They're trying to measure the honesty of the people who work in those places, not people who drop wallets there. And I get your point that local people's reacti

They're trying to measure the honesty of the people who work in those places, not people who drop wallets there. And I get your point that local people's reaction might be different if they thought the owner was a rich foreign tourist who was less likely to return or notice and report a theft. Also yeah sure there may be selection bias due to their choice of test locations, but their methodology seems as good as you can do for a broad international study.