Correcting for Selection Bias in Social Movement Research