This blog goes back to the facial recogntion topic, which at the moment I think is my favourite. I now have a name for it: the cross-race effect. (I didn't make it up, I found the term whilst researching.) But please Ms Zhang, if it's really not going to work, please tell me straight out. I know you said that the global warming experiment would be better, and I really do value your advice, but for some reason I really want to do this experiment and if there's a fair chance that it will work I think I'm prepared to take it, I just don't want to dig myself into a hole.
The facial recognition one would also, I think, be easier to perform. There's no equipment that will be impossible to access or expensive to purchase. At the moment my idea is to only use two different cultures because I wouldn't be able to get large enough groups of 3 or more cultures. The two cultural backgounds I'll use are Asian and Anglo-Saxon. I'm aiming for around 16 people in each group, half boys half girls. All the volunteers will be around my age - say between 14 and 17- because 1) this will remove one of the variables and 2) this is the age group in which I know the most people. I'll use not well-known models (they can't recognise the models) of both cultures as the people they have to distinguish between. This will work well because most models are of the same height and build so one will have to concentrate more on the facial features and because photos of models won't be hard to come by. As yet, I'm not totally sure how to question the volunteers. I might sure them the pictures in a sequence and ask "have I shown you this person before?". One of the experiments that was conducted in Germany (which I posted the other day) showed the volunteers the faces then left a gap of up to 3 weeks before they tested the recognition.
Unfortuanately I haven't been able to find too many examples of how the cross- race effect has been tested. Most sites just mention it in regards to witness-identifiaction. I managed to track down the Meissner and Brigham, 2001 study into the cross-race effect that was mentioned on nearly all of the sites, only to find that I had to buy it before I could read it. Maybe I'll find it a library.
For my experiment, the aim would sound something like "Is a person able to recognise faces from their own culure better than those from another culture?" I think the independant variable is the person and their culture and the dependant variable the effect of culture on recognising faces.