29 October 2011

Do blogs change anything?

I was a blogger before the Internet. In 1990 while at IFPRI I set up and edited the IFPRI Research Bulletin as an internal summary of the research my colleagues were doing. The goal was to help us all find out about each other's work in an easy and quick way and to stimulate debate.

This is one of the key motivators for me to blog: when I see something interesting I want to reflect on it, share it and listen to what others think about it.

But are blogs anything more than mere vanity projects?

Back in August, David McKenzie and Berk Ozler released a paper "The Impact of Economics Blogs". They pose 4 questions of economics blogs:

1. Do they affect the dissemination of economics research? (Yes.)

They note that RePEc (Research Papers in Economics--a collaborative effort of hundreds of volunteers in 75 countries to enhance the dissemination of research in economics) working paper downloads increase by 20-30 fold after a paper has been mentioned on a popular blog. More formally they regress abstract views on papers mentioned in the top 50 blogs and allow for lags and reverse causality and find big impacts of blog mentions on views.

2. Do they affect the reputation of their creators? (Yes.)

Here they use a list of most admired economists (derived from a poll of US academics) and combine that with the top 500 economists in terms of RePEc downloads. The run a probit regression (1=on admired list, 0 not) and try to explain that with variables such as whether the economists regularly blog and where they rank in the RePEc download ratings. It turns out that those who blog regularly are more admired than their RePEc ratings would suggest (although people are surely admired for more than their writing--whether articles or blogs!).

3. Do they change attitudes of readers or lead to their increased knowledge? (Yes.)

Here the authors used the launch of their own blog (Development Impact) on April 2011 to randomise encouragement to read the blog among other researchers. They did a baseline and follow up. As they note their study design has good internal validity (i.e. it is good at assessing whether their blog has an impact) but that they can say less about other blogs (less strong external validity) although they argue that their blog is not atypical of other blogs (although it is a World Bank blog).

They find that those encouraged to read the blog (the treatment group) were more likely to be interested in working as a researcher at the World Bank, had a more positive perception of the World Bank's research quality, were more aware of the authors of the blog, relaxed their perceptions that World Bank staff face censorship over blogs and changed their opinions on the effectiveness of different interventions.

4. Do they influence policy? (Don't know, probably yes)

This is the part of the paper that has the weakest evidence base and is essentially a search for stories which cannot be verified. The bloggers interviewed cannot put their finger on specific policies changed as a result of their blogs, but then again that is not how policy works. A better strategy would be to see how widely policymakers read blogs. But the strong suspicion has to be if policymakers are influenced by research, and if the blogs are any good, the blogs should enhance the likelihood of research being influential).

So, blogging seems to matter. I liked this paper because:

* of the creative use of mixed methods

* it is careful and well done

* it shows that randomised experiments don't have to be intrusive or expensive

* asks an important question in this blogoshere world and shows us that we can actually do experimental research on how policy processes use information and evidence

* it generates research questions for others to pursue

I would have liked to have seen more on whether blogs lead to less diversity in who is talking to who (I suspect the bonding impact amongst the like minded is quite high and the bridging factor across silos less strong), who has the luxury of being able to blog (e.g. institutional support) and how blogs change power dynamics in terms of whose voice is amplified?

The authors end up by asking, if there are all these benefits, why aren't there more bloggers? They suggest that the supply is lower than it should be because the barriers to doing it are large (emotional, time, writing skill etc.).

Now for the paper on Tweets.....


Al Borde del Caos said...

Thanks for the post Lawrence. I was wondering if you know of similar studies done in other fields (sociology, anthropology, development studies) and not only in economics. It would be nice to know... On the other hand, I have just read a post where the question "Why don't scientists share?" is made, and can be usefully linked with your article.

Calestous Juma said...

Blogging has been going on for a long time and predates writing. It can be traced back to grunting. I used to write regular letters to the editors of Kenyan newspapers in the early 1970s and build up a large readership. I now think of it as an early form of blogging but not a significant improvement over the audio blogging from our ancestors. Much of today's blogging seems to belong to the same cave dwellers' genre. Your IFPRI scribbling had more content and foresight.

Berk Ozler said...

Hi Lawrence,

Thanks. I like the penultimate paragraph that describes what more you would have liked to see in such a paper: all three are interesting questions and while I am not sure how well we can address them empirically, some more discussion of these points may well be warranted in a revised version. Sincerely,


Rita Summers said...

I would like to say that your article is awesome. You have good command
macroeconomics term paper