Breakdowns in Scientific Practices: How and Why Some Accepted Scientific Claims May Have Little Actual Support

Abstract

This paper considers the reasonableness of claims made in empirical psychological science. Drawing on validity and institutional theories, our conceptual model views research methods as institutionalized approaches to supporting the (implicit) inferential argument that is used to validate conclusions. Breakdowns occur when researchers falsely believe that a method strongly supports the inferential argument, but where little support is provided. We identify two characteristics of methods that promote breakdowns and show that these characteristics explain breakdowns of two common methods, null hypothesis significance testing and cutoffs for fit indices. Last, we discuss broadly how to reduce breakdowns in scientific practice.