Does collecting repeated blood samples from each subject improve the precision of estimated steroid hormone levels?