1175 words • 5~8 min read

Science blogs and public engagement with science: across the disciplinary divide

Debate over a communications study investigating science blogs and public engagement has recently engaged some of the research subjects and their friends to review the paper, which they’re totally entitled to do. Except that they, as traditional natural scientists, may not have the tools at hand to do justice to such a review. Dr. Isis’ response to the article may have been a bit dramatic but not entirely off the mark.  The study was well-grounded within social studies of science theory, but perhaps not executed or written like a seasoned researcher, which I suspect the author’s not.  So, beyond the fold, my social science perspective on the paper…

The classic image of science participation, citizen science, courtesy birds.cornell.edu

First, kudos to Inna for a good introduction, grounding the study in appropriate literature within science and technology studies, specifically articles discussing the democratization of science and types of knowledge in science creation.  She proposes to add to this literature by investigating blog comments as a means of adding to scientific discourse and discovery.  Whether she accomplishes this goal is debatable, as she later reports that all forms of participation (contribution, digression, expression of attitude or emotion, or call to action) were present in the comments, so it was hard to draw any conclusions about the role of commenters in scientific production.

where grounded theory fits into the grand scheme of research design, thanks informationr.net

What most of the responses thus far have hinted at is the methods chosen for the study.  I can vouch for the fact that they are standard social science qualitative methods.  Most of us have spent countless hours coding as described by the author’s article, generally aided by a computer program called NVIVO.  She used a method of coding commonly referred to as ‘grounded theory’ (yes, we like to make up jargony names for absolutely everything) in which the researcher doesn’t start analyzing with a prescribed set of codes, but instead allows the ideas and concepts that are most prevalent to emerge through the research process.  Ok, all kosher so far.

Where I start to have a problem with her methods begins with how she chose her study blogs – a Google search on “science blogs” and “blogs about science”.  First, this biases the sample toward the science blog network, but she probably didn’t know that prior to work on the study.

Google, a powerful tool but to be used with care

Also, she didn’t specify whether she was attempting to get a representation of widely read and accepted science blogs or a broader sample to cover the spectrum of self-proclaimed science writing in the blogosphere.  If the former, the Google search would have had to been more prescribed and included some method of random sampling (such as taking every 10th result).  If the latter, then the author would have been far better off contacting some blog authors and asking for recommendations of well-respected science blogs or reading the blogrolls and identifying commonly read blogs (this is tantamount to finding key informants, a strategy commonly employed by exploratory social science research).  In this case, the sample size of 11 isn’t really small because all are good representations and you can delve deeper into the content of each blog.

From the sample of 11, first, they aren’t all blogs (Panda’s Thumb) and aren’t all active (BioEthics) but she didn’t discount them because of that.  Also, in order to even out the amount of material analyzed, she took a longer sample (30 days) of less active blogs and short sample (5 days) of more active blogs, meaning that she probably got more content-driven discourse out of the more active blogs covering one or two exciting posts rather than a more representative sample over time.

How different are those first few commenters?

In addition, she only included the first 15 comments in yet another effort to trim the amount of material she had to code and analyze.  I can sympathize with this notion, but question the implications of focusing on the first 15 comments – would that tilt the analysis toward the regular readers of the blog who first read the article?  Would this really matter, ie. are those people different than the ones who join the discussion later?  It’s really a whole different research question, and perhaps one that badly needs to be answered.

In the end, the biggest question asked of any social science study is ‘how is this applicable to a larger scale generalization’?  Some in the field call it big T Theory.  Others are quick to point out that case studies are just that… case studies – and not uniformly scalable to larger conclusions.  I would say that this is the case for the study in question.  The results are certainly accurate for the 11 blogs in focus and can at least be informative to the authors of those studies when considering their readership.

Science sometimes need a bit of discussion to bring everyone into the fold and maybe even create new and better ideas

The conclusions that the comments and discussions are not that different from traditional print media in fostering scientific knowledge creation is also potentially true, on average.  I know a lot of great science bloggers out there, but I also can name quite a few that aren’t so stunning and even more that don’t foster much discussion.  So I’m not exactly surprised by her results and I certainly don’t disagree with them.  In fact, she also identifies most of the readers as people with a scientific background given the content and tone of their comments, which then begs the question why don’t we have more thoughtful discussions?  She’s totally right in concluding that the challenge is out there to make better use of the new tool of the blogosphere to foster scientific thinking.

So quit quibbling over the disciplinary divides regarding methodology, take her conclusions to heart, and get commenting.

~Bluegrass Blue Crab