I was recently quoted in an article for Science News about the relationship between academia and Wikipedia by Bethany Brookshire. I was asked to comment on a recent paper by MIT Sloan‘s Neil Thompson and Douglas Hanley who investigated the relationship between Wikipedia articles and scientific papers using examples from chemistry and econometrics. There are a bunch of studies on a similar topic (if you’re interested, here is a good place to start) and I’ve been working on this topic – but from a very different angle – for a qualitative study to be published soon. I thought I would share my answers to the interview questions here since many of them are questions that friends and colleagues ask regularly about citing Wikipedia articles and about quality issues on Wikipedia.

Have you ever edited Wikipedia articles?  What do you think of the process?

Some, yes. Being a successful editor on English Wikipedia is a complicated process, particularly if you’re writing about topics that are either controversial or outside the purview of the majority of Western editors. Editing is complicated not only because it is technical (even with the excellent new tools that have been developed to support editing without having to learn wiki markup) – most of the complications come with knowing the norms, the rules and the power dynamics at play.

You’ve worked previously with Wikipedia on things like verification practices. What are the verification practices currently?

That’s a big question 🙂 Verification practices involve a complicated set of norms, rules and technologies. Editors may (or may not) verify their statements by checking sources, but the power of Wikipedia’s claim-making practice lies in the norms of questioning  unsourced claims using the “citation needed” tag and by any other editor being able to remove claims that they believe to be incorrect. This, of course, does not guarantee that every claim on Wikipedia is factually correct, but it does enable the dynamic labelling of unverified claims and the ability to set verification tasks in an iterative fashion.

Many people in academia view Wikipedia as an unreliable source and do not encourage students to use it. What do you think of this?

Academic use of sources is a very contextual practice. We refer to sources in our own papers and publications not only when we are supporting the claims they contain, but also when we dispute them. That’s the first point: even if Wikipedia was generally unreliable, that is not a good reason for denying its use. The second point is that Wikipedia can be a very reliable source for particular types of information. Affirming the claims made in a particular article, if that was our goal in using it, would require verifying the information that we are reinforcing through citation and in citing the particular version (the “oldid” in Wikipedia terms) that we are referring to. Wikipedia can be used very soundly by academics and students – we just need to do so carefully and with an understanding of the context of citation – something we should be doing generally, not only on Wikipedia.

You work in a highly social media savvy field, what is the general attitude of your colleagues toward Wikipedia as a research resource? Do you think it differs from the attitudes of other academics?

I would say that Wikipedia is widely recognized by academics, including those of my colleagues who don’t specifically conduct Wikipedia research, as a source that is fine to visit but not to cite.

What did you think of this particular paper overall?

I thought that it was a really good paper. Excellent research design and very solid analysis. The only weakness, I would argue, would be that there are quite different results for chemistry and econometrics and that those differences aren’t adequately accounted for. More on that below.

The authors were attempting a causational study by adding Wikipedia articles (while leaving some written but unadded) and looking at how the phrases translated to the scientific literature six months later. Is this a long enough period of time?

This seems to be an appropriate amount of time to study, but there are probably quite important differences between fields of study that might influence results. The volume of publication (social scientists and humanities scholars tend to produce much lower volumes of publications and publications thus tend to be extended over time than natural science and engineering subjects, for example), the volume of explanatory or definitional material in publications (requiring greater use of the literature), the extent to which academics in the particular field consult and contribute to Wikipedia – all might affect how different fields of study influence and are influenced by Wikipedia articles.

Do you think the authors achieved evidence of causation here?

Yes. But again, causation in a single field i.e. chemistry.

It is important to know whether Wikipedia is influencing the scientific literature? Why or why not?

Yes. It is important to know whether Wikipedia is influencing scientific literature – particularly because we need to know where power to influence knowledge is located (in order to ensure that it is being fairly governed and maintained for the development of accurate and unbiased public knowledge).

Do you think papers like this will impact how scientists view and use Wikipedia?

As far as I know, this is the first paper that attributes a strong link between what is on Wikipedia and the development of science. I am sure that it will influence how scientists and other academic view and use Wikipedia – particularly in driving initiatives where scientists contribute to Wikipedia either directly or via initiatives such as PLoS’s Topic Pages.

Is there anything especially important to emphasize?

The most important thing is to emphasize the differences between fields that I think needs to be better explained. I definitely think that certain types of academic research are more in line with Wikipedia’s way of working, forms and styles of publication and epistemology and that it will not have the same influence on other fields.