Edmund Malesky coauthors reflections on null results
Monday, September 19, 2016
In the academy and well beyond, the problem of null results has become quite significant. Indeed, discussions of null results have made their way as far as TV commentator John Oliver’s recent discussion of science in which he poignantly notes that people generally do not like to hear about null results. And yet, maybe we would all be better off – with more money in the bank – if his headline “Nothing Up With Acai Berries” actually made it to the general public and we embraced it (see NIH).
This is not just a problem in health, the sciences in general struggle with how to engage null results. Social scientists are no exception. Our interest in the topic led to a special issue of Comparative Political Studies, a leading political science journal, in which we solicited results free submissions and conducted the entire review process with all decisions made in the absence of results. (See our original call for papers). This meant that reviewers and editors could not condition their decisions on significant results, thereby allowing a greater opportunity for null results to end up in published manuscripts. This special issue has been accepted and three results free papers will be published along with our introduction.
The exercise demonstrated in practice that papers with developed theories and designs, but that ultimately ended up with null results, could make it through review and into print. One of the three published papers documented statistically insignificant treatment effects in its main experimental interventions. This in itself was a huge success of the special issue—revealing that research may find null results, and allowing readers to learn from them. But the process was also quite instructive on the challenges of evaluating work where null results have a greater probability of being published. We found the authors’, and especially the reviewers’, comments on this process illuminating.
We offer two interrelated suggestions from our pilot to help make null results more prominent in peer-reviewed publications. The first has to do with acclimating reviewers to a new way of thinking about null findings—that they may be meaningful theoretically. The second has to do with helping authors frame their prospective work, so that null results can be read as meaningful contributions.