Published by CQ Press, an Imprint of SAGE Publications, Inc. www.cqresearcher.com Political Polling Do polls accurately measure public attitudes?

mart phones, social media and the Internet have made it easier than ever for people to make their views known, but the new technology can make it S harder for political pollsters to gather and measure public opinions with precision or consistency. They face public suspicions of partisanship, reluctance to provide candid answers and — as cellphone use grows — difficulty reaching respondents by the traditional method of random calls to household landlines. meanwhile, critics charge that the pay too much attention to “horse-race” polls showing who leads in political races Statistician predicted would win the 2012 presidential election in his and not enough to candidates’ policy ideas. The 2014 elections, FiveThirtyEight.com , which aggregates polls rather than relying on a single survey. Critics say in which pollsters miscalled the results of a number of closely media and pollsters put too much emphasis on the so- called “horse race” aspect of election campaigns and not enough on substance. watched races, cast a harsh spotlight on the industry, but pollsters contend their record has improved over the years. Some experts see promise in the increasing use of “opt-in” polls such as those I on the Internet, but the approach is controversial. THIS REPORT N THE ISSUES ...... 123 S BACKGROUND ...... 130 I CHRONOLOGY ...... 131 D CURRENT SITUATION ...... 136 E CQ Researcher • Feb. 6, 2015 • www.cqresearcher.com AT ISSUE ...... 137 Volume 25, Number 6 • Pages 121-144 OUTLOOK ...... 139 RECIPIENT Of SOCIETY Of PROfESSIONAL JOURNALISTS AwARD fOR BIBLIOGRAPHY ...... 142 EXCELLENCE u AmERICAN BAR ASSOCIATION SILvER GAvEL AwARD THE NEXT STEP ...... 143 POLITICAL POLLING

Feb. 6, 2015 THE ISSUES SIDEBARS AND GRAPHICS Volume 25, Number 6 MANAGING EDITOR: Thomas J. Billitteri • Do polls show a persistent Pollsters Finding House - [email protected] 123 political bias? 124 holds Harder to Reach ASSISTANT MANAGING EDITORS: maryann • will and facebook fewer homes have landlines. Haggerty, [email protected], Kathy Koch , [email protected], supplant conventional polling? Scott Rohrer, [email protected] • Does the media’s avid inter - 125 Most See Bias est in polls hurt democracy? in Polls SENIOR CONTRIBUTING EDITOR: Three-fourths of adults say Thomas J. Colin polls are slanted. [email protected] BACKGROUND CONTRIBUTING WRITERS: Brian Beary, 126 It’s Not the Last Straw for marcia Clemmitt, Sarah Glazer, Kenneth Jost, Early Polling Iowa’s Famous Poll Reed Karaim, Peter Katel , Barbara mantel, 130 GOP still embraces it, but Tom Price, Jennifer weeks The first political poll was an critics say it’s a platform for 1824 survey by a Harrisburg, SENIOR PROJECT EDITOR: Olu B. Davis Pa., newspaper. extremists. EDITORIAL ASSISTANT: Ethan mcLeod Spending on Political Dewey vs. Truman 128 INTERN: Sarah King 133 In 1948, pollsters erroneously Polls Rising predicted that Republican Total spending on political FACT CHECKERS: Eva P. Dasher, polling by Republican and michelle Harris, Nancie majkowski Thomas Dewey would defeat Democratic party committees President Harry S. Truman. has risen over the past decade in both presidential Exit Polling 133 media missteps involving exit and midterm elections. polls occurred in 1980 and An Imprint of SAGE Publications, Inc. 1996. 131 Chronology Key events since 1824. VICE PRESIDENT AND EDITORIAL DIRECTOR, HIGHER EDUCATION GROUP: ‘Push Polls’ Cast Negative michele Sordi CURRENT SITUATION 132 Light on Candidates Polling groups say the practice EXECUTIVE DIRECTOR, ONLINE LIBRARY AND REFERENCE PUBLISHING: Presidential Race is manipulative. 136 Experts say pollsters will play Todd Baldwin a prominent role in the 2016 At Issue: presidential election. Copyright © 2015 CQ Press, an Imprint of SAGE Pub - 137 Can “nonprobability” polls lications, Inc. SAGE reserves all copyright and other reliably reflect public rights herein, unless pre vi ous ly spec i fied in writing. Internet Polling opinion? 136 Pollsters are debating the No part of this publication may be reproduced reliability of opt-in Internet electronically or otherwise, without prior written surveys. permission. Un au tho rized re pro duc tion or trans mis - FOR FURTHER RESEARCH sion of SAGE copy right ed material is a violation of Robot Interviewers federal law car ry ing civil fines of up to $100,000. 138 For More Information Some pollsters are consider - 141 Organizations to contact. CQ Press is a registered trademark of Congressional ing using digital interviewers. Quarterly Inc.

Bibliography CQ Researcher (ISSN 1056-2036) is printed on acid-free 142 Selected sources used. OUTLOOK paper. Pub lished weekly, except: (march wk. 4) (may wk. 4) (July wk. 1) (Aug. wks. 3, 4) (Nov. wk. 4) and The Next Step (Dec. wks. 3, 4). Published by SAGE Publications, Inc., Shades of Gray 143 Additional articles . 139 No consensus exists about 2455 Teller Rd., Thousand Oaks, CA 91320. Annual full - the future of political polling. Citing CQ Researcher service subscriptions start at $1,131. for , call 143 Sample bibliography formats. 1-800-818-7243. To purchase a CQ Researcher report in print or electronic format (PDf), visit www.cqpress. com or call 866-427-7737. Single reports start at $15. Bulk purchase discounts and electronic-rights licensing are also available. Periodicals postage paid at Thousand Oaks, California, and at additional mailing offices . POST - mAST ER: Send ad dress chang es to CQ Re search er , 2600 virginia Ave., N.w., Suite 600, wash ing ton, DC 20037.

Cover: Getty Images/Bloomberg/David Paul Morris

122 CQ Researcher Political Polling BY CHUCK MCCUTCHEON

A variety of people and THE ISSUES institutions conduct political polls, often attached to par - efore November’s elec - ticular candidates and parties, tions, many political and most of their surveys are B pollsters thought they kept secret. Others, whose had the races nailed: Incumbent work is made public, are af - Sen. , D-va., was filiated with colleges, think a shoo-in for re-election to the tanks and news outlets. Poll -

U.S. Senate, and ’s con - n sters also help interest groups w o tests for Senate and governor r and political donors hone their B

b

would be so close that runoff o messages or decide to whom B

elections would be needed. / they should give money. h c Those predictions turned t In his 2012 re-election cam - a p s

out to be very wrong. warner i paign, President Barack D -

sweated out a win by less s Obama’s strategists created an e

than 1 percentage point, while m “analytics department” that i T

Republicans prevailed by such d was five times larger than the n

large margins in Georgia that o one Obama used in 2008. The m h

no runoffs were necessary. c department combined statis - i R

Polls in several other states / tics from pollsters with ma - o t

also were off the mark. o terial from other sources, in - h P

“Boy, is that an industry cluding social-media sites such P

that needs some houseclean - A as Twitter. It used that infor - ing,” University of virginia po - Sen. Mark Warner, D-Va., debates Republican strategist mation to run tests predicting Ed Gillespie, left, during the 2014 Virginia Senate race. litical scientist Larry Sabato Polls predicted Warner would easily win re-election, but how people could be influ - said the morning after the he just squeaked by. In another dramatic polling miscue enced by certain appeals. election. 1 — also in Virginia — polls failed to anticipate House Obama’s campaign also The November results — Majority Leader Eric Cantor’s stunning primary defeat by used polls to develop a so - along with polls that failed a virtually unknown college professor. Despite their phisticated understanding of failures, polls are expected to play a key role in the to anticipate House majority 2016 presidential race and other contests. who would vote in key states. Leader Eric Cantor’s GOP pri - In Ohio alone, it collected mary loss in virginia in June polling data from about 29,000 to a virtually unknown college professor Political strategist Patrick Ruffini, a people — an enormous sample that — dramatically illustrated the challenges former technology official at the Re - enabled the campaign to study Obama’s confronting pollsters. Even as technol - publican National Committee (RNC), appeal across a wide demographic ogy has made it easier than ever for has a more dismal view. “If anything,” spectrum. 4 people to give their opinions, consis - he said, “the challenge for polling right The news media make a big deal tently measuring those opinions with now is to keep it from getting any out of political polls. Surveys showing pinpoint precision has grown exceed - worse than it is.” 2 a president’s job-approval rating, or ingly difficult. with the 2016 presidential Polling has become entangled in the results of polls on a controversial race beginning to unfold, the stakes the nation’s prevailing polarized po - topic such as the , are high for pollsters: Their work plays litical climate, with both politicians become prominent news. And polls a crucial role in influencing the flow and the public questioning the validity draw intensive coverage in the months of campaign donations and determining of polls. At a 2011 rally in Iowa, 2008 and years leading up to elections. overall public support for candidates. GOP vice presidential nominee Sarah Election polls require assessments Political polling “is in a state of tran - Palin colorfully disparaged surveys of what groups will show up to vote sition, certainly,” says michael Link, showing that many Americans regarded and in what proportion to other groups. president of the American Association her unfavorably: “Polls? . . . They’re “Election polls are a special breed for Public Opinion Research (AAPOR), for strippers and cross-country skiers,” among public-opinion surveys,” said the field’s professional association. she declared . 3 Cliff Zukin, a Rutgers University public

www.cqresearcher.com Feb. 6, 2015 123 POLITICAL POLLING

matic rather than person-to-person dialing , Pollsters Finding Households Harder to Reach have proliferated, and federal regulations The percentage of households providing usable interviews for Pew bar most robopolls from calling cell Research Center polls conducted via landline telephones fell from phones, a limitation that can end up omitting voters. 36 percent to 9 percent between 1997 and 2012. Success rates Another new type of polling, Internet- for contacting households in which an adult was reached also based opt-in “nonprobability” polls, dropped, from 90 percent to 62 percent. have begun to flourish and are drawing controversy. * (See “At Issue,” p. 137. ) Households Responding to Pew Research Surveys Politicians and interest groups employ (Percentage responding) such opt-in polls to engage followers. 100% But the questions can be loaded. The 80 website of virginia GOP Rep. morgan 60 Griffith, for example, recently featured an “issue survey” asking questions such as, 40 “Do you support or oppose the Envi - 20 ronmental Protection Agency enacting 0 new regulations that make it harder to 1997 2000 2003 2006 2009 2012 use coal as an energy source, killing local Source: “Assessing the Representativeness of jobs and driving up electricity rates?” 9 Households contacted Public Opinion Surveys,” Pew Research Center, Households providing In 2008, long after House Democratic May 15, 2012, http://tinyurl.com/7e8gjrr usable interview leaders refused to consider some liberal Democrats’ proposal to impeach GOP policy professor. “They call for more principle of phone polling. Probability President George w. Bush over the judgments — the art rather than science sampling is based on the concept that decision to invade Iraq, then-California of the craft — on the part of the for the universe of people being sur - Democratic Rep. Pete Stark — a liberal pollster than other types of polls.” 5 veyed, each member has a defined and outspoken Bush critic — drew The isn’t the only place and equal likelihood of being selected admiration from anti-Iraq war groups where election polling can be difficult. to participate in the survey. when he asked in a survey on his After September’s referendum on “The polling field has been address - House website if voters backed im - whether Scotland should become in - ing developments in technology that peachment. 10 dependent from the , have made it harder and harder to Griffith’s and Stark’s polls weren’t analysts noted the final “no” vote was reach people,” says Tim vercellotti, di - intended to be scientific. But pollsters by a much wider margin than polls rector of western New England Uni - say scientific nonprobability polls con - had predicted. 6 versity’s Polling Institute in Springfield, ducted online increasingly are worthy Telephone surveys historically have mass. “It began with [the advent of] of consideration because pollsters can been the main means by which pollsters answering machines, then caller ID, perform statistical adjustments, known gathered public opinion about candi - then cell phones.” The latter “really as “weighting,” to compensate for those dates, issues and consumer products. sort of cut the knees out from under who are not Internet users. But pollsters of all types have had to the industry,” he said, because cell In the old days of landline phone cope with dramatic changes in tele - users are less willing to take calls from polls, pollsters did not need to make phone use. In 2009, the percentage of unfamiliar numbers. such adjustments very often. “when homes without a landline was less than vercellotti and other pollsters say the everyone had a landline and answered 25 percent; by July 2014, it exceeded industry is doing better at incorporating it . . . even a no-name call center could 40 percent. 7 cellphones. But their increasing use has [dial] random digits and get an arguably when nearly all Americans had a come as public response rates to polling landline, pollsters had a relatively easy questions have plummeted, from 36 per - * Nonprobability polls are those in which the par - time designing surveys. These days, cent in 1997 to just 9 percent in 2012, ticipants choose to participate themselves or are however, a “dual-frame sample” — according to the Pew Research Center chosen for their particular demographic attributes, 8 which combines landlines and cell - on the People & the Press . rather than just being randomly selected. Such phones — provides the closest to a In addition, computer-executed polls, polls include so-called “opt-in” polls in which people true probability sample, the bedrock dubbed “robopolls,” that rely on auto - volunteer — generally online — to take part.

