Method Effects and the Meaning of Measurement

Although the idea of a method effect in psychological measurement seems intuitively straightforward - that is, it is said to occur when any characteristic of a measurement procedure contributes variance to scores beyond what is attributable to variance in the attribute of interest - much of the surrounding conceptual vocabulary remains confused. In part, these confusions can be traced to deeper confusion in the human science literature regarding the meaning of measurement. In particular, the thinking of human scientists about method effects has been shaped by (a) received wisdom regarding why method effects are problematic to begin with, and, therefore, what corrective measures are appropriate, (b) the formal and implied semantics of psychometric techniques that have been developed to model method effects, and (c) general philosophical undercurrents that have contributed to the collective understanding of psychological measurement. Notably, tensions between lines of thought that can be broadly characterized as empiricist and realist have contributed to uneven thinking surrounding the concept of a method effect. In this paper, it is argued that it may be possible to formulate an account of what method effects are that is coherent not only across different research traditions in the human sciences, but also with thinking found in other scientific disciplines; however, doing so requires a more explicit commitment to a realist position on measurement than is generally forthcoming from human scientists. By examining these issues, this paper hopes to contribute to semantic clarity regarding not just method effects, but also the meaning of measurement in psychology.