15/21. That's only a little more than half. What's more revealing is when they actually sat the kids down and talked to them, rather than forcing them to pick one or the other.

A lot of them didn't want to pick either. And when they did pick, it mostly came down to a color preference (not a skin color preference). Again, as I said, white/blue/gold/light colors are seen as 'good' whole red/brown/black/purple/orange/dark colors are seen as 'bad'.

Has almost nothing to do with skin color itself. If it did, the kids would pick the whitest ones, not the second-whitest. The phrases enforces a choice. Color association does the rest.

A more fair test would be to have a group of a variety of skin tones (not just a binary choice), and ask "if any" are bad. If so, point out the bad ones.

Be sure to include other stereotypical and associated traits of good/evil things. Stuff like halos/horns, red skin tone, fangs, and a variety of other traits. I guarantee it'll boil down to associations. The things that are similar to stereotypical evil will be labeled as such. And things that are similar to stereotypical good will be labeled as such. With skin tone having very little to do with it.

By presenting two binary options, with the only difference being skin tone, you are forcing the decision to be about skin tone. If you naturally have to choose one, that means one will be selected against.

If the only difference was long/short hair, you'd get the exact same response.

Please tell me what the "ideal" result is. That the white doll one is picked as bad? Is that better? As I said, the phrasing implies you must choose one.

And the results show it was not unanimous in the slightest. A sample size of 21 isn't all that great either, and most seemed to have no problems with either doll, often hesitating before their answers and straining to thing of reasons.

Amusingly, they say one is bad, and their reasoning is because it does bad things. Yet, that's the definition of bad. Of course the bad one will do bad things.

Also of note, one kid couldn't associate with either doll, but strained himself to pick one and come up with a reason, simply due to the nature of the test.

Another common thing (towards the end of the video) is that "different=bad, same=good".

With the test that has 5 pictures on it, the only difference is again, skin tone. Asking "why" is going to only net responses about that. Modify them, and then ask the reasoning. "Because he has horns". "Because he has a halo" will most likely be the reasons. No skin color bias.

The only valid question/response would be "show the color adults like/don't like". Which isn't even a bias on the kid's part, it's just their observation.