124 CQ Researcher representative sample of adults with a tember, the site’s polls put Republican reasonable response rate,” says mark Most Americans See up by more than 7 per - Blumenthal, The Huffington Post ’s senior Bias in Polls centage points at a time when regular polling editor. “media and academic or - polls showed Obama ahead by several ganizations could do even better. very Three-fourths of American percentage points. 16 Romney eventu - little weighting was required.” adults surveyed recently said ally lost by nearly 4. weighting has been shown to im - public opinion polls are biased, In 2004, Democrats were the ones prove the quality of polls. Pew con - with more than half saying they complaining about biased polls. when cluded in a 2012 study that telephone are slanted “in some way.” a Gallup poll showed Democratic pres - surveys that include landlines and proper idential contender John Kerry trailing Pres - weighting “continue to provide accurate ident Bush, the liberal group moveOn.org data on most political, social and eco - Percentage of Respondents took out a full-page ad in The nomic measures.” 11 But if weighting Who Said Polls Are Biased* Times blasting what it called Gallup’s assumptions prove wrong in either “flawed methodology,” adding, “This is 57% phone or Internet polls, they can skew more than a numbers game. Poll results accuracy. Some experts say excessive profoundly affect a campaign’s news cov - and improper weighting plagued too erage as well as the public’s perception many 2014 election surveys. of the candidates.” 17 “The problem, from my point of 22% Some Democrats again criticized the view, is that there’s just so much weight - polls in 2014, before Republicans made 11% ing going on,” Stuart Rothenberg, a 7% 5% large gains in the congressional elec - political forecaster and columnist, said tions. Democratic strategist Brent Bu - at a November forum on polling. Toward Toward In Not Don’t dowsky wrote in the Capitol Hill news - Scientific polls also have become more conser - liber - some biased know paper : “There are so many expensive. Statistician Nate Silver, founder vatives als way razor-thin Senate races that confident of the popular FiveThirtyEight.com blog predictions of which party holds Senate about polls, said the cost of a top-quality * Totals do not add to 100 percent control are, to paraphrase a line from poll “can run well into five figures, and due to rounding. Survey based on Jack Nicholson in “Chinatown,” wind it has increased as response rates have telephone interviews with 562 adults. from a duck’s derriere.” 18 declined.” 12 Source: “Kantar’s ‘Path to Public But people who study polling say Pollsters maintain that their mistakes Opinion’ Poll,” Kantar Group, Sept. 4, that while polls have sometimes under - have been exaggerated and that polling 2013, http://tinyurl.com/kbq8jgk estimated one party’s support over an - accuracy has demonstrably improved. other’s, it’s not because of partisanship. Eric mcGhee, a research fellow at the Do polls show a persistent politi - “There have been years, like 1980 Public Policy Institute of California, a cal bias? and 1994, when the polls did underes - nonpartisan think tank, concluded in for years, politicians and activists timate the standing of Republicans,” said October 2014 that polling in Senate across the political spectrum have ar - Silver, who has written repeatedly about contests has become so much more gued that too many polls are slanted the issue. “But there have been others, accurate since 1990 that “looked at a in favor of the opposition. A 2013 like 2000 and 2006, where they under - certain way,” today could be considered survey by the London-based research estimated the standing of Democrats.” 19 “a golden age of polls.” 13 firm Kantar found three-quarters of In a subsequent article, Silver ex - Pollsters say their surveys get unfairly Americans say most polls are “biased amined Senate polls between 1990 and blamed when candidates lose or are badly toward a particular point of view.” 14 2012 and found the GOP underper - trailing their opponent. “People expect (See graphic, above. ) formed at the ballot box by less than polls to do too much,” says Democratic Before the 2012 presidential election, one-half of 1 percent over what the pollster Anna Greenberg. “Polls can’t pre - some Republicans alleged that polls polls had showed. “Democrats might dict turnout, or measure the effectiveness were stacked against them. GOP blog - argue that a Republican bias has been of a field operation or campaign.” ger Dean Chambers started a website, evident in recent years — even if it As pollsters, academics, the news UnskewedPolls.com, to remove what hasn’t been there over the longer term,” media and others debate political he charged was the deliberate inclusion he wrote. “But the trend is nowhere polling, here are some of the questions of too many Democrats in a projection close to statistically significant. Nor in being raised: of the voting population. 15 In Sep - the past has the direction of the bias

www.cqresearcher.com Feb. 6, 2015 125 POLITICAL POLLING

It’s Not the Last Straw for Iowa’s famous Poll GOP still embraces it, but critics say it’s a platform for extremists.

ne of the nation’s best-known political polls — the moderate, electable candidates to prevail in the Iowa caucus Ames Straw Poll, also called the Iowa Straw Poll — and beyond.” 2 O will be held again this summer after surviving an Iowa Republican Gov. Terry Branstad said last year the poll effort last year to scrap it. had outlived its usefulness and called for ending it. He and first held in 1979, it is a poll of Republican presidential some other Republicans said they were concerned it could candidates taken every four years at an August fundraising event jeopardize the state’s ability to hold the february 2016 political in Ames a year before the presidential election. It is open only caucuses, the first regular voting event of the presidential to Hawkeye State residents regardless of whether they are reg - campaign. 3 istered Republicans. They cast ballots after participating candidates In December, however, a majority of the Iowa GOP’s Central address the audience. 1 Committee favored keeping the straw poll. They asked the Re - The poll — run by the state GOP — has no effect on publican National Committee (RNC) whether the poll violates the actual primary process, but campaigns and the political new RNC guidelines about holding voting events prior to the news media have come to regard it as a critical test of a caucuses. white House aspirant’s organizational ability and grassroots The RNC’s general counsel, John Ryder, said in January the support. poll did not violate those guidelines and could continue as Some critics of the poll say it encourages extremist elements long as it isn’t promoted as an “official event.” 4 The Central of the Republican Party, and some moderate Republicans have Committee then voted unanimously to keep the poll. 5 opted in the past not to take part in it. “Nationally, the Republican The poll’s defenders contend it puts Iowa on the national Party has been hijacked by extreme factions, making it difficult political stage early and brings in money for the state party. to appeal to the emerging voting majorities across the country, “The only thing that we need to scrap is the talk about which is necessary to win the white House,” said James Strohman, there not being a straw poll,” said Iowa U.S. Rep. Steve King, an Iowa State University political science lecturer. By appealing an influential figure in the state’s GOP presidential politics. 6 to the party’s extremists, he said, the poll “makes it difficult for In a subsequent letter to supporters before Ryder’s decision

in the previous election cycle been a 2008 campaign who now runs his own Overall poll numbers can be affected good way to predict what it will look polling company. when candidates publicly release private like in the next cycle.” 20 Last August, Silver’s blog found pos - polling data only when it makes them Even if pollsters add more voters sible examples of nontraditional, opt- look good, pollsters say, even if such from one party to a sample, they say in polls done over the Internet adjusting so-called “outlier” polls don’t reflect the those additions are relatively small from their results to reflect what the website true state of the race. Those polls can a statistical standpoint. described as higher-quality “gold-stan - be folded into political forecasts based “A sample that has 2 percentage points dard” polls. “None of this proves guilt, on an aggregation of multiple polls, more in one survey, Democrat, and an - but it does raise the possibility some skewing the overall result. “we don’t other state has a Republican a point pollsters may be peeking at their neigh - release intentionally misleading surveys,” more, this is all very much within what’s bors’ papers,” it said. 22 says Greenberg, who has done polling called the error margin,” said Lee miringoff, Other pollsters, however, say rep - for Democrats, including New York director of the marist Institute for Public utable members of their profession are mayor Bill de Blasio, Pennsylvania Gov. Polling in Poughkeepsie, N.Y. 21 far more interested in being accurate Tom wolf and several congressional can - Some polling experts cite a phe - than in pleasing one party or the other didates. “If we get an outlier we will nomenon known as “herding,” or ad - or others in their industry. “we all want discourage our candidates from releasing justing one’s data to make it conform to get it right,” vercellotti says. “Con - it when it’s probably not correct.” to the accepted norm in other election scientious survey researchers go through polls, especially if one’s original results this pretty regularly when doing polls: Will Twitter and Facebook sup - are radically different from other polls. You look at your estimates and ask plant conventional polling? “Some pollsters think, ‘It can be eas - yourself, ‘Did I do everything the way University sociologist fabio ier to be wrong with the crowd than I’ve been trained, and is the method - Rojas caused a stir in polling circles in it is to stick your neck out,’ ” says ology sound?’ If the answer is yes, 2013 when, in a Washington Post column Stefan Hankin, a pollster for Obama’s you’ve got to put it out there.” titled “How Twitter Can Predict an Elec -

126 CQ Researcher was announced, King said conservatives “should be outraged “I suspect many candidates in 2016 will make the choice that establishment Republicans are working in secret to end Romney made in 2012,” said Phil musser, a Republican political the Ames Straw Poll.” 7 strategist who served as an adviser to Pawlenty. 10 Some candidates go to great lengths to try to influence the poll. In 2011, Texas Rep. Ron Paul paid $31,000 to buy the — Chuck McCutcheon most-desired booth space outside the convention center in which the vote took place. 8 He ended up narrowly losing to 1 Aaron Blake, “The Ames Straw Poll, Explained,” , Aug. 8, Rep. michele Bachmann. 9 2011, http://tinyurl.com/3ragg49. 2 James Strohman, “Guest Column: Iowa’s GOP Should End Straw Poll,” Strongly conservative Republicans usually have prevailed in IowaStateDaily.com , Sept. 18, 2014, http://tinyurl.com/kzjpbre. the poll over moderates. In addition to Bachmann, poll winners 3 Catherine Lucey, “Branstad Seeks to End Iowa GOP Straw Poll,” The on the far right who subsequently failed to gain political traction , Dec. 15, 2014, http://tinyurl.com/k5y4twb. were television evangelist Pat Robertson in 1987 and Texas Sen. 4 James Hohmann, “RNC Gives Iowa GOP a Green Light for Straw Poll,” Phil Gramm in 1995. , Jan. 8, 2015, http://tinyurl.com/nfwwrdo. 5 James Hohmann, “Iowa Republicans vote to Continue Ames Straw Poll,” A poor showing, on the other hand, has led some candidates Politico , Jan. 9, 2015, http://tinyurl.com/mf3m2v7. to abandon their races. former Tennessee Gov. 6 Emily Schultheis, “will the Iowa Straw Poll Survive?” , dropped out after a fifth-place finish in 1999, while ex-minnesota Aug. 9, 2014, http://tinyurl.com/nutumvy. Gov. Tim Pawlenty withdrew after taking third in 2011. 7 “King: Some Republican leaders want to end Iowa Straw Poll to silence grassroots conservatives,” Storm Lake Pilot Tribune , Dec. 29, 2014, The poll involves only a minority of voters: about 17,000 http://tinyurl.com/kofg5g6. Iowans voted in the straw poll in 2011, compared to the roughly 8 Blake, op. cit. 120,000 who took part in the . Republican mitt 9 “Ames Straw Poll: michele Bachmann Beats Ron Paul — By 152 votes,” Romney — the party’s eventual nominee — chose not to par - , Aug. 13, 2011, http://tinyurl.com/lgg37mh. ticipate in the straw poll, believing he had more of a chance 10 Lucey, op. cit. to have an impact elsewhere.

