Wasn't there a rival thesis done shortly afterwards saying that the figures were wrong because the author had not provided a satisfactory definition of folk?