainly Uncyclopedia) are more fun and less stressful places to edit. So you could put me in your "burnout" group really.
Post a Comment
<< Home
