Employees from banks in the Middle East and Asia are not more dishonest when they think about their job, according to a study recently published in the journal Nature. The new findings cast doubt on the generalizability a previous study, which suggested that business culture in the banking sector favored dishonest behavior.
The new research also highlights some problems that can arise when trying to replicate a relatively well-known study.
“The authors of the original paper posed a really interesting and important question — Does banking culture corrode ethical behaviour? They found some evidence using a novel experimental method, that, as distinct from other professions, the culture of banking did lead to dishonesty,” said study author Zoe Rahwan, a research scientist at the Max Planck Institute for Human Development.
“In addition, I had worked in finance for many years – which helped to provide access to bankers and other financial professionals such as regulators for this research. Further, I had just started my studies in behavioural science at the LSE under Prof. Fasolo (a co-author) — which provoked my interest in the ‘reproducibility crisis’ and the problem of generalizability of important findings.”
The original study, which included 208 bank employees and 133 employees outside the banking industry, randomly assigned participants to one of two experimental conditions. In the experimental group, the participants were reminded of their occupational role while those in the control group were reminded of their non-occupational role in their leisure time.
The participants then completed a coin tossing task that allowed them to increase their income by up to two hundred US dollars if they behaved dishonestly.
Bankers in the control group reported successful coin flips in about 50% of the cases. Bankers reminded of their occupational role, on the other hand, reported 58.2% successful coin flips on average, indicating that they were being dishonest about their success.
Rahwan and her colleagues conducted this experiment again with 1,282 participants from five different populations across three continents. But they found no significant difference in the bankers’ behavior depending on whether they were primed to think about their occupational role or not.
“The key message is banking culture may not necessarily have a corrosive effect on honesty, as was shown in the original study. Among our two (fortunately) larger samples, we found no evidence that banking culture invoked more dishonesty,” Rahwan told PsyPost.
“We believe that banking culture is likely affected by national norms, national banking norms, variation within the banking sector (e.g. investment banking culture differing from commercial/retail banking culture), and of course, variation in individual banking institutions’ cultures (including within same banking types).”
“Indeed, we found some tentative evidence for variation in national banking norms. For example, in the (unnamed) jurisdiction of the original study, doctors were thought to be more honest than bankers. We found no such differences in the Asia Pacific jurisdiction of our largest sample, suggesting that banking culture may not have been so problematic in that jurisdiction,” Rahwan said.
So was the original study bunk? Not necessarily. There are several reasons why the new study may have failed to replicate the previous findings — including the fact that the original study received widespread attention in the media.
“Perhaps the most significant caveat on our work, is that we faced a largely insurmountable selection bias problem in recruiting institutions and their bankers to participate in our studies. That is, we expect (and have some tentative evidence) that only banks that were confident in the soundness of their culture were willing to participate in our study,” Rahwan explained.
“Indeed, we approached 27 institutions to join our study, and only 2 joined our study. As such, it was difficult to perform a truly faithful replication of the original study, despite our best efforts, given the highly-publicised, adverse original findings for banks and bankers.”
“While this is a concern, we did indeed check in our largest sample for effects of familiarity with the original study on behaviour. While we found that ~30% of individuals reported familiarity, this did not account for the smaller and insignificant effect size found. Still, we expect that the banks that chose to join our research project were generally confident of the healthy culture of their organisations, which diminished the probability of reproducing the original finding,” Rahwan said.
“To avoid this selection bias issue, we encourage researchers who use pioneering techniques to reveal novel findings which are likely to attract a lot of publicity, to engage in efforts, where possible, at replication and/or testing of generalizability, ahead of publicizing their results.”
“Better understanding of banking and other professional cultures as it pertains to honesty and other dimensions remains an important and challenging problem for institutions and researchers. We believe there are good opportunities to develop new tools to diagnose and to positively shape organisational cultures,” Rahwan added.
The study, “Heterogeneity in banker culture and its influence on dishonesty“, was authored by Zoe Rahwan, Erez Yoeli, and Barbara Fasolo.