Playing as a black character in a violent video game can tend to increase negative ethnic assumptions, according to a new study.
Researchers from four separate universities conducted studies with white gamers, presenting them with two "violent" video games in which one protagonist was black and the other was white.
"White participants who played a violent video game as a Black avatar displayed stronger implicit and explicit negative attitudes toward Blacks than did participants who played a violent video game as a White avatar," the report's abstract reads.
The research comes amid calls from within the games development community to promote cultural diversity in games with more non-white lead characters.
At the GDC 2014 industry event, the BioWare gameplay designer Manveer Heir was given a standing ovation at the end of his speech on the need for more representative diversity in games.
"Why should we reject stereotypes? Not only is it lazy, but it's fairly boring. We play so many games that use the same stereotypes. I get fed up with the same old story and characters in every game," he told attendees.
Heir did not specify racial diversity as a priority over other marginalised groups, such as women, nor did he imply a desire for more violent games. Yet the study suggests promoting diversity is more than a matter of proportionality.
However, the study's findings come from a small sample of 126 gamers, which may not be enough to convince peers that the results should be considered conclusive.
Additionally, the study appears to feature an exclusively white sample, thereby ignoring the impact had on people of other ethnicities. The abstract also did not cite the potential result from non-violent games, or games where non-violence is an option.
A second experiment, in which 141 test subjects played another violent video game as a Black protagonist, "displayed stronger implicit attitudes linking Blacks to weapons", the research found.
Three US universities contributed to the project (University of Michigan, Central Michigan University, and Ohio State University) along with one academic institution in The Netherlands (4VU University).
The full study can be found in the Social Psychology and Personality Science journal.
The study's methodology was not outlined in the abstract.