This item has the following downloads:
1 MANUFACTURER INFLUENCE ON TONE IN AUTO REVIEWS By GREGORY FINK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS IN MASS CO MMUNICATION UNIVERSITY OF FLORIDA 2013
2 2013 Gregory Fink
3 To my parents, teachers, friends and family who supported me through this process
4 ACKNOWLEDGMENTS I would like to thank my committee chair for taking the time to both teach and guide me through the process of writing this thesis I would also like to thank my committee for providing me with their insights during my proposal and defense. The knowledge I have gained from each of these individuals teachings will help me succeed well into the future. Finally, I would like to thank the University for providing me an environment conducive to my studies.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 ABSTRACT ..................................................................................................................... 8 CHAPTER 1 INTRODUCTION ...................................................................................................... 9 2 LITERATURE REVIEW .......................................................................................... 12 SecondLevel Agenda Setting ................................................................................ 13 Financial Influence .................................................................................................. 17 Product Endorsement in the Media ......................................................................... 20 The Role of the Int ernet and the Gate Keeper ........................................................ 22 3 RESEARCH METHODS ......................................................................................... 29 Diction and Definitions ............................................................................................ 32 Data ........................................................................................................................ 36 Intercoder Reliability ............................................................................................... 39 Empirical Framework .............................................................................................. 40 4 RESULTS ............................................................................................................... 47 Empirical Results .................................................................................................... 47 Testing for Bias in Relation to Awards .............................................................. 47 Testing for Bias in Relation to Advertising Expenditures .................................. 48 Manufacturers advertising expenses, 2011 ............................................... 49 Manufacturers magazine advertising expenses, 2011 .............................. 50 Manufacturers Internet advertising expenses, 2011 .................................. 51 Manufacturers Internet and magazine advertising expenses, 2011 .......... 52 Testing for Bias in Relation to Manufacturer sponsored Events ....................... 53 Testing Car and Driv er for bias from manufacturer sponsored events ....... 54 Testing Motor Trend for bias from manufacturer sponsored events .......... 54 5 DISCUSSIO N ......................................................................................................... 63 Conclusion .............................................................................................................. 63 Future Study ........................................................................................................... 68 APPENDIX: CODEBOOK ............................................................................................. 73 LIST OF REFERENCES ............................................................................................... 74
6 BIOGRAPHICAL SKETCH ............................................................................................ 79
7 LIST OF TABLES Table P age 3 1 Overall advertising expenditures in 2011 by manufacturer ................................. 42 3 2 Advertising expenditures on magazine and Internet in 2011 by manufacturer ... 43 3 3 Variables tested .................................................................................................. 44 3 4 Variables tested by manufacturer ....................................................................... 45 4 1 Results by independent variable ......................................................................... 56 4 2 Results by advertising expenditures in 2011 ...................................................... 58 4 3 Results by magazine advertising expend itures in 2011 ...................................... 59 4 4 Results by Internet advertising expenditures in 2011 ......................................... 60 4 5 Results by magazine and Internet advertising expenditures in 2011 .................. 61 4 6 Results by number of press drives ..................................................................... 62
8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Arts in Mass Communication MANUFACTURER INFLUENCE ON TONE IN AUTO REVIEWS By Gregory Fink August 2013 Chair: Wayne Wanta Major: Mass Communication The relationship between automotive manufacturers and the automotive press is often muddled. Using a content analysis, this study looked at 296 reviews of vehicles by ten major manufacturers in the publications Car and Driver and Motor Trend for the year of 2011 (January thr ough December). Using the theory of secondlevel agenda setting, each review was analyzed for tone and compared with the advertising expenditures of each manufacturer in 2011, as well as the number of press related driving events each manufacturer offered each publication. The results failed to support any financial influence by the manufacturer to affect editorial tone; however, significance was found between tone and press events for one of the tested publications.
