FDR and the Political Use of Polling
In the spring of 1935, the Democratic National Committee (DNC) conducted a public opinion poll in order to assess whether Huey Long’s massive radio appeal posed a threat to Roosevelt’s re-election chances the following year. Using a national sample of nearly 31,000 voters, the DNC poll asked respondents to indicate a preference for Roosevelt, an unnamed Republican, or Huey Long, for president. Although a majority favored Roosevelt, 11 percent of respondents preferred Long in the hypothetical three-way contest. The survey also revealed Long’s broad appeal, one that extended well beyond his native South into the urban centers of the Northeast and Midwest. Seeing the results of the poll, DNC chairman James Farley feared that Long might hold the balance of power in 1936.121
The 1935 poll was the work of Emil Hurja, a deputy director of the DNC whose knowledge of statistics and deep interest in voting produced important innovations in the political use of survey methods.122 In 1928, while working as a mining stock analyst on Wall Street, Hurja tried to convince then DNC chair Jacob Raskob that he could forecast election results using the same principles of sampling he used to test mineral assays. As Hurja later explained:
“You apply the same test to public opinion that you do to ore. In mining you take several samples from the face of the ore, pulverize them, and find out what the average pay per ton will be. In politics you take sections of voters, check new trends against past performances, establish percentage shift among different voting strata, supplement this information from competent observers in the field, and you can accurately predict an election result.”123
In other words, Hurja proposed the same kind of sampling techniques that were then gaining prominence in market research and eventually in commercial polling and other areas of survey research. Raskob, however, dismissed Hurja as a crank.
Undeterred, Hurja offered his services to the DNC again four years later. In 1932, he drafted a memorandum that promised to give the DNC a moving “picture of sentiment” as it evolved over the course of the campaign. By combining all manner of political information—straw polls, census data, historic voting patterns, and newspaper editorials—Hurja claimed he could provide a statistically accurate portrait of Roosevelt’s strength in different parts of the country. Because Hurja recognized the shortcomings of any single source of information, he supplemented traditional straw polls with census data and voting records in order to weight the results and correct for any bias. Armed with a multidimensional picture of the electorate, Hurja promised to “save needless expenditure of campaign funds in districts where [they were] not needed.”124 With the help of publicity director Charles Michelson, Hurja convinced DNC chairman James Farley to hire him as a special consultant for the 1932 campaign.125
Over the next several months, Hurja wrote weekly memos summarizing Roosevelt’s strengths in various parts of the country and among different segments of voters. In what sounds commonplace today but was innovative for its time, Hurja produced a map with red and blue counties in progressively lighter shadings to differentiate those states in play from the ones safely in the hands of the two parties. Displaying tremendous resourcefulness in acquiring information—including house-to-house polls conducted by local bookmakers wagering on the election—Hurja accurately predicted a broad-based shift of urban voters to Roosevelt in 1932.126
Hurja continued to work closely with Farley after the election, employing his statistical skills to direct the distribution of federal patronage and, later, guiding party strategy during the 1934 congressional midterm elections. Again, Hurja used a combination of election results and polling data to develop a more reliable map of Democratic Party strength in various parts of the country. Hurja even polled Works Progress Administration workers to assess public opinion toward the New Deal.127 Hurja’s ability to pinpoint areas of Democratic strength and weakness and to focus campaign resources accordingly helped the Democrats in 1934 pick up nine seats in the House and ten in the Senate.
Hurja showed that poll results could inform political strategy. At a time when his contemporaries such as Roper or Gallup mainly produced election forecasts for public consumption and only dabbled in political polling, Hurja worked within the Roosevelt administration as an adviser and strategist. His ability to translate polls into political action pointed toward a “newly developing science of public opinion ... and the political consulting business that would later grow from it.”128 In the meantime, Hurja’s success attracted the attention of Time, Collier’s, and other popular magazines that described him as a “political soothsayer” and “prophet extraordinary of the Democratic Party.”129 A 1936 profile in the Saturday Evening Post titled “Prof. Hurja, The
New Deal’s Political Doctor” described him as “a great political diagnostician” equipped with “political stethoscopes, popularity meters, and mass-mind indicators.”130 In fact, Hurja’s skills were much more than a campaign tool; they were also a critical device for the promotion of presidential initiatives. “The function of Hurja is to discover what the people are talking about,” the Post explained. “If the conversation is harmful to the New Deal, the next step is to change the conversation.”131 The blending of scientific surveys with White House communications would become a powerful tool for modern presidents.132
Radio, of course, was a key instrument for promoting the New Deal as the refinement of communications techniques increased considerably during Roosevelt’s time in office. The sheer volume of press releases, radio shows, newsreels, and press conferences reflected a belief by Roosevelt and his closest advisers that the success of New Deal programs depended on building public support through a coordinated communications strategy. Stephen Early, Roosevelt’s press secretary, oversaw many of these efforts, including the appointment of more than 100 former journalists to staff press offices in every executive agency.133 However, it was in the use of radio that Early’s influence on the future of political work was most keenly felt.