tion,” he boasted that social media even - dicted the winner in 404 of 435 races, majority of race outcomes,” he said in tually would put pollsters out of work. a success rate of nearly 93 percent. 23 a column. “In most cases, all you need “we no longer passively watch our Researchers at the University of to know is incumbency (or the district’s leaders on television and register our -Amherst noted that in political bent) and the candidates’ parties opinions on Election Day. modern pol - nine tossup Senate races in 2012, the to predict who will win. ” 26 itics happens when somebody com - candidate with the most facebook fol - Steve Koczela, president of Boston’s ments on Twitter or links to a campaign lowers won all but one. 24 In 2010, massINC Polling Group, agreed with through facebook,” Rojas wrote. “In meanwhile, the candidate with the Rothenberg. “Social media changes too our hyper-networked world, anyone greater number of facebook followers fast to give me much confidence that can say anything, and it can be read won in 71 percent of Senate elections, models which worked in 2010 will by millions. This new world will under - and in some cases the social-media work in 2016 or 2020,” he said. 27 mine the polling industry.” site was found to be a better predictor facebook has had some major election Along with several colleagues, Rojas of election results than how much failures. In 2010, Republicans Christine conducted a study concluding that Twit - money candidates raised and spent. 25 O’Donnell, Sharron Angle and meg whit - ter discussions “are an unusually good However, those in the polling in - man amassed many more online followers predictor” of the outcome of U.S. House dustry say the Indiana researchers’ work than their opponents did in the races elections. Drawing on an archive of contained misguided assumptions. Po - for Delaware senator, Nevada senator billions of randomly sampled tweets, litical forecaster Rothenberg noted that and California governor, respectively, but they extracted more than 542,000 that just about 100 of the 406 House races still lost. And in those elections, facebook mentioned a Democratic or Republican in 2010 were seriously competitive. and Twitter were less accurate at pre - candidate in 2010. They found a “strong “Since House re-election rates have dicting who would win the support of correlation” between a candidate’s been over 90 percent in 19 of the past younger voters. 28 “tweet share” and the final two-party 23 elections, you don’t need polls or more recently, two potential 2016 vote, and said that share correctly pre - tweet counts to predict the overwhelming white House aspirants, Democrat Hillary

www.cqresearcher.com Feb. 6, 2015 127 POLITICAL POLLING

Does the media’s avid interest in Spending on Political Polls Rising polls hurt democracy? Some in the political world say the Total spending on political polling by Republican and Democratic news media’s voracious appetite for polls party committees has risen over the past decade during both distorts the political process. much of the presidential (2004, 2008, 2012) and midterm election years (2006, criticism centers on “horse race” journalism, 2010, 2014). Party committees spent almost the same amount on which concentrates more on which can - polling in the 2014 midterms, but in the 2012 presidential year didates are trending up or down than Democrats outspent their GOP counterparts by nearly $10 million. on the issues of the campaign. Polls are often the leading source of Spending on Political Polling, information in horse race articles, even

(in $ millions) by Party, 2004-2014* though they usually lack context, ob - $50 servers say. “The problem is, journalists $42.7 look at the poll and think that’s the 40 story, rather than going out and looking $29.7 at what is really the story,” says Susan 30 $21.0 Pinkus, a former Los Angeles Times polling $17.7 20 $16.9 director who is now a public opinion $9.6 and strategic analysis consultant. 10 Tom Rosenstiel, founder and director 0 of ’s Project for Ex - 2004 2006 2008 2010 2012 2014 cellence in Journalism, agreed that cur - rent political journalism “has increased * Includes spending by national, state and local party commit- Democrat the tendency to allow polls to create tees and congressional campaign committees; data exclude Republican senatorial campaign committees a context for journalists to explain and TOTAL organize other news — becoming the Source: Center for Responsive Politics lens through which reporters see and order a more interpretative news en - Rodham Clinton and Republican Ted unconstitutional; but it was more con - vironment.” The dependence on horse Cruz, have dominated social media. Politi - servative than the overall public reaction race polls, he added, has “reinforced co determined in December 2014 that to Obama’s State of the Union and these tendencies, and further thinned over the previous three months, the for - second inaugural speeches. 31 the public’s understanding toward who mer secretary of State and the Texas Although polling experts say social won and away from why.” 32 senator together accounted for 40 percent media remain too imprecise to supplant The Internet and social media, with of facebook discussions and 47 percent conventional polling, it could be helpful their insatiable demands for new ma - of mentions on Twitter among the 10 in augmenting such polling, they say. terial, have made the situation worse, leading presidential possibilities. 29 for example, they say, pollsters could said Rothenberg, the election forecaster. But even though Clinton has been the learn from Twitter users what poll “Stuff that’s an hour old is ancient overwhelming favorite among Demo crats questions they should ask the broader news,” he lamented. for 2016, Cruz has trailed well behind public. After winning a 2010 re-election race several other Republican candidates in “I think the rise of social media pre - that many pollsters predicted he would polls of voters’ possible preferences, sents an incredible opportunity for those lose, Nevada Democratic Sen. Harry drawing between 4 percent and 8 per - of us who care about watching and Reid — then majority leader — ad - cent in December surveys. 30 measuring and understanding public monished the media for what he called A 2013 Pew study found that reaction opinion,” The Huffington Post ’s Blumen - its overeagerness to report on “mis - on Twitter to major political events thal says. “If you look at the way com - leading” polls. Those polls “are all over and policy moves “often differs a great panies are turning to Twitter, they see the country, they are so unfair, and deal from public opinion as measured it as an early-warning radar: when their you just gobble them up — no matter by surveys.” Twitter reaction was mea - best customers start complaining, they where they came from,” Reid told re - sured as more liberal than the public’s can watch it quickly. There are possibly porters. “You just run with them as if after a federal court ruled that a California analogous opportunities in political opin - they are the finest piece of pastry in law banning same-sex marriage was ion research.” the world.” 33

128 CQ Researcher In 2012, syndicated columnist michael Gerson, a former speechwriter for Pres - ident George w. Bush, also decried the attention given to Silver’s FiveThirtyEight blog, which aggregated and analyzed numerous polls to produce a widely reported accurate forecast in that year’s presidential race. “The most interesting and important thing about politics is not the measurement of opinion but the formation of opinion,” Gerson wrote. “Public opinion is the product — the outcome — of politics; it is not the 34

substance of politics.” n

o

s Despite the attention paid to his site, l

O

t

Silver said his use of polls to estimate t

o a candidate’s percentage-based chances c

S

/ of winning can conflict with the media’s s

e

g hunger for greater certainty. “what’s a

m

I

challenging is that [ FiveThirtyEight ] goes y

t

t against a lot of instincts that some jour - e

G nalists might have, where you want to At a 2011 rally in Iowa, 2008 GOP vice presidential nominee Sarah Palin colorfully put a ribbon on something and say, ‘This disparaged surveys showing that many Americans regarded her unfavorably: candidate is going to win’ or if you don’t “Polls? . . . They’re for strippers and cross-country skiers,” she declared. Polling do that, then you say, ‘Too close to call. has become entangled in the nation’s prevailing polarized political climate, with politicians from both parties and the public questioning the validity of polls. we have no idea. It’s 50-50,’ ” he said. “we don’t exist in a 100-0 or 50-50 world. most things in the world are “Their release of poll results may be done estimates themselves,” said Rutgers’ Zukin, 75-25 outcomes.” 35 selectively and is often a campaign tactic a former AAPOR president. As an example, Experts in polling with backgrounds or ploy.” 38 he cited 2012 polls that showed Obama in journalism say many reporters aren’t University of wisconsin political sci - ahead of Romney, 47 percent to 44 percent, trained in interpreting polls. “The news entist Charles franklin, who studied with an error margin of plus or minus 3 media have long indulged themselves publicly released horse-race polls from percentage points. “If applied to the 3- in the lazy luxury of being both data- 2000 and 2002, found that polls identified point spread, the 3-point margin of error hungry and math-phobic,” said Gary as partisan by National Journal ’s “Hot - would seem to say that Obama’s lead Langer, a former ABC News polling line,” a political website, tended to skew might be as large as 6 (3 + 3), or as little director who now runs his own survey in favor of their preferred candidate. 39 as 0 (3 - 3),” he said. “But when correctly research company. 36 But Greenberg, the Democratic pollster, applied to the percentage point e stimates The National Council on Public Polls, says some reporters have taken the for the candidates, Obama’s support could an organization of leading survey re - edict too far. “I’ve tried to have con - be between 50 percent and 44 percent search firms, has sought to help reporters versations with journalists who tell me, (47 plus or minus 3), and Romney’s be - better understand polls. It publishes an ‘I can’t trust anything you say because tween 41 percent and 47 percent (44 plus online guide, “20 Questions a Journalist you’re a partisan pollster,’ ” she says. or minus 3). Thus the range between the Should Ask About Poll Results.” The Pollsters applaud media outlets for candidates could be from Obama having questions include, “How many people often recognizing the importance of a 9-point lead (50-41) to Romney having were interviewed for the survey?” 37 the sampling error margin, expressed a 3-point advantage (44-47). So, sampling many news organizations do try to as a plus-or-minus number of percent - error is generally much larger than it may avoid reporting on polls that political par - age points. seem, and is one of the major reasons tisans have sponsored. “Be wary of polls “what is less commonly known is that why polls may differ, even when conducted paid for by candidates or interest groups,” the margin of sampling error does not around the same time .” 40 says The Associated Press Stylebook , which apply to the spread between the two Experts commonly cite a tendency many news organizations use as a guide. candidates, but to the percentage point among the media to emphasize a

www.cqresearcher.com Feb. 6, 2015 129 POLITICAL POLLING

done properly. Greg marx, an associate editor at Columbia Journalism Review , said it is especially useful at the outset of a presidential campaign’s primary elec - tion season, when parties are trying to assess their most-electable aspirants. “If reporters wait for the voters to weigh in to take stock of who’s ahead, they’ll have missed much of the story,” he said. 44 BACKGROUND

h

c Early Polling

a

b

r

e

u

A

any politicians have long contend -

h

c

i r m ed that political and legislative

E

/ success springs from an ability to see

s

e

g and implement the public’s wishes. “what

a

m

I I want to get done is what the people

y

t

t desire to have done, and the question

e

G for me is how to find that out exactly,” Polling pioneer George Gallup used what may have been the first scientific President Abraham Lincoln said. 45 political survey in 1932 to predict that his mother-in-law, Ola Babcock Miller, would be elected Iowa’s first female secretary of state. Three years later, Efforts to discern what people wanted he founded the Gallup Organization. In 1936, Gallup correctly predicted that began decades before Lincoln took office. Democratic incumbent Franklin D. Roosevelt would beat Republican Alf Landon The first public was an in the presidential election, after a magazine poll of 2 million mostly well-off 1824 Harrisburg Pennsylvanian survey Americans erroneously predicted Landon would unseat Roosevelt. that correctly predicted Andrew Jackson candidate’s leads in polls even when Newport singled out The New York would defeat John Quincy Adams. How - the lead is within or near the statistical Times ’ online “Upshot” column, which ever, because neither candidate won a margin of error. In a 2013 research in 2014 allowed readers to use its com - majority of the electoral vote, the election paper, an American University student puter model to make their own statistical went to the House of Representatives, found several examples from the 2012 forecasts in Senate races. 42 Newport which picked Adams. 46 campaign, including a report on NBC also cited The Washington Post and Blu - Newspapers continued to supplement “Nightly News” in October that showed menthal’s site on The Huffington Post , their election coverage by interviewing Romney up by 4 percentage points. which dissect various polls and discuss voters as they left polling places. By The network put onscreen, but did what’s happening in the industry. the end of the 19th century, such “straw not make clear verbally, that the margin Langer says the media shouldn’t be polls” were a feature of local as well of error was plus or minus 3.4 per - singled out for criticism for being in - as national papers and magazines. 47 centage points. 41 ordinately poll-hungry. He cites the Some theorize the term came from hold - Certain elements of the media, how - industry, “which has dis - ing up a stalk of straw to see which ever, have become extremely sophis - covered that polling is the new PR.” way the wind blew it . 48 ticated in explaining polls to the public, Public relations companies say polls can By the 20th century, more sophisti - say some pollsters. “You have a lot of highlight the importance of a product cated methods emerged. In 1932, ad - people who are trying to systematically or service and that news releases featuring vertising agency employee George Gallup evaluate the value of various [polling] polls stand out from other releases. 43 used a new technique enterprises,” says frank Newport, editor Even horse race coverage has its de - to predict the election of his mother-in- in chief at Gallup. fenders, who say it can be valuable when Continued on p. 132

