A Lack of Robustness in Robustness Checking from Auspurg and Brüderl

In a comment on our work on ideological bias in the production of research findings (Borjas and Breznau 2026), Auspurg and Brüderl argue that our findings are based on “fragile evidence” and contain a “coding error” that, when corrected, removes the statistically significant results we presented (Auspurg and Brüderl 2026). In this response we address each of their claims carefully. In our published paper, we explicitly acknowledged the limitations of our findings which are based on secondary data and a small sample. After examining Auspurg and Brüderl’s claims, we conclude that they have not presented any new evidence that warrants any correction to our conclusions. We also make a claim of our own. We applaud efforts to examine results in the social and behavioral sciences for robustness. However, we note that Ausprug and Brüderl have published reproductions and robustness checks of more than 10 studies and in every case they find that the original study is not robust. Assuming even a poor rate of robustness in the social sciences, their collective work seems to produce an extremely biased distribution of results, suggesting that their work may itself be biased by something other than checking the robustness of empirical findings.