The Question

Has psychological science made genuine empirical progress over the past 66 years? Smedslund and colleagues (2022) examined this question through a systematic analysis of 1,565 articles published in PsycINFO between 1956 and 2022, focusing on a single metric: explained variance.

What They Found

The average explained variance across all articles was 42.8 percent, and this figure remained constant throughout the entire 66-year period (Smedslund et al., 2022). Psychology, by this measure, has not progressed: studies today explain the same proportion of variance as studies from the 1950s.

The authors argue that this is not accidental. Using latent semantic analysis on a sample of 50 studies, they demonstrate that the factor structures of the constructs involved can be replicated through analysis of the semantic relationships among the items, without any empirical data at all. The statistical clustering that psychological research produces may reflect the semantic structure of the language used to describe psychological phenomena, rather than the discovery of genuine empirical regularities (Smedslund et al., 2022).

The proposed explanation is that researchers select hypotheses with approximately 40 percent semantic truth value: claims that are plausible and interesting but not so obviously true or obviously false as to seem trivially confirmable or unworthy of investigation. Hypotheses at this level of semantic plausibility naturally produce results clustering around 40 percent explained variance through the logic of factor analysis and related methods, regardless of empirical discovery (Smedslund et al., 2022).

Why It Matters

The implication is profound: a substantial proportion of psychological research may be rediscovering what is already implicit in the language used to describe psychological phenomena, rather than learning something new about how the mind works. If findings could largely have been anticipated from semantic analysis alone, then data collection, statistical testing, and the entire apparatus of empirical psychology may in many cases be confirming the obvious while appearing to discover the new (Smedslund et al., 2022).

Reference

Smedslund, G., Arnulf, J. K., & Smedslund, J. (2022). Is psychological science progressing? Explained variance in PsycINFO articles during the period 1956 to 2022. Frontiers in Psychology, 13, 1–18. https://www.frontiersin.org/articles/10.3389/fpsyg.2022.1089089