DAILY NEWSPW comm OF MO 1956 MECHKSAN CONGRESSQONAL ELECTION MPH-6N3 Thais hr the Dome of M. A. AQICHIGAN STATE UNSVERSIW Douglas C. Keéfiey E958 'i-fl‘q II [11 mu; W911" Lin 11)} fl 1121! 12m [w mum ll W276"$ . a 20m ‘1‘ .‘uf'v alt-n“ ‘.~‘ , .‘H HAs V‘v- DAILY NEWSPAPER COVERAGE OF TWO 1956 MICHIGAN CONGRESSIONAL E ECTION CAMPAIGNS By Douglas C. Kelley AN ABSTRACT Submitted to the College of Communication Arts Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of T R OF ARTS 0.4 IIIJ MA: School of Journalism 1958 ”I I -19 1' ‘om . . rgwe “‘9... A-‘ “I U-v ‘4‘ . of 3:137 1 c: .3333? 1‘9f181 31".“ p. 3*:51‘6 t l ‘ . M LA i‘lnl ‘N..Y 12"- V “:6 3.. '1 w of “a: a; “-1. :Cg-Tw ... s' \‘HEi. *T. 1‘:h"8“a“ LI ‘3 lit-1e U; i." ‘e L; ‘1‘: - \fi 0 Is" fls ‘ “V'Lp‘es a h 3:5‘e‘8 of ‘ N \J ‘IAI :4 Douglas C. Kelley The increasingly useful tools of content analysis have previously been used in analyzing newspaper coverage of only 1 congressional election campaign. Although over a dozen reasonably scholarly studies have been published of press coverage of presidential campaigns, when they have found disparity of coverage of the 2 candidates, there has generally been a virtually unanswerable reply that the dis- parity reflected differences in the newsworthiness of the candidates. Closely-matched congressional campaigns are numerous enough that within a short time studies might be made of a fairly large number, under varying circumstances. If such studies were made in marginal districts, where the parties are fairly evenly matched, and if districts chosen included an equal number of incumbents of each party, findings could be of real significance. The effects of incumbency and non-incumbency upon press coverage might also be usefully examined. Newspaper coverage of congressional campaigns is of no little interest pg; g2. Both traditional democratic principles and the present world-wide responsibilities of members of Congress call for voters well—informed of the policy views of opposing congressional candidates. Yet we know very little of what voters learn, through the press or other media, about such candidates. Douglas C. Kelley The present study is limited to coverage by 6 daily papers, for the 6 weeks preceding the election, of theJEEES congressional campaigns in 2 marginal Michigan districts, the 6th (Lansing and Flint) and the 12th (western half of the Upper Peninsula). The 6th District incumbent was a Democrat, the 12th a Republican. Editorial policies ofeflj. 6 papers were pro-Republican, there being no pro—Democratic dailies in Michigan's 7 marginal congressional districts. Our 6 hypotheses, taken together, advanced the propo- sition that among the major determinants of newspaper cov— erage are predispositions on the part of neWSpaper staff members to focus attention on personalities rather than policy issues, on incumbents rather than non-incumbents, and on candidates editorially supported by the paper; but that a fourth influence may be the relative strength of the 2 political parties in a paper's area of major circulation. Major findings were as follows: 1) The 6 newspapers carried little information about local congressional candidates, and very little about their voting records and/or views on foreign and domestic policy issues. The average paper published fewer column inches on both local congressional candidates, in the entire 6 weeks, than it published on sports in a single, typical week-day issue. Mention of foreign policy issues was par- ticularly rare. 4 a” 4...'~.. A—w l n ‘-o v 0 ' 5. ‘ylnI-W‘ .A‘. “1.3 . -pA-tq u-.~.4» A: U... ua‘q cv- I l °9 AA " ' ' ‘5 ‘11:““I‘QV “ ‘QUV~§‘-* ce as 2ch 77?: Douglas C. Kelley 2) Incumbency and at least tacit editorial support by the paper concerned were found to be apparent key fac- tors in a congressional candidate's gaining an advantage in news coverage. Neither incumbency nor editorial support was consistently the more important of the 2 factors, al- though incumbency was more often of greater importance. 5) In their paid political advertising in the 6 pa- pers, the 2 Democratic candidates stressed policy issues, as distinguished from non-policy issues, an average of twice as much as did their Republican opponents. 4) There was found to be no significant correlation between the relative strength of the 2 political parties in a paper's home county and the share of congressional cam- paign coverage given the candidates of the 2 parties. Further study of press coverage of congressional cam- paigns in other marginal districts was recommended, and newSpapers were urged to publish responses of opposing con- gressional candidates to periodic questionnaires concerning their views on important policy issues. _ __—_——_———J - A —_——— DAILY Swf‘l’itté ulchita Appl: DAILY NEWSPAPER COVERAGE OF TWO 1956 MICHIGAN CONGRESSIONAL ELECTION CAMPAIGNS By Douglas C. Kelley A THESIS Submitted to the College of Communication Arts Michigan State University of Agriculture and Applied Science in partial fulfillment of the requirements for the degree of MASTER OF ARTS School of Journalism 1958 The u: Dr. Paul Del“ — — .- ‘— —__ fife, have 1 Yb. Blue:- 3': ACKNOWLEDGEMENTS The understanding patience and helpful suggestions of Dr. Paul Deutschmann and of Cynthia Kelley, the writer's wife, have been greatly appreciated. Thanks are also due ~ Mr. Elmer White, of the Michigan Press Association, for his assistance with the reliability check. ii TABLE OF CONTENTS ACUOWLED GEMENT S 0 0 O O O O O O O O O O O O O O O O 0 LIST OF TABLES O O O O O O O O O O O O O O O O O O O 0 CHAPTER I: CHAPTER II: CHAPTER III: CHAPTER IV: CHAPTER V: CHAPTER VI: INTRODDCTION O O O O O O O O O O O O O O 0 BACKGROUND 0 O O O O O O O O O O O O O O 1956: Presidential 1940: Presidential 1944: Presidential 1948: Presidential 1950: Senatorial 1952: Senatorial 1952: Presidential 1954: Congressional and Gubernatorial 1956: Presidential METHO DS 0 O O O O O O O O O '0 O O O O 0 Scope of the Study The 6th and 12th Congressional Districts Definitions of Terms Hypotheses Methodology Reliability FINDINGS O O O O O O O O O O O O O O O 0 DISCUSSION OF FINDINGS. . . . . . . . . . BIBLIOGRAPHY. APPENDIX. CONCLUSIONS AND RECOMMENDATIONS. . . . . Conclusions Recommendations for Newspaper Action Recommendations for Further Study Coding Card Code Coding Instructions. iii Page ii iv 66 92 101 109 114 Table l. 2. 5. 4. 9. 10. 11. 12. 15. LIST OF TABLES Page Inclusion and Non-Inclusion, by Coders A and B, of the Content of 12 Newspaper Issues . . . . Coding of Campaign Issues by Coders A and B. . . Non-Advertising Items PMblished About Local Con— gressional Candidates . . . . . . . . . . . . Non-Advertising Column Inches Devoted to Local Congressional Candidates, and to Sports . . . Congressional Candidate Advertisements . . . . . Non-Advertising Items Referring to Policy, to Non-Policy and to No Issues . . . . . . . . . References to Policy and Non-Policy Issues in Non-AdvertiSing Items 0 o o o o o o o o o o 0 Treatment Accorded Local Democratic and Repub- lican Congressional Candidates. . . . . . . . Treatment Accorded Local Democratic and Repub- lican Congressional Candidates, Excluding Balanced Items 0 O O O O O O O O O O O O O O 0 Treatment Accorded Local Incumbent and Non-In- cumbent Congressional Candidates. . . . . . . Single-Candidate Advertisements. . . . . . . . . Multi-Candidate Advertisements . . . . . . . . . Rank Orders of Newspapers According to Local Democratic Percentage of Vote, and Demo— cratic Percentage of Space. . . . . . . . . . iv 62- 65 67 68 7O 72 75 78 79 83 89 91 A metal formerly thu increasing> ' 'ce 0111.5 of Fania CHAPTER I INTRODUCTION A member of a prominent American publishing family, formerly the U.S. Ambassador to Italy, has emphasized the increasing world-wide importance of the views and perform- ance of U.S. Congressmen in these words, in a recent issue of Foreig Affairs (Luce, 36, pp. 111-112): Every United States Senator and Congressman functions at one time or other -- and these days with increasing frequency -- inthe field of foreign affairs. Legislators who set about, by voice or vote, to support, amend, cripple or destroy the foreign policy of any given administration are pro- foundly affecting the conduct of our relations with foreign countries, They are, inthe true sense of the word, foreign policy-makers. One has only to consider the impact abroad of certain resolutions ($.55. Suez, Formosa), immigrationand tariff bills, and the vast complex of foreign affairs legislation and apprOpriations, to realize that in the final an- alysis our amateurs in the Congress have more in— fluence on the foreign affairs of the United States, and wield more 'diplomatic' power, than all the prac- titioners of 'organized diplomacy put together. The point which Mrs. Luce and others have made is one of the reasons for the present study. At a time when -- as Barbara Ward of the London Economist has put it -- "America's foreign policy is everybody's destiny,',' it seems appropriate for students of journalism to investigate the extent to which the press is objectiVely reporting the views of congressional candidates on foreign policy and other vital issues. Walter Iflippman has argued (55, p. 50) that the executive branch of government terial and electorate To what ex torate abc propose tc The : course as about thy theorists S‘med u; J have 2 0f tn aisle: ultim denie. mamag It 'hich t‘: preside: mite . government in the western democracies "has lost both its ma- terial and its ethereal powers. The assemblies and the mass electorates have acquired the monOpoly of effective powers." To what extent, then, is the press informing our mass elec- torate about what local candidates for the national assembly propose to do with those powers? The importance of a well-informed electorate is of course as old as democracy itself, and has been lectured about through the years by a long procession of democratic theorists. Some forty years ago Lippman (54, p. 11) ably summed up the argument: All that the sharpest critics of democracy have alleged is true, if there is no steady supply of trustworthy and relevant news. Incompetence and aimlessness, corruption and disloyalty, panic and ultimate disaster, must come to any people which is denied an assured access to the facts. No one can manage anything on pap. Neither can a people. It may be argued that voters don't depend upon their newspapers or other mass. media to furnish information on which they will make their voting booth choices. Indeed, Lazarsfeld, in his study of opinion formation in the 1944 presidential campaign (53, p. xxiii) concluded that”"the ma- jority of people acquired much of their information and many of their ideas through personal contacts with the Opinion leaders intheir groups." But Lazarsfeld continued: "These latter individuals, in turn exposed themselves relatively more than others to the mass media." Thus even though news- paper content about congressional candidates may be of real importance on] opinions of me icance.. is 1 ed it (15, p. We dc will actuai necessity < rflydelegat they trust. for the wit leaders an by making : Perhaps its aren't mm by Personal c: “th from p: Center at the that 26 Per c dent Voters, lens Claim th flaking in at search Center adults Profes ”Warsaw cent °f the v that newspape have sign-fie at least in m Since 1 CORSiderabl importance only for a small segment of the papulation, the opinions of members of that small group are of great signif- icances. As the Commission on Freedom of the Press express- ed 11: (13, p. 28): we do not assume that all citizens at all times 'will actually use all the material they receive. By necessity or choice large numbers of people voluntar— fifuydelegate analysis and decision to leaders Whom they trust. Such leadership doessnot alter the need for the wide distribution of news and Opinion. The leaders are not identified; we can inform them only by making information available to everybody. Perhaps it will be countered that peeple's voting hab- its aren't much influenced either by mass media content or by personal contacts, but almost wholly by predispositions: dating from parental indoctrination. The Survey Research Center at the university of Michigan has found (55, p. 2) that 26 per cent of Michigan adults call themselves indepenr dent voters. Thus better than a fourth of our fellow citi- Qens claim they are willing to assume the burden of decision- making in at least some political campaigns. The Survey Re- search Center also reports (5 ) that only 40 per cent of U.S. adults profess to believe in "straight ticket“ voting. Many a congressional seat hangs on a margin of less than five per cent of the vote, so there would appear to be a good case that newspaper content about congressional candidates may have significant influence on congressional election results, at least in marginal districts. Since 1952, and especially since 1952, there has been a considerable volume of charges that the nation's press is {J permitting its contaminate fr: It is interesti the percentage biased did not lop Poll (21, j ular readers 0. and 9 per cent Uni'versity of P' 1) reWrted readers felt t Stevenson, and t0 Eisenhower. A5 3 me party Dress" ’ Shoe 1957 and t° ““88 the permitting its editorial page Opinions to seep through tb contaminate front-pegs news writing, selection and display. It is interesting to note, however, that there is evidence the percentage of newspaper readers who believed the press biased did not increase between 1940 and 1956. A 1940 Gal- lup Poll (21, pp. 115-119) reported that 27 per cent of reg- ular readers of newspapers felt their paper unfair to F.D.R., and 9 per cent felt their paper unfair to Willkie. A 1956 university Of Michigan.Survey Research Center study (58, p. 1) reported that 24 per cent of the nation's newspaper readers felt the press was unfair (in its news coverage) to Stevenson, and that 5 per cent considered the press unfair to Eisenhower. As a result of the controversy over the alleged "one- party press"; there have been a number of studies published, since 1957 and particularly since 1952, which have attempted to assess the job daily newspapers have done inpresenting news of presidential campaigns. There have also been two studies of press coverage Of U.S. senatorial campaigns (in California, in 1950, and in Wisconsin and Connecticut, in 1952), and a study by a team of graduate students of cover- age Of a congressional and a gubernatorial race (in Wisconp sin, in 1954). In chapter II we shall attempt to summarize the mwthodology and findings Of these studies Of campaign coverage. Despite the attention given press coverage of presi- dential campaigns, and the more limited attention given senatorial c studies of 1 Press tant in and is the writ gressional overall per [zen it is sore Space Ply Simply ness of ca 'l‘Dlly sat in which b maintain t Clinching The candida,“E out the (:01 dates 0f ' Siondl Ca (in news the Party Drags in p to buttre 0n a.essicuaa senatorial campaign coverage, there have been no published studies of press performance in congressional campaigns. Press coverage Of congressional campaigns is impor- tant in and of itself, for the reasons cited above. But it is the writer's view that studies of press treatment of con- gressional candidate news can also shed needed light on the overall performance of the press in political campaigns. When it is found,in studies of presidential campaigns, that more space was given one candidate or the other, editors re- ply simply that space was allotted according tO newsworthi- ness of candidate statements and activity. There is no wholly satisfactory rebuttal, and the result is an impasse in which both the critics and the defenders of the press maintain the righteousness of their position, . but lack clinching arguments to prove it. The writer knows of no evidence that congressional candidates of one political party are consistently, through- out the country, more "newsworthy" than congressional candi- dates Of the other party. Thus, should studies of congres- sional campaign coverage indicate rather consistent favoring (in news space, photos and/or headlines) of candidates of the party editorially favored, critics of the role Of the press inpolitical campaigns would have weighty new evidence to buttress their position. 0n the other hand, should a number Of studies Of con- gressional campaign coverage indicate an approximate over- all balance major partii ance betwee: the two per each of the position to news covers elnost enti presidentia The v Present st». but rather coverage In: research 1: In (:1 “Marne . .VSes of he Chap the preSEn We shall a all balance in news coverage given candidates of the two major parties (or, more realistically, an approximate bal- ance between coverage given incumbent candidates of each Of the two parties, and between non-incumbent candidates of each of the parties) then editors would be in a far better position to argue that what imbalance is found to exist in news coverage of presidential campaigns can be attributed almost entirely to differences in newsworthiness of the presidential candidates' statements and activities. The writer is not suggesting that the results Of the present study alone will perform the function outlined above, but rather that a series Of studies of congressional campaign coverage might serve such a purpose. In any event, further research is needed in this too long neglected field. In chapter II, as suggested above, we shall seek to summarize the methodology and findings Of past content anal- yses of newspaper coverage Of U.S. political campaigns. Chapter III will report the scope and methodology Of the present study, define terms used, and state hypotheses. We shall also briefly describe the nature Of the two Michi- gan congressional districts with.which we shall be concern- ed. Our findings will be reported in chapter IV, and dis- cussed in chapter V. In chapter VI we shall state our major conclusions, and Offer recommendations regarding further study and some possible remedial steps by the press. Cont more and m of content every five PTO! ab0ut of the ce: creased tc in nos 5 rear. 1 Interes calMisha has liars. The 1 ”'3: Presiv on (by the . liahed Study then deVGIOp In 195 . sum;OI storial elec PESulted in [Ls elEfl)°1‘€tte st ”did to date 1:13 1956 Dre CHAPTER II BACKGROUND In 1952 Berelson wrote ( 6, p. 21) that Content analysis techniques have been applied more and more in recent years. In fact, the output of content analysis studies has sharply increased in every five—year interval over the past thirty years. From about two studies a year in the early decades of the century, content analysis studies have in- creased to three a year, then four, eight, twelve, and in.most recent years to an average Of about 25 a.year. Presumably the field is still expanding. Interest in studies of newspaper coverage Of political campaigns has likewise develOped rapidly in the past few years. The first published study of newspaper coverage of a U.S. presidential election was made in 1956, and the sec- 0nd (by the same researcher) in 1944. There was one pub- lished study of 1948 presidential campaign coverage, and then developments came more rapidly. In 1950 a study was made of campaign coverage of a U.S. senatorial election, and in 1952 coverage Of two sen- atorial elections was studied. Charges of "one party press" resulted in seven relatively scholarly and several less elaborate studies Of 1952 presidential campaign coverage, and to date there have been two studies published concern? ing 1956 presidential campaign coverage. More are to be expected, judging from the delay in publication of several of the earlier studies. 1 l The only has been made, four joumalis kee dailies of the two dailie Because of thi “39415118, it findings or 11‘: tom], canpaig the Same as 8] ”Verse. and M W be aim: 01 these Stud: "91'? right t: Please, but 1:] news coverage We pmreren The only study of congressional campaign coverage which has bewn made, to date, was an unpublished 1954 analysis, by four journalism graduate students, of coverage by two Milwau- kee dailies of the Kersten-Reuss campaign. (Also studied was the two dailies' coverage of the gubernatorial campaign.) Because of this lack of material relating to congressional campaigns, it is worthwhile to review the scope, methods and findings of the published studies of presidential and sena- torial campaigns. Problems of methodology were very much the same as are faced in a study of congressional campaign coverage, and many of the findings may reasonably be expec- ted to be similar. (It should be noted that underlying all of these studies there is agreement that newspapers have every right to support, on their editorial pages, whom they please, but that democratic government is endangered when news coverage becomes significantly affected by editorial page preferences.) 