My response variable is deviation from a 50:50 sex ratio, and so the min and max values possible are 0 (only 1 sex) and 0.5 (equal sex ratio). I considered using a hurdle model