9 CHAPTER 1 INTRODUCTION The role of adv ertisers often times bleeds from the pages of the media into the news room as David E. Davis Jr. discovered in his April 1968 article for Car and Driver Turn Your Hymnals to 2002. Driving the thennew BMW 2002, Davis (1968) wrote, The 2002 had a lovely l ooking AM/FM affair neatly slipped into its consoleeasily a hundredandfifty bucks worth of radioand I couldn't pick up a Manhattan station from the far end of the Brooklyn Bridge. Honestly. It was maybe the dumbest radio anybody ever stuck in an automobile, like all Blaupunkt and Becker radios, yet the German car makers for reasons unknowncontinue to use them (p. 86). Davis would admit in September 2009 that his refusal to write a column apologizing to Blaupunkt cost him his job with Ziff Davis Publis hing, the thenowner of Car and Driver (Autoline After Hours, 2009). Automotive publications however not only rely on manufacturers as advertisers; these publications also rely on manufacturers to supply vehicles for testing a practice that has allowed automakers to take advantage of the press and new vehicle consumers. The use of a ringer 1964 Pontiac GTO with a bigger than stock 421 cubic inch V8 engine given to Car and Driver for testing resulted in performance numbers the real car was unable to achi eve (socalspeedzone, 2010). Automotive journalist Jason Cammisa reports that Volkswagen was guilty of putting more powerful Europeanspec engines in Volkswagen Scirocco press cars driven by American journalists when the secondgeneration model of the car w as introduced (Cammisa, 2012).
10 Volkswagens press fleet was called out again late in 2011, when Consumer Reports not iced that the Volkswagen Passat s being reviewed by the press contained equipment unavailable on cars sold to consumers (Linkov). Even Ferrari has dealt with issues of delivering ringers to journalists. British journalist Chris Harris notes in his opinion piece How Ferrari spins that, Ferrari will never admit that its press cars are tuned, but has the gall to turn up at any of the big Eur opean magazines' endof year shindigtests with two cars. One for straight line work, the other for handling exercises (Harris, 2011). Harris point is such: Ferrari prepares its press cars in a manner unlike the models consumers can buy. In an interesting twist, Harris was blackballed from driving Ferrari press vehicles after writing How Ferrari spins (Wert, 2011). Though Harris assertions in How Ferrari spins may be untrue, the ability of manufacturers to pull press credentials over an article the c ompany disagrees with puts into question just how unbiased the automotive media are. In his article Taking Readers for a Ride, Frank Greve (2011) notes that automotive journalists are often gifted free business class flights and expensepaid stays at places like Archerfield House, an exclusive seaside golf resort north of Edinborough. In the instance of the Archerfield House, Volkswagen arranged for the press to stay at the luxury resort during the press launch of the thennew Bentley Mulsanne. Of cou rse gift is a relative term to members of the automotive media. Former Car and Driver editor William Jeanes told John Phillips (2011), As far as manufacturer gifts are concerned, a $25 windbreak er is the intergalactic limit.
11 Yet, in the case of Car and Driver it seems limitations on gifts from manufacturers are limited to tangible items, as Phillips writes about press trips: Asked why he robbed banks, Willie Sutton replied, Because thats where the money is, and the dirty little secret about press trips is that you go, like it or not, because thats where the cars are. Consumers readi ng automotive reviews expect a publication to be free of bias and an accompanying agenda; however, even Pulitzer Prize winning automotive reviewer Dan Neil admits that Pressure for a favorable review is more and more intense (Greve, 2011). In order to test the bias in automotive review s in commercial media, a content analysis will be carried out using major automotive review media publications to test the theory of secondlevel agenda setting that comes to play in these reviews through the use of the manufacturer as both advertiser and vehicle supplier for automotive media. An affective dimension will be used to determine the bias in the commercial automotive media outlets Car and Driver and Motor Trend. Car and Driver is an automotive publication that has established itself as the worlds largest automotive magazine brand (Car and Driver, 2012), while Motor Trend has been in the business of automotive journalism s ince 1949 and has a total audience of just over 6.8 million readers (Motor Trend, 2013). This study examines reviews of vehicles written during the publishing dates of January 2011 through December 2011 for both magazines. Using the theory of secondlevel agenda setting, this study look s at the role tone has on attribute salience in automotive reviews
12 CHAPTER 2 LITERATURE REVIEW The inception of secondlevel agenda setting as a theory can be traced back to two earlier theories highly utilized in media st udies: agenda setting and framing. Though this study focuses on the role of secondlevel agenda setting, acknowledging key studies in agenda setting and framing is important to understand the evolution and significance of agenda setting at the secondlevel Agenda settings origins as a theory can be traced back to the famous Chapel Hill study conducted by Maxwell McCom bs and Donald Shaw (1972). This study attempted to match what Chapel Hill voters said were key issues of the campaign with the actual content of the mass media used by them during the cam paign ( p. 177). Their attempt was to find a link between what issues voters found important and what issues the media had deemed important in its coverage of the 1968 election. McCo mbs and Shaw found that the results of their experiment suggest a very strong relationship between the emphasis placed on different campaign issues by the media (reflecting to a considerable degree the emphasis by candidates) and the judgments of voters as to the salience and importance of v arious campaign t opics ( p. 181). Since the Chapel Hill study many researchers have documented the effects of what Carroll and McCombs (2003) called the sal ience of objects ( p. 38) in the media. Salience can be operationally defined as perceived importance (Weaver, 2007, p. 142) of an object, or the classical assertion that the news tells us what to think about (McCombs and Shaw, 1993, p. 62).
