Social scientists are finding an increasingly useful and stimulating tool in the application of statistical techniques to their problems. As in the employment of any new tool, both the utility and the limitations of this one must be learned. It seems beyond reasonable doubt, however, that quantification of data in the social sciences will become a more widely used and rewarding procedure as time goes on.
Prudence dictates that stress be laid on its limitations. The enthusiasm with which a new tool—toy, some would say—is adopted should not blind the user to dangers which may be implicit in its overuse. One cannot squeeze more juice from an orange than the orange contains, no matter how modern the squeezer. Care must be exercised, too, lest the bitter essence of the rind become mixed with the nourishing juice of the fruit itself.
The present analysis is an attempt, in not too complex a fashion, to make use of such techniques to organize and validate data which might otherwise permit only the broadest sort of generalizations by way of conclusion, conclusions unsatisfactory roughly in proportion to their breadth. The senior author of this article has for more than a decade and a half been interested in the problem of objective measurement of certain aspects of political change in Latin America with particular respect to the sum total of phenomena falling under the rubric of “democracy.” On four occasions, 1945, 1950, 1955, and 1960, he conducted a survey among groups of specialists on Latin America to elicit evaluations which then, with the help of such statistical procedures as seemed useful, were summarized and analyzed.