Since the essence of this study is comparison, virtually all our final statistics were at some stage in the analysis rank-ordered—that is, arranged in order in some scale of comparison. The value of a rank-order depends, of course, upon the range of variation; the smaller the range, the less useful is the rank-order. For example, a rank-order of the simple incidence (ever-never) of masturbation is rather meaningless since the comparative groups are all so near 100 per cent, and their positions in the order are dependent upon minor extraneous factors such as age, the number married, etc.
The general position of a group rather than its exact position is the important thing. It matters little whether, by virtue of some fraction, a group ranks second or third; what does matter is that it ranks high.
When two or three groups of a tripartite larger group (e.g., offenders vs. children, vs. minors, and vs. adults) are contiguous in a rank-order, this contiguity is most unlikely to be chance. The position of one group might be said to substantiate the position of the other.
In addition to noting the positions of specific groups in a rank-order, valuable knowledge may be obtained by looking for general trends: do the offenders whose sexual objects were children tend to concentrate in the rank-order? Do some groups reveal a correlation between rank-order position and increasing age? These and similar questions have been routinely raised and dealt with in our analyses.
*11\161\2*

Leave a Reply
You must be logged in to post a comment.