WHY THE 1936 LITERARY DIGEST POLL FAILED

The Literary Digest poll of 1936 holds an infamous place inthe history of survey research. Despite its importance, no empiricalresearch has been conducted to determine why the poll failed.Using data from a 1937 Gallup survey which asked about participationin the Literary Digest poll I conclude that the magazine's sampleand the response were both biased and jointly produced the wildlyincorrect estimate of the vote. But, if all of those who werepolled had responded, the magazine would have, at least, correctlypredicted Roosevelt the winner. The current relevance of thesefindings is discussed.     

Question Order Effect and the Measurement of Candidate Preference in the 1982 Connecticut Elections

Conflicting poll results from the New York Times and the HartfordCourant concerning the 1982 Connecticut senatorial contest ledto the hypothesis that when voting preferences in two contestsare measured in the same poll, question order will affect candidatestandings. A subsequent poll, based on a split-sample design,verified this hypothesis. The relation of the question ordereffect to party identification, candidate preference, politicalideology, and education is also examined.     

EXPERIMENTS WITH THE MIDDLE RESPONSE ALTERNATIVE IN SURVEY QUESTIONS

A series of experiments shows (1) that people are significantlymore likely to select the middle response alternative on anissue when it is explicitly offered to them as part of the questionthan when it is omitted; (2) that merely mentioning that thereis a middle alternative, in the preface to the question, makesit more likely that respondents will select it, even thoughit is not offered to them as an explicit choice; (3) that theorder in which the middle alternative is presented in the question—inthe middle or last position—can make a significant differencein the results, but that a "recency bias" does not occur invariably,or consistently; and (4) that people who select a middle responsealternative when it is offered would not necessarily answerthe question in the same way as other respondents if forcedto choose sides on the issue.     

The State of the Union for Persons with Disabilities in 2004

ABSTRACT

The following provides a summary of a recent poll taken by Harris and commissioned by the National Organization on Disability (NOD). A summary is offered of each question explored in the poll. The findings suggest that people with disabilities continue to be treated as second class citizens.     

Testing the Permanence of the Permanent Campaign: An Analysis of Presidential Polling Expenditures, 1977 2002

Many presidential observers argue that the modern White Houseis the site of more-or-less permanent campaigning. In a recentPOQ piece, Murray and Howard (2002) [Public Opinion Quarterly66:527–558] explore one indicator of the "permanent campaign,"the extent to which Presidents Carter, Reagan, G.H.W. Bush,and Clinton commissioned independent opinion polls and focusgroups to assist in policymaking and political maneuvering.Murray and Howard suggest that while a sophisticated pollingoperation has been institutionalized in the White House, thereis substantial variation in how much a president uses this operation.In this article, we model presidential polling expendituresover time using monthly figures. We find that presidents donot vary significantly in the average amount spent per monthon polls. There are, however, two recurring patterns of variationwithin presidential administrations: Presidents tend to spendsignificantly more on internal polling during the most intensemonths of a presidential reelection campaign; and polling expendituresincrease over the course of each presidential term. These findingssuggest that there are common forces (e.g., elections, naturaldecline in support) that have driven all presidents since Fordto poll.     

Editor's Note

The masthead for this, the first issue of the 70^th year of publicationof Public Opinion Quarterly, lists some new members of the journal’sAdvisory and Editorial Boards. I want to welcome these scholarsof public opinion to the POQ family. Along with continuing membersof the boards, they perform essential roles in the operationof the journal, and we owe them a debt of gratitude for their     

SELF-INTEREST AND PUBLIC OPINION TOWARD SMOKING POLICIES; A REPLICATION AND EXTENSION

Numerous studies report that self-interest has a non significantinfluence upon various political and social attitudes. In contrast,a recent article by Green and Gerken (1989) reports a Californiastudy showing that cigarette smokers are significantly moreopposed than nonsmokers to public smoking restrictions and tobaccosales tax increases. The present article replicates and extendsthis analysis with data from two different states—Illinois(in 1984, N=458)—and North Carolina (in 1985, N=488)—andusing analysis techniques that differ from and expand upon theCalifornia study. Despite these methodological differ ences,self-interest is again shown to have a significant influenceon opinions concerning public smoking restrictions and taxation,as well as on several additional issues relating to smokingand tobacco.     

Can a Poll Affect Perception of Candidate Traits?

