A lot of people -- and I mean, a hell of a lot -- do research on Wikipedia. They examine everything, from how social networks form on the site, to the life-cycle of editorship, to how Wikipedia responds to hot news, to geographical bias, and more.
A lot of this research gets reported over on Wikimedia Blog, a Web site run by the Wikimedia Foundation to talk about things affecting Wikipedia, Wikimedia Commons, and more. A recent research paper reported on Wikimedia Blog looks at the roles editors see themselves fulfilling. They identified eight roles: Social networker, fact checker, substantive expert, copy editor, wiki gnomes, vandal fighter, fact updater, and Wikipedian (although this last seems undefined). They found that most editors take one up to three roles, although they did not quantify the results. (What percentage take on just one role? two roles? three roles?) They then looked at Wikipedia articles which have been copyedited. They set a start date, and assessed the article's quality. Then, six months later, they assessed the quality again. Did lots of editing by certain kinds of editor-roles mean higher article quality? Yes and no. An article which surges in quality usually has the attention of substantive experts and copy editors. But they then drop out. Fact checkers, wiki gnomes (those who focus on correct templates, categorization, use of 'bots, etc.), and fact updaters tend to swarm over an article only once the surge is over. Their changes may be just as numerous (or more so), but the quality improvement is incremental. There was no attempt to measure how these roles fit into new article creation.
I find this interesting because of my awareness of my own changing role on Wikipedia.
I joined Wikipedia on August 27, 2005. Wikipedia had been launched just four and a half years earlier, on January 15, 2001, and would not surpass 2 million articles until September 9, 2007. That makes me one of the "earliest Wikipedians". (Today, there are 5,143,325 articles in English on Wikipedia.) When I joined Wikipedia, I was primarily interested in labor union topics, and there were massive gaps in American labor articles on the site. I was largely engaged in creating articles from scratch, and only occasionally upgrading stubs. My editing style was deeply influenced by Wikipedia's own evolving standards: Initially, all one needed was to put a list of sources at the bottom of the page. Footnotes were unheard of. The vast majority of labor article holes were filled by 2008, and my attention began to shift to articles about Supreme Court cases, motion pictures, Montana, architectural history, and the District of Columbia.
Here, again, I found myself plugging holes. By now, I was encountering a lot more stub articles and improving them. The more numerous stubs was a consequence of the rapid growth in Wikipedia between 2005 and 2008, and the much higher popularity of the site. But I found that "stub improvement" was essentially creating an article from scratch, because most of the stubs I was encountering where just a few uncited sentences. I was kind of surprised to discover that Wikipedia had been completely revising how its citation system worked behind the scenes and not telling existing editors about it. I found myself making a huge change in the way I cited articles during this period (using the "ref" system) and struggling to keep up with Wikipedia's evolving Manual of Style.
By about 2011 or so, my Wikpedia role underwent a second major shift. For a time, I'd been doing article stubs or "Start"-class articles for Wikipedia. I stopped. It now seemed absolutely pointless to do those. I was coming across so many stub and Start articles which had never, ever been improved -- and they so angered me -- that I (not very consciously) decided it would be better to just do fully cited, fully researched, B-class articles than to fuck around with lower-class edits.
The effect of this has been two-fold. First, the number of edits I make to Wikipedia has fallen dramatically. I now post full-fledged articles, and don't work on them in live edit-space. Second, the quality of my articles is consistently very high. They are fully copy-edited, typo-free (usually), fully cited, and have complete templates in them when I post them. There's no much for wiki gnomes or copyeditors or fact-checkers to do on my articles.
I wonder how widespread my personal experience is among other Wiki contributors. I bet not very common...