13 Like the original Chapel Hill study by McCombs and Shaw, many studies henceforth in agenda settin g involve the news ability to create object salience in the political realm. As McCombs and Shaw (1993) note, agenda setting entered a new phase, during the 1976 election when Weaver, Graber, McCombs and Eyal (1981) extended the idea of agendas into two new domains. One was the agenda of candidate characteristics reported by the media and learned by voters; the other was the larger agenda of personal concerns on which all aspects of politics issues, candidates, and so on are but a single, and usually minor, item (p. 59). This secondary agenda largely introduced the role of secondlevel agenda setting into the theory of agenda setting. A noted aspect of agenda setting and secondlevel agenda setting in the automotive media is the proclivity of awarded endorsements. Car and Driver nominates the 10Best cars available on the market and Motor Trend delivers an of the Year award to the SUV, Car and Truck that the publication deems to be the best vehicle of the model year. Manufacturers honored with thes e awards will in turn advertise the achievement. SecondLevel Agenda Setting The secondlevel of agenda setting is particularly important bec ause it goes beyond presenting to an audience what to think about it also presents to an audience how to think about certain things. Such a look into agenda setting at either the first or secondlevel has not been conducted as extensively in entertainment and review journalism as it has been in political media, n or has extensive research of agenda setting at the fir st or secondlevel been conducted in a medium that requires an author give an evenhanded review of a product in the name of helping consumer purc hasing decisions. This is not to say important studies do not exist in this area of research, it is
14 simply ac knowledging that fewer studies do and notes the importance of this studys research into secondlevel agenda setting in automotive reviews. Before noting the literature that suppor ts agenda setting at the secondlevel, a definitive clarification of framing and secondle vel agenda setting must be acknowledged. David H. Weaver in his 2007 article Thoughts on Agenda Setting, Framing, and Priming looks deeply into the defined operative differences of both framing and secondlevel agenda setting. Weaver notes that frame is a very ambiguous term that can be applied to many different aspects of messages and to many different ty pes of messages ( p. 144). As Weaver concludes: There are similarities between secondlevel agenda setting and framing, even if they are not identical processes But framing does seem to include a broader range of cognitive processes such as moral evaluations, causal reasoning, appeals to principles, and recommendations for treatment of problems than does secondlevel agenda setting (the salience of attr ibutes of an object) ( p. 145 146). Essentially these two terms are tied together very closely, with framing noting a broader more ambiguous term that is often used as a device in approaching the telling of a story and secondlevel agend a setting being more clearly defined as the salience of object attributes or how to think about an object. As Carroll and McCombs (2003) write, citing the work of McCombs and Ghanem at the secondlevel of agendasetting news coverage conveys more than just facts, it also conv eys feelings and tone ( p. 39). This study looks to find how tone affects secondlevel agenda setting in automotive reviews.
