The use of DNA ancestry kits like 23andMe and AncestryDNA is growing rapidly, with 3 million tests sold worldwide. Just like Oprah, Prince Williams and numerous celebrities on PBS’s Finding Your Roots, now you can get a to-the-decimal-place breakdown of how your DNA overlaps with all the major “racial” categories. All it takes is less than $200, a saliva sample, and a six-week wait.

But recent research of ours, published in this month's Personality and Social Psychology Bulletin, suggests that the results you get can have drastic consequences for your degree of racism and hostility to others. 

Whereas groups’ DNA differences were emphasized in both the Rwandan and Bosnian Genocide, peace initiatives have highlighted ethnic groups’ commonalities in an effort to foster peace. This led us to ask: what are the actual effects of finding out that your DNA is either genetically similar or different from another population?

In our first experiment, Jewish and Arab participants read a news article citing a scientific study that either reported high genetic similarities or high genetic differences between both groups. Shortly after, we measured their racist stereotypes towards each other and found that which news article participants read drastically affected their attitudes: when they learned about the genetic differences, rather than the similarities, they characterized the other group as more violent, unfriendly and mean.

In a separate experiment, we tested whether our findings could also affect actual hostile behavior. Here, we led Jewish participants to believe that they were playing a simple computerized game with an Arab opponent sitting in another room. If the Jewish participant won, they could give their opponent a loud blast of noise – up to the intensity of a fire alarm. Strikingly, Jewish participants who had first learned about the genetic differences “punished” their alleged Arab opponent with more intense noise blasts than those who had learned about the genetic similarities.

But can learning about genetic similarities or differences also alter peoples' support for war?  To test this, in a third experiment—run outside the laboratory—we randomly assigned Jewish participants to read one of our various news articles and then rate their support for peacemaking with Palestinians. Here, our results suggested that learning about genetic similarities might be an effective intervention for reducing conflict.

However, when we finally took the study to Israel—a context of ongoing violence and deeply entrenched negative views—we found something quite different. Here, learning about the genetic differences was what was really impactful. In this field experiment conducted on Israeli commuter trains, Jewish Israeli’s supported violence and war-like policies towards Palestinians much more after reading about their genetic differences with Arabs.

Because it is extremely common to get DNA ancestry test results that show zero overlap with other ancestry populations—populations that many of us already hold negative views about—and because genetic differences are also emphasized in genocidal rhetoric, this finding is especially alarming.

We suggest that DNA ancestry services remind us that our ancestry results are actually based on much less than 0.1 percent of our genes. We also suggest that organizations like International Crisis Group and Genocide Watch pay particular attention when propaganda highlights warring groups’ genetic dissimilarities. The popular media should also be highly cautious when reporting on groups’ genetics. News and magazine articles are frequently reporting on the degree of DNA overlap between groups with a history of conflict—Hutu and Tutsi, Jews and Arabs, White Europeans and Roma, Russians and Ukrainians, English and Irish—yet rarely make clear that there is in fact no genetic basis for race.

Until then, when encountering information about how our DNA is different from other populations, we must remind ourselves that these variations are in fact minuscule. If we fail to, it can have drastic consequences.