Elmo Burns Roper was born in Hebron, Nebraska, on 31st July, 1900. He ran a jewelry business with his brother in Iowa in the 1920s. As Michael Wheeler, the author of Lies, Damn Lies, and Statistics: The Manipulation of Public Opinion in America (2007) has pointed out: "Roper began not as a researcher but as a salesman for a jewelry manufacturer; he sold its line to retail stores. One of Roper's functions was to pass judgment on possible items for the coming year's line. When he was first hired, he was very impressed by his colleagues who could simply look at something and say whether or not it would sell. He did not have the confidence or the experience to make decisions of that sort, however, so he made it a practice to take samples out to jewelers he had sold to in the past and ask them what would move and what would not. He would then return to his home office and tell the company which items should be pushed and which dropped."
Roper went to work for the Traub Company. In 1933, he joined forces with Paul T. Cherington and Richardson Wood to establish Cherington, Wood, and Roper, a marketing research firm. The partnership broke-up the following year and Roper established his own company. In 1935 Henry Luce commissioned Roper to run public opinion polls for Fortune Magazine.
The 1936 Presidential Election was the first where polls were published in newspapers during the campaign. The Literary Digest sent out 10,000,000 questionnaires to voters. It predicted that Alfred Landon (57%) would defeat Franklin D. Roosevelt (43%). It later emerged the magazine reported its results based on only 15% of the questionnaires set out. It seems that Landon's supporters were much more willing to send back their questionnaires and therefore distorting the results.
George Gallup used a very different strategy: "The sampling procedure described is designed to produce an approximation of the adult civilian population living in the United States, except for the persons in institutions such as prisons or hospitals... The places (where people were interviewed) were selected to provide broad geographic distribution within states and at the same time in combination to be politically representative of the state or group of states in terms of three previous elections." Gallup's survey showed Roosevelt winning with 55.7% of the vote. However, this was 4% less than Roosevelt actually achieved and it was Roper's poll that came closest to predicting the result (only 0.9% out).
In 1940, President Franklin D. Roosevelt hired Roper to assess public opinion of Lend-Lease prior to its implementation. In 1942 he was hired by William Donovan to be the deputy director of the Office of Strategic Services. Roper helped convince George Marshall and Dwight Eisenhower of the importance of opinion research in the armed forces. In the later stages of the Second World War Roper worked with the Office of War Information.
Roper was highly critical of the work of George Gallup who he considered to be too close to the Republican Party and in the 1944 Presidential Election he gave considerable support to Thomas Dewey in his campaign to be the party's candidate. The British Security Coordination, who had an agent, David Ogilvy, working for Gallup claimed that the "Gallup Poll did not prove a reliable guide to the Presidential election of 1944. This was so, largely because Gallup is himself a Republican and a staunch supporter of Dewey. As William Stephenson learned, there was little doubt that Gallup deliberately adjusted his figures in Dewey's favour in the hope of stampeding the electorate." Ernest Cuneo told William Stephenson: "Dewey is one of Gallup's principal clients... Dewey is calling up Gallup so often they have to have a clerk to answer him."
Elmo Roper stated after the 1944 Presidential Election: "I think prediction of elections is a socially useless function. Marketing research and public opinion research have demonstrated that they are accurate enough for all possible commercial and sociological purposes. We should protect from harm this infant science which performs so many socially useful functions, but which could be wrong in predicting elections, particularly in a year like this."
After the war Roper founded the Roper Center for Public Opinion Research at the University of Connecticut. In 1947 Roper appointed Louis Harris as his assistant. However, he later established his own company, Harris and Associates, that provided competition to Roper and George Gallup. Roper was critical of Harris when he joined the campaign team of John F. Kennedy. Roper attacked those "so-called public opinion researchers," who allow their polls to be exploited "rather openly for propaganda purposes."
Roper was also a syndicated newspaper and radio columnist, and during presidential elections he often appeared on network television. According to his biographer Roper was a "tireless activist for liberal causes" and was active in the civil rights movement. He was also a member of the National Urban League, Planned Parenthood and the US Citizen's Commission on NATO.
The principal task of Louis Harris, Kennedy's personal pollster, was to churn out polls which would keep the bandwagon rolling. Sorensen describes one aspect of Harris' work. "Equally important, however, were the results of privately financed and conducted polls which were primarily for the senator's information though given to friendly politicians and columnists."
The late Elmo Roper, who apparently never missed a chance to criticize Harris, his former employee, was more explicit about just how these private polls were used. He bitterly attacked those "so-called public opinion researchers," meaning Harris, who allow their polls to be exploited "rather openly for propaganda purposes." Roper recounted an incident which shows just how willing Harris was to have polls used this way. "In one case polls were even leaked to an opponent showing he was seven points ahead. When he called me up long distance to ask me why I thought he would be shown a poll that Mr. Kennedy had paid for and that Mr. Kennedy had not yet seen, I said, "Well, has it occurred to you that they would like to see you let up on your campaigning a bit?"
Hubert Humphrey, the victim of much of this pollsmanship, today recalls Kennedy's and Harris' work with more envy than resentment. "Kennedy used the polls with a master's touch. Lou Harris was doing a lot of it for him. The polls were always being planted in the newspaper columns, Scotty Reston's and everybody else's. The information was made available; here it is, here's how our man is doing. When you're running high like that, and hard, you get the reputation as a front-runner."
The most basic kind of commercial polling, and the kind which has the oldest roots, is market research. Many of the pioneer public opinion pollsters got their start by testing the markets for various products; from there they branched into other fields. The career of the late Elmo Roper was typical of this pattern. Roper began not as a researcher but as a salesman for a jewelry manufacturer; he sold its line to retail stores.
One of Roper's functions was to pass judgment on possible items for the coming year's line. When he was first hired, he was very impressed by his colleagues who could simply look at something and say whether or not it would sell. He did not have the confidence or the experience to make decisions of that sort, however, so he made it a practice to take samples out to jewelers he had sold to in the past and ask them what would move and what would not. He would then return to his home office and tell the company which items should be pushed and which dropped.
This was market research in its crudest yet perhaps most useful form. Roper quickly had the best track record of all the salesmen. As his son, Burns Roper, puts it, "He wasn't shooting from the hip, just saying what he himself liked or disliked." The senior Roper's success convinced him that the routine he followed could have a much broader application. He began doing similar surveys for other businesses with different kinds of products. At the outset, he did most of the interviewing and worked on an intuitive, subjective basis.
Roper's clients began to ask that people's preferences be broken down according to their age, sex, or geographic location. This required more work, so Roper had to hire other people to do the interviewing, which in turn mandated written questionnaires. Methodology grew out of need and was progressively refined. At the core, however, Roper's function was to determine who would buy what.
There is nothing pernicious about this kind of market research; in fact, it is very useful in any complex economy. Whether one has great faith in the efficiency of the free enterprise system or favors a planned economy, there is a need for information which will help supply meet demand. Good market research can tell a manufacturer that a certain product just is not wanted and thus save him, and society, the cost of its production. Likewise, if demand is higher than anticipated, good market research can tell a company to produce more and thus take advantage of economies of scale.