130 CQ Researcher Chronology

1948 candidate Patrick Buchanan are later 1800s Polls emerge to with polls showing Thomas Dewey revealed to have tried to influence measure political opinions. leading President Harry S. Truman exit pollsters. in the presidential election, pollsters 1824 stop conducting interviews weeks 2000 In first opinion poll, Harrisburg or months before vote and fail to uses Pennsylvanian correctly predicts forecast Truman’s victory. data to project — before florida’s presidential candidate Andrew polls close — that vice President Al Jackson’s defeat of John Quincy • Gore would win that state and the Adams, but the House of Repre - presidency. The U.S. Supreme Court sentatives elects Adams after nei - ultimately declares George w. Bush ther candidate wins majority of 1960s-1980s the winner. electoral vote. Political polls and pollsters gain higher profile. 2004 • Gallup poll shows white House 1960 contender John Kerry trailing Presi - Democratic presidential contender dent George w. Bush, leading the 1930s-1950s John f. Kennedy relies heavily on left-wing group moveOn.org to Scientific polls become popular poll data to help him defeat Re - blast Gallup’s “flawed methodology.” but have shortcomings. publican Richard m. Nixon. 2007 1932 1968 Statistician and Nate -agency employee George Pollster Robert Teeter becomes ad - Silver issues predictions about the Gallup predicts his mother-in-law’s viser to Nixon and later to Republi - upcoming presidential election and election as Iowa’s first female secre - can presidential candidate George subsequently founds the influential tary of state in what the company H. w. Bush. FiveThirtyEight.com blog, which is he later started says may have been picked up by the first scientific political survey. 1976 before moving to ESPN. Democratic pollster Patrick Caddell 1935 urges presidential candidate Jimmy 2008 Gallup founds the American Institute Carter to focus on creating an image Pre-primary polls in New Hampshire of Public Opinion. of trustworthiness to counter post- put presidential contender Barack watergate skepticism of politicians. Obama ahead of Hillary Rodham 1936 Clinton, but she wins easily. Experts Literary Digest polls 2 million peo - 1981 blame an overstatement of support ple but incorrectly predicts that Richard wirthlin joins Reagan ad - for Obama among whites. Republican Alf Landon would beat ministration as first semi-official Democratic President franklin D. staff pollster. 2012 Roosevelt; Gallup correctly calls Alleging that polls are stacked the election for Roosevelt after • against them, Republicans heavily conducting his own, more repre - criticize Silver for concluding that sentative poll of 50,000 people. Democrat Obama’s statistical chances 1990s-2000s of election are greater than Republi - 1941 Major polling mistakes draw can rival mitt Romney’s. National Opinion Research Center, criticism of the industry. a private nonprofit group, is found - 2014 ed as the first noncommercial 1996 five months after being rebuked for polling agency. Three television networks incorrectly failing to predict the GOP primary cite exit polls to project a third-place loss of House majority Leader Eric 1947 finish for Republican Bob Dole in Cantor, R-va., pollsters draw even American Association for Public Arizona’s presidential primary. fol - more criticism for miscalling races Opinion Research founded as pro - lowers of Republican presidential in several states and underestimating fessional pollsters’ organization. the GOP’s strength at the polls.

www.cqresearcher.com Feb. 6, 2015 131 POLITICAL POLLING

‘Push Polls’ Cast Negative Light on Candidates Polling groups condemn the practice as manipulative.

hey might seem legitimate, but so-called push polls accused of using pushing polls. In the 2014 race for Suffolk aren’t actually polls — they’re tools of parties and political County, N.Y., comptroller, Republican candidate John Kennedy T activists willing to engage in dirty tricks. In a push poll, accused the campaign of his Democratic opponent, James Gaugh - members of the public respond to what they believe is a valid ran, of conducting a push poll that misrepresented the work telephone-based survey, only to be asked leading and negative Kennedy’s wife, Leslie, did as his legislative aide. He said the questions about a politician. The intent is to try to influence Gaughran campaign portrayed it as sheer nepotism, when in opinion rather than to gather information. fact she put in long hours and was well regarded among con - major polling organizations strongly condemn the practice, in - stituents. 5 cluding the National Council on Public Polls, a coalition of those Gaughran defended his campaign’s efforts, saying Leslie groups, which calls push polling “political manipulation trying to Kennedy’s work was a legitimate campaign issue. 6 But Gaughran hide behind the smokescreen of a public opinion survey.” 1 lost, and Republicans said it was because voters recognized the A prominent example of push polling came in a 2013 special attacks as misguided. 7 U.S. House election in which South Carolina Republican mark many states have banned the practice. But in October, New Sanford — a former governor whose career had been derailed Hampshire’s Supreme Court upheld a lower court ruling that by revelations of an extramarital affair — was running against said state limits on push polls do not apply to federal candidates. Democrat Elizabeth Colbert Busch, a sister of television humorist The case stemmed from a 2010 claim by the state attorney Stephen Colbert. A group supporting Sanford, called SSI Polling, general’s office that Republican Rep. Charlie Bass’ campaign reportedly conducted a poll in which questions included: “what had asked negative questions about Bass’ Democratic rival, Ann would you think of Elizabeth Colbert Busch if I told you she mcLane Kuster. 8 had an abortion?” and “what would you think of Elizabeth In the U.S. Senate, outgoing Democrat Colbert Busch if she had done jail time?” and what would you in November introduced an unsuccessful bill to expand the think of Elizabeth Colbert Busch if I told you she was caught national Do Not Call Registry, which allows consumers to block running up a charge account bill?” 2 telemarketers’ calls, to ban push polls. 9 None of the accusations was true. Sanford — whose campaign denied knowledge of the calling effort — won the election — Chuck McCutcheon less than a week after the media and political reported on the calls. 3 1 Sheldon R. Gawiser and G. Evans witt, “20 Questions a Journalist Should According to the liberal website ThinkProgress.org , the Connecticut Ask About Poll Results,” National Council on Public Polls, undated, http://tinyurl.com/mj52bbg. market research firm SSI said it had been involved in placing calls 2 Scott Keyes and Adam Peck, “Dirty Tricks: mysterious Conservative Group to South Carolina voters with similar content on behalf of an Sending Out Push Polls In South Carolina Special Election,” ThinkProgress.org, unidentified client. A company official said her version of the call may 2, 2013, http://tinyurl.com/br48ep6. script did not contain any questions about abortion but that other 3 “Colbert Busch ‘Abortion’ Push Poll Rocks SC-1,” FitsNews.com , may 1, 2013, http://tinyurl.com/nbuqslr. versions of the script could have been used. 4 , Courage and Consequence: My Life as a Conservative in the South Carolina was the scene of another push-polling controversy Fight (2010), p. 151, http://tinyurl.com/mycxgtm. during the 2000 presidential campaign. Arizona Republican Sen. 5 Rick , “Suffolk’s 12th LD: Kennedy blasts ‘lying’ in Democratic message,” John mcCain and his aides complained voters received calls de - Newsday , Sept. 1, 2014, http://tinyurl.com/lw8udb3. 6 David m. Schwartz, “Legis. John Kennedy’s hiring of wife an issue in scribing mcCain as a “cheat and a liar and a fraud” and intimating Suffolk comptroller race,” Newsday , Oct. 13, 2014, http://tinyurl.com/mtfo4rt. the untrue accusation that he had fathered an illegitimate child. 7 Rick Brand, “Kennedy’s Big Boost at Home,” Newsday , Nov. 9, 2014. Karl Rove, the political strategist for mcCain’s rival, George w. 8 David Brooks, “State Supreme Court says push polls for federal candidates Bush, was widely cited by Democrats and journalists as being are OK, but not for state offices,” The Telegraph (Nashua, N.H.), Oct. 16, behind the calls. Rove has adamantly denied the charge. 4 2014, http://tinyurl.com/o3mr62z. 9 “Begich Introduces Bill to Expand ‘Do Not Call Registry’ and Protect Alaskans’ But Republicans are not the only ones who have been Privacy,” Alaska Business Monthly , Novvember 2014, http://tinyurl.com/q9f6527.

Continued from p. 130 the American Institute of Public Opinion said it was “about right” and 9 percent law, Ola Babcock miller, as Iowa’s first in Princeton, N.J., and began writing said “too little.” 50 female secretary of state in what the a syndicated column, “America Speaks.” Polls played a key role in the 1936 company he started says may have been The first column was about government presidential election. Literary Digest the first scientific political survey. 49 spending: 60 percent of respondents polled 2 million Americans — mostly Three years later, Gallup founded said it was “too great,” while 31 percent well-off people who read magazines,

132 CQ Researcher owned cars or had phones — and pre - idency in 1960, setting the stage for spirit in our country.” The speech led dicted erroneously that Republican Alf subsequent white House contenders. Republicans to depict Carter in the Landon would beat Democratic incum - Pollsters eventually began acquiring 1980 campaign as a pessimist in contrast bent franklin D. Roosevelt. Gallup’s own reputations as savvy strategists, pro - with Republican Ronald Reagan. 56 poll of 50,000 people, which was more viding advice on what issues were representative of the overall population, popular or unpopular. said Roosevelt would win. Roosevelt’s In early 1976, with the country reel - Exit Polling victory helped establish Gallup as the ing from the watergate scandal and nation’s leading polling figure . 51 the resignation of President Richard m. olls caused controversy on Election “my dad thought that polls were Nixon, pollster Patrick Caddell became P Day that year. By then, news media absolutely vital to a democracy,” said a trusted confidante of Democratic can - “exit polls” of voters leaving polling George Gallup Jr., who followed his didate Jimmy Carter. After Carter’s elec - places had become standard. But NBC father into the business. “He felt that tion, Caddell used his data to advise News projected Reagan’s victory, based polls were extremely important because Carter on issues such as the 1977 treaty on such polling, nearly three hours it removed the power from lobbying to cede U.S. control of the Panama before the polls had closed on the groups and from smoke-filled rooms Canal to Panama. 55 west Coast, fueling speculation about and let the public into the act.” 52 By 1941, the National Opinion Re - search Center became the first noncom - mercial polling organization. A private, nonprofit group, it is now based at the and receives fund - ing from other nonprofits as well as government agencies. Six years later, the American Association for Public Opinion Research (AAPOR) was founded as the professional organization of pollsters . 53

Dewey Vs. Truman

o

t

o

he timing of polls was a major f

r

e

issue in the 1948 presidential elec - p T p

o

tion between President Harry S. Truman P

/

s

and Republican Gov. Thomas Dewey e

g of New York. The Roper poll, founded a

m

I

by pollster Elmo Roper just after world y

t

t war II, discontinued surveys in Sep - e

G tember, while Gallup stopped a few President Harry S. Truman gleefully holds up the iconic Chicago Tribune weeks before Election Day. Both pre - headline mistakenly announcing his 1948 re-election defeat by New York Gov. dicted a Dewey win, ignoring the pos - Thomas Dewey. Major pollsters had discontinued their surveys weeks before sibility that people might change their Election Day, leading to the paper’s mistake. After that, pollsters continued surveying potential voters until closer to Election Day to account for minds as the election drew closer. those who changed their minds at the last minute. But Truman prevailed, and the photo of him holding up a Chicago Tribune Two years later, Caddell implored whether the premature news of Rea - mistakenly contending he had lost has Carter to give a speech addressing the gan’s victory discouraged many would- become an iconic image. Gallup and country’s “crisis of confidence” stem - be voters from casting ballots. other pollsters publicly apologized and ming from soaring inflation, oil prices News organizations subsequently refined their methods, including polling and other factors. In Carter’s televised agreed not to project any voting results closer to Election Day. 54 address on July 15, 1979, he said: “The in a state until its polls had closed. Democrat John f. Kennedy relied solution of our energy crisis can also Nevertheless, media mistakes involving heavily on polls in winning the pres - help us to conquer the crisis of the exit polls continued.

www.cqresearcher.com Feb. 6, 2015 133 POLITICAL POLLING

In 1996, three television networks Tom Bradley’s 1982 gubernatorial bid in As washington lobbying boomed in — CNN, CBS and ABC — cited exit California, in which the black Democrat the 1990s and 2000s, so did the use of polls to project a third-place finish for led in pre-election polls only to lose to polls by partisan interest groups to support Republican Bob Dole in Arizona’s GOP Republican George Deukmejian, who is their policy positions. Such groups often presidential primary, which pundits pre - white. 58 commissioned a poll and then used the dicted would deal a serious blow to One of the biggest polling miscues findings to publicly frame the issues, pri - the Kansas senator’s campaign. Dole arose in the 2000 presidential election. vately help shape lobbying tactics or actually came in second and later ended voter News Service, the polling sta - both. The infamous “Harry and Louise” up winning the GOP nomination. It tistics group that had worked on the television ads in 1994, which were seen as helping turn public sentiment against President ’s health-care reform proposal, were shaped by Republican polls. The Health Insurance Association of America, an industry lobby group, ran the ads depicting a middle-class married couple worrying about rising health care costs and bureaucracy. “Organized interests that can afford to conduct such surveys have a leg up on lobbying legislators, for the polls give them critical policy and political data that are both precise (as to num -

h

c bers) and targeted (state or district),”

s

t

e

i University of Kansas political scientist

D Burdett Loomis said in a 2003 paper.

n

i

v

e “Especially when lawmakers are rela -

K

/

s tively indifferent on an issue, such in - e 61 g formation can be powerful.”