During presidential elections, poll results frequently are presentedin the news. Reporters use these polls to tell the public whatit thinks about the presidential candidates. We argue that pollingresults tell the public what it should think about the presidentialcandidates as well. This study outlines how a character traitthat is not usually used to assess presidential candidates wasput into play during the 2004 presidential campaign. By repeatedlyascribing "stubbornness" to incumbent president George W. Bush,Democratic challenger John Kerry may have prompted this trait’sinclusion in a Los Angeles Times summer 2004 survey. The poll’sevidence that the public saw Bush as more stubborn than Kerrythen produced an attribute agenda-setting effect that strengthenedthe link between that term and Bush. Using data from the NationalAnnenberg Election Survey, we argue that the news coverage ofthis Los Angeles Times poll increased the salience of the trait"stubborn" in assessing President George W. Bush during Juneof the 2004 presidential campaign.     

Editor's Note

This issue inaugurates AAPOR’s relationship with OxfordUniversity Press as publisher of Public Opinion Quarterly. During2003,     

Trade unionism in Canada and the United States: a reply to Bowden

Gary Bowden's (1989) comprehensive report on opinion poll questions bearing on Canadian trade unionism is presented in the context of a criticism of my earlier comparative discussion of the reasons why union density in Canada is much higher than in the United States. He contends that my emphasis on value differences in trying to account for the greater strength of unionism (and socialism) in Canada is wrong since, countering my expectations, survey findings do not demonstrate that Canadians are more sympathetic to unions than Americans. These results, however, do not surprise me or serve as a basic challenge to my analysis (Lipset, 1986). In my recent effort to evaluate the many differences in institutions and values in the two countries, I, too, indicate:
Curiously, given the large difference in union strength across the border, questions inquiring about attitudes toward unions, their power, and the outcomes of strikes that have been posed by the Gallup polls, Decima and CRI, Harris, NORC, and the Wright-Myles studies do not show Canadians as more pro-union than Americans. Some even find the population in the north less approving of unions and more concerned about union power, which in fact is much greater there than in the northern [sic] nation. (Lipset, 1989; 1990a: 69). [I must apologize here for an obvious error, northern clearly should have read southern].     

Job Mobility of Chippewa and Poor White

Abstract

Using a 1978 national Gallup poll, we test the hypothesis that social marginality is a predictor of belief in paranormal phenomena. Being female and not being married generally correlate as hypothesized with paranormal belief, but age, low education, being Black, and being unemployed generally do not. The marginality hypothesis has a questionable theoretical foundation and fails to predict a variety of beliefs consistently.     

FAMILY SIZE IDEALS OF CANADIANS: A METHODOLOGICAL NOTE*

A partir de données recueillies par le sondage d'opinion Canadian Gallup sur le nombre d'enfants par famille considéré comme idéal, cet article suggère que la moyenne de 4.2 enfants par famille, statistique de 1960 fréquemment citée, est faussement exagérée. L'auteur compare les données du sondage de 1960 à celle d'un sondage antérieur (1957). Elle compare aussi les données de récents sondages canadien et américain sur la dimension idéale de la famille. Using data from Canadian Gallup Poll questions on the number of children considered ideal, this paper suggests that the frequently cited 1960 figure of 4.2 children is erroneously inflated. The discussion compares responses in 1960 to responses of an earlier 1957 poll. Comparisons are also made between American and Canadian responses to recent questions on ideal family size.     

Likely (and Unlikely) Voters and the Assessment of Campaign Dynamics

Only in recent years has the "likely voter" technology beenextended to polls well in advance of an election. In the caseof the 2000 U.S. presidential election, CNN/USA Today/Galluptracking polls indicated considerable fluctuations in likelyvoter preferences, greater than among the larger pool of registeredvoters surveyed. This article explores how Gallup’s likelyvoter model exaggerates the reported volatility of voter preferencesduring the campaign. Much of the reported variation in candidatepreference reported by Gallup in that election is not due toactual voter shifts in preference but rather to changes in thecomposition of Gallup’s likely voter pool. The findingshighlight dangers of relying on samples of likely voters whenpolling well before Election Day.     

A User's View of the Polls

The treatment of polls and surveys by the New York Times haschanged from casual to extensive use, both in its internal decision-makingand its political and social reporting, with many obvious benefits.But some implications of poll proliferation, especially theirdistorting effects on the political process, raise troublingthoughts.     

