Well, some certainly do, but as always, it is a game of sampling, so individual errors are not as important as proportions. If we take the article at face value

Well, some certainly do, but as always, it is a game of sampling, so individual errors are not as important as proportions. If we take the article at face value, proportions differ drastically between very different cultures so there must be something in the culture that biases the proportions towards one outcome or the other systematically. Of course, it's always possible any individual study contained unknown and unaccounted biases. Funding in science is still relatively small, the number of people participating few and science itself is hard.