First a thought on the visualization exercise: being able to quickly visualize a generic individual for a given phrase suggests bias (why'd you assign those traits?). But, I'm hesitant to say the same about visualizing a specific individual. Sure, bias likely plays a role in all my interactions, but the simple fact that I consider X person "annoying," and think of them when so prompted doesn't seem to suggest bias. (The fact that I unconsciously assume that anyone who reminds me of X is also annoying shows one way that bias forms, but that idea didn't really come across in the intro video)
I've taken many IAT tests over the years, and my general observation is that if I stumble, I tend to take a while to recover my flow. As a result, I occasionally get "strong preference" results that I ascribe mostly to chance. Today, for example, I got a "Strong preference for light-skinned" and a "moderate preference for African Americans"
I found the "Native IAT" odd. I didn't start with any association between the pictures and their concepts, so I was pretty slow and not very accurate in the first few rounds, but got progressively better. It didn't so much feel that I was making Foreign vs American associations so much as learning an arbitrary categorization. I'd prefer it just have the locations (so I could read them better). But, I only got a "slight preference"...
Similarly some parts of the "Asian IAT" - there were a couple locations that I had switched, and some faces I had difficulty categorizing. I ended up with a moderate Asian=American preference.
Overall, I don't feel like I learned anything about myself (especially not by trusting the results), but I answered all the survey questions, so hopefully my data will be useful in some small way