Social scientists are increasingly publishing in academic journals rather than books, suggesting a weakening of critical cross-disciplinary collaborations, a decade-long assessment of US research universities has concluded.

Book publications across 12 social sciences disciplines fell by between 34 per cent and 54 per cent during the 2010s, while journal article publications grew by as much as 64 per cent, according to data compiled by the research arm of the consulting firm Academic Analytics.

The findings??in the journal?Plos One?and presented as evidence that social scientists ¨C as they embrace the analytical tools of the hard sciences ¨C may be losing valuable connections to the theoretical underpinnings of the humanities.

¡°We are witnessing the marginalisation or minoritisation of humanities-type research in the social sciences,¡± said one co-author of the analysis, Anthony Olejniczak, director of the Academic Analytics Research Center.

ÁñÁ«ÊÓƵ

The study touches on a long-running and?contentious debate,?largely seen?as pitting social scientists who increasingly bring advanced statistical tools to their assessments, against academics who regard human beings as too behaviourally varied and inconsistent to allow for rigorous data-based conclusions and predictions.

Dr Olejniczak said that neither he nor his study took direct sides in that debate. But the growing preference?for academic journals among social scientists, he said, could mean that their statistics-heavy analyses are increasingly being developed without the theoretical underpinnings provided by more traditional understandings of the humanities.

ÁñÁ«ÊÓƵ

¡°The way that social scientists communicate is becoming more like those who are using technological solutions to larger problems,¡± Dr Olejniczak said. ¡°Books are not just very long journal articles, in a nutshell, and different ways of knowing require different ways of telling,¡± he said.

The study was based on data from the 2010s at 280 US research universities. It covered work from more than 1,500 academic departments in 12 social sciences disciplines: anthropology; criminal justice and criminology; economics; educational psychology; geography; international affairs and development; political science; psychology; public administration; public policy; social work and social welfare; and sociology. The study¡¯s comparisons involved peer-reviewed journal articles found in the CrossRef database and academic book publications listed by the distribution company Baker & Taylor.

Statistics-laden social science research has shown?significant?growth in the past couple of decades,?probably because of the increased availability of large data sets of human activity, Dr Olejniczak said.

The heavy shift from books to journal articles appears to be strong indirect proof that shift, as data-based research ¡°lends itself to hypothesis-driven article format publications¡±, he said. ¡°Books, on the other hand, tend to be transdisciplinary.¡±

ÁñÁ«ÊÓƵ

A switch by researchers toward fewer books and more journal articles also could save universities money, said Diana Hicks, a professor of public policy at the Georgia Institute of Technology who studies the data underlying science and technology policy. University libraries can¡¯t afford to buy every book from an academic publisher, Professor Hicks said. ¡°If books are narrowed down to authors who really need the depth available in the longer form,¡± she said, ¡°that would be a good thing, I think.¡±

paul.basken@timeshighereducation.com