Heterogeneity, not opportunism, is what organizes results from a replication of “Self-interest and fairness: self-serving choices of justice principles”

Manuscript title: Heterogeneity, not opportunism, is what organizes results from a replication of “Self-interest and fairness: self-serving choices of justice principles”
Invitation to Review date: 17.06.2026
Review submission date: 08.07.2026
Review type: original submission / revision / other

I have reviewed the revised submission of the manuscript that I had reviewed in May 2025 at the time of its initial submission. I read through the response letter and read the article again. I opened some of the materials and did not run any code. The authors addressed the points in a clustered way (1, 5, 6/9, 7, 8/12/17 …). Comment 2 was addressed in the replies to the other reviewer, comment 4 was addressed, and comment 3 was addressed in the new section 3 (conceptual framework and replication hypotheses).

I find the Open Science Transparency Table very helpful. The new justification of the replication, sample size, frozen repositories, and deviations are all very good. Statcheck found seven cases of test statistics and p-values all of which were consistent with computed p-values.

Overall, I thank the authors for addressing all of my remarks. My only remaining remark is about the language and tone: I found some parts of the text difficult to read. Specifically the last phrase of the conclusion’s first paragraph: “The aggregate data reject single linear justice principles precisely because multiple unyielding, homogeneous subpopulations are holding their ground simultaneously, demonstrating that what originally appeared to be situational flexibility was an artifact of limited statistical precision.“ Maybe that is because I am used to psychological and meta-scientific descriptions. I recommend that you make the text a bit simpler (and add an AI use statement – can say “no AI was used”).

1. I strongly recommend to tone the conclusion down (other examples: “require a fundamental reassessment“, “This structural persistence provides a definitive bridge reconciling the laboratory evidence”, “transform replications into major engines for theoretical advancement”). I would like you to address that a failed replication does not imply that the original finding was a false-positive. There are many other reasons for failed replications.
2. I suggest that, instead of interpreting significances between the original and replication findings, you conduct statistical tests. For example, tests that are now significant can be so because of the high precision. This would be possible by computing standardized effect sizes such as Cohen’s f from the F tests (see https://lakens.github.io/statistical_inferences/17-replication.html for examples). However, I understand if you prefer adhering to the SCORE success criteria.

Sincerely,
Lukas Röseler