NEW YORK (AP) — Scientists say they were unable to confirm a highly publicized 2014 study that suggested banking culture can promote dishonesty.

New lab experiments with bankers in the Middle East and Asia Pacific turned out differently, maybe indicating that the original finding doesn’t apply worldwide, researchers reported Wednesday in the journal Nature.

In response, the authors of the original work pointed to several differences between the two studies and stressed that they never intended to imply that all banks have a problematic business culture.

Instead, the point of the 2014 study is that such a culture could be part of the reason for episodes of fraudulent or unethical behavior, Alain Cohn and colleagues of the University of Zurich wrote in Nature.

In fact, the original study focused on employees of a large international bank that was involved in several high-profile legal disputes involving “problematic business practices,” they wrote. They did not identify the bank or its location because they had promised confidentiality, but said it was involved in trading, asset management, investment banking and private banking.

The original work, covered by The Associated Press and other news agencies, subjected 128 employees of that bank to a laboratory test of honesty. Each repeatedly flipped a coin and reported the results, knowing that certain outcomes would earn them money. The question was how often they’d lie about the results.

Participants who had just been asked about their work at the bank showed evidence of cheating, while other bank employees, who hadn’t been asked about their work, did not. That indicates the employees were basically honest, but that thinking about their jobs triggered dishonesty. That suggests that banking culture favors dishonest behavior, Cohn and his colleagues concluded in 2014.

The new study repeated that experiment with 620 commercial bankers from a large institution in the Asia Pacific region and 148 commercial bankers from a medium-sized institution in the Middle East. It found no statistically significant effect of asking employees about their work ahead of the coin-flipping.

Zoe Rahwan of the Max Planck Institute for Human Development in Berlin and co-authors of the new study suggested several possible explanations for the disparate results.

One is that the banking culture sampled in the original study is different in other places, they said. Another is that only “ethical” banks let their employees participate in the new study, following widespread press coverage of the 2014 results. The new work was done in 2015 and 2016.

Cohn and colleagues agreed that banks with history of misconduct were probably less likely to let their employees participate in the new work, resulting in a sample different from the one they tested.

That and some procedural differences between the studies — one studied a bank involved in trading and investment, the other studied a commercial bank, for example — suggest the new work can’t address whether the 2014 study applies to different institutions or countries, they said.

___

The Associated Press Health and Science Department receives support from the Howard Hughes Medical Institute’s Department of Science Education. The AP is solely responsible for all content.