Using Evidence Typologies to Evaluate the Evidential Value of Published Sociological Research

• In Event: 30821 - Open Refereed Roundtables
  In Refereed Roundtable (90 minute): Table 12

Abstract

Most sociologists agree that scientific claims should be supported by a robust body of empirical evidence before they are (provisionally) accepted as true. But standards for what counts as good evidence vary widely, meaning that claims that rest on very different evidentiary bases—and potentially reflect empirical reality to widely differing degrees—can nonetheless all be presented as being “strongly supported by the evidence” and work their way into established sociological knowledge. I argue that we can make significant progress in addressing this issue by adopting the practice of evidence grading that is currently used in the medical and (some) applied behavioral sciences. Like pre-registration (another adoption from medicine), evidence grading is a methodological tool that helps researchers apply best-practice principles in a systematic way. Evidence grading provides guidelines that can help researchers both clarify the claims being made in studies and rigorously evaluate the quality of evidence for those claims based on statistical principles known to produce accurate, replicable results. I illustrate the utility of evidence grading by assessing the quality of the evidence supporting the claim that personal values affect voting for left/right political parties.

Author

• Andrew Miles, University of Toronto