1956: Presidential E. 0. Stene of the University of Kansas was the first to publish findings 'in the field of campaign coverage con- tent analysis. He studied 1956 presidential campaign cov- erage by 21 dailies scattered over the United States. For the final month of the campaign he measured the number of column inches devoted to each of the candidates, the "atten- tion value" (prominence of position and size of headlines) of relevant : favorable (w candidate me stories acoo lines accord tene more space ; and l (livid 9'21“) the Save a majc didate who one eICGPtj 'hich R0081 fin-dine (g nedrEr of Diet didate of relevant items, and whether they were favorable or un- favorable (with neutral items counted as favorable to the candidate mentioned). He used a point system to weight stories according to their position, and to weight head- lines according to their size. Stene found that 15 of the 21 papers studied gave more space and attention to Landon, 5 more to Roosevelt, and 1 divided its coverage almost 50-50. He reported (53, p. 214) that "all except one of the twenty-one dailies gave a majority of their political news space to the can- didate Who was supported through the editorial pages." (The one exception was a GOP daily in a strongly Democratic city which Roosevelt had visited.) Also of interest is Stene's finding (ibidg that: In most cases the attention value of news was nearer to a fifty per cent division than the number of column inches. On the other hand the division of picture space was much more favorable to the can- didate supported through the editorial columns. 1940: Presidential There was no scholarly study made of press coverage of the 1940 presidential campaign, but Harold Ickes reported (25, pp. 9-15) that during October, 1940, he had "checks made," on two days, of five "representative" dailies (the New York Times, Chicago Tribune, Pittsburgh Post-Gazette, Rochester Times and.the Los Aggeles Times). "The checks," —¢ Ickes {om presidenti; 40 per cm 96 per cm of the worz Icke: given a to candidate . - and 548 per cent t( °r Bedtime: mm? in Ha °*' eech otl tines 38 m1 tines as m; Seer. gem“ 9201 large, the 10 Ickes found, "showed that the news space assigned to the two presidential candidates varied from 60 per cent for one and 40 per cent for the other in the best of the newspapers, to 96 per cent for one and 4 per cent for the other in the case of the'worst."' (Ickes' ”worst" was the Chigago Tribune.) Ickes said the papers examined ”were found to have given.a total of 2,245% inches of their news space to one candidate -- the one whom a minority of the people favorede -- and 548 inches to the other." Thus the ratio» was 86.6 per cent to 15.4 per cent. Ickes added that another check of Baltimore §gg_coverage of a Willkie rally and a Roosevelt rally in Madison Square Garden in New York, within a few days of each other, revealed that the Willkie rally received three times as much space as the Roosevelt rally, and that five times as many g2; staff members were assigned to cover it. Secretary Ickes concluded, however, that "despite dan- gerous exbeptions and grave shortcomings, we have, by and large, the best press in the world.” In another essay in the same Ickes—edited.volume ( 9, pp. 51-61) Irving Brandt, contributing editor for the St. Louis gtggrgiggg, argued that the press in the 1940 campaign ignored much pro-Democratic and anti-Republican.news, and played up anti-Democratic and pro-Republican news. 125'.- : Presidential In 1945 Stene published his second study of presiden— I tial campaign studied15“pro to! then for preceding the noted space, 8 ever, a single factors inV011 non-candidate p. 260) that ' with the Pres 8711"Prislngly 0f Stem Roosevelt and £115 gave n eClitorially 1 P. 264), "the '3? Or the o1 ll tial campaign coverage, on the 1944 campaign. This time he studied 15 "prominent daily newspapers" (all in large cities), 4 of them for 4 weeks preceding the election, 10 for 12 days preceding the election, and 1 for 6 days. As in 1936, he noted space, attention value, and direction. This time, how- ever, a single index figure was computed on the basis of all factors involved. Items dealing with the President in a non-candidate role were excluded, but Stene reported (54, p. 260) that "the number of items excluded because they dealt with the President rather than the presidential candidate was surprisingly small." 0f Stone's 15 dailies, 11 were pro-Dewey, 2 mildly pro— Roosevelt and 2 strongly pro-Roosevelt. He reported that all 15 gave more favorable news attention to the candidate editorially favored. "It might be argued," he added (ibid., p. 264), "that there is no intent to color news reports one way or the other, but that reporters and editors present news as they see it. So it would be said that the newspaper editors present what they, in their best judgment, consider important, and that news reports are favorable without pur- poseful discrimination. Even if that were true, it would not change the fact that the news attention varies accord- ing to the editorial policies of the papers." 1348: Presidential Martin Millspaugh studied the coverage given the 1948 presidential the homing § all of which ceding the e text and pho items were c didate, or a cartoons was Millet “15th fa “5131 news P351; devote than to new Siderably ] than to nev 12 presidential campaign by four Baltimore, Maryland newspapers: the Morning §_u_n, Evening 5.29.: Eggs-39gb. and Ago—American, all of which editorially favored Dewey. For the week pre- ceding the election he measured space (including headlines, text and photos) devoted to the presidential campaign. All items were classified as favorable to one or the other can- didate, or as neutral. Space devoted to editorials and to cartoons was separately measured. Millspaugh reported (40, p. 125) that the Morning §E§ slightly favored Truman in news space, the Evening Sun gave equal news space to Truman and Dewey, and the Eggggggg§,and Afro-American heavily favored Deweyin news space. The Egg- §g§§_devoted more space to pro-Dewey editorials and cartoons than to news of both sides; the 5 other papers devoted con- siderably less space to their pro-Dewey editorial material than to news. 1 0: Senatorial Next came the first study of coverage of a U.S. sena- torial campaign -- the 1950 Nixon-Douglas campaign in Cali- fornia. The findings were reported by Qalen Rarickin anun- published Stanford University M.A. thesis (48) which the Stanford University Library reports is unavailable. Sum- maries of the study's methods and findings are available, however, in a Journalism Quarterly article by Chilton R. Buah (10) and in an Editor and Publisher article by Camp- bell Watson t The st1 for Journali: headlines an: nia dailies papers) from 1)” a press to September Releva idea, and he ample) pro, ”Wales. Sj Republican 1 (Bought De cPatio leade OVer E statement We to be Pro—h: 01‘ HOt stat! 13 bell Watson (61). The study,made by the Stanford University Institute for Journalistic Studies, covered all news items (including headlines and photo captions) in 12 representative Califor- nia.dailies (7 "independent;? 2 Democratic and 5 Republican papers) from September 1 to November 7, 1950. Unfortunate- ly, a press release announcing the study was issued prior to September 1, so publishers were alerted. Relevant “statements" (expressions denoting a complete idea, and having a relevant referent)were coded as (for ex- ample) pro, neutral or anti-Nixon, pro, neutral or anti- Douglas. Six referents were used: 5 Republican (Nixon, Republican Party, Republican leaders) and 5 Democratic (Douglas, Democratic Party, Truman Administration and Demo- cratic leaders). Over 8,000 statements were coded. If the source of a statement was a candidate or his supporters, it was presumed to be 'pro-hinl' or "anti-his Opponent'; if the source was neutral or not stated, a Judgment was made by the coder as to direc- tion. Just under 35 per cent of the statements coded were neutral. For each candidate, in each paper, percentages were computed to indicate the per cent of references to him/her which.were favorable, neutral and unfavorable. A "favored index" was computed for each paper, by subtracting the net favorable or unfavorable score for the less-favored candi- *l‘fig V‘ \Jo-ne .H ‘J Teri!" .c- ”pan new -“Ae- ’uv. CI ' n‘fl""0 ‘50-“ is alnost Rar- §V_J fi ,.. t :3 1"; now-#- UV“ “1:33 a 1‘5”: .1 9’- ‘ 2“." mg“ the c to 14 date from the comparable figure for the more-favored can- didate. Thus the conclusions offered were based not on how much Space was given each of the candidates, but on the relative frequencies of favorable and unfavorable state- ments about each of them. The average "favored index,? for all 12 papers, was pro—Nixon 29.2. 0f the favorable statements, Douglas received 45.2 per cent and Nixon 54.8 per cent. Of unfavorable state- ments almost 70 per cent referred to Douglas, and 50 per cent to Nixon. Only 14 per cent of the statements coded were unattributed, and the unattributed statements gave Nixon a favored index of 5.8. Of the 11 papers which editorially supported a candi- date (2 for Douglas, 9 for Nixon), 9 gave a favored index to the candidate editorially supported. These 9 included the 2 Douglas papers and 7cnfthe Nixon papers. (Two Nixon papers and the neutral paper gave Douglas a favored index.) Lionel Horwitz wrote an unpublished M.A. thesis on "Political Advertising in the 1950 Wisconsin Senatorial Campaign" (24), but the University of Wisconsin Library indicates the study is unavailable. 1952: Senatorial A study of press coverage of 1952 Wisconsin and Con- necticut senatorial campaigns (18) was published in 1954, by political Eichigan Ste dltes' use c with press c included 11 tral; tOgetI lisconsin) a dailies (no: Per cent of lies include m day, am issue, for 1 tion. The I Entire PMS: All el: and the dim coded (1569 Senate Beat; Any item WI 6:11an 0f : iniiCated p: favorable" A Hum Stag 3 1" 1) 1:. larly 15 by political scientists LeRoy Ferguson and Ralph Smuckler of Michigan State College. They were concerned with the candi— dates' use of policy and non-policy issue appeals as well as with press coverage of the campaign. In Wisconsin the study included 11 dailies (4 pro-McCarthy, 2 pro-Fairchild, 5 neu- tral; together, about 70 per cent of daily circulation in Wisconsin) and 30 weeklies. In Connecticut there were 9 dailies (none of them pro-Democratic; comprising over 65 per cent of daily circulation in Connecticut) and 13 week- lies included. About half the dailies were coded every sec- ond day, and the remainder of the papers were coded every issue, for the 6 weeks preceding the November 4, 1952 elec- tion. The papers were selected to be representative of the entire press of the two states. An elaborate IBM card code covering the use of issues and the direction of items was used, and over 3,500 items coded (1369 in Wisconsin, and 2147 in Connecticut, where 2 Senate seats were at stake due to the death of an.imdumbent). "Any iten which reported criticism of a candidate or indicat- ed lack of support was counted as unfavorable. If an item indicated praise or support of a candidate it was considered favorable" (gpgg, p. 61). A number of the findings of the Ferguson and Smuckler study are relevant to our present concerns: 1) Policy issues (those "which have in recent years faced decision by government policy makers, particue larly.Congress") appeared "much less frequently than --A.—.__ non-pol. (Of all appeals per cen 2) ”i: of sac? to be theig: 3) “l foreil (p. \H Polic nectj 4) 1 P011. (1:. 5) ul>l 'Peeg 5) thz Emu a1; 0&1 530 n E l6 non-policy themes in the press in both states" (p. 28). (Of all tppeal items, 28 per cent were policy issue appeals; of all candidate-originated appeal items, 46 per cent were policy issue appeals.) 2) “Analysis of the individual campaigns in the press of each state does not show the Democratic candidates to be consistently more policy issue oriented than their Republican opponents" (pp. 51-52). 3) "Domestic issues were used more frequently than foreign policy questions in the press of both states" (p. 32). (In Wisconsin, 36 per cent of references to policy issues were to foreign policy issues; in Con- necticut, 48 per cent.) 4) ”Only rarely were the same foreign or domestic policy issues emphasized by Opposing candidates" (p. 45). 5) ". . .Eeferences to personal qualifications made up by far the largest portion of the non-policy ap- peal in the press” (p. 45). 6) In Wisconsin "Senator McCarthy‘s name appeared in three and one half times as many news items as the name of his opponent, Thomas Fairchild. . .McCarthy attention was at least three times as great as Fair- child attention in all but one of the dailies, and in some the ratio rose to as much as seven to one” (p. 60). "Separate analysis of front page items in.the Wisconsin press re tion to dominanc front pa attentic ceived c lines t1 cent of unfavora lines me 7) "Wh: an absoj °°mpare in thei Conclus (Thus a Paper, news fr Wag 91:13 ‘88 was the paE or edit 17 press revealed an even larger preponderance of atten— tion to McCarthy. . .more than four to one. . .McCarthy dominance of news attention is most clearly evident in front page headlines where 85 per cent of candidate attention centered on McCarthy" (p. 61). "McCarthy re- ceived. considerably less unfavorable attention inhead- lines than in news items. . .For example, while 50 per cent of the McCarthy front page attention items were unfavorable, only 26 per cent of the front page head- lines mentioning McCarthy were unfavorable" (p. 62). 7) "While it does not appear possible to construct an absolute index of partiality, it is possible to compare the content of several individual newspapers in their coverage of a particular campaign and to draw conclusions as to their relative partiality" (p. 65). (Thus a "margin of favoritism" was computed for each paper, by subtracting the percentage of Democratic news from the percentage of Republican news.) "As was expected. . .the pro-McCarthy rank in news cover- age was directly related to the editorial policy of the papers” (p. 64). "Furthermore, this relationship of editorial position to news coverage was [in Wiscon~ sin] more consistently obvious in front page headlines than it was in news generally" (p. 66). 8) ". . .The Connecticut Democrats received more fav- orable references than their Republican opponents in 18 what was an editorially Republican press" (p. 72). However, "analysis of the combined Connecticut dailies tends to support the Wisconsin finding that headlines reflect editorial policy more clearly than does gener- al news coverage" (p. 74). Although both Democrats re- ceived more items than their Republican opponents, the Democratic incumbent Senator fared better (60-40 per cent) than the Democratic non-incumbent did against a second Republican non-incumbent (55-47 per cent). (Ferguson and Smuckler do not suggest the possibility, but in the writerls view the "non-incumbent" Democrat's 53—47 advantage over his non-incumbent Republican opponent was probably in large measure due to the fact that the Democrat Eg§,an.incunbent Con ressman, and thus had many of the advantages which normally accrue to an incumbent running for re-election. Neither of the two Republican candidates was holding public of— fice at the time of the campaign nor had ever held. office, to this writer's knowledge. 9) ". . .It seems reasonable to conclude that those elements which would contribute to the enlightened. popular control of public policy were less than ad- equately presented in the press during the campaigns . . .The relatively small attention given policy issues and the almost complete lack of Joining of issues by the candidates would not appear to contrib- a»-__- h._ l9 ute to the role of the election as an occasion for the exercise of popular control over broad policy questions. Furthermore, the findings in Wisconsin indicate that even those issues and themes that did receive emphasis in the press were presented in a manner which might not give the reader enough im- partial information to make a free choice between alternatives. The content of the press appeal to the electorate would tend to turn the attention of the voter away from policy questions and toward DOD? policy themes, particularly the personalities of the candidates" (pp. 77-78). 1952: Presidential The presidential campaign of 1952 marked the high point, to date, of charges and countercharges regarding press coverage. On October 24, 1952, 96 American authors, including 24 Pulitzer Prise-winners andaNobel Prize-winner, in a Joint statement (Editor and Publisher, 45, pp. 7 and 59) attacked the "unrelenting attempt by the majority of our country's newspapers and.magazines to play down one of the candidates" for the presidency. "Partisanship has been allowed to soak through from its legitimate place on the editorial pages and in signed columns. IDthas been allowed to infect the very writing of the news.and, even more, the way it is displayed. Even photographs have been used to mislead. " crate, in. inadequatq stories 1‘; 13% meant A1101 ”1" of t1 1) rep: dai] to t head P011 GOP hOw inen 111R *9 20 The 96 authors' statement was based on ahurried study (by 4 of the authors: John Hersey, John Steinbeck, Herman Wouk and Cleveland Amory) of 26 newspapers in 6 politically doubtful states (Ohio, Massachusetts, Michigan, Illinois, Pennsylvania and California) for an unstated period of time. The public statement included no description of methodology. Only 4 of the 26 papers were found to be fair in their cov- erage. The others were charged (Blumberg, 8, p. 18) with "slanted writing of news stories, prejudiced use of photo- graphs, unfavorable placement of stories concerning Demo- crats, inadequate volume of coverage of Democratic news, inadequate general coverage of Democratic news, omission of stories favorable to Democrats and use of headlines reflect— ing meaning of stories inaccurately." Among other spokesmen for critics of the 1952 campaign role of the press were: 1) Jean Begeman, who in a N_e_! Republic article (5) reported the findings of a study of 21 representative dailies in large and small cities, from September 1 to the end of October. The study was concerned with headline and news play, photographs and editorial policy. The conclusions: there was bias toward the GOP (no figures were reported as to how much or in how many of the papers) in regard to number and prom- inence of items and headlines, headline editorializ- ing, and use of photos. 21 2) Miss Begeman also reported (ibid.) that a survey by Frontieg Magazine of 4 San Francisco dailies for 2 weeks early in the campaign revealed more coverage for the GOP. 5) Some months after the election, the prominent radio and television news analyst Eric Sevareid de- clared (52): The truth must be faced: Dozens of ex- cellent new3papers, with a record of honorable news handling, cheated in their allotment of news and picture Space as between the two can- didates. I do not believe the election result would have been any different, even had abso- lute fairness ruled, but that does not erase this blot on the record. . . 4) Erwin' Canham, editor of the Christian Science Monitor, wrote (12, p. 111) that ". . .enough evi- dence is at hand to indicate that a considerable nump ber of American newspapers did not adequately live up to their own precepts in the 1952 campaign." He re- minded his readers, however, (ibid., p. 114), that "except in a few newspapers in a few countries reach- ing a few people, there are great news inadequacies throughout the world." The first extensive reply to critics of the role of the press in the 1952 presidential campaign came in the form of a one-day study by the Associated Press of the front pages of 115 dailies. There were 52 morning papers, and 63 even ning papers; the day chosen was Oétober 10, a day when it is reported both major candidates as well as their chief cam- tre' -. 11‘ DC , ‘ as r .3“ U -. a -* el nu f G» a» do ya”. at a V l e S 27;?- ,1 Q the: T in n+- fiA‘ U stor , as M H“ link S\\V' r ce vdt ad do Delitic' “.2 care acc ‘3 241 / H k U. 22 paigning supporters were in action. Stories on both the candidates and their active supporters were counted and their prominence noted. Egipqp g Publisher reported (45, p. 54) these findings: Thirty-two per cent of the newspapers examined gave equal treatment to either Eisenhower and Steven- son or Eisenhower and Truman. Forty-seven per cent gave the Democrats the dominant coverage. Twenty-one per cent gave the edge to Eisenhower. .. . .a major- ity of newspapers, which supposedly have been support— ing the Republican candidate in their editorial col- umns, have been leaning over backwards so far to be fair in their news coverage that they have actually been giving 'the breaks' to the Democratic candidate. In a November 15, 1952 "Post-Election Round-up", Edi- tor § Publisher reported (46, p. 10): Newspapers all over the country are toting up their political campaign news coverage with the same care accorded advertising linage in order to disprove politicians' charges of news bias by a so-called 'one- party press.