A ‘Non-election’ in America? Predicting the Results of the 1970 Mid-term Election for the U.S. House of Representatives

Abstract

One of the reasons for paying attention to the results of American mid-term elections is the hope that they will tell us something about the standing of the President's party with the electorate. But the electoral verdict is notoriously hard to interpret. Since the President is not himself standing for re-election, the verdict has to be inferred from the results of House and Senate races in which the national mood of the electorate may be obscured by local and temporary factors. This will be especially true in Senate races, with only some 33 seats at risk; but even in House elections, with some 435 seats at risk, a grave problem arises when one comes to compare the results with those of the preceding Presidential election. In every mid-term election since that of 1934, the party of the President, whether it be Republican or Democratic, has lost some of the seats it had won at the previous Presidential election. The net loss has been as low as four seats in 1962 and as high as seventy-one seats in 1938, but it has always occurred. Some loss to the President's party is considered to be ‘normal’ at mid term, and it is only to the extent that the actual loss diverges from the normal loss that implications can be drawn to the President's standing with the electorate.