A national study of medical residents training to become emergency department doctors found that while men and women started on equal footing, by the end of their third and final year of training, the men on average scored higher than the women in all evaluation areas.

Researchers from the University of Pennsylvania and the University of Chicago medical schools suspect the culprit may be gender biases -- possibly unwitting -- in the medical training process.

"We think it's likely that the gender differences in our study are at least partially driven by systemic, implicit bias," according to a statement by study colead authors Daniel O'Connor, a Penn medical student, and Arjun Dayal, a University of Chicago medical student.

In the study, published in JAMA Internal Medicine on Monday, men and women started with similar skills and knowledge, according to the authors.

During the first year of emergency medical training, women on average scored somewhat higher than the men in 15 of 23 areas, which included areas such as diagnosis, multitasking, and accountability. By the end of training, the men on average outscored the women in all 23 evaluation categories.

"Such a uniform trend, with comparable evaluations at the beginning of residency, strongly implicates bias rather than diminished competency or skill," the authors said.

Past studies have suggested the possibility of gender bias in medical training, they noted.

The new study included 33,456 direct-observation evaluations for 359 emergency medical residents at eight sites in the United States. The residents were evaluated by 285 faculty members from July 1, 2013, to July 1, 2015. The scores given by female evaluators were similar to those given by the men, according to the authors.

While senior residents might be expected to display assertiveness and independence, the authors noted that female residents were sometimes penalized for displaying those traits, the authors noted.

For example, O'Connor said, one woman was called on to be "more assertive" in one of her evaluations. However, the same resident was criticized in another evaluation for arguing with the attending physician, being very confident in her diagnosis, and having "a hard time entertaining other possibilities."

Another woman was criticized for not being independent enough, he noted. But the same woman was also faulted by an evaluator for a "negative attitude toward supervision."

O'Connor, in an interview, said a follow-up study is awaiting publication and others were anticipated to further investigate the presence of bias in the medical training evaluation process.

"We hope that further research … will help us better understand the mechanisms behind these trends so that effective interventions can be designed and implemented to promote gender equity in medicine," the authors said.