‘ Most of them are reporting their news stories about Eisenhower and Stevenson about equal, but when Truman's whistle-stop campaign stories are added to the Stevenson total, the news space was dominated by the Democrats. Most interesting of the several examples cited in the Editor §_Publisher article were these two (ibid.): The Cincinnati (Ohio) Volunteers for Stevenson made a survey of the city's three newspapers. It re- ported each candidate received 50 per cent of the play in the Cincinnati Enquirer and the Cincinnati Post and that the Cincinnati Times-Star gave 56 per cent for Eisenhower and 44 per cent for Stevenson. The Baltimore SEQ reported that from October 1 to November 4, it printed168 letters of which 121 dealt with election issues. It said 87 took definite sides in the presidential election, with 58 for the Stevenson-Sparkman ticket and 49 for Eisenhower-Nixon. 'The published Eisenhower-Nixon letters were 56.5 per cent of the total, and the figures show that Eisen- hower-Nixon captured 55.8 per cent of Maryland votes;? the Sup said. In torial p (15. p- l tor g callec exten: gree c carpal A reel sional propos sive 9 major object commit tion. cial c be 'fe Whe: 13% in th holds 0th “Willing At reSolutio. “waign - 11001. or Ciety or ‘ 23 In the May 25, 1955 New Leader, Irving Dilliard, edi- torial page editor for the St. Louis Post-Dispatch, wrote (15 9 P0 17): The votes were not yet fully counted when Edi- tor g Publisher, the newspaper world's trade weekly, called in its November 15 issue for 'an impartial, extensive, scientific study to reveal the exact de- gree of fairness or lack of it in this Presidential campaign.‘ The demand for a survey spread quickly. A week later, Sigma Delta Chi, the national profes- sional journalistic fraternity, adOpted a resolution proposing that the fraternity sponsor a comprehen- sive survey with the help of one of the country's major foundations, the details of the 'thorough and objective analysis' to be worked out by a special committee. It looked as though we were getting ac- tion. But now, six months later, we learn the spe- cial committee has reported that a survey would not be 'feasible' . . . When another six months had passed, Dilliard was writ- ing in the Nieman Reports (16, p. l4)that ". . .the press holds other institutions up to searching scrutiny, but is unwilling to have the same scrutiny applied to itself." At the next (1955) convention of Sigma Delta Chi, a resolution urging that a study of press performance in the campaign be undertaken as soon as possible did not reach the floor of the convention. In the interval the American So- ciety of Newspaper Editors and the American New3paper Pub— lishers Association had each turned deaf ears to pleas for an analysis of 1952 campaign coverage (American Society of Newspaper Editors, 2, pp. 184-185, and Blumberg, 8, p. 25). Although some members of these professional organizations took the attitude that what newspapers published was no one 's business b1 parently p1 Signs Delta erly to we: collective which we 11: Deep: ride study hails in so: scholarly ; ited to a eisc0) whi Covered bu reSpective Sidn October 1 21, Steven 24 business but the editors' and publishers', most of them ap- parently professed to agree with the report of the special Sigma Delta Chi committee (Dilliard, 15, p. 17), that: "Pr0p- erly to weight. . .material objectively to arrive at a sound collective judgment of bias or lack of bias is a problem for which we have found no feasible solution." Despite the demise of efforts for a thorough, nation- wide study of press coverage of the 1952 campaign -- or per- haps in some cases because of that demise -- 7 more limited scholarly studies were made and published. Three were lim- ited to a single city or state (Florida, Wisconsin, San Fran- cisco) while 4 were nation-wide in scope (although 2 of them covered but 8 papers; the others covered 51 and 55 papers, respectively). Sidney Kebre studied 54 of Florida's 57 dailies, from October 1 through November 5. (Eisenhower had the support of 21, Stevenson the backing of 8, and 5 were neutral). He noted, for each political party, in each paper, the number of news stories, column inches, headlines (and size of same), front page stories (and position of same), photos, editor- ials and articles by national and local columnists. Kobfe found (28) that although there was wide varia- tion among the papers, an average of 51 per cent of the news stories were Democratic and 49 per cent Republican. Repub- licans were favored in size of headlines, number of editor- ials and number of articles by columnists. Democrats were 25 slightly favored in column inches, number of front page stories, position of front page stories, and number of photos. Kobre found that about half the papers studied did not favor, news-wise, the candidate editorially favored. He advanced no possible explanations for this departure from the findings of most previous studies. Since Kobre's has been the only study of newspaper campaign performance in a solidly Democratic environment, however, it may be'mrit the explanation lies in a hesitancy -- conscious or uncon- scious —- on the part of newspaper staff members to risk running counter to strongly held traditional views of a large majority of readers. Kobre concluded (ibid., p. 169): It would be practically impossible for many pa- pers to strike an exact equal balance in news treat- ment and give a 50-5O play. The news just doesn't break that way, and certain stories may have more sig- nificance and human interest than others. Butsirange such as 52-48 per cent, can be established. . .Beyond these upper and lower limits, in all fairness to the public, the paper should not deviate. Charles E. Higbie studied the front pages of 14 Wis- consin evening dailies for the 8 weeks immediately preced- ing the election. He does not report (25) his criteria for choosing those 14 papers, nor how many were independent po- litically; he mentions that only one was Democratic. He counted, for each party (and each particular leader) the number of items, column inches, and photos, noted the page position of each item and analyzed headlines and news story corded ve GOP presi headlines Eisenowe held a re Sparkman dential a it was fc °°mPared likeness His EOre c105 ial POlic KObrels I In dential < Blumb e rg 26 leads. Higbie found (gigwp. 56-57) that ". . .although major Democratic and Republican national figures were ac- corded very nearly the same amount of front page space, the GOP presidential candidates appeared more often inthe major headlines and in the news pictures of the campaign. . . Eisenhower and Nixon subject matter heads in top position held a revealing 65 to 57 Per cent lead” over Stevenson- Sparkman heads. He added (3339, p. 60) that "when presi- dential and senatorial candidates were all counted together, it was found that Republican likenesses appeared 152 times compared to 81 times for Democratic candidates. Eisenhower's likeness appeared in 75 cuts to 49 for Stevenson." Higbie commented that his finding that headlines are more closely related than is total news coverage to editor- ial policy "conforms in general" with the findings of Kobre's Florida study. In the first hard-cover book in the field of presi- dential campaign news coverage analysis (£3), Nathan B. Blumberg reported the findings of his study of 55 repre- sentative U.S. dailies for the 50 days preceding the elec- tion. Editorially, 26 of the papers were pro-Eisenhower, 7 were pro-Stevenson and 2 were neutral. Front pages were studied for all 50 days, and inside pages on 5 days chosen "be- cause they are well-spaced and represent days on which there were no especially important political developments which would favor one party." Blur l evaluai statist of phoi lines a statist issues, to see ences 1 Display lines)1 it? 36 ed. j circuns paper's dates) alterir 27 Blumberg explained (ibid., pp. 54-55): To arrive at the conclusions in this study, evaluation first was made of the categories of the statistical table: number of stories, column-inches of photograph space, number of multi-column head— lines and total column-inches of space. Then the statistical results of inside page coverage in three issues, including the same categories, were examined to see whether there were any significant differ- ences between front page and inside page treatment. Display (primarily banner headlines or top play head— lines), position of stories and examples of partial- ity partiality is never defined] were then apprais- ed. inally, consideration was given to special circumstances (for example, the visit to a news- paper's city by only one of the presidential candi- dates) which would justify a subjective judgment altering the statistical findings. He added (ibid., p. 58): Many subjective decisions, some of them made only after months of pondering the problems, were necessary before the conclusion of this survey. After Blumberg had made both quantitative and qualita- tive ;guggments, he concluded that of the 26 pro—Eisenhower papers, 15 showed no partiality, 4 did show partiality in favor of their candidate, and 7 showed only inconclusive evidence of partiality. Of the 7 pro-Stevenson papers, Blumberg determined that 2 showed no partiality, 2 showed partiality in favor of their candidate, and 5 showed only inconclusive evidence of partiality. Of the 2 papers which were editorially neutral, one showed no partiality and one gave insufficient evidence to warrant a conclusion of par- tiality. Thus Blumberg concluded that only 6 of the 55 news- papers examined provided evidence of partiality in their news columz charges of rents as t1 took an edi amount of 1 ported, am Blunt elusive" e1 one notes 1 not only £3 Democrats 1 lished the Plate on p, it 9:25 To: If doubt a] “Slit wen 28 news columns. In thereby "clearing" most of the press of charges of bias, he minimized the importance of such state- ments as this (ibid, p. 45): "Of the 55 newspapers which took an editorial stand on the campaign, 22 gave a greater amount of front page coverage to the candidate they sup- ported, another factor which appears significant." Blumberg's decisions that certain papers gave "incon- clusive” evidence of partiality seem open to question when one notes that classified with such papers was one which not only gave the Republicans twice as much space as the Democrats (Ike visited the city, says Blumberg) but pub- lished the following eight-column skyline over its name plate on page one: "Ike Will Arrive at Montgomery Stadium At 9:25 Tomorrow Morning; Let's Be There To Say”Howdy.'" If doubt about such a paper existed, it would appear it might well have disappeared when it was found the paper also ran a three—column photo of Eisenhower with the cap- tion: "The friendly, down-to-earth personality that has endeared Dwight D. Eisenhower to all America- is reflected in this portrait. . ." (ibid” p. 48). Robert Batlin, advancing the dual hypothesis that news bias did exist in 1952 campaign coverage but was not as great as in earlier campaigns (1+), studied 5 San Fran- cisco dailies from October 1 to election day, 1896, and for the same period in 1952. He chose, in each case, 5 papers with the largest circulations. Two of the 5 papers studied '1: among the ic nominee 1952, by z The as was us; smarizec their lea; eluded a ¢ Batl 18% the 1 58.2, for balanc. 11 (In each 1 his Senat. (avenge . 29.2. Bat °°1umn in Demcrats in 1895. Alt 1 29 studied in 1896 -- the 2 which were Republican -- were still among the top 5 in 1952. The third, which supported Democrat— ic nominee William Jennings Bryan in 1896, was replaced, by 1952, by a third GOP paper. The method employed was that of analyzing "statements," as was used in the 1950 California Senate campaign study summarized above. The study also: covered all photos and their length in column inches, evaluated display, and in- cluded a qualitative analysis. Batlin's hypothesis was supported by his findings. In 1896 the news coverage imbalance indexes ranged from 59.5 to 58.2, for the party editorially favored. In 1952 the im— hbalance index (average of the 5 papers) was down to 21.4. (In each case, in 1952, it was pro—GOP.) The 1950 Califor- nia senatorial campaign's pro-Nixon index of imbalance (average of 12 papers), it is interesting to recall, was 29.2. Batlin found that photo coverage in 1952 was 1,751 column inches for the GOP, and 825 column inches for the Democrats, but points out that this was less extreme than in 1896. The same is true of display favoritism. Although Batlin found that the 1952 papers devoted less space than had the 1896 papers to campaign news and to campaign photos, there were indications that the gualigz of coverage improved. In 1896 41 per cent of all statements were favorable or unfavorable, and 59 per cent neutral; in '. N e w -. 50 1952 84 per cent were neutral. In 1896 reporters and ed- itors were the sources for 57.1 per cent of directional statements; in 1952 the figure was 12.6 per cent. Also of interest (ibid., p. 505): foreign policy and national security, the second-ranking theme in 1952, accounted for almost 22 per cent of all directional ma- terial, but in 1896 it embraced only 6.11 per cent. Malcolm W. Klein and Nathan MacCoby studied 8 dailies in 4 cities (a Republican and a Democratic paper in each city) for from October 1 through November 5, 1952. They examined only the front pages and inside page continuations of front page stories; only stories on the presidential can— didates themselves were considered, and only photos refer- ring to news stories already included. The analysis was both quantitative and qualitative. "Bias" was defined (27. p. 289) as "a differential, larger than could be expected by chance alone, between the proportional front-page cover— age allotted the two candidates by the two sets of papers." Klein and MacCoby found that l) the 2 sets of papers gave the editorially supported candidate more coverage than the opposed candidate; 2) placement of stories and numbers of photos both reflected bias, although captions and cut- lines were unbiased; and 5) in the total of 571 items in- cluded in the study, 519 biased remarks were found —- more often applying to the favored than to the opposed candidate. They added, however (ibid., p. 295), that "it should be care- fully note was subjec The type 0 of such fa Gran judged by treatment, heir cont its lack o Pric his findin istic idea cfindidates handling 0 the I"Tort exWart pan epapers s Pric 51 fully noted that it has not been stated M set of papers was subject to more bias, nor what degree of bias exists. The type of analysis used does not permit the determination of such facts." Granville Price studied campaign coverage by 8dailies judged by a panel to include: 1 paper strongly GOP in news treatment, 2 mildly GOP, 5 balanced and 2 mildly Democratic. Their content was analyzed during a 2 week period chosen for its lack of any sensational political news on either side. Price has reported (47) chiefly his method rather than his findings. He rejected equality of space, as an unreal- istic ideal in the light of the varying newsworthiness of candidates, and advocated, instead, comparing different papers' handling of political campaign news. He sought to compare the reporting or non-reporting of 55 "events" (judged by an expert panel to be 11 Democratic and 24 Republican), by the 8 papers studied. Price did not report in the article cited his findings regarding coverage or non-coverage of the 55 events, but did indicate (£919, p. 456): Partisan direction was assigned to 247 pictures: 95 Democratic and 154 Republican. This was a heavier Republican proportion than was found in news stories. Only one paper carried more Democratic than Republican pictures. The paper which readers had called strongly Republican had nearly four times as many Republican pictures as it did Democratic. Arthur E. Rowse, a copy editor for the Boston Traveler, studied the way 51 of the largest U.S. dailies handled the story of Richard Nixon's year-'round political fund when that story brol called it paign,“ a: one of the In a his findir September her 28, ti closure re daily Cire Circulatie their pub: tbmough 0‘ Row: Nixon sto: 18) Paper; ass°¢iati. my 4 of 32 story broke during the 1952 campaign. Editor g Publisher called it the "biggest news break in the 1952 election cam- paign," and it was listed by Associated Press editors as one of the top 10 stories of 1952. In a book entitled Slanted Ngwg (49), Rowse reported his findings of press bias. The 51 papers he studied (from September 18, the day the Nixon story broke, through Septem- ber 28, the day a second story on the Stevenson fund dis- closure reached its close) represented 27 per cent of total daily circulation in the U.S. The 51 were among the 57 top- circulation.dailies; 6 of the 57 had been excluded because their publishers were already represented in the study, through other papers. Rowse reported that the United Press distributed the Nixon story in time for most Thursday afternoon (September 18) papers to carry it. By early Thursday evening all press associations were carrying it. Rowse found, however, that only 4 of the 15 evening papers he studied put the story on page 1 that Thursday. One was the Chicago 23111 Eggs, carry- ing the Peter Edson column which broke the story (along with the £93 1931; M and Los Angeles M Eggs); the other 5 were the only 5 pro-Stevenson papers of the 51 studied, IFour other papers ran the story inside Thursday afternoon, and 5 waited until Friday to pick it up. (All but 1 of the 15 papers had united Press service, and several had received Edson's column, according to Rowse.) or t Friday on omitted t? page l 11:11 In e found the placement that page; Nixon sto: than the e Pens stud ment of t? about Ere ROWSQ 18f 33 0f the 18 morning papers studied, 8 ran the story Friday on page 1; 7 ran it Friday on an inside page, and 5 omitted the story altogether, 2 of them not running it on page 1 until Sunday. In contrast, Rowse found that 26 of the 51 papers found the later Stevenson fund disclosure worthy of page 1 placement the same day that story broke. Rowse emphasized that papers which waited until the second day to carry the Nixon story tended to stress Republican rebuttals rather than the original charge. He concluded that of the 51 pa- pers studied, only the N2! £23k Timgg was fair in its treat- ment of the 2 stories. In then proceeding to generalize about press bias, on the basis of his limited findings, Rowse left the realm of objective research. 1954: Congressional and Gubernatorial Four Marquette University journalism graduate students studied coverage by the Milwaukee Journal and thelMilwaukee Sentinel of the 1954 Kersten-Reuss race for Congress in the 5th District (in Milwaukee), and the Hohler-Proxmire race for the Governorship of Wisconsin, the same year. All 4 students planned Master's theses, and 5 of the theses (14, 59, 60) are summarized below. The fourth thesis, in which Thomas W. Brennan was to have analyzed the 2 papers' edi- torials relating to the 2 campaigns, is not available from the Marquette University Library. can, pen of news 54 Robert H. Weber used a column inch method of analyz- ing coverage of both campaigns by both papers. For the period from September 15 (the day after the primary) to November 2, Weber studied the source (Democratic, Republi— can, neutral or editorial), the direction and the referents of news story column inches, news story leads and heads; he studied the positioning of stories (noting positions of stories on the first pages of the first and second sections of the 2 papers); and he noted pictures published. Cover— age of a special congressional investigating committee bearing the name of the incumbent Republican Congressman (Kersten) was noted separately, as were the source and di- rection of "single source" stories (those in which all ma- terial was attributed to 1 source). Finally, he studied the reporting or non-reporting of events, noting stories appearing i‘n-one paper but not in the other. In the gubernatorial race no stories regarding Gov- ernor Kohler's official activities were included; in the congressional race most stories relating to the incumbent Congressman's official activities were connected with the congressional committee mentioned above, and were tabulat- ed separately. No stories about future speeches or events of any kind were included in the study -- an exclusion which the present writer feels was unjustified. Weber reported (62) he found that in both races both papers gave more news coverage to the candidate editorially 55 favored. The Milwaukee Journal was pro—Democratic in most of Weber's numerous categories and sub-categories of cover- age, and the Milwaukee Sentinel was pro—Republican in all of them. In 15 of 16 categories, the Sentinel was more strong-— ly pro—Republican than the Journal was pro-Democratic. For example: in the congressional race, 76.5 per cent of the Sentinel's relevant column inches were pro- Republican or anti—Democratic, and 15.5 per cent pro-Dem— ocratic or anti-Republican. Corresponding figures for the Journal were 52 per cent pro-Democratic or anti-Republican, and 58 per cent pro—Republican or anti—Democratic. 