a

m

I

During the 2008 Democratic presi -

y

t t dential primaries, a new election-polling

e G mistake surfaced. Pre-primary polls in Polling mistakes in Florida in the 2000 presidential election added to confusion over the outcome of the contest between Democratic Vice President Al Gore and New Hampshire put then-Illinois Sen. former Gov. George W. Bush of Texas. The polling organization Voter News Obama ahead of New York Sen. Hillary Service projected, before Florida’s polls closed, that Gore would win that state, Rodham Clinton by an average margin giving him enough electoral votes to become president. Six hours later, of more than 8 percent. But Clinton the service declared Bush the winner, then later shifted again and declared prevailed easily. the race too close to call. The Supreme Court ultimately declared Bush the winner of the state — and the presidency. Andrew Kohut, then president of the Pew Research Center, blamed the dis - later emerged that followers of Dole’s earlier Arizona primary, used exit poll crepancy on the “long-standing pattern opponent, Republican , data to project, before florida’s polls of pre-election polls overstating support had tried to influence exit pollsters by closed, that Democratic vice President for black candidates among white vot - actively seeking them out in order to Al Gore would win that state, giving ers, particularly white voters who are shape the speculation in favor of their him enough electoral votes to become poor. ” 62 A Stanford University researcher candidate. 57 president. Six hours later, the service later discovered a twist on the Bradley meanwhile, the increasing entry of declared Bush the winner, then later effect, in which white poll respondents African-Americans into politics gave rise shifted again and declared the race were more likely to say they would to a phenomenon known as the “Bradley too close to call. 59 The issue of who back Obama if the person interviewing effect” — the tendency of voters to tell won florida went to the Supreme Court, them was thought to be black. 63 pollsters that they are undecided or are which ultimately declared Bush the The Supreme Court’s landmark Citi - likely to vote for a black candidate, but winner of the state, and thus the pres - zens United v. Federal Election Commis - to actually vote for a white opponent. idency. voter News Service was re - sion decision in 2010, which abolished The term came from Los Angeles mayor placed with a new outfit in 2004. 60 many federal restrictions on campaign

134 CQ Researcher spending, introduced a new entity — respondents by calling only publicly in the fall. In Kansas, many pre-election so-called “super PACs,” which can make listed landlines. To accurately reflect the polls showed Gov. Sam Brownback and unlimited donations as long as the money voting public, Newport said, Gallup Sen. Pat Roberts in serious electoral trouble, does not go directly to candidates. 64 should also have called voters with un - but both Republicans won easily. Chapman Super PACs for both parties began com - listed numbers . 69 Rackaway, a political scientist at fort Hays missioning their own polls, which drew Other polls by organizations unaffiliated State University, blamed herding. Pollsters media attention. for example, two months with a political party have fared better. “want to be close to each other, so if before Kentucky’s 2014 Senate GOP pri - mcGhee, of the Public Policy Institute of they’re wrong, they’re no more wrong mary, a super PAC supporting Sen. mitch California, examined publicly available than everybody else,” he said. mcConnell of Kentucky released a poll Senate polls conducted one week before Brownback’s , showing that he held a nearly 40-point Election Day from 1990 to 2012. He found mark Dugan, said public polls — as lead over challenger matt Bevin. 65 that both the volume of polls, and their opposed to his campaign’s internal sur - In the 2012 presidential race, accuracy, had markedly increased. veys — underestimated how many Re - FiveThirtyEight blogger Silver became Pollsters’ mistakes have been mag - publicans would vote. 73 a target of Republicans’ hostility. Even nified because they have often come In addition to erring in the virginia when polls showed Romney ahead of in high-profile elections. In the 2013 Senate and Georgia Senate and gover - Obama, Silver issued forecasts based virginia governor’s race, regarded as nor’s races, many pollsters failed to fore - on polls showing Obama’s statistical a key barometer for 2014’s elections, see the defeat of Democratic Sen. Kay chances of winning remained greater. most pollsters did not adequately gauge Hagan in North Carolina. And even Silver ended up precisely predicting Republican Ken Cuccinelli’s support. polls that predicted wins for GOP Senate that the president would receive 332 Although Cuccinelli had trailed in pre- candidates Tom Cotton in Arkansas and Electoral College votes, 62 more than election polls by double digits, he lost Kentucky’s mcConnell failed to anticipate he needed to win. 66 to Democrat Terry mcAuliffe by fewer how decisive their victories would be. Heading into Election Day, Romney’s than 3 percentage points. 70 Republicans ended up with 55 seats to private polls showed him ahead in Seven months later, Cantor’s primary regain control of the Senate. florida, Colorado and other key bat - loss in virginia to college professor “The performance of the [pollsters’ tleground states, but he won just one Dave Brat sparked massive criticism of forecasting] models was strikingly sim - — North Carolina. 67 In a subsequent polls. Although several pre-election polls ilar,” wrote ’s John Cas - post-election study, the Republican Na - were wrong, the one drawing the most sidy. “On the eve of the election, all tional Committee’s Growth and Op - attention was by Republican pollster of them indicated that it was unlikely portunity Project — a panel of key John mcLaughlin that had shown Cantor, that the Republicans would get 54 seats party officials — said 70 percent of the House majority leader, up by 34 per - or more. Several of them suggested the GOP pollsters it surveyed agreed centage points. Cantor lost by double that the probability of this happening their party’s official polling was inferior digits, and mcLaughlin blamed a variety was less than 25 percent.” 74 to the Democratic Party’s. of factors, including higher-than-expected One factor, pollsters agree, is that polling A significant problem was the GOP- voter turnout and harsh attacks late in is far easier to do during presidential- financed pollsters’ failure to include the campaign on Cantor’s stance on election years than in midterm-election adequate numbers of Hispanics and immigration. 71 years because in a presidential year, more younger voters. more than one-third But some experts wondered if other of the electorate is engaged and shows of GOP pollsters said Republican sur - factors were at work. GOP pollster and up to vote. That makes it easier to predict veys that failed to properly model voter political consultant frank Luntz called their behavior and obtain their responses turnout were a “very important” factor the miscue “a mind-blowing modern- ahead of time. “The midterm Polling Curse in the inaccuracy. 68 day ‘Dewey Beats Truman’ moment” struck us all” in 2014, said polling forecaster But it wasn’t just the GOP’s internal and said polls do not always provide Sam wang. 75 polls that proved wanting. Gallup’s Elec - all of the information needed to gauge Pollsters also said inaccurate polls tion Day survey showed Romney ahead an election. drowned out more accurate ones in of Obama, 49 percent to 48 percent. “without qualitative insight — talking their forecasting models. The most ac - Obama won, 51 percent to 47 percent. with voters face to face to judge their curate polls, however, did not depend Newport, the Gallup editor, later cited mood, emotion, intensity and opinion solely on aggregating poll data but several problems, including shortcom - — polls can be inconsequential, and also made their own adjustments. They ings in the company’s model of likely occasionally wrong,” Luntz said. 72 included The Washington Post ’s Election voters and the practice of finding survey High-profile misjudgments continued Lab and FiveThirtyEight.com . 76

www.cqresearcher.com Feb. 6, 2015 135 POLITICAL POLLING

But pollsters pushed back against governor, who has yet to formally an - One such debate occurred when criticism that the election was a disaster. nounce his white House plans. CBS News and The New York Times FiveThirtyEight.com asked 17 prominent “It seems to me that in New Jersey, announced last year they would begin pollsters, “In light of Tuesday’s results, people are over the Bridgegate scandal, using YouGov panels of voters who did election polls do well this year at and nationally, it is not even on the express their opinions online as part depicting the electorate’s views?” Ten public’s radar,” Geller said in a speech. of their election poll models. YouGov, answered yes; six said no; one said, “I’m not speaking to whatever the a United Kingdom-based survey re - “Some did and some didn’t.” 77 future may hold.” 80 However, a Quin - search firm, uses the online panels as In a later statistical analysis, mcGhee, nipiac University poll in January showed an alternative to traditional telephone- of California’s Public Policy Institute, Christie with a rating of 46 percent based surveys. ( See sidebar, p. 132. ) said the 2014 polls weren’t quite as approval/48 percent disapproval among David Leonhardt, editor of The Times ’ on-target as in prior years. But he said: New Jersey voters. “It’s his worst score “Upshot” site, said the new partnership “we are still clearly in a world of better in almost four years, so Bridgegate would result in “a significantly expanded accuracy.” 78 continues to harass Gov. Christie,” said panel. . . . You get a more accurate maurice Carroll, the Quinnipiac poll’s picture of the horse race when you assistant director. 81 look at it from several angles.” 84 Another potential GOP candidate, However, pollsters who prefer phone CURRENT wisconsin Gov. Scott walker, in No - polling were sharply critical. Link, from vember downplayed a poll of wisconsin the American Association for Political voters showing that only 42 percent of Opinion Research (AAPOR), took the SITUATION respondents said he would make a unusual step of issuing a public state - good president. That figure was 4 per - ment saying YouGov’s survey methods centage points below that of wisconsin have “little grounding in theory” and Presidential Race Rep. Paul Ryan, the 2012 GOP vice- questioning whether the news orga - presidential nominee. “In the end, any nizations would adequately disclose ollsters are expected to play a poll right now is ridiculous,” walker said. their methodologies. 85 P prominent role in the 2016 pres - “You look over the past four or five In response, Andrew Gelman — a idential election. Clinton, the consensus elections, people who poll high at the political scientist and director of Co - Democratic frontrunner, in January beginning are not the people who end lumbia University’s Applied Statistics hired Joel Benenson, whose polling up being the nominees.” 82 Center who says Internet polls have for Obama was considered a key factor However, the fact-checking website merit — wrote a sarcastic post on a behind his two white House wins. Politifact.com looked at past early pres - blog with the headline: “President of In her failed 2008 race for the pres - idential election polls and found that American Association of Buggy-whip idency, Clinton had relied on Demo - five contenders who did well at the manufacturers takes a strong stand cratic pollster mark Penn, whom she time did become the nominee. It rated against internal combustion engine, ar - ultimately replaced when she fell be - walker’s assertion “mostly false.” 83 gues that the so-called ‘automobile’ has hind and tried to shake up her cam - ‘little grounding in theory’ and that ‘re - paign. Obama used Benenson as part sults can vary widely based on the of a team of pollsters, an approach Clin - Internet Polling particular fuel that is used.’ ” 86 ton is said to be interested in duplicating Link told CQ Researcher in January if she decides to run. 79 ow much attention should be ac - he did not intend to criticize YouGov Pollsters are at odds over whether a H corded to polls using nonprob - and that his statement “was not as effective controversial decision by New Jersey ability sampling — such as opt-in sur - or efficient as it should have been. with Republican Gov. Chris Christie’s admin - veys conducted on the Internet — is the issues it raised, the silver lining was istration last year to close a portion of one of the biggest ongoing debates in that it should generate a dialogue.” the George washington Bridge for po - the polling world. An AAPOR task force on nonprob - litical reasons is hurting his possible “Our future may involve methodolo - ability sampling issued a report in 2013 presidential bid. On the Republican side, gies where your gut says, ‘This’ll work,’ concluding that no single consensus Adam Geller, a pollster for Christie, made but the mathematical foundation of prob - method exists for conducting it. 87 But news in December when he asserted ability sampling isn’t there,” western Link says his organization’s annual con - that the controversy — dubbed “Bridge - New England’s vercellotti says. “You’re ference, scheduled for may, will include gate” — was not hurting the New Jersey seeing really spirited debates about that.” Continued on p. 138

136 CQ Researcher At Issue:

Canyes “nonprobability” polls reliably reflect public opinion?