15 In a similar approach to secondlevel agenda setting, Dietram A. Scheufele (2006) notes five factors that may potentially influence how journalists frame a given issue: social norms and values, organizational pressures and constraints, pressures of interest groups, journalistic routines, and ideological or political orientations of journalists ( p.109). Each of these can be seen in automotive review journalism as well. Looking particularly at pressures of interest groups, automotive journalists must often succumb to the pressure of advertisers. As noted above in this study s introduction, David E. Davis Jr. felt advertiser pressure to write a column apologizing for his views on the Blaupunkt radio in his article Turn Your Hymnals to 2002 ( 1968). Similarly, Chris Harris argues that manufacturer pressure and fear of being blacklisted from press events is in large part why automotive journalists do not call Ferrari out for providi ng ringer press cars (2011). Yet, one cannot ignore the ideologicalorientations of a journalist (Scheufele, 2006, p. 109) in the automotive medium. Though it can be assumed that most automotive journalists attempt to remain fair and balanced, personal ideology can often times lead to the framing of a story that supports the authors personal automotive preferences. What Scheufele fails to acknowledge in his list of potential influencers of journalist frames is the role of the publisher, as Warren Breed noted in his study Social Control in the Newsroom: A Functional Analysis (1955) In this study Breed concludes that newsroom policies (and thus the means of delivery and presen tation that can
16 influence both framing and secondlevel agenda setting) are learned by osmosis ( p. 328). This osmosis can be traced back to numerous aspects of the newsroom, but Breed particularly notes the role of editors and publishers in establishing a secondlevel of agenda setting. As Breed notes, when an executive sees a clearly anti policy item, he bluepencils it, and this constitutes a lesson for the staffer (p. 332); this form of editorial notation allows journalists of a publication to lea rn the newsroom policies toward agenda setting at the first and secondlevel. Breed concludes that it is the publisher who influences such policies that encourage a publication s newsroom agenda; however, Breed fails to look into the psychology of why a publisher is setting this agenda. Since the latter part of Breeds study (finding out the motivations behind publisher decisions) was not carried out, it is hard to conclude what drives a publisher s editori al decisions; however, as David E. Davis Jr. not ed on Autoline After Hours (2009), it was not stereomanufacturer Blaupunkt that asked him for an apology column for his negative remarks of the stereo in the BMW 2002, but the publisher of Car and Driver at the time, Ziff Davis Publishing. From such an ex ample we can infer that the publishers who influence the agenda of a newsroom are themselves influenced by their advertisers. Such a policy not only implies the encouragement of secondlevel agenda setting, but also a potential for bias. Bias then is how s econdlevel agenda setting and framing differ most. Framing, James W. Tankard Jr. (2001) argues goes beyond notions of pro or con, favorable or unfavorable, negative or positive Framing adds the possibilities of additional, more
17 complex emotional responses and also adds a cognitive dimension (beliefs about objects as well as attitudes) ( p. 96). Financial Influence Robert Entmans (2007) study on media bias reveals how monetary means can suppor t bias in a publication. In the political realm he notes that, If anything, research cited previously suggests a net advantage for conservatives acros s a range of issues and groups Even if journalists working for the national media tend to be predominantly (though moderately) liberal (Pew Research Center for Peopl e and the Press, 2004), that could be outweighed by for skilled media management (Hacker and Pierson, 2005), the limits that the campaign man, 2005), and the strong 1994; Bennett, 2007) ( p. 170). Entmans study supports the role of monetary bias in the media. He notes that the frames of journalists are gener ally liberal, but the agenda being set by the publisher often encourages reporters to work outside of their own frames and support a conservative agenda due to the monetary influence of the Republican P arty. Money, though, doesnt only bias the political articles we read, as television is also a proven medi um in which the media succumbs to advertisers as well. Keith Brown and Roberto Cavazos (2005) found that even when adjusting for audience size and demographic composition, advertisers pay a premium for advertising spots on sitcoms, and pay a discount for advertising on darker programming like news magazines and police dramas. As a result, the broadcast networks air a disproportionately high amount of sitcoms ( p. 18).
18 Such a finding reveals the potenti al for advertisers to set the media agenda. In essence, television networks desire to make more money from advertising. In order to do this, the networks allow the advertisers to influence the agenda; in this case it sets the agenda of a predominant skew of light sitcoms over dark dramas. The authors cite HBO as a counter example to the content based agenda advertisers encourage networks to follow: HBO deliberately empl oys darker themes, and more complex story lines to consciously differ entiate themselves from network television. Since HBO is a pay channel, this certainly indicates that broadcast networks systemically underserve large groups of viewers; viewers willing to pay directly fo r different programming ( p. 20). In a similar study Champlin and Knoedler (2002) found advertiser influence not only in the kinds of programming put on, but also in the content within the show. This secondlevel of agenda setting has been well accounted for, as the authors note that, 'When the Project on Excellence in Journalism surveyed local TV news directors in 2000, one third suggested that they had received pressure to avoid negative stories about advertis ers, while 75 percent of investigative reporters and editors at television stations reported that advertisers had tried to influence their content (FAIR 2000, 1) (p. 463). Unfortunately, the current market for automotive media largely forces automot ive review publishers to rely on advertising revenue in order to produce content; however, the very nature of the magazine industry encourages users to pay a fee to read the material. As such, magazines are, in theory, financially culpable to the needs of both its advertisers and its readers.