0f the "single source" stories on the congressional race, 49 per cent of the Sentinel's were from GOP sources, and 15 per cent from Democratic sources; 52.5 per cent of the Journars: were from Democratic sources, and 29 per cent from GOP sources. The one category in which the Democratic Journal's coverage was more lop-sided than that of the Republican Sentinel was reporting or non-reporting of events. The Journal omitted 19 GOP-source stories the Sentinel pub- lished (and only 6 of Democratic-source), whereas the Spa- tiggl ommitted only 18 Democratic-source storieS'mreJOurnal published (and 6 of GOP-source). Weber concluded that since the papers' coverage of the same events differed so consistently, bgth could not have been upbiased. He added (ibid., p. 79) that "its possible t: treatment 1 al ways. it? .11 56 possible both were wrong. The theoretical ideal of equal treatment was at least approached by the Journal in sever- al ways. The Sentinel's coverage never approached equal- ity." John H. Crowley did a Bush type statement analysis (14) of the Journal's coverage of the 2.campaigns, and Richard J. Voelkel did the same type of study (59) of the Sentinel's coverage. Both studies included stories on the official (non—campaign) activity of the incumbent Congress- man and incumbent Governor. Both students notad'mresources, referents and direction of all statements. The findings of the 2 statement analysis studies sup- ported Weber's finding that the Sentinel gave the Republi- can candidates an even better break than the Journal gave the Democratic candidates. Crowley reported, for example, that in the Journal, 56.7 per cent of all relevant state- ments were pro-Democratic 0r antirRepublican, whereas 25.4 per cent were pro-Republican or anti-Democratic. Voelkel reported that of all relevant statements in the Sentinel, 47.6 per cent were pro-Republican or anti-Democratic, whereas but 12.7 per cent were pro-Democratic or anti- Republican. Crowley and Voelkel found that in virtually all categories of coverage, each paper's news coverage favored the candidates editorially favored. Robe: Dispatch, e gig; nagazi I: Democrat tines d1 the Rape forty-fl and the times (1: Republie We natural Ram 11 In ae there haVe coverage p1 J3me: PennsylVan; the 00’6on or 24 rem were not :2 mm Were 57 1956: Presidential Robert Lasch, editorial writer for the St. Louis Post- Dis atch,‘ wrote in the November, 1956 issue of The Progres- sive magazine(30)a In the seven papers I read, Stevenson and the Democrats received t0p position on Page One thirty times during a two-week test period. Eisenhower and the Republicans received top position on Page One fortybfour times. On the inside pages, Stevenson and the Democrats received top position forty—eight times during the same period. Eisenhower and the Republicans got it ninety-five times. Was it coincidence, political bias, or the natural result of the 'ins' making more news? I leave it to you. In addition to the rather hasty study reported above, there have been two more ambitious studies of 1956 campaign coverage published thus far. James W. Markham and Guido H. Stempel, III, of the Pennsylvania State University School of Journalism, studied the October 1 to November 7 presidential campaign coverage of 2# representative Pennsylvania dailies. (Sunday issues were not included in the study.) The 24 papers included 18 which were pro-Eisenhower, 2 pro-Stevenson, and 4 neutral. Stories dealing with "non—campaign activities" of the Presi- dent were not included. The campaign news output of the wire services reaching the 24 papers was also analyzed. The number of column inches for each party, size and direction of headlines, issue frequency (22 issues were coded) and the number and size of pictures were noted. 58 Markham and Stempel reported (58) that the GOP was given 55 per cent of space, 56 per cent of headlines and that Demo- crats had 51 per cent of issue references. They added.(3£gg., p. 16): "For none of these three measures can we be sure that the State's [129] papers were not divided 50-50. The differences obtained from the sample may merely reflect normal fluctuations of a sample from its universe." Of photo space, they reported, Republicans had 65 per cent, and of front page space, 70 per cent. Also, a table in- cluded in their report (ibid,,Table II, p. 19) reveals that of front page headlines wider than 2 columns, Republicans had 65 per cent. Markham and Stempel concluded (ibid., p. 57) that; "The evidence, therefore, provided no [sic] basis for the hypoth- esis that the papers slanted and displayed the news in the direction of the political leanings of their editoriali pages.” They stated there was no significant difference be- tween.the two Stevenson papers' coverage and the 18 Eisen- hower papers' coverage of the campaign, but also reported that the Stevenson papers gave the Democrats 49 per cent of space and the Eisenhower papers gave them but 45 per cent. (No such comparative totals of phgtg space in Stevenson and Eisenhower papers are given.) The findings for each of the 24 papers studied are listed, but the editorial preferences of the individual papers are not stated, so the reader may not see for himself what differences existed. 59 It should also be noted that Markham and Stempel re- ported (ibign p. 50) that the output of the wire services was as a whole slightly Democratic in both space and issues. No suggestion is offered as to how or why that ratio became changed to 45 per cent Democratic space versus 55 per cent Republican space, in the 18 pro—Eisenhower papers studied. The poor design of the Markham and Stempel study made almost inevitable their conclusion that there was no evidence that papers' editorial views affected news coverage. Eflbparb ity of coverage would.have had to have been very great in- deed before findings regarding a sample of 24 papers would have offered conclusive evidence that Pennsylvania's 129 dailies weighted their news coverage in the direction of editorial preferences. It can also be argued that -- even if the study design is accepted -- the authors‘ conclusion was unjustified when their own figures had revealed that in nearly every category of coverage the number of papers favor- ing the Republicans was much larger than the number favoring the Democrats. There is little likelihood that these re- peated evidences of disparity of coverage reflected simply "normal fluetations of a sample" from a universe character- ized by parity of coverage. Finally, Markham and Stempel's findings seem particu- larly questionable as a result of an article in Journalism Quarterly (37) in which they summarized the methodology of their study. Despite the fact that the pamphlet report of 40 their findings had clearly stated (58, p. 9) that "four measures of coverage were used" (column inches, display, frequency of themes, and size and direction of pictures), in the Journalism Quarterly article Markham and Stempel reported (57, p. 187) that "three measures of coverage were used." ”Results of the three content measures correspond closely despite some difference in the averages," they added (ibig,,p. 189), completely ignoring the fact that there had been a fourth measure, and that it did 29;: correspond closely The fourth measure had indicated that Republican photo space was nearly double Democratic photo space (65 per cent Repub- lican, 55 per cent Democratic). Markham and Stempel found (58, p. 24) that "the de- fense issue. . .turned out to be the major issue of the cams paign, getting almost exactly 25 per cent of the issue ref- erences for the papers studied. Foreign policy was second, getting almost 11 per cent, plus the 4 per cent that went to the Suez issue. . ." William H. Taft, of the journalism faculty of the University of Missouri, studied October 1 through November 7 coverage by 5 Missouri dailies of the 1956 presidential campaign. His findings were reported in.2hg Missouri Egggg have (56). The 5 papers Taft studied included all 5 of the "in- dependent" Missouri dailies, l of the 16 which supported Eisenhow can thus parity 0. more lib 1y balan. Sh presiden' Photos, 1 also rec: pers 817m Ta neWSpape- 41 Eisenhower, and 1 of the 15 which backed Stevenson. (It can thus be argued that the study was "loaded" in favor of parity of coverage, assuming that independent papers are most likely than editorially committed papers to offer even- 1y balanced campaign news coverage.) Stories dealing with the 2 presidential and 2 vice- presidential candidates were measured and headlines noted. Photos, cartoons, editorials and letters to editors were also recorded -- and it was found that only 1 of the 5 pa- pers studied published any letters concerning the campaignl Taft's conclusion (ibid, p. 16) that ”if these five newspapers are representative of the publications inMissouri, ’the 'one-party press' charge would have little place in our Journalistic picture" is difficult to understand in the light of his own statement of findings. He found that the 5 inde- pendent papers and the pro-Eisenhower paper all published more Republican than Democratic news, and that the pro- Stevenson paper published more Democratic than.Republican news. Of total column inches of campaign news in the 5 pa- pers, 44 per cent dealt with both parties, 51.4 per cent with the Republicans, and 24.4 per cent with the Democrats. No headline figures were reported. There will probably be additional studies of coverage of the 1956 campaign published within the coming year or two. One such possibility died aborning, however, when a proposal (26) by a group of researchers under the sponsor- 42 ship of Sigma Delta Chi and the Council on Communications Research (of the American Association for Education inJour- nalism) failed to elicit desired support from a selected group of publishers and editors. The preposal outlined an extensive study of press coverage of the 1956 presidential campaign. Two other studies should be at least briefly mentioned. , Charles Zuegner of the Elgin, Illinois M Courier-Nils re- ported.(65) he found "political bias was definitely present in the coverage of the first 100 days of the Eisenhower Ad- ministration" in both a Democratic and a Rephblican paper (the St. Louis ngt-Dispatch and the Kansas gipz,§§gg). ”Where the §E§£ was syppathetic, the Past-Dispatch was criti— cal." The staff of the Columbia University Bureau of Ap- plied Social Research similarly found (57) considerable vari- ation.in.the coverage of 9 New'York City papers of the 1948 Edward U. Condon security case. Strongly Republican papers published a much higher percentage of statements critical of Condon, in their news stories on the case, than did Democra- tic and independent papers. A review of the foregoing indicates that of some 17 reasonably scholarly studies of campaign coverage, l£2found that well over half the papers studied gave more news cov— erage to the candidate or party editorially favored. Thus the bulk of the findings of the "experts" would appear to support the 27 per cent of the nation's newspaper readers who feel Researci of both E: anusri: opposing imbalanc lines 51 prove 11 used in 118 inc} tistics the use favorabj of the . 45 who feel (according to the University of Michigan's Survey Research Center) their newspapers do not give candidates of both parties a fair shake in the news columns (58, p. 1). Most of the students whose work we have attempted to summarize agree that a 50-50 division of news space between Opposing political candidates is an unrealistic ideal. An imbalance in Space and/or number of stories, photos or head- lines suggests the possibility of press bias, but does not prove its existence. Two approaches have therefore been used in attempting to escape the pitfalls of simply "count— ing inches." They have been the comparing of coverage sta- tistics of papers of varying political points of view, and the use of the Bush method of comparing the percentages of favorable and unfavorable statements published about each of the candidates. Despite increased use of these 2 criteria for the existence or nonexistence of "bias," however, there does seem to be a continuing assumption that an unhealthy situ- ation arises when candidates of one political party are consistently given more news space, items, headlines and prominence than candidates of the other party. CHAPTER III METHODS Scope of the Study Content analysis of hundredscflfissues of newspapers, or of any other extensive communication material, is so time-consuming a task that rather severe limitationsxmrat be placed on the sc0pe of a study such as the present one, if it is to be feasible for a single part-time researcher. In delineating the scope of the present study, it was first determined that it should be concerned with coverage of ggggpessional rather than presidential or senatorial cam— paigns. (Coverage of only one congressional campaign had been the subject of content analysis, and as suggested in chapter I, the very number of congressional campaigns means there are greater possibilities for comparison of neWSpaper performance under varying circumstances than is the case in regard to coverage of presidential campaigns. The writer's greater familiarity with Michigan jour- nalism and politics led to the choice of congressional dis- tricts in this rather than another state. It was felt de- sirable to choose "marginal" districts (districts in which the winning candidate has -- in at least one recent elec- tion -- won 55 per cent or less of the total vote), and to choose sue}; ,- ‘JiOIl Of the have severa oresuably c 45 choose such districts outside the area of major circula- tion of the Detroit dailies. In "safe" districts, of which each of the major parties have several in Michigan, press performance in the campaign presumably can have no decisive effect on the outcome of the election, and thus editors may have less incentive1x>slant their news coverage. In marginal districts, on the other hand, press performance might well affect the outcome. For reasons stated below it was decided to code all issues of the papers to be studied, during the 6 week pe— riod selected. Thus it was concluded that congressional districts within the area of major circulation of the De- troit dailies should not be included in the study. (With- in a given amount of time, more newspapers could.tmzstudied if the bulky Detroit dailies were excluded.) Thisdecision knocked out 5 of the state's 7 marginal congressional dis- tricts (the 7th, north of Detroit, the 17th, in suburban Wayne County, and the 18th, Oakland County, to the west of Detroit). The 4 remaining marginal districts included 5 with Republican incumbents (the 9th, extending up the western edge of the Lower Peninsula, the llth, comprising the eastern half of the Upper Peninsula and the northern tip of the Lower Peninsula, and the 12th, the western half of the Upper Peninsula), and 1 with a Democratic incumbent (the 6th, near the center of the Lower Peninsula). In order to in- clude an incumbent of each party, the 12th and 6th congres- 46 sional districts were chosen for study. Neither district has a pro-Democratic daily neWSpaper, but neither do any of the other marginal districts in Michigan; the state's sole pro-Democratic daily, the Marshall Evening Chronicle, is located in a safely Republican congressional district. Since we did not use a sampling procedure, we will not be justified in generalizing about the performance of all Michigan dailies on the basis of our findings. It was felt, however, that detailed description of press coverage of the campaigns in the 2 districts chosen would be more fruitful than a less detailed description of press cover- age in a larger number of Michigan's 18 congressional dis- tricts. As the Markham and Stempel study (58) of 1956 cam- paign coverage by a sample of 24 of Pennsylvania's 129 dai- lies demonstrated (see above, pp. 55—58), there are inherent and apparently unavoidable pitfalls when sampling tech- niques are used in selecting papers for content analysis of campaign coverage. Disparity of coverage of Republican and Democratic candidates must be found to be very great, in a sample of papers, before the researcher can be reason- ably assured that it accurately reflects coverage by the entire group of papers. The 6th and_l2th Congpessggnal Districts The three-county 6th district is one of the largest 47 in the state in terms of papulation (over 470,000, accord- ing to the 1950 census); the 12th district, which includes 8 counties, is one of the largest in area, but is the smallest by population (about 178,000, the 1950 census re- ported). The 6th district is the home of more auto workers than any other congressional district in the country; Gen- eral Motors plants are the backbone of industry in both Flint and Lansing, the district's 2 largest cities. About three-fourths of the pepulation of the district lives in the Flint and Lansing metropolitan areas (roughly one- fourth in and around Lansing and about one-half in and around Flint). Lansing is located in Ingham County, which is still fairly safely Republican despite a rising Deno- cratic vote in Lansing. Flint's large Democratic vote dominates Genesee County. The third, still predominately agricultural county, Livingston, is Republican by nearly 2 to 1. The 6th District's incumbent congressman, in 1956, was Democrat Don Hayworth, a former Michigan State Univer- sity (East Lansing) speech professor. He had been the Democratic nominee in 1952, but had lost to Republican Kit Clardy, who served in Congress during l955-‘54 in a manner which most observers felt helped elect Hayworth in 1954. In 1956 the Republican nominee was Charles E. Chamberlain, the young first-tern prosecuting attorney in Ingham County. ». ~ . V _ . I O \ V I 1' p . . . . . . ,- . - . . _ , , ~ -—. .‘ . r e F. \ - I -.,- . . . 1 Q . . . _ - .«u ' v . - . —- .x , .., .‘ ,A V n f _. ‘ . , ‘ s ‘V ' . v 1 . . A \ .- v -— r ‘ e ' ' 1 ‘ A g. " . _. ”J r . ‘ '. -. ‘ ~ ._. .. l _ . I P . \ e e g - r. f \. I " . _ 1 . "' (".v 'e . ' ' “ ! f 3 .. . .- I. . \. .a . . .- u . (HJy " . . ._. .. U ~ L. f ’ e K ‘. , n ‘. 1‘ I ' " e~ . - a, I _~ .- _ - I , I ~_» . .3- ‘. ,. . (.. , 9 .y t' ’ {.. . l n ‘ l I- . , ‘ _ —- ’ D . n . ' I . . . ’ . ' X \I . .4. L‘I/7 r 1‘ ‘ . l . . . ‘ e < - . .,- \_ . ‘ . .' 1 . . '.. a , - ’V . ‘ .- . . \J x . ‘ Q . 1 » _. r . . . . _ \ . i. ‘ ’L ' .. 7 ‘ . .. : . . _— . . L“ , ' e e _ . - . ' f - . a - J . . \f » -. . e '. ‘ u ' ' . ‘.l J \ _ .g . . .1“ w . . . . . 1 u .- - r _ . * "J'QLJf‘S ‘ ~ -. a “.‘ ‘ . . . , 1 .I -\ . . . . a , I ~ . . v I , ‘ . e . a — - ' . .- i . ,a , - ,. ' ~ I a. .7 .. - . p . l x ' " .~r _ T5 , x '- . 7 z . . ‘ l v .11 w .. -.--. I . ' . o ‘ 48 Hayworth had support from the leadership of the United Auto WOrkers, despite a widely publicized disagreement with the union over farm policy, and Chamberlain had quiet but effective support from corporation and bank officials. The 6th district's 2 dailies are the Lansing §§g§g Journal (66,254 circulation), one of the 5 Federated Pub- lications papers in Michigan and Indiana, and the Elia; Journal (94,568 circulation), one of the Booth chain of 9 Michigan dailies. Each is read throughout its own county, and each is read in parts of Livingston County, where Detroit and Ann Arbor dailies also have considerable cir- culation. The 12th district, part of Michigan's "North Woods," has in recent years had a chronic unemployment problem and a declining population as a result of the depletion of its best copper, iron and timber resources, which in earlier years afforded dependable employment for the Danish, Fin- nish, Cornish and Italian immigrants who papulated the area in the 1890's and after. Mining, lumbering and the tourist trade are still the chief industries, for although the area is largely rural, agriculture is limited to dairy- ing, potatoes and hay, and is of minor importance. Efforts to promote the economic development of the area are the focus for considerable political discussion within the district. The 12th district's 4 dailies are located in the 5 _ .l s r I . . I, I I I . i I . . \ . . I . I . h . _ I y . . - e s I . . I I 4 . s I . . I o s . I _ , I . , . .., e . . . _ _ \ _ I .I . . I ‘ I I I I ,. It . _ . . . . I - I I . . , .. . , e I; l - p . . . _ a O I 4\ l I . e 9 I e . I \ _ . I. . I . V I \ I . e I . I . r I V .. , . _ . l . e . .I I . . . . I. ' I. e . . n. . I , I . . . . .I. I _ ‘ u . I . . I . V I I. . . 4 I . .. I . . .l . i . r . . . . I ; _ I . . . I. y I I . . e . I I I . . . . t o l I ' .aI. I I I I . b I h '- . u a l.