On the Comparison of Two Methods of Combining Independent Tests