Measuring the ''personal vote'' of members of Congress''

This paper uses a pooled time series/cross sectional research design to measure the ''personal vote'' of three groups of members of Congress. The personal vote is defined here as that part of a candidate's vote share that is unique to him or her. It is measured using a legislator-specific intercept for each legislator in an equation predicting vote share. The other variables in the equation account for the effects of the election year environment and the tenure of the representative. Three cohorts of representatives are analyzed: 1) those whose careers ended before 1966, 2) those whose careers began after 1965, and 3) those whose careers began before 1966 but ended after. The mean legislator-specific intercept is nearly twice as large for the first group as for the second; it is 29 percentage points larger for the first than the third. A second stage regression explains these intercepts using legislator characteristics, party affiliation, region of origin, and number of terms served. The term effects increase at an increasing rate as tenure increases for all those whose careers end after 1966. Personal vote accumulates nearly twice as fast for those first elected after 1965 as for those first elected before 1966.