T HE FOREGOING analysis of voting studies was written before the appearance of a new work from the Michigan Survey Research Center, The American Voter.88 The intent of this work and its general approach are similar to those of The Voter Decides, and its merit consists in the extent to which it is informed by political considerations and to which it carries the search for causative factors.
Instead of being content to identify "psychological" attitudes, similar to those in The Voter Decides -- and which Rossi criticized as being largely tautological -- the authors of The American Voter look upon the six "psychological" attitudes 89 or "attitude forces" as "intervening variables linking behavior with a host of antecedent factors" (p. 120). That is, they recognize that voting choice is governed by what we have called political opinion, and they seek to discover the antecedent causes of this opinion, which they hope to state in the form of deep-seated laws of social behavior. Though their purpose is the same as that of Berelson, Lazarsfeld, and McPhee, they are critical of the "social approach" to the study of voting behavior because of the ephemeral character of its findings, among other reasons.
A correlation between the fact of being a Negro and the casting of a Democratic ballot gives us interesting information, yet information pitched at a low level of abstraction. Generalizations of this sort tend to fall by the wayside with the passage of sufficient time, if not reformulated in more general terms. In the case of Negroes, for example, there is evidence to indicate that not more than a decade or two ago the relationship was reversed, with Negroes tending to favor the Republican Party. And it seems entirely plausible that the relationship might become reversed again in the fairly near future, without upsetting any very deep-seated "laws" of social behavior.
The question concerns the character of these laws and whether they will____________________