published the first sophisticated~L quantitative study of Latin American military behavior in 1967. NoR subsequent statistical analysis of Latin American politics has been more widely cited, praised, or reprinted; moreover, no superior quantitative research on the causes of military intervention has emerged during the last fourteen years. Indeed, many textbooks on Latin American politics now commonly state Putnam's two major conclusions as verified empirical truths; students are taught to accept as proven facts the propositions that higher social mobilizationl levels inhibit military intervention and that traditionally noninterventionist military institutions are likely to remain so. More recent theoretical works by Huntington (1968), O'Donnell (1973O'Donnell ( , 1976, and others, however, have cast considerable doubt on both of these assumed relationships. In addition, the imposition of military rule during the 1970s by formerly noninterventionist armed forces in highly socially mobilized Uruguay and Chile indicates the existence of some anomalous cases which seem to contradict Putnam's conclusions.In short, both theoretical revisions and Latin America's actual experiences since 1967 suggest the value of reexamining Putnam's two most generally accepted findings by means of a new quantitative test. This paper will review Putnam's research, discuss the theoretical and empirical justifications for a reanalysis, and retest Putnam's principal propositions using comparable methodological procedures and a new data set. The final portion of the paper will offer both substantive and methodological conclusions on the issues raised by comparing the two sets of test results.
PUTNAM'S ANALYSIS AND THE LITERATUREQuantitative studies were quite new to Latin American politics in 1967. Martin Needler (1966: 616-26) had performed some early statistical analyses but Putnam's study of military intervention was far more comprehensive and sophisticated.