Two waves of telephone interviews with a sample of 141 voters were used in a study of political communication effects during the 1974 congressional election in the Illinois 24th Congressional District. Seven variables specified by the combined-effects model were derived from the interviews and factor analyzed. Two factors were found and interpreted as direct effects and limited effects. The direct-effects factor included items measuring early and late cognitions about the democratic candidate, mass media exposure, and interpersonal discussion. The second factor, limited effects, consisted of early and late attitudes toward the democratic candidate and intention to vote. The first factor was interpreted as consistent with an agenda-setting model of political communications effects, and both factors were interpreted as consistent with a combined-effects model. (AA)