[LIS-Forum] How scientists can stop fooling themselves over statistics