A multidimensional conceptualization of party identification

The standard treatment of party identification makes several untested assumptions, especially that citizens can identify with only a single party and that political independence is just the opposite of partisanship. A more general possibility is that several attitudes must be taken into account: attitudes toward the Republican party, the Democratic party, political independence, and political parties generally. A literal reading of the usual party identification is consistent with this multidimensional interpretation. Citizen ratings of the two parties turn out to be virtually uncorrelated, as are ratings of independence and political parties, confirming this multidimensional view. Strength of identification and strength of independence are separate in this model, which explains some of the anomalies in the current literature, including intransitivities in relationships with other variables and weak correlations involving independence. New questions included in the 1980 CPS National Election Study support this interpretation and provide a new understanding of political independence.