Abstract
Previous studies have highlighted the inherent subjectivity, complexity, and challenges associated with research quality leading to fragmented findings. We identified determinants of research publication quality in terms of research activities and the use of information and communication technologies by employing an interdisciplinary approach. We conducted web-based surveys among academic scientists and applied machine learning techniques to model behaviors during and after the COVID-19 pandemic. Using model-agnostic explanations, we identified the determinants of research publication quality across 66 activity models. These models reflect the variety of behaviors among academic scientists during and after the COVID-19 pandemic. Our two-fold perspective distinguishes between research activities of academic scientists who increase research publication quality and those who maintain it. Notably, our findings reveal a diversity within activity models in shaping research publication quality. Academic institutions can apply our approach to analyze research staff behavior, stimulate activities, and ensure alignment with institutional objectives, thereby fostering individual and team complementarity.