DOUGLAS RIVERS GARY LANGER CHIEF SCIENTIST , Y OU GOV ; P ROFESSOR OF FOUNDER AND PRESIDENT , L ANGER POLITICAL SCIENCE , RESEARCH ASSOCIATES ; F ORMER ABC STANFORD UNIVERSITY NEWS POLLING DIRECTOR

WRITTEN FOR CQ RESEARCHER , FEBRUARY 2015 WRITTEN FOR CQ RESEARCHER , FEBRUARY 2015

very public opinion poll conducted during the 2014 e cannot be confident that nonprobability samples midterm elections used a nonprobability sample of vot - validly and reliably represent broader public atti - e ers. Some pollsters claim to have probability samples w tudes. They fall short in the two key tests of scien - because they used random digit dialing (RDD), but they’re con - tific inquiry: theoreticism and empiricism. In both, by contrast, fusing a probability sample of phone numbers with a probabili - probability-based methods excel. ty sample of voters. These are not the same thing. Probability-based survey research — conducted, for example, In probability sampling, every voter has a known positive via random-sample telephone or face-to-face interviews — is probability of selection. At one time, RDD did provide based on probability theory, which provides grounds for infer - something close to a probability sample. Before the era of ence to a larger population. The theory allows for random cellphones, answering machines and , most nonresponse, a reason that probability-based samples, even households had a single landline telephone. when it rang, with low response rates, are consistently accurate when mea - they answered it and took surveys. Nowadays, people have sured against relevant benchmarks. multiple phone numbers, screen their calls and — if reach - Nonprobability sampling has made a comeback of late with the able — aren’t willing to take surveys. Response rates are advent of the Internet and social media, offering fast, large-scale, often under y10 percent, soe that the prs obability that a voter inexpensive datan collection — but ito lacks a theoretical basis. most ends up in a poll isn’t the probability that their phone num - prominent are opt-in Internet panels, in which individuals sign up ber was selected. to click through questionnaires online in exchange for cash and If you don’t know the probability of selection, then it’s not gifts. There’s evidence of a cottage industry of professional respon - a probability sample, and you should stop pretending that you dents, signing up for multiple panels, perhaps under assumed have a probability sample. we already weight our data to identities, to increase their prize-winning opportunities. match census demographics because some hard-to-get cate - Above and beyond validation of identities, quality control gories of voters are underrepresented in phone samples by a often is lacking. full disclosure is rare, including details of factor of 10 or more. sampling, sample weighting and respondent selection proce - This is where so-called nonprobability methods and Internet dures. And such surveys often make unsupportable claims, in - panels have advantages. Large panels of respondents have cluding the assertion of a calculable margin of sampling error. been recruited to take surveys. Instead of randomly selecting Empirically, academic studies have suggested that such surveys panelists for a poll, they are chosen to be representative of produce inconsistent results in terms of accuracy, time trends and the population. If the panel is large and diverse enough, it’s relationships among variables. In 2010, the American Association feasible to obtain a sample with smaller skews than an RDD for Public Opinion Research concluded, “There currently is no phone sample. This means that less weighting is needed to generally accepted theoretical basis from which to claim that sur - correct for skews in the sample. vey results using samples from nonprobability online panels are The theoretical basis for making inferences from a non - projectable to the general population.” It advised researchers to probability panel and for weighting an RDD phone are “avoid nonprobability online panels when one of the research identical, since both rely upon adjusting the sample for re - objectives is to accurately estimate population values.” spondent characteristics. How well they work is an empirical Attempted quantitative analysis of social media also raises matter. concerns. These vast data sets may include, for example, multiple At YouGov, we’ve used an online opt-in panel to produce postings (and repostings), created via orchestrated campaigns estimates for nearly every Senate and governor election as using computerized “bots.” valid demographic information — well as presidential vote by state for almost a decade. The even nationality — generally is unavailable. Selecting relevant average error for Senate races has ranged from 1.8 percent posts and accurately discerning their meaning are highly chal - in 2010 to 3.5 percent in 2008 (with 2014 being 3.2 percent). lenging given the use of slang, abbreviations, acronyms, irony State level presidential estimates had an average error of and sarcasm. And, again, a theory to support inference is absent. 1.3 per cent in 2012 and 2.5 percent in 2008. These error All survey methods — probability and nonprobability alike rates were about the same or slightly less than phone — should be subjected to ongoing, independent evaluation. sampleno s in the same states over the same period. for this to be accomplished, full disclosure of relevant informa - tion on sampling, weighting and data production is essential.

www.cqresearcher.com Feb. 6, 2015 137 POLITICAL POLLING

Continued from p. 136 voting population at large. The New York “There are some robopolls out there “an entire focus” on nonprobability Times in October 2014 began running with the recorded voice of a real person sampling, “so we can all come together articles based on the panel’s work . 89 asking the question — there have been and learn from each other and learn The RAND panel’s final pre-election experiments with avatars administering from the new techniques as well as result in 2012 was closer to the election’s polls to you,” Keeter said. “But there are learn from the new methodologies.” actual outcome than that of many other issues here, too. what’s the accent? Does Similar efforts already are underway. polls. 90 she remind you of any old girlfriend or an ex-wife, and does that have an effect on your answers? for sensitive questions like drug use, are people more or less likely to tell Cortana the truth? This is why we need experiments like this.” 91 The public’s disdain for robopolls can affect their quality. Two researchers compared robopolls of a president’s job- approval rating with live calling and In - ternet surveys. Although they found phone calls and online surveys produced

e

k “quite similar” results, they found that

a

L

robopolls had far lower participation

a

s i rates resulting in “a significantly higher

L

/

s estimate of the disapproval rate of the

e

g

a president and a significantly lower es -

m

I 92 timate for ‘no opinion.’ ”

y

t

t

e The Research Association,

G Pollsters are expected to play a prominent role in the 2016 presidential election. the trade group representing the opin - Former Secretary of State Hillary Rodham Clinton, the consensus frontrunner ion and marketing research industries, for the Democratic nomination — if she decides to run — in January hired asked the federal Communications Joel Benenson, whose polling was considered a key factor in Commission (fCC) in January to con - Barack Obama’s two white House wins. sider relaxing federal regulations to allow survey, opinion and marketing In November, the online survey company research calls to cell phones. Surveymonkey announced it would begin Robot Interviewers Existing regulations that forbid the collaborating with the research organi - use of automatic dialers or “robocallers” zation westat and the Pew Research any in the polling industry are to cell phones limit cell phone polling, Center. The groups will study the science m watching software giant pollsters say. Currently, only human be - behind adjustments and weighting used Corp., which drew attention in Sep - ings can call cell phones. The association with nonprobability Internet sampling tember when it started an online survey contended in comments to the fCC along with several other sampling meth - website called microsoft Prediction Lab, that the rules “unnecessarily and perhaps ods. “Having no singular framework for in which users can submit predictions unintentionally limit researchers’ ability nonprobability sampling is limiting the on politics and other subjects. to gather unbiased, reliable and accurate insights market researchers and opinion The company said Cortana, the digitally public views that cost-effectively rep - pollsters can deliver,” said mike Brick, based smartphone assistant similar to resent all demographics.” 93 a westat senior statistician. 88 Apple’s Siri, eventually could conduct Consumer groups, however, are ex - Some pollsters say finding and re - the interviews as an alternative to human pected to continue their ardent oppo - taining online panels of voters who truly questioners used in phone polling. sition to allowing robocalls on cell reflect the public at large can be extremely Scott Keeter, the Pew Research Cen - phones. They say such calls are an in - expensive and time-consuming, but they ter’s director of survey research, ap - trusion and could open the door to un - agree it merits further investigation. plauded the company’s work. He said wanted and marketing pitches . 94 Since 2003, the RAND Corp., a non - replacing a human pollster with a digital many states regulate robocalls, pri - profit research organization, has used one was not entirely unprecedented, marily by restricting when they can such a panel that seeks to replicate the but cautioned that it needs ample study. occur. Several specifically ban polling-

138 CQ Researcher related calls at certain times of the day for now, though, many pollsters say into the DNA of political coverage and or night. 95 telephone polling will remain dominant encourages people to think more about Bill mcInturff, partner and cofounder in the political world. “In the short uncertainty,” he said. 101 of Public Opinion Strategies, a survey run, there isn’t anything as good as a whatever polling changes are pro - research firm in Alexandria, va., is among live call to a cell phone or landline,” posed, western New England’s vercel - those exploring the potential of cell Democratic pollster Greenberg says. “I lotti predicts that they will provoke in - phone use. His firm conducted the first don’t see that going away.” tense discussion. “The industry is filled cell-only national political survey in Terry Casey, a political science pro - with really smart, determined people 2013, and he says cell polls can help fessor at Rose-Hulman Institute of Tech - — feisty, argumentative people,” he illuminate the views of younger voters, nology in Terre Haute, Ind., agrees says. “what makes it interesting is to who are avid cell phone users. with Greenberg. “I would be surprised see how we all adapt.” Though mcInturff says a cell-only sur - if what we do in terms of polling is vey isn’t representative of the population radically different in 10 years,” Casey at large, he adds: “Once you understand says. “more often than not, [pollsters] Notes you’re doing it for targeted audience re - do get the results right.” search, the kind of depth you can get The Times ’ Leonhardt said he foresees 1 Chuck Ross, “Larry Sabato: Polling Industry through a mobile-designed survey is a greater reliance on text messaging to Needs Some ‘Housecleaning,’ ” The Daily Caller , really powerful.” cell phones. “If text messaging becomes Nov. 5, 2014, http://tinyurl.com/n2gnal7 . nearly universal and occurs mostly 2 Emily Cadei, “Political strategist Patrick Ruffini through peoples’ mobile phone, surveyors forecasts the GOP’s future,” USA Today , Oct. 23, could use random-digit texting as a way 2014, http://tinyurl.com/mexkzo4 . to poll people, for instance,” he said. 99 3 Jake Gibson, “Palin: Polls are for Strippers OUTLOOK and Cross Country Skiers,” .com, Hankin, the former Obama pollster, says the industry needs to broaden its Sept. 3, 2011, http://tinyurl.com/pzow7j9 . 4 michael Scherer, “How Obama’s Data Crunch - ability to capture public ideologies. His Shades of Gray ers Helped Him win,” CNN.com, Nov. 8, 2012, firm, Lincoln Park Strategies, uses more http://tinyurl.com/lq8y4hp . layers in its questions and sampling 5 Cliff Zukin, “Sources of variation in Pre- o consensus appears to exist about than many traditional firms, enabling Election Polls: A Primer,” American Association N the future of political polling. Poll - it to score the electorate into 16 different for Public Opinion Research website, Oct. 24, sters in the field tend to be less willing demographic clusters that can be used 2012, http://tinyurl.com/mxxtdrb . to try new approaches than corporations to tailor messages to voters. 100 6 Carl Bialik and , “why Pollsters and other entities that rely on surveys. “There’s way too much one-dimensional Think They Underestimated ‘No’ in Scotland,” “Political pollsters are not usually viewing of how opinion works,” he FiveThirtyEight.com , Sept. 19, 2014, http://tiny the industry leaders,” said Christina says. “You’re either A or B, pro-choice url.com/mmqlftj . 7 matthews, a Republican pollster in or pro-life. But most people live their Robert Channick, “40% of homes now without a landline,” Chicago Tribune , July 8, 2014, Alexandria, va. 96 lives in shades of gray. what [pollsters] Some scholars wonder if the rise of http://tinyurl.com/mx5fhrg . really need to be asking is, ‘what’s 8 “Assessing the Representativeness of Public “big data,” the technological ability to your ideology, and where do you place Opinion Surveys,” Pew Research Center for the swiftly vacuum numerous details about Obama or Romney on that spectrum?’ People & the Press, may 15, 2012, http://tinyurl. people’s habits, will someday overtake when you start overlaying these things, com/7e8gjrr . traditional polling. Such data can be really 5 percent of the public are people 9 “2015 Congressional Issues Survey from Con - gleaned and analyzed faster than through who would, by definition, be an actual gressman morgan Griffith,” Office of Rep. mor - telephone polls, with new features — moderate or an actual independent.” gan Griffith, http://tinyurl.com/lfke2jb . such as tagging the data with Global FiveThirtyEight ’s Silver, meanwhile, 10 “Congressman Pete Stark’s Online Survey: Positioning System coordinates — greatly says he hopes his approach of aggre - Should Congress Impeach Bush?” warIsA broadening its reach . 97 Crime.org, http://warisacrime.org/node/30892 . gating a wide number of polls — in - 11 “I foresee a world in which we’ll Pew Research Center, op. cit. stead of placing too much importance 12 blend surveys covering full populations Nate Silver, “Is the Polling Industry in Stasis on any single poll — continues to Or In Crisis?” FiveThirtyEight.com , Aug. 25, . . . with continuous-time, sensor and gain currency. 2014, http://tinyurl.com/lrssa2m . other data on subsets of the population,” “I hope that — probably not by 13 Eric mcGhee, why The Polls Are Better Than said Georgetown University Provost next election cycle, but by 2024 or Ever at Predicting Senate Races,” The Washington 98 Robert Groves, a social statistician. 2028 — it’s just kind of incorporated Post , Oct. 28, 2014, http://tinyurl.com/o646py4 .

