Coding Results. That there were problems involved in coding BV513 becomes evident from Tables 6.15a and 6.15b, which admit that nearly 30 percent of the parties could not be coded for their orientation toward civil liberties. This high failure rate was due both to the parties' failure to address the issue in their policy statements and to writers' -neglect to consider it in their evaluations of party policy. The mean codes for AC513 are either the lowest or among the lowest for the issue orientation variables, but there was no relationship between the variable codes and our confidence in assigning codes. It should be noted, finally, that the parties' commitment to civil liberties was often only superficial. The practice of civil liberties, passionately defended by a party out of power, not infrequently fell by the wayside as government policy after its erstwhile champion had ridden it into office.

The previous thirteen variables in the issue orientation cluster have been operationalized in a manner that permits the ready investigation of their conformity to an underlying left-right continuum. The merit in this approach remains to be determined through intensive analysis of the intercorrelations among the indicators, which is the task for another time and place. One method of determining the validity of our procedure, however, does lie within our grasp. The method of "concurrent validation" holds that measurement validity is established if the proposed measure conforms to some outside criterion whose own validity is either established or presumed. To demonstrate concurrent validity of our operationalizations of the issue variables, we need to obtain high correlations with other, presumably valid, ratings of parties on the left-right dimension.

Unfortunately, there are few such comprehensive ratings available in the literature. Students of comparative politics speak freely of parties as being located on a left-right or liberal-conservative continuum, but few attempt to be explicit, systematic, detailed, and comprehensive in their comparisons. However, two usable sets of cross-national party evaluations were located to provide validating criteria for our own measurements. The first set is furnished by experts from the U.S. State Department; the second from experts within the Soviet Union.

For more than 20 years, the State Department's Bureau of Intelligence and Research has classified parties as "Communist," "Non-Communist Left," "Center," and "Conservative" in its annual report of World Strength of Communist Party Organizations. In addition to providing detailed information on the membership and strength of communist parties throughout the world, this publication reports election results and legislative representation for the major parties in each country, with the parties classified in one of the four categories mentioned. Although the State Department appears not to have used "right" or "rightist" as a category by

go to page 74