You also have to be very careful how you decide to present this, some biases out there result in dangerous situations. I’d be cautious about suggesting participants “pass” as Muslim in some areas for example. And therein lies the real problem with this type of experiment, the results can be dangerous.

You could always replicate a study I read some years ago but apply it locally. The researcher completed employment applications at a variety of businesses. Everything was essentially equal except that the applicants had clearly identifiably ethnic names. The result was that the people with ethnic names didn’t get interviews.

Another researcher I read had people of different sex/race/socio-economics park on the side of the road with the hood up to see who got helped. The hot blonde got a lot of attention. The poor looking guy got left to fend for himself.