- . . . . . De , . ' . . .. I , ; .tr . . , I I I- . I I . \ . . I. _. I. . I . I . . I 1 e I . , r . I ,. . J . I . I . oI O. . . . . e u I I. . .a. . . _ . . \ I) I . . .. I. . I (I l\ . v r . . . . I s . n . I . . n. . . . . . . . I . I . . I . I b. . I . . I . . I . I . I t .\ . r. r — I .I. a. t. . Is . v I. . I I . \I I . I .I. . a I u I I I Pt I . c I I. I I . . v . v . .. . I ‘ - I I. . . a . I . II. I \ , u,. , . I .I. I. . .I. I I u q . . . o. . I . _ . , . s. 6 I I I I . .. A I s . e , t l- . . _ A e. . a .‘ o a - ' . . . a s r , . . , e e I . , r If I . k I I I. H a . I. 4 . _ _ I , 0 e. 49 largest cities, Marquette (population 17,000), Ironwood (11,500), and Iron Mountain (9,700) and one of the largest villages, Houghton (5,800). The papers are the Marquette Mining Journal (14,527 circulation), the Ironwood Daily glgbg (8,998), the Iron Mountain News (10,241), and the Houghton 133141 Mining Gazette (10,317). The district's Republican incumbent congressman, John Bennett, an Ontonagon attorney, was first elected in 1942, and has since then onl3r once (in 1944) been defeated. His defeat was by his predecessor in Congress, Democrat Frank Hook, who was also the Democratic nominee against Bennett in 1946 and 1954. The Democratic candidate in 1956 was a young auto salesman from Ironwood, Joseph S. Mack, who had nOt had prior experience in government. Definitions of Terms Certain terms used in ensuing; chapters are intended to have these definitions: gtgm: a complete news or feature story, separate photo, editorial, cartoon, letter to the editor or advertisement which mentioned a local congressional candidate in more than 1 sentence, or which otherwise qualified for inclu— sion in this study (see coding instructions, page 118). Eggscandidate item or headline: an item or headline which indicated praise or support for a local congressional can- didate, without doing likewise for the candidate's opponent. 5O Neutral itgm g; headline: an item or headline classified either to the Democratic or the Republican candidate but which indicated neither praise or support for him.ng£ criticism or lack of support for him. Anti-candidate 31339; 9; headline: an item or headline which indicated criticism or lack of support for a local congres- sional candidate, without doing likewise for the candidateks opponent. Balanced EEEE.2£ headline: an item or headline which con- tained "neutral" mention of both local congressional candi- dates, or "pro" mention of both of them, or "anti" mention of both of them; or a headline which mentioned neither local congressional candidate, by name or title, but appear- ed over an item included in this study. (In coding, such headlines mentioning neither candidate were designated as not directly relevant to a local congressional candidate nor to the congressional race.) Headline: the entire heading over an item, including the secondary or sub-headline, if any, but not including any "article dividing line" headlines between paragraphs or sections of the item. Multircolumn headline: a headline at least part of which was more than 1 newspaper column in width. Separate phgtg: a photo and caption published separately, and not as part of an adjoining news, feature or other item. Single-candidate advertisement: a paid advertisement for 51. a local congressional candidate, or for such a candidate and the presidential candidate of his party, but which mentioned no other candidate of that party. Multi—candidate advertisement: a paid advertisement which mentioned a local congressional candidate, but also listed various other candidates of his party. Reference 32 an issue: mention (by word, phrase, or sen- tence) of an issue, in any item included in this study. The recording unit1 was the complete item; thus any number of mentions of a single issue, in a single item, were re- corded asl reference. Policy issues: campaign appeals concerning specific ques- tions which have been in recent years (or are now expected to be) subject to decision by government policy makers, particularly Congress. (Examples: social security bene- fits, foreign aid, farm price supports, civil rights legis- lation.) In coding, policy issues are classified as domes- tic policy issues or foreign policy issues. Egg-policy issues: campaign appeals other than policy issues; appeals based on personal attributes of candidates, identification with special interest groups or with other political figures, support for or opposition to certain g 1"The recording unit may be defined as the range of text'for which the occurrence of a symbol is tabulated with the unit weight of 1, even if it occurs more than once in ghefiggecified text." Geller, Kaplan and Lasswell, 22 , 52 ideals, etc. (Examples: honesty, deception by Opponent, support of or association with presidential or vice—presi— dential nominee of own party, support of farmer's interests.) Hypotheses The following 6 hypotheses were.dedided upon after a review of the findings of other studies of campaign cover- age by daily newspapers. In addition, it was decided that data descriptive of the extent of congressional campaign coverage by 6th and 12th district dailies should be gath- ered and reported. These hypotheses, taken together, advance the prop- osition that among the major determinants of newspaper coverage of political campaigns are predispositions on the part of newspaper staff members to focus attention on: 1) personalities and other, related "non-policy" matters, 2) incumbent candidates,-who already have community status,znni 5) candidates editorially supported by the paper. A fourth influence may also be at work, however, and may sometimes run counter to numbers 2 and 5: newspaper staff membersnmy' also -- consciously or otherwise-n-want to avoid seriously offending the generally held political sentimentsrflfthe paper's readers. Hypothesis 1 In daily newspaper non-advertising items less atten— 55 tion is devoted to congressional candidates' views (and in the case of incumbents,_their voting records) on specific policy issuesipthan to non-policy issues of the campaign. (19, p. 48) that "Tim; magazine prides itself that 'our subscribers can understand the event in terms of the per- sonality who caused it.'" Flesch expressed doubt that personalities really cause events but added (ibid.) that "there is no doubt about one thing: human interest makes for easier reading. Scientific tests have shown that people are better at reading about other people than about anything else." According to Flesch (gpgg., p. 56), Timg style contains about 8 personal references per 100 words. Thus it seems reasonable to hypothesize that daily news- papers too, like Timg, tend to stress personalities. As 2 students of political communication have recentlywmritten (Lang and Lang, 29, p. 109): In political life, the complexities of issues, the contests between conflicting messages, and such projective distrust as may result, place an undue premium on familiar qualities that can be directly experienced. 'Sincerity' and 'personal warmth' are among the traits. . . This hypothesis is also advanced on the basis of the findings of the Ferguson and Smuckler study of newspaper coverage of 1952 senatorial campaigns in Wisconsin and Connecticut. Ferguson and Smuckler reported (18, p. 28) that policy issues appeared "much less frequently than non-policy themes in the press in both states." Only 28 54 per cent of all appeal items, Ferguson and Smuckler found, were policy issue appeals, although the percentage was higher (46 per cent) among candidate-originated appeal items. This is a matter of considerable importance, especially to those who believe independent voters should and to some extent do make their choices on the basis of candidates' views "on1flua issues." If but a small share of the content of neWSpaper coverage of congressional campaigns concerns candidates' views on important issues, then democratic theory may be quite remote from democratic practice. Hypothesis 2 Eggpe is a:;elationshippbetween the editorially ex— pressed political preferences of daily newspapers and the extent of non-editorial,l non-advertisingcovepage_given Democpatic and Republican candidates fq;_gongress by sugh ggwspapeps.pand th;§_pelationship is more_pppnounced;;p numbepsof headlinesi photogpaphs and frontgpage items, than it is in total numtfips of itemsflL op in column inches. As stated above, most past studies of political cov- erage have found that well over half the papers studied gave greater coverage to the political party editorially favored than to the party not favored. A number of studies have found that disparity of L"Editorial" here refers only to newspaper editor— ials, and is not intended in the broader sense. 55 treatment is more pronounced in headlines, photographs and front-page items. Higbie, for example, in his study of front- page coverage of the 1952 presidential campaign by 14 Wis- consin dailies (of which only 1 was editorially pro-Demo- cratic) found (25, p. 56) that ". . .although major Demo- cratic and Republican national figures were accorded very nearly the same amount of front page space. . .when the t0p heads in size and placement were surveyed as to candidate preference based on this nearly equal balance, Eisenhower and Nixon subject matter heads in tOp position held a re— vealing 65 to 57 per cent lead." To cite another example, in the 1956 Markham and Stempel study of the performance of 24 Pennsylvania dailies (only 2 of them pro-Democratic) it was found (57, p. 16) that the greatest disparity of coverage was in photo space, of which the GOP was given 65 per cent. The same study found that although the GOP was given but 56 per cent of all headlines, it had 65 per cent of all fggpt page direc- tional heads wider than 2 columns, and 70 per cent of fpggt page photo space. Thus it is felt to be of interest and significance to seek data on whether coverage of copgres- sional campaigns is weighted for the side editorially fav- ored, and most heavily weighted in terms of headlines, photos and front-page items. Hypothesisgj There is a positive relationship between congression— 56 al incumbency and the extent of coverage_given congres— sional candidates in daily newspapers. This hypothesis is based on general knowledge of the news—making advantages of being ”in". Also, Ferguson and Smuckler found (18, p. 72) that ". . .the Connecticut Dem- ocrats received more favorable references than their Repub- lican opponents in what was an editorially Republican press." The Democratic candidates for the 2 Connecticut Senate seats were both incumbents in a sense: one was an incumbent senator, and the other an incumbent congressman, with many of the advantages of an incumbent senator's sta- tus. As we have suggested above, the Ferguson amiSmuckler Connecticut findings clearly point to the greater coverage which can generally be expected to be given an incumbent over'a.non-incumbent, party affiliations aside. It isfelt to be worthwhile to test whether this incumbency advantage applies at the congressional candidate level, and what effect it has when the incumbent is not editoriallyhfavored. Hypothesis 4 There is a relationship between the editorially ex- ppessed political preferences of daily newspapers and the direction of opinion regardinngemocratic and Republican congressional candidates expressed in letters to the edi— tor columns of such newspapers. Only l of the published reports of campaign coverage 57 studies has offered data on this point; thus it seems de- sirable to investigate whether there is a tendency'imlselec- tion of letters, as reportedly there is in selection of news, to favor the party favored editorially. The one re- port of a break—down by political direction of letters to the editors has been quoted above (Editor & Publishep, 46, p. 10); the Baltimore §pp reported having published Repub- lican and Democratic letters in almost exactly the same ratio as the people of Maryland cast their votes in the presidential campaign of 1952 (Eisenhower-Nixon letters were 56.5 per cent of the total published, and their votes were 55.8 per cent of the Maryland total). The Spp editors did not indicate whether they attributed the close parallel to their own intuitive selection of letters or simply to the ratio of Republican to Democratic letters received. Hypothesisp5 There is a relationship between Michigan congression- al candidates'_party_affiliations and the extent to which, in their paid advertising, references are made to specific pplicyfissues. Ferguson and Smuckler reported (18, pp. 51-52) that "analysis of the individual :senatorial campaigns in the press of each state [Wisconsin and Connecticut] does not show the Democratic candidates to be consistently more policy issue oriented than their Republican Opponents." Nevertheless in 5 of 4 cases they reported, Democratic can— 58 didate advertisements did stress policy issues considerably more than did Republican candidate advertisements. (And in news items, in 2 of 5 cases the Democrats stressed policy issues more than did the Republicans.) It is the writer's belief, based on personal observation, that the Democratic Party in Michigan is considerably more policy issue-orient- ed than are either of the parties nationally, or in most other states. The hypothesis is thus intended to test the assertion by many leading Michigan Democrats that they and their candidates are more concerned than are their oppo- nents with policy issues. othosis 6 There is a:pelationship between the:pelative local strength of the Democratic and Republican parties and the extent of coverage given Democratic and Republican candi- dates for Congress in daily newspapers: the greater the local strength of thepparty of a candidate who is editor- ially Opposed, the more nearly equally divided is the cov- areas. The earliest indication of possible support for this hypothesis was in the report of Stene's study of 1956 pres- idential campaign coverage by 21 dailies. Stene found (55, p. 214) that all but 1 of the 21 papers gave a majority of news space to the candidate supported editorially. The one exception was a pro—Republican paper in a strongly Democra- tic Cit—yo 59 More recently, Kobre found (28), in his study of 54 Florida dailies' coverage of the 1952 presidential campaign, that although 21 of them were pro-Eisenhower, most of the pro—Eisenhower papers did pg: give Eisenhower more news space than Stevenson. Kobre's findings might be interpret- ed as supporting this hypothesis; clearly, more evidence pro or con would be of interest. Methodology It was determined that fill the content of ell issues of the 6 papers during the 6 weeks should be examinedqllest a few lengthy or important articles be missed. This was felt to be especially important in connection with Hypoth- esis 1, concerning the amount of attention given candidates‘ views on policy issues. A paper might have given lengthy coverage to candidates' views on policy issues in a single feature story, and such an item might have been missed if only a sampling of issues of the paper had been examined. Specially prepared punch cards similar to McBee Key- Sort cards were used in recording data; 1 card was used for each item coded. (A sample card may be found in the appendix.) The date of the item and its length in column inches was written on the card. Punches recorded the re- mainder of the desired data, with the exception of the de- tails of certain issues, which were also written on the card. The code and coding instructions are included in 60 the appendix. The writer did all the coding, thus limiting possible biases to his own. He followed as closely as possible the coding instructions which are included in the appendix. All newspaper items which might include mention of a con- gressional candidate were scanned, even though the headline did not suggest relevancy. An item was coded if it includ- ed mention of a local congressional candidate in more than 1 sentence, or if it qualified in certain other ways listed in the coding instructions. A reliability check was run on the entire contents of 5 per cent of the newspaper issues; results are reported below. The basic coding unit was the gpem, or complete news or feature story, separate photo, editorial, cartoan,letter to the editor or advertisement. A statement analysis along the lines of the Bush method employed in the 1950 Califor- nia senatorial campaign coverage study might also have been desirable, but was not used on grounds of economy of time and effort. Waples and Berelson (60, p. 17), and more re- cently several others, have advised that the close parallel between the direction of sentences and the direction of whole articles "suggests the greater economy of classifying content in the large, rather than in detail." In coding of issues, however, the coding unit was the reference to an issue: the mention, by word, phrase, or sentence, of a policy or non-policy issue. Fithin a single 61 newspaper item, repeated references to the same issue were not counted again; thus either a bare mention of social security 23 5 paragraphs on it was counted as a single ref- erence to social security (a policy issue). Reliability A reliability check was run on the entire contents of 12 newspaper issues, 5 per cent of the total of 228 newspaper issues examined in the study. The 12 issues were completely re—examined by a journalism graduate stu- dent (Coder B) who favors the political party not favored by the writer (Coder A). Inclusion of items Coders A and B were in agreement on inclusion or non-inclusion of 99.99 per cent of the content (measured in column inches) of the 12 issues. 62 TABLE 1 INCLUSION AND NON—INCLUSION, BY CODERS A AND B, OF THE CONTENT OF 12 NEWSPAPER ISSUES Per cent of content coded as relevant by both A and B .47 Per cent of content coded as relevant by neither A nor B 99.52 Per cent of content coded as relevant by A but not by B .00 Per cent of content coded as relevant by B but not by A .01 Total: 100.00 Coder B coded as relevant 15 items which had been sim— ilarly coded by Coder A. He also coded as relevant,hewever, 2 items (both of them Republican candidate advertisements) totalling 5% column inches, which had escaped notice by Coder A. (Advertisements are more easily missed than other items, in the writer's opinion, because -- since congres- sional candidate ads seem consistently crowded and lacking in white Space —- advertisement headlines are less readily noticed than news, feature and other headlines.) If Coder A missed a proportionate number of column inches of advertising in the 216 issues not re-examined by Coder B, then a total of 70 column inches of advertising (both single-candidate and multi—candidate) was missed by Coder A. Seventy inches constitutes less than 5 per cent 65 of the total of 2,705 column inches of advertising coded by Coder A. Coder A's measurement of the 15 items coded as rele- vant by both coders totalled 245 5/8 column inches; Coder B's total for the 15 items was 245%. Direction Berelson (6, p. 175) attributes to Albert Reiss the suggestion that a distinction can be made between "gross reliability," applying to ell the coded categories, and "net reliability," applying only to those categories, often cumulations, used in the report of a content analysis. The "net reliability" figure for directional coding of the 15 items coded by both coders in the present study is 100 per cent. The "gross reliability" figure is 88 per cent. (In coding the 15 items there were 26 decisions to be made as to direction: the direction of the pgdy ofennfll of the 15 items, and the direction of the headline of each of them. 0n 25 of these decisions the coders were in agreement. Twice the coders disagreed as to whether an advertisement headline should be coded as "pro-Republican" or "neutral Republican," and there was a third case ofsnufll disagreement over a Democratic advertisement headline. The pro-Republican, neutral Republican and anti-Democratic figures have been combined in our report of findings, and likewise the pro-Democratic, neutral Democratic and anti- Republican figures. Thus "net reliability" of directional coding remains l00 per cent.) Issues There was 100 per cent agreement between coders as to whether references to issues were to policy or to non- policy issues (i.e. no issue reference was coded as "policy' by one coder and as "non-policy" by the other), and 79 per cent agreement as to the words, phrases and sentences to be coded as references to particular issues. 