Replicated successfully: When an ostensibly prosocial action also serves some private goal, the action is viewed as less virtuous than a purely selfish act

The tainted altruism effect: a successful pre-registered replication. Valerie Alcala, Kendra Johnson, Caroline Steele, Juanshu Wu, Donglai Zhang and Harold Pashler. Royal Society Open Science, January 26 2022. https://doi.org/10.1098/rsos.211152

Abstract: Newman and Cain (Newman, Cain 2014 Psychol. Sci.25, 648–655 (doi:10.1177/0956797613504785)) reported that observers view a person's choices as less ethical when that person has acted in response to both altruistic and selfish (commercial) motivations, as compared with purely selfish interests. The altruistic component reduces the observers' approval rather than raising it. This puzzling phenomenon termed the ‘tainted altruism’ effect, has attracted considerable interest but no direct replications in prior research. We report direct replications of Newman and Cain's Experiments 2 and 3, using a larger sample (n = 501) intended to be fairly representative of the US population. The results confirm the original findings in considerable detail.

4. General discussion

The replication crisis has generated concern and worry across the field of psychology for the last decade. Too few replications have been conducted to provide a broad sense of the validity of the psychological literature as a whole, but the results to date have not been terribly encouraging. Canny observers have long noted that even in the presence of the usual bias toward publishing positive effects, there can also be a ‘reverse publication bias' that might favour the publication of failures to replicate more than successful replications like the current paper. This can potentially distort the literature in the opposite direction from the classic publication bias effect [8–10]. Clearly, in order to see reality accurately, careful direct replications need to be conducted frequently (at least for influential findings) and published without regard to their outcome.

To the best of our knowledge, this study is the first direct replication of any of the experiments in Newman & Cain's [1] study on tainted altruism. Using a larger sample size and a US sample intended to be fairly representative of the US population at least with respect to age, sex and ethnicity, we readily reproduced the main results in the original studies quite closely.