Alejandra Lopez

Who they are

Her team did the thing this publication says should be done with a compromised result: they went and earned it again, in the open, with eight times the sample, and reported it smaller.

Two terms are worth glossing. A direct replication repeats the original procedure as closely as possible, rather than testing the idea in some new setting. Preregistered means the design, the sample size and the analysis were filed publicly before any data were collected, so the authors could not quietly choose the version of the analysis that worked best. Both are what make her paper usable where the original is not. See Preregistration.

The study she replicated came from Brian Wansink's laboratory; that page records why the original is not used as evidence here. The senior author on her paper is A. Janet Tomiyama, who appears elsewhere on this site as a co-author on Mann and colleagues (2007) at Set Point Theory, though she has no page here yet.

Key papers

Lopez, A., Choi, A. K., Dellawar, N. C., Cullen, B. C., Avila Contreras, S., Rosenfeld, D. L., & Tomiyama, A. J. (2024). "Visual cues and food intake: a preregistered replication of Wansink et al. (2005)", Journal of Experimental Psychology: General, 153(2), 275 to 281. First author. Four hundred and sixty-four participants. The effect held at about half the original size, and people eating from the refilling bowl again did not believe they had eaten more than the others.

Key ideas

A compromised finding is not automatically a false one. The original bottomless-bowl paper cannot be cited, because the laboratory it came from cannot vouch for its data. The underlying effect nonetheless survives, at a smaller size, because somebody tested it properly.

The gap between what people ate and what people believed they ate is the durable part. That is the result her replication preserves, and it is why the finding belongs next to Meal Memory rather than to any claim about willpower.

Bigger sample, smaller effect, is the normal shape of a good replication. Early studies with small samples tend to report larger effects than later ones with large samples. See The Replication Crisis.

Concepts on this site

Research Misconduct, the rule her team's work resolves in practice

Meal Memory and The Portion Size Effect, the mechanisms her result speaks to

Preregistration

The Replication Crisis

YMYL, the handling rule for this subject

---