0f 52 refer- ences to issues coded by one or both coders, 41 were iden- tically coded by coders A and B. 65 TABLE 2 CODING OF CAMPAIGN ISSUES BY CODERS A AND B Policy issues Non-policy issues All is- fDomestic Foreign Personal Others sues policy policy Identically Coded by A 8 5 14 16 41 and B Coded only by A 0 2 2 l 5 Coded only by B l 0 5 0 6 Totals ‘ 9 5 21 17 52 Per cent of issues iden- tically cod- ed by A & B , 89 60 67 94 79 CHAPTER IV FINDINGS The 6 newspapers studied devoted an average of 199.5 column inches each to all non-advertising items about either (or both) local congressional candidates, during the 6weeks ending with election day. This is roughly equivalent to one and one-eighth eight-column newspaper pages, or 9 full length columns. In addition, the 6 papers published an average of 274 column inches each of advertising for the Republican candi- date, and an average of 176.9 column inches each of adver- tising for the Democratic candidate. There was a wide variation between papers in the num- ber of news stories published, with less variation in the numbers of feature stories and of photographs published. In 5 of the papers news stories were the most numerous items; an average of 12.6 were published in each of the 6 papers during the 6 weeks. Feature stories averaged 4.5 per paper during the 6 weeks, and separate photos, 5.8. Photos accompanying news or feature stories averaged 1.8 per paper during the 6 weeks. Three of the 6 papers ran no editorials on the local congressional candidates, and 4 papers ran no letters to 67 the editor pertaining to such candidates. The average pa- per published 22.6 non-advertising items of all kindsaflxnfl: local congressional candidates (not including 1.8 photos accompanying news or feature stories). TABLE 5 NON-ADVERTISING ITEMS PUBLISHED ABOUT LOCAL CONGRESSIONAL CANDIDATESa O) m 0 +3 «4 H «P H F") ’D 55 (H ’ C ,‘C.’ m I) m 4.3 -r-I on m o .23 ;,i <1) :3 <1) :41: C30) F1 $451+ '13 mgmo,® H :uH ac) .0 0:3 r4 A; c5 5 ,CE 2‘7 c) L5 .23 .3, t 1 :1 £1 :2- co 2.; ,3"; :1 H O :4: r‘1 O t: H C3 ',.'. r--{ U :5 (._l C) .C", :4 Q4 0 LT. “ c“, C A; :1 C... u 3.4 2.2". O O .H :1; a U or! '1 mi 0 H , ' o o ._4 Lansing Z State Journal 0 0 {2 1.5 ' 94 5 15.7 Flint Journal 15 199.8 i4 18.7 9 150 0 0 Total, 6th District 15 199.8 6 20.2 14 224 5 15.7 Houghton Daily Mining Gazette '6 188 4 2.7 6 517 7 15.6 Iron Mountain . News 8 211.5 1 .5 6 1512 l 7.2 1 Marquette ; Ironwood Daily 10 e 5 216.5 1 1.8 4 7 409 2 11.1 Total, 12th District 26 854.5 8 7.2 24 A555.5 13 49.1 aColumn inches of only those portions of multi-candi- date advertisements which dealt with the congressional can— didate exclusively. In the following 6 sections are reported our findings relevant to each of the 6 hypotheses first stated in chapter III. 71 Hypothesis 1 In daily newspaper non—advertising_items less atten— tion is devoted to congressional candidates' views (and in the case of incumbents, their votingrecords) on specific policy issues, than to non-policy issues of the campaign. The hypothesis is supported by our findings regard- ing 5 of the 6 papers, as indicated in Table 6. Five of theESpapers published more items in which only non-policy issues were referred to than items in which policy issues were referred to. The average paper published 5 policy issue items, 12 non-policy issue items, and 5.6 items in which no issues were referred to. (Such items often simply announced future speeches or party meetings.) It is note- worthy that in the case of the one exception, the Lansing State Journal, the total of items in which n2 issues were referred to exceeded the total of items in which refer- ences were made either to policy 9; non-policy issues. 72 TABLE 6 NON-ADVERTISING ITEMSa RFFERRING T0 POLICY, TO NON-POLICY AND T0 N0 ISSUES Items re- Items re- Items ferring to ferring to referring policy non-policy to no issues issues only issues Lansing State Journal 15 5 20 Flint Journal 6 12 10 Houghton Daily Mining Gazette 1 11 1 Iron Mountain News l 10 0 Marquette Mining Journal 5 21 2 Ironwood Daily globe _ 6 l5 1 Average of all I 6 newspapers 5 12 5.6 "Including items not directly related to the campaign. Table 6 might conceal the fact that a newspaper gave lengthy coverage to candidates' views on a wide range of national policy issues, in but a few giggg, despite the fact that most items did not contain references to such views. Table 7 has been compiled, therefore, in order to explore that possibility. The unit in Table 7 is the w- gggg to an issue rather than the entire newspaper item, which is the unit in Table 6. The data presented in Table 7 support the hypothesis even more strongly than that presented in Table 6. Here 75 all 6 of the papers are seen to have published fewer ggfgg- gaggg to policy issues than to non-policy issues. The av- erage paper published nearly 5 times as many references to non-policy issues as to policy issues. Table 7, it will be noted, indicates that only 4 re- ferences were made, in all 156 non—advertising items coded, to foreign policy issues. These 4 were very brief, slogan- ized references to support for Eisenhower‘s foreign policy (2 references), support for his Middle East policies, and support for a strong U.S. defense. TABLE 7 REFERENCES TO POLICY AND NON-POLICY ISSUES IN NON-ADVERTISING ITEMSa References to policy issues References to . non—policy Foreign Domestic issues Policy Policy 1 Lansing State Journal 5 22 44 Flint Journal 1 7 59 Houghton Daily Mining Gazette 0 5 28 Iron Mountain News 0 5 51 Marquette MiniQSHEEEEEEI 0 7 79 Ironwood Daily Globe l 12 64 Average of all ':: 6 papers .8 9.7 50.8 aIncluding items not directly related to the campaign. 74 Hypothesis 2 There is a relationship between the editorially ex- pressed_political preferences of daily newspapers and the extent of non-editorialinon-advertising coverage_given Democratic and Republican candidates for Congress by such newspapers, and this relationship is more pronouncedpig numbers of headlines, photogpaphs and front—page items, than it is in total numbeps of itemsgop in column inches. In compiling the data relevant to this hypothesis, the following items were omitted from consideration: a) all items classified as balanced, since they could not readily be objectively judged to favor either candidate (the numbers of balanced items have nevertheless been list- ed in Table 8, for whatever light they may shed on overall press campaign performance); b) all items not directly re- lated to the campaign, all but 1 of which were reports of "non-partisan" announcements or actions by the 2 incum- bent congressmen, and thus were presumably published as part of the newspapers' coverage of government rather than as part of campaign coverage; and c) obviously, all editor- ials and all advertisements. As stated in chapter III, anti-Democratic items are counted as Republican, and anti- Republican, as Democratic. l"Editorial" here refers only to newspaper editorials, and is not intended in the broader sense. 75 No distinction is made between the political prefer— ences of the 6 newspapers, for it was found that 5 of them gave outright or tacit endorsements to both Eisenhower and the Republican candidate for Congress (there were 4 out- right endorsements of Eisenhower and 2 outright endorse- ments of GOP congressional candidatesl). The sixth paper, the Houghton Daily Mining Gazette, editorially favored, just short of tacit endorsement, both Eisenhower and the GOP candidate for Congress, even though it gave an out— right endorsement only to the GOP candidate for the State Senate. Thus all 6 newspapers are classified as pro—Repub- lican in editorial policy. Our findings support to a considerable extent the first portion of the hypothesis. The 6 papers published 52 items on the 2 Republican candidates for Congress, and 2 57 items on the 2 Democratic candidates ; thus the Repub- licans were favored over the Democrats by 58 per cent to 1One of the papers which gave what must be termed an outright endorsement to the Republican candidate for Congress was the Lansing State Journal, which we have not reported as having published an editorial on either con- gressional candidate. On November 4 the State Journal called for the election of "a congress containing a major- ity of Republicans in both the house of representativeseuml the senate," but made no reference to either of the local congressional candidates, by name or title. Thus the edi- torial was not counted as an item for the purposes of this study. The Ironwood Dailijlobe was the second papervdflmfli made an outright endorsement of a congressional candidate; it Specifically endorsed the Republican candidate. 2Excluding balanced items, items not directly re- lated to the campaign, editorials and advertisements. 76 42 per cent. Five of the 6 papers published more items on the Republican candidate than on the Democratic candidate; the sixth paper (the Iron Mountain News) published 2 items on each of the candidates. The ratio varied from 15 items (Republican) to 14 items (Democratic) in the Lansing Spgtg Journal, to 6 (R) to 1 (D) in the Houghton Daily Mining Gazette. The average ratio, for all 6 papers, was 8.6 (Republican) to 6.1 (Democratic). (It should be noted in passing that of items not di- rectly related to the campaign, the average 12th District paper published 2.7 about the incumbent Republican Con— gressman, whereas the average 6th District paper published but .5 about the incumbent Democratic Congressman.) The 6 papers published 447.6 column inches on the 2 Republican candidates, and 297.8 column inches on the 2 Democratic candidatesl; thus the Republicans were favored over the Democrats by 60 per cent to 40 per cent. Three of the 6 papers published from nearly twice as many to over 10 times as many column inches on the Republican as on the Democrat; the column inch totals of the other 5 pa- pers favored the Democrats by much smaller margins. The average ratio, for all 6 papers, was 74.6 column inches (Republican) to 50.7 column inches (Democratic). 1Excluding balanced items, items not directly related to the campaign, editorials and advertisements. 77 Data presented in Table 10 (p. 85) is also highly relevant here. By comparing treatment accorded Democra- tic and Republican incumbents, and Democratic and Repub- lican non-incumbents, it was found that in 4 of 5 cate- gories the Republicans fared better regardless of incum- bency or non-incumbency. In items, column inches, head- lines and photos, the Republicans received higher percent- ages of total coverage than the Democrats. Only in per- centages of front page items did the Democratic incumbent lead the Republican incumbent, and the Democratic non- incumbent the Republican non—incumbent. 78 - ‘_——-- *— -‘ --.w‘.‘- -_— ._. ‘4 smear eBsd quoa; 'mGG-I4UE B dOO, I ‘;qu0J§mgggug§§gi .4 (D o c: a; smear eBad quoa; O ’w' T' '- -:v- 1~ 137.. 22141.0 3;? ) Til ,’ "(WI L111 -— - -J.‘ mWQ fl 4‘.) ID: I ». 1"1 n .L.L.LI\) !'|"'\ i», V. 1* '1 ALL.) ._3 uk- ."1' .. .-_—'-—.--——- 509 tans 3 'mea '* 5* ‘3 r' '4 9 80:10qu L': {O H O h") °meq-Ique B dOO osoqoqd peouatag, (J C; C) c c; o O osoqoqd , ' _ 609—14078 4? omaq‘ L’J <1' 0 (\1 P.) CQ [0 SQUIIPBGH o fined—14113 ? (I09 0 t0 (O 02 (D b’) :34 L!) SGUTIP‘QGII "PGOWI'BE H c: t0 [0 in (\‘z Ckl C. r‘l r"! H {D SGTIIIPBGQ . C; ' N JOB-131T? =8 mac: ,8. “ O N H 0 LC‘ V“ L‘ KC) ‘1‘ (D C C . O . . . 88110111.: WHIOO t5.) b L‘ 0‘ H, :4 I: " r’l {Q C) r‘ H mac: mus =3 dos H H H (\‘2 CC (1) l0 [0 N.) SOLICIT; to U; to a.” 0.2 <24 t0 IIIIInIOO pGOUBI’eq m u‘: N <11 LO 0: <14 C72 L0 ‘73 KO C; § 891.1011? WIOO DC.) 0 C\2 LO 1» LO Q (IDS-1:411? ’8 ”119G :1 b cu as an L0 «o 8111841: LO r—s 02) CO CD 0;: (D ‘ ' ‘WGG-IQUQ a dOD ._. r—i LI) $111941: pSOUBI‘Bg b b (\1 t0 :3 C; H 8111941; v C H cu c; cu to dos-mm B Wm H H tnLCD H r" r4 rq H a n ri f -‘~ (. 'H 23 OH (:3 m .I._ :: C" F) U:- S... (.4 C) 'D .1. z. m N 4A 9.; a) o 44 :4 L3 : I C": r: O C). .. (37 o c: c: U :3 ca 44 :2: § 3 (3)1‘3' ml 0 O (D {L {7.1 L4 ‘5 Q. 43 Cd) 2"; mm 1) o :2, on a) 43 ,'T'_‘] 3". c; k 0 7' ) so +5 {1 ih—l g: m: :4 q) c: m 5:: ‘..'{1 H L: C: C) i," Q) Q \;_. J '> g- ‘c ’ O 'H 7“ (D n“? I‘.‘ :3 L I?) E: me; Hi" , 7 ago {-14 511 (See next page for Table 8 footnotes) 79 aThis table excludes items not directly related to the campaign, editorials and advertisements. bApparent discrepancies between totals of items and totals of headlines are due to inside page continuation headline S 0 0Both separate photos and those accompanying other items (news or feature stories). TREATMENT ACCORDED LOCAL DEMOCRATIC AND REPUBLICAN TABLE 9 CONGRESSIONAL CANDIDATES, EXCLUDING BALANCED ITEMSa .‘I‘ .D (0 co 0'3 W +3 .p L a . - 9 O c C m p a rI +‘ +3 C: O 44 +3 o, O 9 C3 3-1 :4 ..—-| .r-I o o r'—~ #4 CH C... 04 C1 0—4 ° 04 ° :11: [ll ' C14 ' .1; in CI g: :1 L; C) if C) if (1‘ a) CD a) (’7 i3 (S (I) (1‘. (I) . I Q I 5‘. I CI I Q I 5‘. H I «4 I 0H I or-I I H I -+—) ”'4 44’ 'H 4—) H 4:) H 43 U) 0r“' m 5: .LJ :3 43 C! m 44‘ co :1 43 :13 w {-5 “:5 C1. .‘3 C :3 l) .C. (D :3 5:; :5 a) (2' (D I ~33 CS 5.1 ~ m :1 CU .0 c3 4: ‘08 08 m 0’) 9809' 0H (‘0’ Ob "—4 H I S a) aw) 44 Sri S S 0 o ,5: .1". of? SC“- 0 - m o.) ‘34; C14 t“: 0 CL 0 E; 03 04 c3 i:- p« =15 1) 04 ‘ ) p C2 a) I”? C" 71 D f) c) '1‘- 1) c; Q rU ; a Q C: Q «'4 rim—4 {4:41 C3 :2: {LEI C‘. é—I I‘ (1': Per cent of cate- gory, total of all 6 papers 42 58 4O 6O 56 64 47 55 55 45 .Bbr cent of cate- gory, total of 2 6th Dist. papersc 48 52 50 50 59 41 55 47 100 0 Per cent of cate- gory, total of 4 12th Dist. papers 55 67 5O 7O 15 85 41 59 17 85 a . . This table also excludes items not directly related to the campaign, editorials and advertisements. b items (news or feature stories). Both separate photos and those accompanying other °The Lansing State Journal and the Flint Journal. 80 It will be noted that theE’larger circulation papers (the Lansing State Journal and the Flint Journal), in the Democratic-incumbent 6th Congressional District, published nearly exactly the same number of items and column inches about each of the candidates, in contrast to the 4 smaller circulation papers in the Republican-incumbent 12th Congres- sional District. (In front page items, headlines and photo— graphs, the 6th District papers favored the Democrat some- what, although the Republican was slightly favored in total items; the column inches were almost exactly evenly dividedJ Despite the fact that Ironwood was the home of the 12th Dis— trict Democratic candidate, the Ironwood Daily Globe didrun: publish a single news story about him. It published.22items (both separate photos) about him, and 9 items about his in— cumbent Republican Opponent. (It should be remembered that the latter figure does not include those "not directly re- lated to the campaign" items about the Republican —- there were 5 of them -- which are attributable to his incumbency). The Houghton Daily Mining Gazette published but 1 item about the Democratic candidate (a 2.5 inch news story) and 6 about the Republican. The Iron Mountain News, on the other hand, published 2 items about each of the candidates, and in column inches favored the Democrat by better than 2% to 1. As the percentage figures at the bottom of Table 8 and in Table 9 indicate, our findings do not support the 81 second part of the hypothesis, concerning greater dispar- ity of treatment of the 2 candidates in numbers of head- lines, photographs, and front-page items. Although the 12th District papers ggg favor the Republican more strong- ly in number of front page items and number of headlines than they did in number of items and number of column inches, their photo coverage favored the Republican.1§§§ strongly than their item and column inch totals. And in the 6th District, the Democratic candidate was given a 5 to O advantage in front page coverage, and slight advan- tages in numbers of headlines and photos, despite the very slight advantage given the Republican in total number of items published. It is noteworthy that the numbers and the column inch totals of balanced items were, on the whole, far smaller than the numbers and column inch totals of direc- tional items. Of the total of 27 balanced items (in all 6 papers), 18 were feature stories and 9 were news stories. Of the 9 balanced news stories, 5 appeared in the Lansing State Journal and 4 in the Flint Journal; none were pub- lished in the 4 12th District papers, which largely relied, for their balanced items, upon feature stories reprinted from the 5 Detroit dailies, each of which sent a political reporter to the 12th District to report on the congression- al race there. It may appear, from the figures in Table 8, that 82 both candidates were mentioned in most of the headlines, for the average paper ran 12.5 balanced headlines in con- trast to 2.6 Dem. & anti-GOP headlines, and 4.5 GOP &anti- Dem. headlines. Of that average of 12.5 balanced head- lines, however, an average of 9.5 were balanced simply be- cause they directly referred to neither candidate, by name or title, nor to the congressional race (in either the head or sub-head). Hypothesis 5 There is a positive relationship between congression- gggincumbency’and the extent of covepage givancongpessiongl candidates in daily newsnapers. Our findings fully support the hypothesis with respect to every category of coverage. The scope of the study in- cluded an incumbent of each party and a non-incumbent of each party. As indicated in Table 10, the incumbent Repub- lican fared better than the non-incumbent Republican in percentage of itemsl, percentage of column inches, percent— age of headlines, percentage of photos and percentage of front page items. Similarly, the incumbent Democrat fared better than the non-incumbent Democrat in each of these re- spects. lBoth percentage of related items and percentage of items not directly related to the campaign. TABLE 10 TREATMENT ACCORDED LOCAL INCUMBENT AND NON- INCUMBENT CONGRESSIONAL CANDIDATESa Democratic Republican Candidates Candidates t I t ,‘ E3 f: A a; A o I -1-) 0 Q 0 +3 L5 Us 43 -I~‘ £3 +3 +9 5:: :3 .rI s: ca :5 C) C, w: 0 if] Q) «‘4 C) 'r_‘_* C) ,—.‘J C1 1:: E) C. é-I ,0 H +4; a H r: A z; 23 :3 3:: 2:: a at c; H :3 (3) C) Q) C: H :;~’ the Pvt H‘“ Per cent of items, all papers in his District 55 48 52 67 Per cent of column inches, all papers in his District 30 5O 5O 70 Per cent of headlines, all papers in his District 15 59 41 85 Per cent of photos,b all papers in his District 41 55 47 59 Per cent of front page items, all papers in 4 his District 17 100 I O 85 aThis table excludes balanced items, items not di- rectly related to the campaign, editorials and advertise- ments. b It is also worthwhile to examine how items not di- rectly related to the campaign were distributed among in- cumbent and non-incumbent candidates. Both separate photos and those accompanying other items (news or feature stories). Such items are of considerable importance to the candidate, for they call his name to the attention of the voters in a more advan- tageous way than do items clearly connected with the cam- paign. Voters tend to build up resistance to claims and counterclaims obviously aimed at winning votes, but may be caught off-guard, and influenced, by items which appear to be "non-political," and thus more "reliable" indications of what the candidate is ”really like." A Here, too, our findings support the hypothesis, for of the total of 15 items (in all 6 papers) which were not directly related to the campaign, 12 dealt with some aspect of the incumbent's conduct of his office. They reported his work in behalf of bringing natural gas to the district, or in furthering the district's industrial development, or. his announcement of plans for a new post office building in the district, etc. In reality, of course, these items were not unrelated to the campaign; the incumbents may even have timed them so as to appear during the campaign period. Of all items coded as not directly related to the campaign, the 2 incumbents had 110.7 column inches motheir opponents' 7.7 column inches, 5 front page items to their opponents' O of same, and 8 headlines to their Opponents' l. The sole "not directly related" item about a non-in— cumbent candidate described remodelling of his office (the Ingham County Prosecutor's office in Lansing). 85 Hypothesis 4 There is a relationship_between the editorially ex- pressedppolitical preferences of daily_newspapers and the direction of opinion regarding Democratic and Renublicgp congpessional candidates expressed in letters to the edi- tor columns of such neWSpapers. As noted in the statement of findings relevant to Hypothesis 2, above, all 6 of the papers editorially sup— ported -- to varying degrees —- the Republican candidates. Thus if Hypothesis 5 is to be supported, our findings should indicate that published pro-Republican—candidate (and anti-Democratic—candidate) letters outnumbered pub- lished pro-Democratic-candidate (and anti—Republican—can— didate) letters. Our findings are inconclusive. Only 4 letters to the editor concerning the congressional candidates or cam- paign were run by the 6 papers, during the entire 6 weeks. The Flint Journal ran 1 pro-Republican-candidate letter, 1 anti-Democratic-candidate letter, and 1 from the Repub- lican candidate himself. Thus the findings concerning 1 paper support the hypothesis. Of the other 5 newspapers, however, only the Marquette Mining Journal published a letter to the editor relevant to either of the local con- gressional candidates, and it, by implication, was mildly critical of the incumbent Republican. 