Pollsters Survive Truman's Victory
Truman's Victory in 1948 Threw Opinion Pollsters into Survival Mode!
In the aftermath of Harry Truman’s re-election in 1948, George Gallup and the “scientific pollsters” were stunned by their embarrassing prediction that President Truman would lose. All three of the top pollsters in the United States at the time predicted that Thomas Dewey would defeat President Harry Truman by a margin of between 5 to 15 percentage points. The scientific pollsters blew it in a big way.
Gallup and Roper Lost Business
According to George Gallup, Jr., the fiasco was a significant blow to the industry and a major embarrassment to his father, but Dr. Gallup did not let it defeat him. However, it puzzled him. He is reported as stating immediately after the election that he didn’t know what happened, but he had the confidence to find out what went wrong and do so quickly. Indeed, Gallup had to move quickly because around 30 or more of the newspapers that carried his syndicated poll, were cancelling their contracts.
The American Institute of Public Opinion had grown substantially since 1935, and eventually contracted 200 newspapers for Gallup’s weekly column. But, Gallup had to visit many of the newspapers right after this election because his credibility was on the line. In order to keep their business, he had to explain to the publishers what went wrong.
Also, Elmo Roper, faced financial disaster. Years afterward, Roper’s son is quoted as stating that he and his father, who were both Democrats, were caught in a dilemma: We saw our man winning, but our company was going down the tubes.
Investigations into the Polling Errors
After Gallup had finished analyzing his mistakes, he understood that a major lesson was to continue polling all the way to Election Day because much happened in the final two weeks after they had stopped their surveys. The scientific pollsters had quit polling with 14 per cent of the voters still undecided! He realized that it would have been wise to conduct a last-minute poll. He also learned that undecided voters did not behave in the way he estimated, and ultimately concluded that the undecided will usually side with the incumbent.
In 1949, a panel of Social Science Research Council (SSRC) conducted some investigations of such polling errors in the 1948 presidential election. The conclusions of the SSRC were that: 1) there were errors of sampling (use of quota sampling technique) and of interviewing; and 2) there were errors that involved the polling enterprise's failure to adequately address the issue of the undecided voters.
The SSRC ultimately urged the the polling industry to revise and make improvements to their efforts with respect to their methods of obtaining samples and to interview voters up until the finish of future presidential election campaigns. Today most professional pollsters use the random sampling method of polling. Such professionals also do not abandon voters before Election Day. In fact, we now are all accustomed to the exit polls which bug voters right after leaving their polling places.
The Industry Survives
Although the outcome of the incorrect prediction was a huge blow to the young polling industry, Gallup and the others did not give up their belief in the validity of scientific sampling. Actually, Dr. Gallup turned this negative event into a learning experience and developed the argument for more extensive political polling. We have him to thank for introducing the concept of exit polls. As we know, opinion polling continued to grow and is now thriving in the U.S.
Source: Moore, David W. The Superpollsters: How They Measure and Manipulate Public Opinion in America (1995)
More in History
Latest Articles