Radio offered distinct advantages over newspapers for communicating administration policies. By 1933, for example, the estimated number of radio listeners surpassed the combined circulation of the largest daily newspapers.134 In addition, the radio offered a way to circumvent hostile newspaper coverage; radio afforded a more direct form of public communication. As Early explained, newspapers “cannot misrepresent nor misquote.”135 Finally, and most important, the government enjoyed privileged access to valuable airtime. Although broadcasters operated under a relatively lax regulatory environment, they still depended on government licenses to operate. The Radio Act of 1927 stipulated that broadcast licenses be granted for reasons of “public interest, convenience, or necessity.” However, as historian Douglas Craig has noted, actual federal oversight of radio was minimal. The 1927 act established a system of self-regulation later codified in the Communications Act of 1934 and the creation of the Federal Communications Commission. Fearful of further regulation, most broadcasters obliged government requests for free airtime as a way to fulfill their public service mandate.136
Capitalizing on these advantages, the administration issued a steady stream of radio broadcasts, including 20 talks by the president, 17 by Eleanor Roosevelt, and 107 by members of the cabinet in the first ten months of the New Deal alone.137 Many federal agencies produced their own weekly radio programs for the national networks or distributed recordings for use by local and independent stations. By 1936, a total of 5 ,000 government recordings were in circulation.138 Stephen Early coordinated many of these efforts, securing airtime for administration broadcasts and tailoring programs to meet the interests and expectations of radio listeners.139 Easy access to network airtime became a valuable resource for Roosevelt as he headed into the 1936 election. Much to the consternation of the Republican Party, the networks deemed many of the president’s speeches that year as nonpartisan in nature and therefore aired them free of charge. As Douglas Craig summarizes, radio “added a powerful new advantage to incumbency.”140
Of course, the centerpiece of Early’s radio strategy was Roosevelt himself, notably the “fireside chats” and other radio addresses he delivered while in office. The president possessed an appealing radio persona that created a personal relationship with the public, but it is important to emphasize how hard Early and those around him worked to make the most out of the president’s gifts. Broadcasts were timed to maximize audience and effect, often scheduled between nine and eleven in the evening, when East Coast listeners were not yet in bed and those on the West Coast were returning home from work. Care was given so that a broadcast would not conflict with popular radio programs. Content was carefully considered to fit within both current events and larger administration goals. Advance publicity and press releases built up audience anticipation. In the case of the first fireside chat on March 12, 1933, Early even arranged for spot advertisements throughout the day promoting the speech. Words were painstakingly chosen and drafts endlessly rehearsed by Roosevelt in order to perfect the rhythm and pace of his speech. During recording, special paper was used to dampen the sound of rustling pages, and Roosevelt even wore a bridge to cover a missing tooth so he would not whistle when he spoke.141
In the days after a fireside chat, Early would make a careful study of its effects, noting editorial and newspaper coverage of the speech as well as telegrams and letters from listeners themselves.142 In addition, Early tracked the size of Roosevelt’s radio audience using available surveys and audience measures. In a span of fourteen radio addresses between June 1936 and February 1942, the number tuning in to hear the president grew from 6 to 60 million homes. Audience surveys also showed that the fireside chats attracted the occasional radio listener as well as those who regularly tuned in, helping Early appreciate the distinctive nature of this format and prompting him to reserve its use for special issues or important administration goals.143
Surveys and other instruments of audience measurement offered a new way to gauge the president’s popularity: through his radio appeal. More broadly, innovations in the political use of radio illustrate how the development of media and polling went hand in hand. As radio expanded the reach of political communication, survey methods offered a way to measure its effects. Developing a portrait of the listening audience, who was tuning into the president and when, was an important first step in assessing the political value of radio. However, reliable measures of audience size merely begged the question about the influence of radio communication on listeners. This, after all, was the question that motivated market researchers and social scientists alike, whether radio was used as a tool of advertising or as an instrument of political propaganda. For the Roosevelt administration, in particular, tracking the size of the radio audience was part of a deeper concern about whether the president and his staff could shape public opinion through crafted communication.144 In seeking an answer to this very question, the Roosevelt administration turned to a variety of experts who straddled the commercial and academic worlds of public opinion research.
