ABSTRACT
Six components were discovered that, together, completely explained the outcome of each election. The components were: attitude toward each candidate, issue concerns, party identification, late deciders and, in some elections, an “adjuvant factor.” Bar graphs are shown for each election, which allow the reader to see just how much each of these factors helped push a candidate to one side or the other of a 50–50 result. The length of five of the components was measured by multiplying the regression coefficient of the non-standardized Model Equation (see Introduction) by the mean of the component variables. The adjuvant factor was determined by the intercept term (the “a” term) in the multiple regression equation. The length of the components varied from election to election; sometimes it was a candidate that was most important, sometimes both candidates, sometimes an issue, sometimes party identification, and sometime the late deciders. But it was the net total of all components taken together that estimated the election result. That result, based on my method, was within 1 percent of the actual vote in all ten elections studied.