Response-Order Effects and the Early Gallup Split-Ballots

A meta-analysis of split-ballots conducted by the Gallup Organizationin the 1930s, 1940s, and early 1950s shows that response-ordereffects were generally small in magnitude when averaged acrossa great variety of topics and questions—and as comparedwith many of those reported in the response-effects literaturetoday. When analyzed by various question characteristics, theresults provided some support for predictions derived from currentcognitive models of response-order effects, particularly thosefrom satisficing theory. As predicted, questions asked orallywere more likely to generate a statistically significant patternof recency effects if the response alternatives or the questionsas a whole were longer rather than shorter. Other predictedpatterns of primacy and recency effects failed to materialize,however, perhaps largely because of the inherent design limitationsand partial confounding of question attributes in any such secondaryanalysis of archival survey data, but perhaps, too, becauseof simple chance variations. The data from these early experimentsnonetheless provide a partial, albeit limited, test of rivalhypotheses and explanations of response-order effects in theliterature.     

AAPOR Award

TThe American Association for Public Opinion Research Presentsthe AAPOR AWARD For Exceptionally Distinguished Achievementto Andrew Kohut As a brilliant media commentator on the nature and interpretationof poll findings, Andy Kohut has become the public face of opinionresearch to millions of Americans. Andy is     

The Polls Trends: Public Reactions to Global Health Threats and Infectious Diseases

Over the past two decades, newly emerging infectious diseaseshave developed into major global health concerns, sparking intensemedia coverage, and triggering fears of a global outbreak amongpublic health experts and authorities. This article focuseson trends in American attitudes toward these newly emerged infectiousdiseases by analyzing poll data over the past 6 years aboutissues relating to avian flu, severe acute respiratory syndrome,West Nile virus, and anthrax. The polls show that Americans’attention to news coverage seemed to be event driven, peakingwhen there were new human or animal cases, and decreasing rapidlywhen the diseases seemed to have been contained. Americans’perceptions of threats were usually the highest in the earlystages of major outbreaks. The public became more complacentwhen the outbreaks seemed to be under control. Both behavioralchanges and general knowledge remained largely constant, suggestinga limited impact of the various informational and awarenesscampaigns by governmental agencies in the wake of these pandemics. Abbreviations: ABC, ABC News • ABC/WP, ABC News/Washington Post • AP/IPSOS, IPSOS-Public Affairs for Associated Press • CBS, CBS News • CBS/NYT, CBS News/New York Times • GALLUP, Gallup Organization • GALLUP/CNN/USA, Gallup Organization for CNN/USA Today • HARRIS/TIME/CNN, Harris Interactive for Time and CNN • ICR/HARVARD, International Communications Research for Harvard School of Public Health Project on the Public and Biological Security • ICR/HARVARD-RWJF, International Communications Research for Harvard School of Public Health Project, Robert Wood Johnson Foundation • IPSOS/DC, IPSOS-Reid for Dittus Communications • NORC, National Opinion Research Center • OD/FOX, Opinion Dynamics for Fox News • PSRA/KAISER, Princeton Survey Research Associates for Henry J. Kaiser Family Foundation, Harvard School of Public Health • PSRA/NEWS, Princeton Survey Research Associates for Newsweek • PSRA/PEW, Princeton Survey Research Associates for Pew Research Center • PSRA/PMFMI, Princeton Survey Research Associates for Prevention Magazine/Food Marketing Institute     

GALLUP AND FORTUNE POLLS

This section contains a compilation, topically arranged, ofpoll results released by the American Institute of Public Opinion,by Fortune, and by the Canadian Institute of Public Opinion.The AIPO results cover the period from October through December1941. (Previous aipo questions were reported in the July 1938,October 1939, and all subsequent issues of the QUARTERLY.) TheFortune questions are those which appeared in the October andDecember issues of the magazine, together with those taken fromthe Fortune Management Poll which appeared in the November issue.(Previous Fortune questions were reported in the March 1940and all subsequent issues of the QUARTERLY.) Releases from theCanadian Institute of Public Opinion are included for the firsttime. (The first cn>o report was issued November 29, 1941.) Under each topic, all of the Institute data are given in chronologicalorder, then all of theFortune material, also in chronologicalsequence. Dates appearing in connection with AIPO questionsare those carried in the date lines of Institute releases tosubscribing newspapers; dates following Fortune questions indicatethe issue of the magazine in which the information appeared.Institute questions are designated by AIPO; Fortune questionsby FOR; "DK" stands for "don't know"; "no op." for "no opinion." In considering these poll data, the reader should bear in mindcertain salient points of reference set forth on pages 75 and76 of the March 1940 issue of the QUARTERLY. The QUARTERLY wishesto express its appreciation to George Gallup and the AmericanInstitute of Public Opinion and to the editors of Fortune andElmo Roper for their cooperation in making these survey resultsavailable in convenient form to other students of public opinion.