Meta-analysis, ideals of objectivity, and the reliability of medical knowledge

Abstract

This paper focuses on the ideals of scientific objectivity as they emerge in discussions concerning meta-analyses and medical research. Stegenga (2011) has argued that meta-analyses fail to be objective because conducting them involves making judgments. I show that his reasoning is based on the so-called procedural ideal of objectivity, which can be questioned: This ideal is unattainable and does not capture some of the problematic issues of medical research. By introducing a case in research on selective serotonin reuptake inhibitors, I demonstrate why the so-called social view on objectivity succeeds better in accommodating 1) the way in which scientific research necessarily involves judgments, 2) the possible risks involved in research, and 3) the influence that the institutional context has on research activities. Adopting this ideal of objectivity helps us better appreciate the virtues of meta-analyses and pinpoint which practices threaten the reliability of meta-analyses' results.