19 A similar study looked at the role advertisers play on media bi as. In this study the authors acknowledged that most media are either paid for by subscribers and advertisers or solely by advertisers. This study discovered t hat when media companies only source of revenue [are] advertising fees, newspapers choose to mi nimally differentiate in their reporting strategies. By eliminating any slanting in their reporting and appealing to readers with moderate beliefs, the newspapers can increase their readership, which allows them to command higher fees from advertisers (Gal or, Geylani & Yildirim, 2010, p. 26). The researchers found that when revenues come from advertising in addition to subscription fees, the additional advertising market introduces three different effect s on the behavior of newspapers. The authors first note the existence of advertising revenues puts downward pressure on subscription fees, as newspapers intens ify competition for subscribers. This results in publishers choosing to increase polarization in order to alleviate this competition. The authors also find that as publishers attempt to increase their readership in order to attract advertisers, they may moderate slanting in order to appeal to readers having moderate beliefs Finally the authors note that in order to reach an audience that wil l be more receptive to the messages of advertisers, the advertiser may induce newspapers to seek greater distinctiveness and bias, and by doing so, offer a better match betw een advertisers and readers ( p. 26). Though this studys subject matter involved newspapers in particular, a similar patt ern of revenue and competition affects the automotive media industry, as well. According to this study, we can infer that bias will be more prevalent if an automotive
20 publisher receives more funding by its advertiser s than its subscribers. This is a growing trend, as online material by automotive publications generally does not require user fees but still does rely upon advertising. While a reliance on advertisers and subscribers by the automotive media industry may i ncrease bias according to the findings in the above study, Ellman and Germano (2009) found that monopolistic media conditions automatically lead to advertiser influence on reporting, while media industry competition in duces maximal accuracy ( p. 699). Us ing the results of both the study by Gal or, Geylani and Yildirim, as well as the study by Ellman and Ger mano, it can be infer red that the competitive nature of the automotive press should serve to limit bias. Product Endorsement in the Media Chiang and Knight (2011) looked into the role of political endorsements by the media and the effects this had on agenda setting. Like an auto manufacturer boasting of an award by the press, an endorsement by a popular media company for a political candidate can set the agenda of voters for who m they should cast their ballot for on Election Day (in the case of car buyers, the agenda would be related to which vehicle a consumer should purchase). Yet, the authors discovered that although there is the potential for an agenda to be set by endorsements voters generally rationally account for the credibility of any endorsements ( p. 817). The authors further concluded that, these results suggest that voters are sophisticated and attempt to filter out any bias in medi a cover age of politics ( p. 817).
21 However the authors did note that, Endorsements are also more influential among moderate voters and those more likely to be exposed to the endorsement ( p. 817). Such a statement infers that endorsements are less influential to individuals who are well informed or individuals who are presumably set in their way s politically. The results of this study can correlate to car buyers and the media agenda set by the automotive press. As noted earlier, publications such as Car and Driver nominate the 10Best cars available on the market each year and Motor Trend provides three vehicles (an SUV, a car and a truck, respectively) with its of the Year award, an award for being the best vehicle in its respective segment for a given model year. While automotive enthusiasts who pay attention to the industry and its products will likely be less influenced by the agenda setting at the first or secondlevel by the press, the average car buyer (which can be equivocated to the moderate voter) is m ore likely to be influenced by an endorsement by an automotive publication, as they are not as well vers ed on the subject matter This relays the importance of this study, as a biased automotive media can influence the purchase decisions of the average car buyer by simply handing out awards t o their major advertisers or by reviewing manufacturer supplied press vehicles that are ringer models of the actual vehicle available to the public. The real i ty of using awards in advertising is that bias as a result of agenda setting at the first and secondlevel moves from beyond the physical or digital pages of the consumer oriented automotive media, and into the advertisements of the profit oriented automobile manufacturer. In this regard Golan, Kiousis and McDaniel (2007) found significant evidence of an agendasetting effect at the first level and mixed evidence of a secondlevel agenda-
