Let the reader be forewarned: what he gets out of this book will depend on thoughts which he is bound to have--say, somewhere in the middle of the third or fourth chapter-- about matters that have nothing to do with either national leadership or foreign policy. For these substantive topics are treated, in the pages that follow, in terms of quantified data provided by a relatively small sample of a large population. If the reader, like the author at the outset of the inquiry, has never had occasion to analyze aspects of the foreign-policy- making process quantitatively, he will no doubt wonder about the meaningfulness of such an approach, about the relevance of these particular data to such general topics, and, indeed, about reliable knowledge itself--what it is and how it is obtained.
Clarification of my thoughts about reliability occurred during the process of coding and tabulating the responses of hundreds of national leaders. To cumulate and organize a mass of data is to be reminded of the variability and complexity of human behavior. The final product of quantitative analysis--tables of data systematically arrayed in columns and rows--suggests an orderliness and simplicity about human affairs that can be profoundly deceptive. But the experience of actually sorting data into columns and rows makes one sensitive to the fact that a variety of factors underlie the distributions which seem so simple in the neatly organized tables.
More specifically, in handling the data I was impressed with the vast difference between tracing a single actor or unit through a particular period of time and accounting for many of them under varying conditions. The single actor or unit does A, B, or C, and what he does is thus unmistakable and can be interpreted accordingly. In the case of many actors, however, nothing is so clearcut. Some do A, others do B, still others do C, and probably a few will pursue a course between A and B or between B and C. Furthermore, within