86 Hypothesis 5 There is a relationship between Michigan copgression— al candidates' party affiliations and the extent to which, in their_paid advertising, references are made to Specific policy issues. Our findings support the hypothesis. The 2 Democra— tic candidates made more references to policy issues, in their advertisements, and relatively fewer references to non-policy issues, than did the 2 Republican candidates. (Only single-candidate advertisements are held to be rele-' vant here, inasmuch as multi-candidate advertisements are usually prepared and paid for by party committees rather than by the congressional candidates or their own campaign committees. Table 12, below, reports separately the numbers and references to issues of mgltlecandidate advertisements which mentioned a congressional candidate.) In the 6th Congressional District, of references to issues in the Democratic candidate's advertisements, 4£fiper cent were to policy issues. Of references to issuesimlthe 6th District Republican candidate's advertisements, only 11 per cent were to policy issues. In the 12th Congressional District, of references to issues in the Democratic candidate's advertisements,397per cent were to policy issues; of references to issues in the Republican candidate's advertisements, 25 per cent were to policy issues. 87 It is noteworthy that foreign policy issues were al- most wholly ignored in the candidate advertisements. Of the 175 single-candidate advertising references to policy issues, only 6 were to foreign policy issues, and of those 6, 4 were simple references to "peace" by the 6th District Republican candidate. Two were references to support for "funds to fight international Communism," by the 6th Dis- trict Democratic candidate. Neither of the 12th District candidates ever mentioned foreign policy issues in their advertisements. The most frequently mentioned policyhissue, in advertisements in both districts, was liberalization of social security benefits. 88 TABLE 11 SINGLE CANDIDATE ADVERTISKENTS Democratic Candidate Republican Candidate I ICII)r ‘3‘.) P3 (H. :2: O c... o c) O :‘1 O H O O O ..-: 4.3 7') 44 H +1 -H H x. H o m :1. gem 0 U) a) m m m <1) :3.» In c a) m a. a) '7: - ° "' I :14 " H I} "' I :5 ._5 m 0,-le 6-; :1 m ’a m CH H m CH 5:1 to y; :5 a) o C o m .5 ”J- <1) 0 U) 0? O m :21, u .11 C14 :2 H r“; :3 1 CLH ;‘ ,1 t: H Lansing State Journal 0 O O 5 2 l2 Flint Journal 15 52 45 9 2 20 Total, 6th District 15 52 45 14 4 52 Houghton Daily Mining Gazette 6 25 52 6 14 40 Iron Mountain News 8 22 47 6 12 42 Marquette Mining Journal 7 21 52 5 12 41 Ironwood Daily Globe 5 25 45 7 12 58 Total, 12th District 26 _ 89 154 24 50 161 89 TABLE 12 MULTI—CANDIDATE ADVERTISEMENTS Democratic Republican c... O :5; Ln. 0 _ E; C 43 (“or-i U .‘__) U "—1 .p, .1 . -«'—' r! :4 :5 Lean 0 ca :4 ('5 mm o c‘) ”:1 m o :3 090.. .1) ”3 1’3 2 :3 or :7 To) (tul'm Léa QEm car-IF: ain't”. . "3 P) (‘5 m 77 O U) "‘5' 3) O :7) G.) O m :7-” 'zn4ti I.rib5 Fist t: +4 C‘tstfi Lansing State Journal 2 O O 5 l 10 Flint Journal 4 O 5 O O 0 Total, 6th District 6 O 5 5 l 10 Houghton Daily fining Gazette 4 O O 7 O 5 Iron Mountain News 1 O O 1 O 5 Marquette Mininngournal 2 O O 5 l 10 Ironwood Daily Globe l O O 2 2 8 Total, 12th District 8 o 1 o I 15 I 3 26 aReferences to issues anywhere in multi-candidate advertisements were counted if they seemed clearly intended to apply to the congressional candidate as well as to other candidates listed. 9O Hypothesis 6 There is a relationship between the relative local strength of the Democratic and Republican parties and the extent of coveragepgiyen Democratig and Republican candi- dates for Congress in daily newspapgps: the greater the local strength of the party of a candidate who is editor- ially opposed, the more nearly equally divided is the cov- erage. Our findings do not support the hypothesis. IDlTable 15 the 6 papers have been ranked, first according to the local strength of the Democratic Party, and secondly, according to percentage of campaign coverage space given the Democratic congressional candidate. A Spearmann rank order correlation coefficient has been computed, and found to be .20, not a significant degree of correlation. RANK ORDERS OF NE LOCAL DEMOCRATIC PERCENTAG. AND DEMOCRATIC PERCENTAGE EOE SPACE 91 TABLE 15 WSPAPERS ACCORDING TO 30E VOTE, Dem. % of Dem. % of Difference votea spaceb in rank Ironwood Daily Globe 1 (59%) 5 (18%) 4 Iron Mountain , News 2 (55%) 1 (72%) 1 Flint Journal 5 (55%) 2 (51%) l Marquette Mining Journal 4 (52%) 4 (55%) O Houghton Daily Mining Gazette 5 (48%) 6 ( 8%) l Lansing State Journal 6 (45%) 5 (50%) 5 I 8For the Michigan Secretary of State, in the preced- ing election (1954), in the county in which the newspaper is published. The vote for Secretary of State irigenerally accepted as the best index of party strength, at least in Michigan at this time. , b The percentage is of total space given both Demo- cratic and Republican Congressional candidates, excluding balanced items, items not directly related to the campaign, editorials and advertisements. CHAPTER V DISCUSSION OF FINDINGS Perhaps it is not surprising that we have found the average paper, among the 6 studied, to have published fewer column inches on the local congressional campaign, in the entire 6 weeks, than it published on sports in a single, typical week-day issue. The chief purpose of most news- papers, after all, is not adult education, or evangelism in behalf of the democratic process, but the making of profit. Thus we can reasonably attribute the lack of more adequate coverage of congressional campaigns to the absence of a public demand strong enough to impress editors with the desirability of expending the time, money and effort required to more fully cover such campaigns. Yet at the same time, the writer heartily concurs with the advice James Russell Wiggins, now managing editor of the Washing- ton Post g Times-Herald, offered fellow editors in 1942 (64', p. 165): I would not leave out the news that is 'inter— esting', but I confess that newspapers as a whole have not made anywhere near the effort which society has a right to require of them to make the important news (that often is dully treated) just as interest- ing as the news which has inherent reader—appeal. Why isn't this done? Primarily because it is easier to exploit a story with obvious reader appeal than it is to exploit a story that requires painstaking work, varied sources of information, imagination and 95 inventiveness. Probably most newspapers lack suffi- cient personnel to treat public affairs as they should be treated. Time, long hours of it, must be added to basic intelligence, imagination and in- genuity, if much news in the field of public affairs is to be made as exciting in print as it is in fact . . .I think editors owe it to readers to do a lot more digging for the drama in the news about public affairs than they have in the past. Perhaps if they exploited the state-house run with staffs as big as those on the sports desk, they would, in the course of decades, find the interest in public affairs as great as the interest in foot- ball. There would be no need to worry about the fate of a democracy should that condition come to pass. Perhaps it never will, but newspapers, in my Opinion, should exert themselves a lot more to lead public attitudes in that direction. Hypothesis 1 It was found that references to non-policy issues were nearly 5 times as frequent as references to policy issues. By way of comparison, Ferguson and Smuckler, in their study of senatorial campaigns in 2 states in 1952, found (18, p. 29) that "over 72 per cent" -- or nearly 41K) 1 -— "of the total appeal references were to non-policy themes of various types." Thus there is a possibility can- didates and/or newspapers may feel called upon to stress policy issues somewhat more strongly inncampaign for the U.S. Senate than they do in a race for a seat in the House of Representatives. This is apparently most strongly the case in regard to attention devoted to foreign policy issues, for although Ferguson and Smuckler found (ibi§., p. 520'that over 42 per cent of all policy issue references were to 94 foreign policy, our own finding in the present study was that less than 2 per cent of non-advertising references to policy issues were to foreign policy issues, and that just over 4 per cent of single-candidate advertisement refer- ences to policy issues were to foreign policy issues. Here again, as in the matter of the paucity of over- all coverage given these 2 congressional campaigns, the "explanation" appears to be simply a lack of any convic- tion on the part of editors that their readers are gppgp- ggpgg in congressional candidates' views on policy issues. (Candidates themselves are no doubt also partially respon-- sible for the lack of emphasis on policy issues, but if editors wanted to publish policy issue material, reporters could easily turn up the candidates' views, or lack of them.) It's easier to write about candidate A's handsome smile and what long hours he campaigns than aboutlhisvfiews on foreign policy, or -- if he's an incumbent -- how he voted last year on tidelands oil. In the absence of com- pelling reasons for coveringsuch items as the latter 2, they are apparently very likely to be ignored. Hypothesis 2 Five of the 6 newspapers studied, it was found, pub- lished more items on the Republican than on the Democratic candidate; 5 of the papers gave the Republican candidate more column inches, (Items "not directly related to the 95 campaign" were excluded from these totals, it should be remembered.) In the 6th Congressional District, the Repub- lican had 52 per cent of all directional (other unmaneutral) items; in the 12th District, the Republican had 67 per cent. We cannot soundly conclude, on the basis of these findings, that there necessarily was political blag, on the part of any editor or editors. The case which most strong- ly suggested the presence of bias was that of the 12th Dis- trict Democratic candidate's home town daily (the Ironwood Daily Globe), which failed to publish a single news story about him. In that case it would appear there was either strong bias, or an incredible lack of initiative in devel- Oping local news possibilities. In general, however, our findings relative to this hypothesis gighp'be simply the result of chance and, in the 12th District, the result of an active, effective in— cumbent Republican congressman, experienced in producing news releases,being pitted against an inexperienced Demo- cratic candidate unaware of the likeliest ways of breaking into print. There ig, however, undeniable support, in our findings, for the possibility of bias in the direction of the editorial leanings of the paper involved. As the num- ber of studies of congressional campaigns under varying circumstances grows perhaps we will have more conclusive evidence. The writer suspects that further studies of other 96 congressional campaigns will confirm the second portion of Hypothesis 2 (concerning greater disparity of treatment of the 2 candidates in numbers of headlines, photographs and front page items than in totals of items and of column inches), even though our findings do not support it. The 6th District Democratic candidate's good showing in numbers of headlines, photos and front page items is largely attri— butable to the fact that in support of his candidacy there were 5 visits to major cities in the district by major national figures; the correSponding figure for his Republi- can opponent was 5. (For the Democrat, Stevenson visited Lansing and Flint, Kefauver visited Flint, and Governor Clement visited Lansing and Flint; for the Republican, Nixon visited Lansing and Flint,enx1Governor Dewey visited Flint.) On} the basis of our findings, it can be assumed that if graph of the candidates had been favored with 5 such visits by national figures, press coverage would have been weighted more heavily in favor of the Republican candidate. In the 12th District, the Democratic candidate's bet- ter showing in photo coverage (41 per cent of total publish— ed) than in item or column inch totals (55 and 50 per cent, reSpectively), was the result of 5 of the 4 papers in the district publishing 2 pictures the candidate distributed: 1 of Mrs. Roosevelt and himself at breakfast, and l of Sen- ator McNamara and himself talking with a patient in a local veterans hospital. Both the pictures were taken in the 97 district; both were highly newsworthy. Thus, in both the 6th and 12th districts, unusual circumstances (wdfixfllhelped the Democrats) accounted for the nature of our findings relative to the second part of Hypothesis 2. It should be noted that the methods employed in the present study have not enabled us to report gl; indica- tions of bias observed in campaign news coverage by the 6 papers involved. Among a number of such indications of bias not reported thus far, the most disturbing was a ten- dency on the part of the Lansing State Journal to publish single-source stories reporting Republican attacks on Dem- ocratic candidates, but to incorporate Republican rebut- tals into all stories reporting Democratic attacks on Re- publicans. Hypothesisp5 We found that in every category of coverage, the in- cumbent Republican fared better than the non-incumbent Re- publican, and the incumbent Democrat better than the non- incumbent Democrat. The reasons for this being the case are generally understood among newspapermen, politicians and political scientists: the incumbent has ready access to facts and speech material (his own paid staff, plus the services of the Legislative Reference Service of the Libra- ry of Congress); he has certain official announcements to 98 make and can regularly take official actions that are news- worthy; in addition to his paid staff he has officialoffice and office machine allowances which make the mechanical as— pect of the production of news releases less of a problem; and he speaks from a position of prestige. In the light of our findings relative to the first part of Hypothesis 2, it is interesting to compare the rel- ative importance in terms of extent of coverage of a candi- date's being an incumbent yg. his being editorially favored by the newspaper concerned. Neither factor was found to be consistently more important, no matter which of a variety of ways of compiling figures was used. In total number of itggg, the 2 Republican candidates led the 2 Democrats by a wider margin than the 2 incumbents led the 2 non—incum- bents. In total column inches, the Republicans led the Democrats by almost exactly the same margin by which the incumbents led the non—incumbents. In numbers of headlines and of photos the incumbents led the non-incumbents by con- siderablywi‘d'er margins than the Republicans led the Demo- crats, and in numbers of front—pgge items, by a consider- ably wider margin than the Democrats led the Republicans. .Hypothesis 4 There is little to discuss regarding our findings on Hypothesis 4. The rarity of published letters to the edi— tor about congressional campaigns clearly suggests that Q 99 editors haven't stirred up much interest in the subject, through either their news or their editorial columns. Hypothesis<5 It was found that in both districts the Democratic candidate's advertisements contained a higher percentage of references to policy issues (as distinguished from non— policy issues) than did the Republican candidate's. ILead— ers of the Democratic Party in Michigan pride themselves on what they term their Party's "programmatic" orientation -- its greater concern with program and principles, and its lesser preoccupation with patronage and personalities, compared with the Republican Party in Michigan and with both parties in most other states. Our findings ggy be interpreted as supporting that claim. On the other hand, it may simply be that in 1956 GOP congressional candidates felt they could win on Eisenhower and other "non-policy" issues, and had no need for discussion of policy issues, with all its attendant pitfalls. Further study of Michi- gan congressional campaigns under differing circumstances is needed for a more reliable conclusion. It should be recalled that Ferguson and Smuckler found (18, p. 51) that Democratic candidates stressmipolicy issues more than did Republican candidates in 5 of 7 cases (including both advertisements and news stories). Thus 100 further study might possibly reveal a nationwide tendency for Democratic candidates to be more policy issue-oriented than Republican candidates. Hyppthesis 6 Regardless of our finding that the relative local strength of political parties does not appear to affect press coverage of congressional candidates, further study of this relationship would seem desirable. The findings of Stene (1956) and of Kobre (1952), summarized above,vflunl considered together with our own findings, suggest that Hy- pothesis 6 might be supported if further studies were made of a group of newspapers including but not limited to pa- pers in Southern states and in very large Northern cities. In both the latter situations the setting would be likely to be strongly Democratic; the findings of such a study might not coincide with our present finding. CHAPTER VI CONCLUSIONS AND RECOMMENDATIONS Conclusions Our major conclusions are as follows: 1) The 6 neWSpapers studied carried little informa- tion about local congressional candidates, and very little about their voting records and/or views on foreign and domestic policy issues. The average paper published fewer column inches on both local congressional candidates, in the 6 weeks preceding the election, than it published on sports in a single, typical week-day issue. During the entire 6 weeks the average paper published but 5 non-adver- tising ipgpg in which congressional campaign policy issues (foreign or domestic) were mentioned, and published nearLy 5 times as many references to ppp—policy issues asto policy issues. Mention of foreign policy issues was particularly rare, occurring less than once (.8) in the average paper's total 6 week non-advertising coverage of the local congres- sional campaign. 2) Incumbency and (at least tacit) editorial support by the paper concerned were found to be the apparent key factors in a congressional candidate's gaining an advan- tage in news coverage, in the 6 papers studied. Each of 102 the incumbents was given a larger share of total coverage than the non-incumbent of his own party. In 4 of 5 cate- gories of coverage the Republican incumbent fared better than the Democratic incumbent, and the Republican.ppp- incumbent fared better than the Democratic non-incumbent in the same 4 out of 5 categories. Neither incumbency nor editorial support was consistently the more important of the two factors, although incumbency was more often of greater importance. 5) Throughout their paid political advertising in daily papers, the 2 Democratic candidates stressed policy issues, as distinguished from nun—policy issues, an average of more than twice as much as did their Republican opponents. 4) There was found to be no significant correlation between the relative strength of the 2 political parties in a paper's home county and the share of congressional cam- paign coverage given the candidatescnfthe 2 parties. Recommendations for Newspaper Action Self-righteous demands for exact parity of coverage for congressional candidates of each of the major parties appear to be fruitless. Not only is such an ideal unreal— istic in terms of accepted standards of journalism, which stress newsworthiness rather than concern for political or other labels, but it is also unrealistic in terms of human 105 nature. Editors and publishers are human, and as Walter Lippman.has observed (54, p. 21), “we are peculiarly in- clined to suppress whatever impugns the security of that to which we have given our allegiance." Lippman also wrote (ibid., p. 64) that it "may be bad to suppress a particular opinion, but the really dead- ly thing is to suppress the news," and there are many dis— tinguished journalists who feel his concern is relevant today.