Game Over
Your Score: 500
Average reaction time:
Black Armed:712.92ms
Black Unarmed:835.16ms
White Armed:754.8ms
White Unarmed:719.96ms
According to this, I have a bias (by milliseconds) against African American men. But then I took it a second time, and this time, I knew that people would be bouncing all over the pages and I had an idea of what the objects looked like. Big difference - now they're about even:
Game Over
Your Score: 565
Average reaction time:
Black Armed:654.32ms
Black Unarmed:752.4ms
White Armed:663.2ms
White Unarmed:749.8ms
Frankly, they cheated, IMHO, by telling me I was "too slow" several times. What does that mean - I'm actually looking for the object, instead of basing my judgment on color? I got this a lot with another test, the one where you're supposed to associate harmless or harmful objects and black or white faces. "Too slow"?! Is it "too slow" to not shoot when I'm not sure? That would, to me, indicate a lack of bias, but I guess they don't count it because it doesn't fit their preconceptions.
What I also have an issue with is the shooting of completely innocent-looking targets. I'm going to hesitate to shoot a kneeling businessman. But it told me I took "too long". Why am I being asked to shoot people who aren't actively pointing guns at me and have innocuous expressions on their faces?
Clothes, too. A disproportionate number of the black men had oversized clothing. At least two had oversized white T shirts, which can (in some areas) be an indicator of gang activity, or at least affiliation with thug culture (even if only via popular music, culture, etc.).
In short, it's rather flawed.
I'd be curious to see this test in several forms:
a) High contrast object/background and low contrast object/background (a HUGE issue with this test)
b) All business/professional clothing and all casual/baggy clothing
c) Female characters
Your Score: 500
Average reaction time:
Black Armed:712.92ms
Black Unarmed:835.16ms
White Armed:754.8ms
White Unarmed:719.96ms
According to this, I have a bias (by milliseconds) against African American men. But then I took it a second time, and this time, I knew that people would be bouncing all over the pages and I had an idea of what the objects looked like. Big difference - now they're about even:
Game Over
Your Score: 565
Average reaction time:
Black Armed:654.32ms
Black Unarmed:752.4ms
White Armed:663.2ms
White Unarmed:749.8ms
Frankly, they cheated, IMHO, by telling me I was "too slow" several times. What does that mean - I'm actually looking for the object, instead of basing my judgment on color? I got this a lot with another test, the one where you're supposed to associate harmless or harmful objects and black or white faces. "Too slow"?! Is it "too slow" to not shoot when I'm not sure? That would, to me, indicate a lack of bias, but I guess they don't count it because it doesn't fit their preconceptions.
What I also have an issue with is the shooting of completely innocent-looking targets. I'm going to hesitate to shoot a kneeling businessman. But it told me I took "too long". Why am I being asked to shoot people who aren't actively pointing guns at me and have innocuous expressions on their faces?
Clothes, too. A disproportionate number of the black men had oversized clothing. At least two had oversized white T shirts, which can (in some areas) be an indicator of gang activity, or at least affiliation with thug culture (even if only via popular music, culture, etc.).
In short, it's rather flawed.
I'd be curious to see this test in several forms:
a) High contrast object/background and low contrast object/background (a HUGE issue with this test)
b) All business/professional clothing and all casual/baggy clothing
c) Female characters