Abstract

The biomedical, life, and social (BLS) sciences are experiencing a rapid development of data analysis techniques. Meanwhile, there is growing worry that students are not being sufficiently prepared for modern research through quantitative techniques teaching. Demands for changes to undergraduate and graduate quantitative research method courses have been sparked by these trends. We contend that such reform should be founded on data-driven insights concerning the application of analytical techniques within and across disciplines. About 1.3 million publically accessible research articles were examined as part of our assessment of the peer-reviewed literature to track the interdisciplinary mentions of analytical methodologies over the previous ten years. In order to detect trends in analytic method mentions shared across disciplines as well as those specific to each discipline, we used data-driven text mining analyses to the “Methods” and “Results” sections of a significant subset of this corpus. We discovered that the most frequently mentioned statistical techniques in research articles in the fields of biomedicine, life science, and social science are the t test, analysis of variance (ANOVA), linear regression, and chi-squared test. Between 2009 and 2020, however, the proportion of published literature that mentioned these techniques fell. On the other hand, the overall share of scientific publications has significantly increased for multivariate statistical and machine learning approaches, including artificial neural networks (ANNs). Additionally, we discovered distinct clusters of analytical techniques related to each BLS research field, such as the application of structural equation modelling (SEM) in psychology, survival models in cancer, and multiple learning in ecology. We talk about how these results affect research techniques and statistics education, as well as disciplinary and inter-disciplinary collaboration.