www.cqresearcher.com Feb. 6, 2015 139 POLITICAL POLLING

14 Scott Clement, “Survey Says: Polls Are Bi - 28 Schlesinger, op. cit. Relations,” visible Strategies Communications, ased,” The Washington Post , Sept. 4, 2013, 29 Hadas Gold, “#Hillary #TedCruz Rule,” Politi - http://tinyurl.com/m32ve8m . http://tinyurl.com/mza82wa . co , Dec. 4, 2014, http://tinyurl.com/lc6728e . 44 Greg marx, “In Defense of [the Right Kind 15 Jeff Zeleny, “Poll: Obama Holds Narrow 30 “white House 2016: Republican Nomination,” of] Horse Race Journalism,” Columbia Journalism Edge Over Romney,” The New York Times , PollingReport.com , http://tinyurl.com/m6e6dsy . Review , Sept. 6, 2011, http://tinyurl.com/nwwgc7c . Sept. 14, 2012, http://tinyurl.com/k7wz8qn . 31 Amy mitchell and Paul Hitlin, “Twitter Re - 45 “History of Opinion Polling,” NOw with 16 Ruby Cramer, “Conservatives Embrace Alternate action to Events Often at Odds with Overall Bill moyers, June 14, 2002, http://tinyurl.com/ Polling Reality,” Buzzfeed.com, Sept. 24, 2012 , Public Opinion,” Pew Research Center, march 4, op8s7aw . http://tinyurl.com/p5amuse . 2013, http://tinyurl.com/cvqq5hz . 46 “23d: The 1824 Election and the ‘Corrupt Bar - 17 Jason Stanford, “50 Shades of Crazy: Skewed 32 Tom Rosenstiel, “Political Polling and the gain,’ ” USHistory.org, http://tinyurl.com/6q4cmlb . Polls,” The Huffington Post , Sept. 27, 2012, http:// New media Culture: A Case of more Being 47 “History of Opinion Polling,” op. cit. tinyurl.com/ntv9hst . Less,” Public Opinion Quarterly , vol. 69, No. 5, 48 “How Did the Straw Poll Originate?” Inno 18 Niall Stanage, “Dems: Don’t Trust the Polls,” 2005, http://tinyurl.com/oppsesm . vateUs.net, http://tinyurl.com/mfblrbr . The Hill , Oct. 13, 2014, http://tinyurl.com/nuqwyoz . 33 Adam Nagourney, “Reid Assails Polls That 49 “George Gallup and the Scientific Opinion 19 Nate Silver, “Poll Averages Have No History Predicted His Loss,” The New York Times , Poll,” The first measured Century, PBS.org, of Consistent Partisan Bias,” The New York Times , Nov. 3, 2010, http://tinyurl.com/24v8utd . www.pbs.org/fmc/segments/progseg7.html . Sept. 29, 2012, http://tinyurl.com/pxfjebz . 34 michael Gerson, “The Trouble with Obama’s 50 Sarah van Allen, “George Gallup, Twentieth- 20 Nate Silver, “The Polls might Be Skewed Silver Lining,” The Washington Post , Nov. 5, Century Pioneer,” Gallup.com, Dec. 29, 1999 , Against Democrats — Or Republicans,” 2012, http://tinyurl.com/c4dz38j . http://tinyurl.com/pz8utqu . FiveThirtyEight.com , Oct. 15, 2014, http://tiny 35 Gautam Hathi, “fiveThirtyEight founder Nate 51 “five biggest political polling mistakes in url.com/l7ke8yj . Silver tackles election predictions, website goals,” U.S. history,” National Constitution Center, 21 “marist Institute for Public Polling Director The Chronicle (Duke University), Nov. 24, 2014 , Oct. 2, 2012, http://tinyurl.com/m4d8xsr . Lee miringoff,” HughHewitt.com, Sept. 14, 2012 , http://tinyurl.com/ozjrf2t . 52 “George Gallup and the Scientific Opinion http://tinyurl.com/mb4y7n2 . 36 “Tracking Polls,” On the media, march 26, Poll,” op. cit. 22 Harry Enten, “Are Bad Pollsters Copying 2010, http://tinyurl.com/no28du2 . 53 “History of Opinion Polling,” op. cit. Good Pollsters?” FiveThirtyEight.com , Aug. 11, 37 Sheldon R. Gawiser and G. Evans witt, “20 54 “five biggest political polling mistakes in 2014, http://tinyurl.com/k4fp8bn . Questions a Journalist Should Ask About Polls” U.S. history,” op. cit. 23 fabio Rojas, “How Twitter can predict an (Third Edition), undated, National Council on 55 Alexander Heard and michael Nelson (ed - election,” The Washington Post , Aug. 11, 2013, Public Polls website, http://tinyurl.com/24a246 . itors), Presidential Selection (1987), p. 212, http://tinyurl.com/m5wa8vs . 38 Norm Goldstein (ed.), The Associated Press http://tinyurl.com/nwrd7gg . 24 matthew macwilliams, Edward Erikson and Stylebook (2013), p. 214. 56 “Carter’s ‘Crisis of Confidence’ Speech,” Nicole Berns, “Can facebook Predict who wins 39 mark Blumenthal, “what makes a Poll Par - PBS.org, http://tinyurl.com/3b5qpmj. Also see the Senate in 2014?” Politico , Jan. 20, 2014 , tisan?” National Journal , march 22, 2010, http:// D. Teter, “Iran between East and west,” Jan. 26, http://tinyurl.com/ls8kpp7 . tinyurl.com/o2dzvnn . 1979, Editorial Research Reports , available at CQ 25 Jennifer Schlesinger, “Popularity Contest: 40 Zukin, op. cit. Researcher Plus Archive. Can facebook and Twitter Predict Election 41 Alex Kaplan, “media Coverage of the Pres - 57 Bill Carter, “Networks Admit Error in Arizona Results?” ABC News.go.com, Nov. 25, 2010, idential Horse Race and of Presidential Polls,” Primary Calls,” The New York Times , feb. 29, http://tinyurl.com/2eqp3bz . American University research paper, Spring 2013 , 1996, http://tinyurl.com/qhmgx3z . 26 Stuart Rothenberg, “Twitter Can’t Yet Predict http://tinyurl.com/kdpr449 . 58 Brian mcCabe, “fiveThirtyEight Glossary,” Elections,” Roll Call , Aug. 14, 2013, http://tiny 42 Josh Katz, wilson Andrews and Jeremy Bow - The New York Times , http://tinyurl.com/pazjfon . url.com/m2387j9 . ers, “Elections 2014: make Your Own Senate 59 Kate Pickert, “A Brief History of Exit Polling,” 27 Steve Koczela, “Twitter Doesn’t Have Pollsters forecast,” The New York Times , http://tinyurl. Time , Nov. 4, 2008, http://tinyurl.com/jwubfm5 . Running Scared,” Commonwealth, Aug. 20, com/mkslt86 . 60 “five biggest polling mistakes in U.S. history,” 2013, http://tinyurl.com/m8jsahu . 43 “Creating a Buzz with web-Powered Public op. cit. 61 Burdett Loomis, “from Hootie to Harry (and Louise): Polling and Interest Groups,” Brookings About the Author Institution, Summer 2003, http://tinyurl.com/ mcl6qey . Chuck McCutcheon is a freelance writer in Washington, 62 Andrew Kohut, “Getting It wrong,” The New D.C. He has been a reporter and editor for Congressional York Times , Jan. 10, 2008, http://tinyurl.com/ Quarterly and Newhouse News Service and is co-author of lo49xxc . the 2012 and 2014 editions of The Almanac of American 63 Steven Shepard, “Researcher Examines vari - Politics and Dog Whistles, Walk-Backs and Washington ation of the ‘,’ ” National Journal , Handshakes: Decoding the Jargon, Slang and Bluster of may 18, 2012, http://tinyurl.com/lfajma5 . 64 American Political Speech. He also has written books on for background, see the following CQ Re - and nuclear waste. searcher s: Kenneth Jost, “ Debates,” may 28, 2010, pp. 457-480; Thomas

140 CQ Researcher J. Billitteri, “Campaign finance Reform,” June 13, 2008, pp. 505-528; Christina L. Lyons, “Nonprofit Groups and Partisan Politics,” Nov. 14, 2014, FOR MORE INFORMATION pp. 961-984. 65 American Association for Public Opinion Research , 111 Deer Lake Rd., Suite Phillip m. Bailey, “mitch mcConnell Holds 100, Deerfield, IL 60015 ; 847-205-2651 ; www.aapor.org . The main association for wide Lead Over matt Bevin, Super PAC Senate public opinion and survey-research professionals. Poll finds,” wfPL.org, march 10, 2014, http:// Annenberg Public Policy Center , 202 S. 36th St., Philadelphia, PA 19104-3806 ; tinyurl.com/me7m3q8 . 215-898-9400 ; www.annenbergpublicpolicycenter.org . University of Pennsylvania- 66 Andrew Hough, “Nate Silver: Politics ‘Geek’ based group that conducts a wide-ranging national election survey. Hailed for Barack Obama’s U.S. Election fore - FiveThirtyEight.com , 147 Columbus Ave., 4th floor, New York, NY 10023 ; cast,” The Telegraph (UK), Nov. 7, 2012, http:// www.fivethirtyeight.com . website founded by statistician Nate Silver that dissects tinyurl.com/o4vp693 . polls and examines political polling. 67 mark Halperin and John Heilemann, Double Gallup Organization , 901 f St., N.w., washington, DC 20004 ; 202-715-3030 ; Down: 2012 (2013), pp. 456, 455. www.gallup.com . major polling organization that focuses heavily on politics. 68 “Growth and Opportunity Project,” Repub - lican National Committee, p. 38, http://gop Pew Research Center , 1615 L St., N.w., Suite 700, washington, DC 20036 ; 202- project.gop.com . 419-4300 ; www.pewresearch.org . Nonpartisan “fact tank” providing information on issues, attitudes and trends shaping the United States. 69 Scott Clement, “Gallup Explains what went wrong in 2012,” The Washington Post , June 4, Princeton Survey Research Associates International , 600 Alexander Rd., Suite 2013, http://tinyurl.com/o4a64pt . 2-2, Princeton, NJ 08540 ; 609-924-9204 ; www.psrai.com . Independent company 70 James Hohmann, “why Terry mcAuliffe that does polling for a variety of clients, including the news media. Barely won,” Politico , Nov. 6, 2013, http://tiny Roper Center for Public Opinion Research , 369 fairfield way, Unit 1164, url.com/kby7mat . Storrs, CT 06269-1164 ; 860-486-4440 ; www.ropercenter.uconn.edu . University of 71 Shane Goldmacher, “Eric Cantor’s Pollster Connecticut-based research organization that studies political trends. Tries to Explain why His Survey Showed Cantor Up By 34 Points,” National Journal , June 11, 82 Tom Kertscher, “Those who poll well early 90 Nate Silver, “which Polls fared Best (And 2014, http://tinyurl.com/m2tub5f . don’t become presidential nominee, potential worst) in the 2012 Presidential Race,” The New 72 frank Luntz, “why Polling fails,” The New contender Scott walker says,” PolitiFact Wisconsin , York Times , Nov. 10, 2012, http://tinyurl.com/ York Times , June 11, 2014, http://tinyurl. Nov. 24, 2014, http://tinyurl.com/mgrpnto . aorrju4 . com/k4j95jq . 83 Ibid. 91 Alan Schwarz, “microsoft Begins a Push 73 Bryan Lowry, “why Kansas Election Polls 84 Chris Cillizza, “The New York Times rocked Into the Polling world,” The New York Times , were Dead wrong,” The Wichita Eagle , Nov. 5, the polling world over the weekend. Here’s Sept. 29, 2014, http://tinyurl.com/kngo5b3 . 2014, http://tinyurl.com/l5pz8fa . why,” The Washington Post , July 31, 2014, http:// 92 Jan van Lohuizen and Robert wayne Samohyl, 74 John Cassidy, “why Didn’t Anyone See the tinyurl.com/q8ld2dj . “method Effects and Robopolls,” Survey Practice , GOP Tidal wave Coming?” The New Yorker , 85 Peyton m. Craighill and Scott Clement, vol. 4, No. 1, 2011, http://tinyurl.com/obmegkb . Nov. 7, 2014, http://tinyurl.com/qzgx5v6 . “How the polling establishment smacked down 93 “Excluding mR from the TCPA: mRA Asks 75 Sam wang, “The Polls failed to Predict a The New York Times,” The Washington Post , the fCC to Allow marketing Research Calls Republican Landslide. Here’s why,” The New Re - Aug. 4, 2014, http://tinyurl.com/msbnhcl . to Cell Phones,” marketing Research Associ - public , Nov. 5, 2014, http://tinyurl.com/q36bwmm . 86 Andrew Gelman, “President of American ation, Jan. 2, 2015, http://tinyurl.com/pfbfk6a . 76 John Sides, “Election Lab on track to forecast Association of Buggy-whip manufacturers 94 mitch Lipka, “If You Hate Robocalls, You’ll 35 of 36 Senate races correctly,” The Washington takes a strong stand against internal combustion Really Hate This Idea,” Daily Finance.com , Post , Nov. 5, 2014, http://tinyurl.com/mwdzymq . engine, argues that the so-called ‘automobile’ Jan. 16, 2015, http://tinyurl.com/kkxx7f8 . 77 Carl Bialik, “most Pollsters we Surveyed Say has ‘little grounding in theory’ and that ‘results 95 “Political Robocall Laws,” The Stratics Group Polling Did well This Year,” FiveThirtyEight.com , can vary widely based on the particular fuel website, http://tinyurl.com/nbjcjhy . Nov. 7, 2014, http://tinyurl.com/qx6ovlq . that is used,” Statistical Modeling, Causal In - 96 Steven Shepard, “Sorry, wrong Number,” 78 Eric mcGhee, “The 2014 Polls were Still Really ference and Social Science blog, Aug. 6, 2014, National Journal , July 12, 2012, p. 25. Good at Predicting winners,” The Washington http://tinyurl.com/qgk9l73 . 97 for background, see Tom Price, “Big Data Post , Dec. 12, 2014, http://tinyurl.com/nq32g4s . 87 Drew Desilver, “Q/A: what the New York and Privacy,” CQ Researcher , Oct. 25. 2013, 79 maggie Haberman, “ brings Times’ polling decision means,” Pew Research pp. 909-932. in Robby mook, Joel Benenson for likely team,” Center, July 28, 2014, http://tinyurl.com/lfysh5f . 98 Robert Groves, “A future world of Blended Politico , Jan. 7, 2015, http://tinyurl.com/q9688fu . 88 “Surveymonkey Announces Collaboration Data,” “The Provost’s Blog,” Georgetown Uni - 80 Elizabeth Swain, “Christie pollster Geller with Leading Research Organizations to Ex - versity, July 30, 2014, http://tinyurl.com/q9q5z6c . says ‘people are over the Bridgegate scandal,’ ” plore Online Non-Probability Sampling,” news 99 Cillizza, op. cit. PolitickerNJ.com , Dec. 3, 2014, http://tinyurl. release, Nov. 24, 2014, http://tinyurl.com/okj 100 Tess vandenDolder, “Democratic Pollster com/qanj4d7 . vdgb . Stefan Hankin Explains the math Behind a 81 Kevin mcArdle, “Bridgegate is a ‘traffic night - 89 “About the RAND American Life Panel,” winning Campaign,” DCInno.Streetwise.com, mare that never ends’ for Christie,” NJ1015.com, The New York Times , Oct. 9, 2014, http://tinyurl. June 24, 2014, http://tinyurl.com/l5euny8 . Jan. 21, 2015, http://tinyurl.com/p3rdmnl . com/q3pf28n . 101 Hathi, op. cit. www.cqresearcher.com Feb. 6, 2015 141 Bibliography Selected Sources

