The “gun embodiment effect,” or the likelihood that someone holding a gun perceives that another person also has one, may explain police shootings of unarmed individuals, a study from Colorado State University researchers suggests.

The findings of CSU psychology faculty Jessica K. Witt and Jamie E. Parne, along with Nathan L. Tenhundfeld of the University of Alabama, appear in Cognitive Research: Principles and Implications. In their experiment, 212 students at CSU stood five feet away from a television screen, holding a pretend gun or a spatula. On the screen were movies of a masked individual holding a gun or a shoe.

Researchers calculated “false alarm” rates indicating when the participant believed there was a gun when it was actually the shoe. Those participants holding a spatula were slightly more likely to correctly see the shoe than those pointing the fake gun.

“But if you have this small effect, and put it on a national scale, and you talk about how many people have guns in this country, even these small effects are important,” Witt told the campus news service. “For example, if 100 officers wielding guns interact with 10 unarmed people a day for 100 days, in these 100,000 interactions, our data suggest [there] will be 1,000 misperceptions of an unarmed person as holding a gun.”

The study also found a slower reaction time for people holding guns versus spatulas.

“In other words, performance at detecting a neutral object suffered when holding a gun compared with holding a spatula,” the authors concluded.

Witt made a similar discovery nearly a decade ago, and the CSU study attempted to replicate those findings with a larger sample size.