Evan Weingarten

Who they are

He is a marketing researcher. The meta-analysis is held in the University of Pennsylvania's marketing papers collection, and the University of Pennsylvania Wharton School's faculty directory carries profiles for both Weingarten and his co-author Hutchinson. The vault holds no further biography and no confirmed current post, and none is claimed here.

The paper sits at the border between social cognition and consumer research. The ease-of-retrieval effect began as a social psychology result about how people rate themselves, and it has since been taken up in consumer research, the field in which this meta-analysis was written. A meta-analysis that sorts the evidence is useful to both fields because each tends to cite the original 1991 experiment as though the question were closed.

Key papers

Weingarten, E., & Hutchinson, J. W. (2018). Does ease mediate the ease-of-retrieval effect? A meta-analysis. Psychological Bulletin.

The analysis covers 142 papers, 263 studies and 582 effect sizes. Its abstract states four findings. On average the standard few-versus-many manipulation exhibits a medium-sized effect. In standard-paradigm conditions about one third to one half of the total effect is mediated by subjective ease, which supports the standard explanation but suggests that other mediators are present. There is evidence of publication bias that reduces the standard effect by up to one third. And moderators that fall outside the standard design produce smaller, often reversed, effects, while several theory-based moderators produce strong effects that match the theory.

Key ideas

The paper's title asks a mechanism question. Does ease mediate the effect? The answer is partly. The assumption underlying a generation of availability research was that the felt difficulty of retrieval carries the signal. The meta-analysis finds that it carries some of it, and that other processes carry the rest. This is the reason this publication describes the ease-of-retrieval effect as a modest regularity and not a law.

The publication bias finding matters in the same direction. The studies that reach print tend to be the ones that found the effect, so the pooled average overstates it. Correcting for that lowers the estimate by up to a third, in the authors' words. The estimate that remains is still positive and medium-sized.

The moderators result is the most useful for practice. Effects outside the standard few-versus-many design are smaller and sometimes reversed. A communicator who borrows the finding to design, for example, a review request or a list of reasons should expect it to be fragile.

The paper has also drawn a critical response. A blog reanalysis by Ulrich Schimmack in 2021 argues that the pooled estimate depends heavily on a handful of very small studies and that weighting by sample size brings the bias-corrected estimate near zero. That reanalysis is not peer reviewed and this page has not verified its numbers, so it is noted as a dissent only.

---