The effect on heritability estimates of different data exclusion policies for post weaning fat depth were examined using industry data for Maternal, Merino and Terminal sheep. The exclusion policies which had the most positive effect on heritability were mean fat depth for the contemporary group, and standard deviation of fat depth for the contemporary group. Exclusion based on mean weight of the contemporary group showed some effectiveness for light animals. Excluding records based on mean age of the contemporary group had little to no effect on heritability estimates.  In order for fat depth to be analysed most appropriately animals should be measured when they have had the opportunity to express the trait. It is appropriate to exclude contemporary groups from analysis where the variation in fat depth for a contemporary group is low.