As noted previously, Princeton was an important locus of activity where academic and commercial survey researchers worked in close proximity to one another. Princeton was home to the first university research center on public opinion established by Hadley Cantril as well as the journal Public Opinion Quarterly he cofounded. Princeton was also the center of the commercial polling business, including the American Institute of Public Opinion founded by George Gallup in 1935. Within the confines of this one university town, a network of academic and commercial pollsters interacted regularly, perfecting their methods and promoting their infant industry.
These connections helped spawn early experiments in the use of poll-tested messages by the Roosevelt administration. Hadley Cantril, in particular, was a key figure in this regard. “Perhaps more than any other social scientist,” Jean Converse has noted, “[Cantril] bridged the gap between commercial polling and social science.”145 With a foot in both worlds, Cantril understood that polls had to be deciphered in politically useful ways for them to be of any use to politicians. During his work on behalf of the Roosevelt administration throughout the 1940s, Cantril interpreted poll results with an eye toward the political needs and communication strategies of the president, translating the scholarly insights of the social scientist into the digestible format of the pollster’s memo.
Cantril began work for the Roosevelt administration in 1940, when his former college roommate, Nelson A. Rockefeller, asked him for help surveying public opinion in several South American countries. Rockefeller was then head of the Office of Inter-American Affairs (OIAA), a somewhat shadowy agency the United States established to check Axis influence in South America. Propaganda was an important part of this strategy, and it was here that Cantril’s expertise on radio and its effects proved valuable.146 In cooperation with George Gallup, Cantril set up a nonprofit research institute, American Social Surveys, to conduct polling operations abroad. With funding from the newly established Executive Office of the President, Gallup and Cantril sent a crew of twenty interviewers to Brazil to collect information on radio listening habits. Jean Converse described the effort as “an intelligence operation disguised as market research,” enabling Cantril to gauge the extent of Nazi sympathies and the potential reach of US propaganda.147
Cantril’s research in South America led to more work for FDR.148 With the conflict in Europe raising debates at home about US neutrality, Cantril supplied the president with memos explaining the results of Gallup surveys about potential American involvement in the war. Roosevelt took particular interest in the results of a July 1940 poll in which 37 percent of respondents thought the United States should help Britain, while 59 percent responded that the United States should avoid war altogether. At Roosevelt’s request, Cantril included the same question in subsequent polls, eventually developing a trend report for the president that he updated as new polls became available. As Cantril later recalled, “Nothing interested [Roosevelt] more than the trend charts,” especially those that showed a growing sense among Americans that involvement in the war was inevitable.149 By the time the United States entered the war in December 1941, White House requests for polling data had exceeded the resources available to Cantril either through the Rockefeller-funded Office of Public Opinion Research or through the continued generosity of George Gallup.150
Providence intervened in the form of a financial angel named Gerard Lambert. A Princeton graduate himself, Lambert made his fortune building his family’s business, which produced the popular mouthwash Listerine. In fact, it was Gerard’s talent for marketing, pitching Listerine as a cure for “chronic halitosis” (a term he invented), that transformed the brand into a mass-market phenomenon. By the 1920s, Lambert had amassed a vast fortune and retired to the stately mansion he built for himself and his family in Princeton, two miles from the university gates. It was there that Lambert met George Gallup, who in turn introduced Lambert to Hadley Cantril.151 Shortly after the Japanese attack on Pearl Harbor, Lambert and Cantril met to discuss the growing need for survey work on behalf of the president, and they agreed to establish a polling operation of their own, independent from the funding limitations of the Rockefeller Foundation or the field staff and facilities of the Gallup Organization. Their creation, the Research Council, Inc., operated out of Cantril’s Office of Public Opinion Research at Princeton on punch card equipment on loan from IBM. Within months, interviewers began polling a nationwide sample of 1,200 respondents. Lambert covered the costs. “At the end of each month,” Cantril recalled, “I sent Lambert’s New York office a report of the amount of money spent during the month, and a check was returned to the Research Council immediately.”152
