Page:CIAdeceptionMaximsFactFolklore 1980.pdf/11

 inserted and/or indirect evidence used as a surrogate. Despite these difficulties, the evidence for many of the conclusions drawn in this paper is sufficiently strong that the analysis is robust to even substantial errors of omission/commission. Both strategic and tactical level engagements on land, sea, and air are included.

In analyzing a subset of these data in his manuscript, Stratagem, Dr. Whaley presented numerous cross-tabulations, sorts, counts, trends, etc. as raw or summarized data but omitted various statistical tests of hypothesis. This omission was deliberate, reflecting two considerations:


 * First, the major battles constituted nearly an exhaustive sample.4 If the population is viewed as finite, i.e., only those battles that actually took place in this time period, then statistical tests are unnecessary as the variances of all estimates are essentially zero. This was the assumption in Whaley's original analysis. However, if the battles themselves are regarded as a sample from a larger population — i.e., battle situations that might have occurred, then statistical methods are appropriate and, indeed, are necessary.