It's excellent methodology because they asked indirect questions. Coming out and asking someone "Hey, are you a racist?" wouldn't be effective.
Asking someone in an unrelated economic experiment something like "Would you mind if your neighbors were black?" is more accurate because it tricks people into revealing their true biases.
But this didn't ask them would you mind if your neighbours were black. This made them pick what they're most worried about. That question can only be asked in context of my country. If there's almost no change of a foreigner living next to me but there's a high chance of crime in my area even though I'm massively racists and would actually care about more about a foriegner living next to me (all things being equal) because of my circumstances I'm actually more worried about crime atm.
Asking someone in an unrelated economic experiment something like "Would you mind if your neighbors were black?" is more accurate because it tricks people into revealing their true biases.