In fact, Gerard Lambert was much more than just a benefactor. He also brought considerable experience in marketing and advertising, as well as early attempts to use public opinion data in the conduct of campaigns. After a disappointing tour in Washington working for the Federal Housing Administration, Lambert returned to Princeton, where he devoted his energies to defeating Roosevelt in the 1940 election. As Lambert recounted in his memoir, “My first step was to go to Dr. George Gallup. ... From him I borrowed some men and began to run small public-opinion polls of my own.”153 Specifically, Lambert tested different versions of political speeches to see how slight changes in their substance would affect their public appeal, much as he had done when perfecting advertising copy for Listerine. Sometime in 1939, Gallup introduced Lambert to Thomas Dewey, then the New York City district attorney, who asked Lambert to conduct opinion surveys in his bid for the 1940 Republican presidential nomination. Lambert agreed. Again with help from some of Gallup’s top assistants, Lambert set up polling operations for Dewey. After Dewey failed to garner the nomination, Lambert continued his public opinion work for the Republican nominee, Wendell Willkie, but he found the candidate to be uninterested in his advice. Lambert claimed that his pre-election polls predicted Willkie’s loss to Roosevelt by two percentage points.154
A little over a year later, with the United States at war, Lambert found himself back in Washington. Dedicating himself to the war effort, Lambert helped Cantril establish a polling operation that reported regularly to members of Roosevelt’s inner circle. According to historian Richard Steele, “The relationship [with the White House] grew as the war progressed, and although it was never Roosevelt’s only source of public opinion, it was probably his most important.”155 Their work ranged widely, including memoranda on public views toward lend-lease, governmental secrecy, and the progress of Allied forces on the Italian front. Cantril and Lambert probed domestic issues as well, such as what the public thought of full employment policies. Their services could be called upon at a moment’s notice, as when the White House wanted to gauge the reaction among Catholics to the possible bombing of Rome.156
Across the range of issues they examined, Cantril and Lambert often emphasized how public opinion polls recommended particular communication strategies by the president. In a December 1942 memorandum on production targets for war materiel, Cantril and Lambert advised “a policy of understatement and overperformance” when FDR communicated to the public.157 At the request of Sam Rosenman, FDR’s speech- writer, Cantril and Lambert poll-tested different versions of a radio address designed to explain federal subsidy programs to farmers. On other occasions, Cantril and Lambert used poll results to recommend a particular rhetorical style or tone. A February 1944 memorandum suggested that FDR would achieve greater support from the public if he used an upcoming radio address to ask for cooperation with his aims rather than criticize those who opposed him. Another time, Lambert and Cantril reported that “our studies show a wide desire on the part of the people to have the President admit a few human failings and to minimize sarcasm in his speeches.”158 In a message to Congress on the war’s progress shortly thereafter, Roosevelt acknowledged that “mistakes have been made,” but he also noted that, on the whole, the country had met its wartime challenges successfully.159 Time magazine cheered the speech, adding caustically that such admissions of fallibility from FDR “have the rarity of pearls in a restaurant oyster.”160
Cantril and Lambert took great care crafting their memos to the president. Both understood that the style in which they presented their findings would influence how their recommendations would be received. As Cantril himself noted, “A great deal of valuable material social scientists uncover or create is presented in so academic or slipshod a fashion that no busy person is going to waste time digging out what may be of significance.”161 As they prepared a report for the president, Lambert and Cantril reviewed and revised, paring down their memos to at most two or three pages. Major findings and conclusions were highlighted, and poll results were summarized with a graph or chart. Cantril recalled that “we always tried to remember that the President ... would completely lose interest if we became verbose or technical.”162 It was here that Lambert’s experience in business and advertising contributed to the collaboration by helping to convey detailed findings of various surveys in a simple, easily understood fashion. There was an art as well as a science to polling, especially when interpreting technical findings in politically meaningful ways.
In sum, polling offered a new source of information and a new kind of expert who could distill social scientific information in politically useful ways. Both challenged older sources of political intelligence. As Gerard Lambert reflected on his experiences as a political adviser, “This technique of discovering people’s attitude by public-opinion surveys is a controversial subject. In general those who are vociferous in opposing it have one thing in common. They derive their prestige or their living by expressing their own opinions.” Lambert focused much of his scorn on newspaper editors, columnists, and other “high priests of opinion.”
For these traditional interpreters of public sentiment, “a survey robs them of their oracular authority.” Polling, Lambert concluded, “is an encroachment upon their own domain.”163