Books those online, adjust their results to track the findings of higher-quality polls. Craig , Richard , Polls, Expectations and Elections: TV News Making in U.S. Presidential Campaigns , Lexington Books , Gelman , Andrew , and David Rothschild , “Modern Polling 2014 . Needs Innovation, Not Traditionalism,” The Washington A San Jose State University journalism professor discusses Post , Aug. 4, 2014 , http://tinyurl.com/n2aqp3d . how polls often drown out other aspects of presidential cam - Two polling experts discuss the controversy over The New paign coverage. York Times ’ move to online polling.

Erikson , Robert S. , and Kent L. Tedin , American Public Sabato , Larry J. , “Skewed: Why Americans Hate the Polls,” Opinion (9th edition), Pearson , 2014 . Politico , Oct. 6, 2014 , http://tinyurl.com/nojbwfa . The textbook examines the role of public opinion in the The director of the University of virginia’s Center for Politics electoral process. discusses problems in political polling.

Goidel , Kirby , ed., Political Polling in the Digital Age: Shepard , Steven , “Why The GOP Still Doesn’t Have Its The Challenge of Measuring and Understanding Public Act Together on Polling,” National Journal , Sept. 12, 2013 , Opinion , Louisiana State University Press , 2011 . http://tinyurl.com/kwgggmc . An assortment of pollsters and others discuss how polling A journalist chronicles the Republican Party’s post-2012 is changing, especially with improvements in technology. efforts to improve its polling.

Silver , Nate , The Signal and the Noise: Why So Many Pre - Reports and Studies dictions Fail — But Some Don’t , Penguin Press , 2012 . The statistician and founder of FiveThirtyEight.com examines “Assessing the Representativeness of Public Opinion Sur - the science of forecasting in politics and other fields. veys,” Pew Research Center for the People & the Press , May 15, 2012 , http://tinyurl.com/7e8gjrr . Articles A research group concludes that telephone polls that include landline and cell users can be accurate. Bialik , Carl , and Harry Enten , “Why Pollsters Think They Underestimated ‘No’ In Scotland,” FiveThirtyEight.com , Barber , Michael J. , et al. , “Online Polls and Registration- Sept. 19, 2014 , http://tinyurl.com/mmqlftj . Based Sampling: A New Method for Pre-Election Polling,” Two journalists explore how polls failed to correctly predict Political Analysis , Summer 2014 , http://tinyurl.com/omj4ch4 . the results of a referendum on Scottish independence. four political scientists outline a new survey method that produces a representative sample of the electorate. Cassidy , John , “Why Didn’t Anyone See the GOP Tidal Wave Coming?” The New Yorker , Nov. 7, 2014 , http://tiny DiGrazia , Joseph , et al. , “More Tweets, More Votes: Social url.com/qzgx5v6 . Media as a Quantitative Indicator of Political Behavior,” A journalist examines how polls overlooked significant Re - Social Science Research Network , Feb. 21, 2013 , http://tiny publican wins in November’s elections. url.com/kk795et . Indiana University researchers argue that Twitter can be an Clement , Scott , “Gallup Explains What Went Wrong in accurate vehicle for predicting U.S. House elections. 2012,” The Washington Post , June 4, 2013 , http://tinyurl. com/o4a64pt . Rosenstiel , Tom , “Political Polling and the New Media The polling organization assesses how its polls missed fore - Culture: A Case of More Being Less,” Public Opinion Quar - casting President Obama’s 2012 re-election victory. terly , Vol. 69, No. 5, 2005 , http://tinyurl.com/oppsesm . The founder and director of the Project for Excellence in Connelly , Marjorie , “Push Polls, Defined,” The New York Journalism studies how the news media’s focus on polls has Times , June 18, 2014 , http://tinyurl.com/kwkeegz . led campaigns to use polls as a political tool. A reporter explains push polls and how they differ from legitimate surveys about politicians’ negative advertising. Zukin , Cliff , “Sources of Variation in Pre-Election Polls: A Primer,” American Association for Public Opinion Research , Enten , Harry , “Are Bad Pollsters Copying Good Pollsters?” Oct. 24, 2012 , http://tinyurl.com/ph8gt9u . FiveThirtyEight.com , Aug. 11, 2014 , http://tinyurl.com/ A Rutgers University professor and former president of the k4fp8bn . American Association for Public Opinion Research explains A journalist looks at whether nontraditional polls, such as how pollsters conduct their surveys.

142 CQ Researcher The Next Step: Additional Articles from Current Periodicals

Bias Cohn , Nate , “Explaining Online Panels and the 2014 Midterms,” The New York Times , July 27, 2014 , McGhee , Eric , “What if election polls are biased against http://tinyurl.com/ldro42r . winners?” The Washington Post , Dec. 8, 2014 , http://tiny The New York Times and CBS News partnered with YouGov, url.com/mbbcfkr . a global market research firm, to more accurately poll An analysis of Senate election polls from 1990 to 2014 Americans before the 2014 midterm elections. suggests that surveys in close races may be subtly biased against the winners in those years, not a political party. McCormick , John , “Polls Misdirected Midterm Narratives,” , Nov. 6, 2014 , http://tinyurl.com/os9 Ranganathan , Meghana , “Where Are the Real Errors in pnyc . Political Polls?” Scientific American , Nov. 4, 2014 , http:// As media outlets, universities, special interests and consulting tinyurl.com/k83l2fh . groups increasingly conduct their own political polls, inaccurate Some political polls may be inherently biased due to sampling outcomes and skewed data are becoming more prevalent. that relies on factors of convenience, such as picking re - spondents who have phones or cars. Social Media

Silver , Nate , “Here’s Proof Some Pollsters Are Putting Moody , Chris , “How the GOP used Twitter to stretch election A Thumb On The Scale,” FiveThirtyEight.com , Nov. 14, laws,” CNN, Nov. 17, 2014 , http://tinyurl.com/lmazaub . 2014 , http://tinyurl.com/mr9upsd . Republican groups used anonymous Twitter accounts to A study shows polls tend to “herd” results, or align them share internal polling data, a potential violation of federal closely with one another, which contributes to inaccurate campaign finance laws. election predictions. Nuzzo , Regina , “The Future of Election Forecasting,” Scientific Cellphones American , Oct. 14, 2014 , http://tinyurl.com/oeuk67h . Polling companies may rely on user data from social media Blumenthal , Mark , and Ariel Edwards-Levy , “Fewer Than outlets such as Twitter and flickr, as well as microsoft’s Xbox Ever Americans Still Use Landline Phones,” The Huffington gaming system, to survey millennials in future election cycles. Post , July 8, 2014 , http://tinyurl.com/luzttf8 . The rising number of wireless-only households poses prob - Oates , Mary , “How , Facebook and Tumblr are lems for pollsters who are restricted from making automated trying to increase voter turnout,” Tech Times , Nov. 4, calls to cellphones under federal law. 2014 , http://tinyurl.com/q2der4x . Social media platforms such as Google, facebook and Fung , Brian , “Pollsters used to worry that cellphone users Tumblr are encouraging users to vote through search functions would skew results. These days, not so much,” The Wash - that locate nearby polling stations. ington Post , Nov. 3, 2014 , http://tinyurl.com/mq9otgo . As decreasing landline use threatens to skew samples and CITING CQ RESEARCHER data, pollsters are minimizing the costs of conducting cellphone surveys by eliminating calls to ineligible voters. Sample formats for citing these reports in a bibliography include the ones listed below. Preferred styles and formats Thee-Brenan , Megan , “How the Rise of Cellphones Affects vary, so please check with your instructor or professor. Surveys,” The New York Times , July 10, 2014 , http://tiny url.com/lo87rs3 . mLA STYLE Because of declining landline use in American households, Jost, Kenneth. “Remembering 9/11.” CQ Researcher 2 Sept. more polling companies are contacting respondents on their cellphones to maintain an accurate demographic sample. 2011: 701-732. News Outlets APA S TYLE Jost, K. (2011, September 2). Remembering 9/11. CQ Researcher, Byers , Dylan , “Fox News accused of breaking exit poll 9, 701-732. rules,” Politico , Nov. 4, 2014 , http://tinyurl.com/lxpknu3 . fox News broke rules set by the National Election Pool, CHICAGO STYLE a consortium of media pollsters, by broadcasting exit poll Jost, Kenneth. “Remembering 9/11.” CQ Researcher , September data from the New Hampshire midterm election two hours 2, 2011, 701-732. before the polls officially closed.

www.cqresearcher.com Feb. 6, 2015 143 In-depth Reports on Issues in the News

Are you writing a paper? Need backup for a debate? Want to become an expert on an issue? for 90 years, students have turned to C?Q Researcher for in-depth reporting on issues in the news. Reports on a full range of political and social issues are now available. following is a selection of recent reports:

Civil Liberties Education Health/Safety Religion and Law, 11/14 Race and Education, 9/14 Reforming veterans’ Health Care, 11/14 Abortion Debates, 3/14 Paying College Athletes, 7/14 E-Cigarettes, 9/14 voting Controversies, 2/14 Dropout Rate, 6/14 Teen Suicide, 9/14 whistleblowers, 1/14 School Discipline, 5/14 Understanding Autism, 7/14 Religious Repression, 11/13 Regulating Toxic Chemicals, 7/14 Environment/Society Treating Addiction, 5/14 Crime/Law Central American Gangs, 1/15 Police Tactics, 12/14 Robotic warfare, 1/15 Politics/Economy Campus Sexual Assault, 10/14 Global Population Growth, 1/15 European Unrest, 1/15 Transnational Crime, 8/14 Protecting the Oceans, 10/14 Nonprofit Groups and Partisan Politics, 11/14 Racial Profiling, 11/13 Housing the Homeless, 10/14 future of the GOP, 10/14 Domestic violence, 11/13 future of Cars, 7/14 Assessing the Threat from al Qaeda, 6/14 Upcoming Reports Infectious Diseases, 2/13/15 Urban Gentrification, 2/20/15 Presidential Power, 2/27/15

ACCESS CQ Researcher is available in print and online. for access, visit your library or www.cqresearcher.com.

STAY CURRENT for notice of upcoming CQ Researcher reports or to learn more about CQ Researcher products, subscribe to the free email newsletters, CQ Re - searcher Alert! and CQ Researcher News : http://cqpress.com/newsletters.

PURCHASE To purchase a CQ Researcher report in print or electronic format (PDf), visit www.cqpress.com or call 866-427-7737. Single reports start at $15. Bulk purchase discounts and electronic-rights licensing are also available.

SUBSCRIBE Annual full-service CQ Researcher subscriptions—including 44 reports a year, monthly index updates, and a bound volume—start at $1,131. Add $25 for domestic postage.

CQ Researcher Online offers a backfile from 1991 and a number of tools to simplify research. for pricing information, call 800-818-7243 or 805-499-9774 or email [email protected].