- Editor James A. Wechsler of the New York Post more recently wrote (65, p. 9): What concerns me about the American press is not what most of these papers stand for, but what so many of them refuse to stand for. It is not what they say that seems to me most alarming; it is what they fail to say, and what they refuse to debate. And this, I think, goes to the heart of our modern disease, which is the reluctance to engage in 'con— troversy'. . . In a recent philosophical piece for the Saturdgy 327 view (51, pp. 8, 50) Editor Louis B. Seltzer of the Cleve- land Press wrote: NewsPapers, many of them.built to greatness on the tradition of fearless reporting, are only going through the motions of covering beats or waiting for the news releases to be thrown through the transom . . .It's much easier to hire wire services. . . We have seen that this lack of adequate news cover— age is clearly evident in regard to congressional cam- paigns. There is no easy solution to the problem. Part of what is needed is simply more energetic, conscientious reporting -- and the stimulation, by opinion leaders, of 104 more public demand for it. But in addition, it is the writer's view that some variation of the "battle page" would be a highly desirable addition to daily newspaper coverage of congressional campaigns. A systematic, side- by-side presentation of Opposing candidates' views on ma— jor policy issues would not be unprecedented, for George Seldes reported (50, p. 588) 20 years ago that in 1956 the Npg;12p§‘2gi;y_figp§ "published both a 'Presidential battle page' in the Roosevelt-Landon campaign and an 'economic battle page' on labor." In a recent syndicated column (41), Raymond Moley praised an organization in California which he reported is active in publicizing the voting records, views and activi- ties of the state's Congressmen of both parties. Why can't daily newspapers do the.job, and obviate the pggg for such organizations? Specifically, the writer would have daily papers pub- lish congressional candidates' responses to periodic ques- tionnaires, (The questionnaires, along the lines of similar forms used throughout the country by the League of Women Voters, would ask the candidates' views on various key policy issues. As a public relations device, and as a means of pro- tecting himself against charges of bias in "loading" the questionnaires, an? editor might set up an advisory com- mittee of respected local citizens not prominently identi- 105 fied with either political party, to assist in the prepara- tion of the periodic sets of questions. Representatives of the clergy, of the bar association, the League of Women Voters and of a local college or university political sci- ence department might be invited to serve on such a commit- tee. The committee might weed out defanntOry; and mislead- ing questions, but would seek to stimulate honest contro- versy on basic issues. It is the writer's view that such an effort to great- ly enlarge the gpppg of newspaper coverage of congressional campaigns, along lines that would assure a reasonably fair division of coverage (as between the two candidates), would be far more fruitful than seeking only a more equal divi- sion of present, inadequate coverage. Lippman has declared (54, p. 47) that "the task of selecting and ordering the news is one Of the truly sacred and priestly Offices in a democracy. For the newspaper is in all literalness the bible of democracy, the book out of which a people determines its conduct." Our not—so priest- ly‘editors will be failing their flocks if they do nothing to give the growing number of independent voters more in- formation on the foreign and domestic policy views and records of candidates for the most powerful legislative body in the world. 106 Recommendations for Further Study As we have suggested above, in chapter V, hypotheses 2, 4, 5 and 6 (stated on pp. 54-59) merit further study. Studies of newspaper coverage of other congressional races would provide more evidence as to whether the disparity in news coverage accorded candidates of the two parties is or is not generally greatest in numbers of headlines, photos and front page items. Further studies could turn up a sufficient number of relevant, published letters to edi- tors to suggest the presence or absence of bias in their selection by editors. Studies of congressional campaign advertising in other states could confirm or deny, on a national scale, our Michigan finding that -- in 1956, at least -- Democratic congressional candidates stressed poli— cy issues more than did the Republican candidates. Also, our finding that there appears to be no significant rela— tionship between local party strength and press coverage accorded congressional candidates of the two parties -- ppig finding could be tested under varying circumstances. Of greatest importance, however, is the possibility that objective appraisal of press coverage of a number of congressional campaigns could give us a reasonably conclu- sive answer to the question so heatedly discussed in recent years: do most daily newspapers slant their news coverage in favor of political candidates they support editorially? 107 As suggested in chapter I, there is no wholly satis- factory rebuttal when it is claimed that greater press cov— erage given presidential candidates of a certain party re- flects not bias, but simply the greater newsworthiness of certain of the candidates. Coverage of a sufficient number of congressional campaigns could be studied to escape that difficulty. Presumably, congressional candidates of one party are not consistently, in dozens of districts, more newsworthy than their opponents in the other party -- es- pecially in marginal districts where the 2 parties are of about equal strength. There are fewer than 200 marginal congressional.dis- tricts (districts in which the winning candidate generally gains 55 per cent or less of the total vote). If the jour- nalism schools of several universities were to jointly un- dertake the task - or if Sigma Delta Chi or the Council on Communications Research were to Sponsor it -- a care— fully planned study might be made of daily newspaper cover- age of 1958 and/or 1960 congressional campaigns in a large sampling of the country's marginal congressional districts. In the writer's view, such an undertaking would have far greater potential value than equal effort spent on study of presidential campaign coverage. Disparities in coverage given opposing presidential candidates are far more easily explained in terms pppgp than editors' bias than would be disparities -- if such were found -- in cov- 108 erage given a dozen or more congressional candidates of each party (an equal number of incumbents of each party) running in closely contested districts. Conversely, if coverage accorded such congressional candidates were found to be roughly equal, defenders of the role of the press in political campaigns would have gained a greater point than ‘would be represented by a finding of approximately equal coverage for opposing candidates in a single presidential race 0 l. 2. 5. 4. 5. 7. 8. 9. 10. 11. 12. 15. 14. BIBLIOGRAPHY Adler, Kenneth P. Can Bias be Measured? The Quill, April, 1955, p. 8. American Society of Newspaper Editors. Problems of Journalism: Proceedings of the 1955 Convention (Wash- ington, D. Go), Pp. 184.1850 Associated Press. Dispatch, Nov. 2, 1956, from Ann Arbor, Mich. Batlin, Robert. San Francisco Newspapers' Campaign Coverage: 1896, 1952. Journ. Quar., summer, 1954, PP. 297-5030 Begeman, Jean. The One Party Press Pays Off. New ReEUbO, NOV. 17, 1952, PP. 17-210 Berelson, Bernard. Content Analysis in Communications Research. Glencoe, Ill., Free Press, 1952. Berelson, Bernard, et. a1. Voting. Chicago, Univ. of Chicago Press, 1954. Blumberg, Nathan B. One Party Press? Lincoln, Nebras— ka Press, 1954. Brant, Irving. The Press-for-Willkie Club. Freedom of the Press Today, ed. Harold Ickes, New York, Vanguard Press, 1941, pp. 51-61. Bush, Chilton R. The Analysis of Political Campaign News. Journ. Quar., Spring, 1951, pp. 250-252. Campbell, Angus, et. al. The Voter Decides. Evanston, Ill., Row, Peterson and Co., 1954. Canham, LErlii. The American Press and the 1952 Presi- dential Election. Confluence: Anplnternatippal Forum, March, 1955, pp. 108-116. Commission on Freedom of the Press. A Free and Respon- sible Press. Chicago, Univ. of Chicago Press, 1947. Crowley, John H. A Study of the Treatment by the M11- waukee Journal of the 1954 Kersten-Reuss and Kohler- 15. 16. 17. 18. 19. 20. 21. 22. 25. 24. 25. 26. 27. 110 Proxmire Campaigns. Unpublished Master's thesis, Journ. Dept., Marquette Univ., Milwaukee, Wis., 1955. Dilliard, Irving F. Is it a One—Party Press? New Leader, May 25, 1955, pp. 16-18. Dilliard, Irving F. The Press and the Bill of Rights. Nieman Reports, Jan., 1954, p. 14. Estabrook, Robert H. Press Performance in the Campaign. Nieman Repprts, Jan., 1955, pp. 10-12. Ferguson, LeRoy C., and Smuckler, Ralph H. Politics in the Press: An Analysis of Press Content in 1952 Sena- torial Campaigns. East Lansing, Governmental Research Bur., Mich. State Coll., 1954. Flesch, Rudolph. The Art of Plain Talk. New York, Harper & Bros., 1946. Gabel, Milton, and Gabel, H. W. Texas NeWSpaper Opin- ion. Pub. Qpinion Quar., 1946, pp. 57-70, 201-215. Gallup, George H. The People and the Press. Freedom of the Press Today, ed. Harold Ickes, New York, Van- guard Press, 1941, pp. 115-119. Geller, A., et. al. An Experimental Comparison of Four Ways of Coding Editorial Content. Journ. Quar., Winter, 1942. PP. 562-570. Higbie, Charles E. Wisconsin Dailies in the 1952 Cam- paign: Space vs. Display. Journ. Quar., Winter, 1954, pp 0 56-60. Horwitz, Lionel. Political Advertising in the 1950 Wisconsin Senatorial Campaign. Unpublished Master's thesis, Journ. Dept., Univ. of Wis., 1950. Ickes, Harold (ed.). Freedom of the Press Today. New York, Vanguard Press, 1941, pp. 1—15. Isaacs, Norman E., et. al. A PrOposal for a Study of the Role and Performance of the American Daily News- paper in the 1956 Presidential Election Campaign. Minneapolis, mimeographed, 1955. Klein, Malcolm W., and MacCoby, Nathan. NeWSpaper 0b- jectivity in the 1952 Campaign. Journ. Quar., Summer, 1954, pp. 285-296. 28. 29. 50. 51. 52. 55. 55. 56. 57. 58. 59. 40. 41. 111 KObre, Sidney. How Florida Dailies Handled the 1952 Presidential Campaign. Journ Quar., Spring, 1955, pp. 165-169. ~ Lang, Kurt, and Lang, Gladys Engel. The Television Personality in Politics: Some Considerations. Pub. Opinion Quar., Spring, 1956, pp. 105-112. Lasch, Robert. I See by the Papers. The Progressive, Nov., 1956, pp. 17-20. Lasswell, Harold D.,,et. al. The Politically Signifi— cant Content of the Press: Coding Procedures. Journ. Quar., Spring, 1942, pp. 12-25. . Language of Politics. New York, George W. Stewart, Inc., 1949. Lazarsfeld, Paul F., et. al. The People's Choice: HbW' the Voter Makes Up His Mind in a Presidential Campaign. New York, Columbia Univ. Press, 2nd edn., 1948. Lippmann, Walter. Liberty and the News. New York, Harcourt, Brace and Howe, 1920. . The Public Philosophy. Boston, Little, Brown and COO, 1955. Luce, Clare Booth. The Ambassadorial Issue: Profes- sionals or Amateurs? Foreign Affairs, Oct., 1957, pp. 105—121. Markham, James W., and Stempel, Guido H. III. Analysis of Techniques in Measuring Press Performance. Journ. Quar., Spring, 1957, pp. 187-190. Markham, James W., and Stempel, Guido H. III. Pennsyl- vania Daily Press Coverage of the 1956 ElectDMlCampaign. University Park, Penn., pamphlet, School of Journ., Phnn. State Univ., 1957. Michigan Press Association. Michigan Newspaper Rate- book. East Lansing, Mich. Press Service, Inc., 1956. Millspaugh, Martin. Baltimore Newspapers and the Pres- idential Election. Pub. Opinion Quar., Spring, 1949, pp. 122-125. Moley, Raymond. The National Observer. Syndicated newspaper column, McClure Newspaper Syndicate, New 42. 45. 45. 46. 47. 49. 50. 51. 52- 55. 54- 55- 56. 57. 112 Mott, Frank Luther. NeWSpapers in Presidential Cam- paigns. PDb. Opinion Quar., Summer, 1944, pp. 548-567. . Has the Press Lost its Punch? The Rotarian, 0013., 1952, Po 150 Newspaper Bias? Ed. & Pub., Oct. 25, 1952, p. 54. Page 1 Display Contradicts Authors on Press Bias. Ed. & Pub., Oct. 25, 1952, pp. 7, 59. Post-Election Roundup: News Held in Balance. Ed. & Pub., Nov. 15, 1952, p. 10. Price, Granville. A Method for Analyzing NeWSpaper Campaign Coverage. Journ. Quar., Fall, 1954, pp. 447- 458. Rarick, Galen. California Daily Newspaper Reporting of the 1950 U.S. Senatorial Campaign: A Content Analy- sis. Unpublished Master's thesis, Journ. Dept., Stan- ford Univ., Palo Alto, Calif., 1951. Rowse, Arthur E.- Slanted News. Boston, Beacon Press, 1957. Seldes, George. Lords of the Press. New York, Julian Messner, Inc., 1958. Seltzer, Louis B. What's Happening to Our NeWSpapers? Saturday Rev., April 28, 1956, pp. 7-8, 50-52. Sevareid, Eric. The Big Truth. Minneapolis, pamphlet, School of Journ., Univ. of Minn., and Twin Cities Local, Amer. Newspaper Guild, 1955. Stene, E. 0. Newspapers in the Campaign. Soc. Sci., Spring. 1957. pp. 213-215. . NeWSpapers in the Presidential Campaign. Southwestern Soc._Sci, Quar., March, 1945, pp. 258-264. Survey Research Center, University of Michigan. Poli- tical Party Identification in the State of Michigan. Ann Arbor, hektographed, 1956. Taft, William H. Does Missouri Have a 'One-Party Press'? The Missouri Press News, Feb. 1957, pp. 6, 15—16. Trial by Newspaper. Scientific Amer., Feb., 1949, pp. 16-21. 58. 59. 60. 61. 62. 65. 65. 115 Two-Thirds of Newspaper Readers Believe the Press. The Publishers Aux., Sept. 7, 1957, p. 1. Voelkel, Richard. A Study of the Treatment by the Milwaukee Sentinel of the 1954 Kersten-Reuss and Kohler-Proxmire Campaigns. Unpublished Master's thesis, Journ. Dept., Marquette Univ., Milwaukee, Wis., 1955. Waples, D., and Berelson, B. What the PeOple Were Told. Chicago, mimeographed, Grad. School, Univ. of Chicago, 1941. Watson, Campbell. Nixon—Douglas Contest Coverage Measured. Ed. & Pub., Jan. 27, 1951, p. 12. Weber, Robert H. A Study of the Treatment by the Milwaukee Journal and Milwaukee Sentinel of the 1954 Kersten—Reuss and Kohler—Proxmire Campaigns. Unpub- lished Master's thesis, Journ. Dept., Marquette Univ., 'Milwaukee, Wis., 1955. Wechsler, James A. The One Interest Press: Reflec- tions on Sacred Cows, Elephants and Donkeys. The Pro- gressive, May, 1957, pp. 9—11. Wiggins, James Russell. The Function of the Press in a Modern Democracy. Journ. Quar., Summer, 1942, pp. Zuegner, Charles. A Study of Political News in Two Major Dailies. Journ.Quar., Spring, 1956, pp. 222—224. APPENDIX 115 CODING CARD "9’)D))W‘IIVWWDDDWDDDDDD fli981fl|8€4filmrlf fl'Ei » 9v9vlurst & d5|£ n 251$ s 9515 a: as as Q2 «9E as 4:: ~92: as: so: as 68676665|64636261605958 JJJJJJJJJJ /' o A 316 5 b “\‘\ )W ,'N 23:n|n zHI9 mlrlls w hllB n 1110 9 8» *9‘1 ‘I‘W w “l \ \ ‘\‘\ w w “\'\‘w 1 _._l 4 \ CODE Hole Number 1 Lansing State Journal 2 Flint Journal 5 Houghton Daily Mining Gazette 4. Iron Mountain News 5 Marquette Mining Journal 6 Ironwood Daily Globe 18 During first 5 weeks 19 During last week (Oct. 51 or later) 20 Begins on front page 21 Begins on other page 22 News story 25 Feature story 24 Photo accompanies news or feature story or adver— tisement 25 Photo (separate) 26 Editorial 27 Cartoon 28 Letter to editor 29 Single-candidate advertisement 50 multi-candidate advertisement 51. Body of item refers to Dem. candidate, pro 52 Body of item refers to Dem. candidate, neutral 55 Body of item refers to Dem. candidate, anti 54 Body of item is balanced 55 Body of item refers to Rep. Candidate, pro 56 Body of item refers to Rep. candidate, neutral 57 Body of item refers to Rep. candidate, anti 58 Multi-column.headline 59 Single-column headline 40 Headline mentions Dem. candidate, pro 41 Headline mentions.Dem. candidate, neutral 42 Headline mentions Dem. candidate, anti 45 Headline is balanced 44 Headline mentions Rep. candidate, pro 45 Headline mentions Rep. candidate, neutral 46 Headline mentions Rep. candidate, anti 117 Domestic Policy issues (write specifics on card) Foreign Policy issues (write specifics on card) Personal issues (write specifics on card) Support of or association with presidential or vice-presidential nominee of own party Support of or association with national program Support of or association with gubernatorial nominee or state program.of own party Support of small business' interests Opposition to big business domination of government Opposition to organized labor domination of govern- Promotion of congressional district's interests Other issues (write specifics on card) No-issue advertisement, with picture of congression- No-issue advertisement, without picture of congres- Item not directly related to the campaign Neither the candidate nor a member of his family 47 48 49 5O 51 of own party 52 55 Support of economy in government 54 Support of free enterprise 55 56 Support of farmers' interests 57 Support of labor's interests 58 Opposition to socialism 59 Opposition to communism 6O 61 ment 62 65 64 a1 candidate 65 sional candidate 66 67 is in photo coded 68 Headline not directly relevant to a local congres- sional candidate nor to the congressional race (does not mention candidate by name or title, and does not refer to the campaign) CODING INSTRUCTIONS In the up or left corner of the card (with hole 1 at the lower right write the date of newspaper issue, and beneath the date, the column inch total for the item. (If the item begins on the front page but is continued on an inside page, the column inch totals for the two pages should be noted separately.) I. Inclusion, non—inclusion: 1) 2) 5) 4) 5) Include an item if it mentions (by name or title) a local congressional candidate in more than one sen- tence, pp if it: -- mentions such a candidate in the first sentence, -— mentions such a candidate in the headline, -- is a photo in which a local congressional candi- date is recognizable and is identified in the caption, or a photo which the caption specif- ically relates to a local congressional candi- ate, —- is a letter to the editor signed by a local con- gressional candidate, -- is a letter to the editor in which what is said about a local congressional candidate is the main point of the letter, or -- is an advertisement in which a local congressionr a1 candidate's name is mentioned. Include an item mentioning a local congressional candidate in only one sentence, g3; -- the mention of the candidate is "pro" or "anti", or -- EEe one sentence is at least one column inch in length and is wholly about the congressional cans didat e 0 Include an item about a member of a local congres- sional candidate's family if it mentions the candi- date's name or title and if it otherwise qualifies (according to these instructions) for inclusion. So long as the candidate's name or title is men- tioned once, mention of the relative's name is an acceptable substitute for further mention of the candidate's name or title.) Do not classify an item as a front page item if the only mention.of the congressional candidate is on an inside continuation page. Do not include a photo accompanying an item unless it is specifically related by its caption to a local congressional candidate. II. III. IV. 119 Column inches: 1) 2) 5) 4) When a local congressional candidate is mentioned in an item chiefly about other subjects, measure only the paragraphs mentioning the congressional candidate 0 In editorials and other items which may be printed. in extraawide columns, increase the column inch figure so as to account for the additional space. Include headline and photo space in column inch figures. In multi-candidate advertisements, measure only the congressional candidate's share of copy, photo and white space; don't include column inches of headlines which refer to other candidates as well as to the congressional candidate. Direction: 1) 2) 5) 4) 5) When an item or headline is both anti-one candi- date and pro-the other, make a judgement as to which is predominant. When an item or headline is both.neutral one can- didate and pro or anti-the other, classify it to the latter candidate. When an item or headline refers to only one local congressional candidate by name, classify it to him even though it may also refer to his Opponent less directly. Consider a prediction of election victory to be a pro-candidate statement. When determining whether the body of a multi-can- didate advertisement is neutral or pro the con- gressional candidate, consider all content which appears intended to apply to the congressional candidate. 6) When determining whether the headline in a multi- candidate advertisement is neutral or pro the con- gressional candidate, consider all headlines which appear intended to apply to the congressional can- didate 0 Issues: 1) Don't include issues reported as mentioned by per- sons other than the congressional candidate, unless a specific connection between such issues and the candidate is made. V. 120 2) When references to domestic policy, foreign policy, personal or "other" (see code, hole 65) issues are found, each such issue should be listed on the card, beneath the column inch figure. 5) In multi-candidate advertisements include all issues which appear intended to apply to the con- gressional candidate. Advertisements: l) Classify advertisements mentioning a local congres- sional candidate and the presidential candidate of his party (but no other candidates of their party) as single-candidate rather than as multi-candidate advertisements. 2) When determining whether a multi-candidate adver— tisement headline is single or multi-celumn,l consider only the widest headline which directly refers to the congressional candidate. Ti(w “‘3 .g .1- "IIMIIIIIIAIIM