22 se tting function ( p. 441) w hen studying political ad campaigns. Though this study deals with salience in political advertising and not in the news media, it is important to note how salience can be transferred from the media to advertising. The authors finding of only mixed object attribute salience in their study of political advertising would seem to make for an interesting f uture study. From the abovementioned study by Chiang and Knight (2011) it seems it is possible that voters might be similarly influenced by endorsements in advertising. In the case of the study by Golan, Kiousis and McDaniel (2007) we c an infer that these voters may use their rational e to negate the effects of secondlevel agenda setting in a political ad as a means to sway their vote, though the object itself (agenda setting at the first level) becomes a topic the voters choose to think on. It can be i nferred that automotive advertising that uses media endorsements will also likely show evidence of agenda setting at the first level with mixed evidence of secondlevel agenda setting in advertising to new car buyers. Where second level agenda setting is present in automotive advertisements, car buyers have the potential to use their rational e to negate the endorsed content in the advertisement by the manufacturer The Role of the Internet and the Gate K eeper Outside of the print medium, the study of agenda setting at both the first and secondlevel has also been studied in the digital realms of the Internet As McCombs (2005) states, the Internet is the new frontier for research on these traditional agenda setting affects (p. 544). McCombs explains that the rise of the Internet has caused some social observers [to] predict the end of agenda setting as audiences fragment and virtually
23 everyone has a unique external media agenda that is a highly individualized composite constructed from this vast wealth of online news and information ( p. 544). While in theory this is a legitimate argument, in real ity it is not the case. McCombs uses James Hamilton as a source in noting that the five largest American newspapers Wall Street Journal, USA Today, New York Ti mes, Los Angeles Times, and Washington Post account for 21.5 percent of the circulation among the top 100 daily newspapers (p. 545). Subsequently the top five newspaper websites which includes three of those newspapers, USA Today, New York Times, and Washington Post in addition to the Detroit News and Seattle Times account for 41.4 percent of the total links found on the Internet to the top 100 newspapers (p. 545). Despite the rise of the Internet, individuals are still leaning toward the agendas of only a few media companies. As Marilyn, Wanta and TzongHorng (2002), discovered when measuring the influence of agenda setting by The New York Times, Associated Press, Reuters, Time Magazine and CNN against online Internet discussions on electronic bull etin boards that threefourths of the issues examined in the study showed a clear agenda setting relationship between media coverage and Internet discussions (p. 459). Such a finding reveals the influence these major media outlets have on the news agenda. While the findings by Marilyn, Wanta and TzongHorng show the news medias ability to set the agenda, Esrock and Leichty (1998) discovered that relatively few corporations use the medium [of their web page] for either public consultation or to help set the public agenda (p. 317). Such findings support the notion that a corporation, or automakers, best way to influence agenda setting is through the media. This study
24 attempts to draw a correlation between the agenda setting power of the media by determining if corporate spending on advertising is able to influence how the media portrays a vehicle manufacturers products. Another way the media is able to set the agenda is through their role as a gate keeper. As David Manning White (1950) found in his study The Gate Keeper: A Case Study In the Selection of News another aspect of agenda setting (a term he did not use since his study stems before the Chapel Hill study) is that of the gate keeper In Whites study, the gate keeper was a wire editor who chose which wire stories to run in the newspaper he worked for White concluded that in the case of his subject known as Mr. Gates in his position as gate keeper the newspaper editor sees to (even though he may never be consciously aware of it) that the community shall hear as a fact only those events which the newsman, as the representative of his culture bel ieves to be true ( p. 390). Similar to Breeds (1955) findings, White concludes that those in charge set the agenda. T hough there may be mor e information available today due to the Internet, those in charge of the most trafficked sites are the gate keepers who set the agenda of the public. This is important t o note as the two publications in this study publish their material in the form of a printed magazine, as well as in the online realm Two studies penned by Jonathan Reuter testing for bias in review journalism must be noted for their relation to this study of secondlevel agenda setting in automotive reviews Testing review bias in the r atings of wine, Reuter (2009) discovered that Overall, I find little consistent evidence of bias. At worst, the tests for biased ratings
25 suggest that Wine Spectator rates wines from advertisers almost one point higher than wines from nonadvertisers ( p. 19). Of note is that the studys control subject is a wine reviewer funded entirely by subscribers, while the other subject analyzed, Wine Spectator uses both subscribers and advertisers for funding. T he fact that the bias found in Wine Spectator is inco nsistent and generally unsubstantial supports the findings by Gal or, Geylani and Yildirim (2010) as well as Ellman and Germano (2009) who se studies concluded that when both subscriber fees and advertising account for funding, as well as when there is me dia competition, bias is limited. Though some wines can rival the price of a new car, generally the finances required to purchase a bottle of wine are merely fractions of the cost of a new vehicle. As such, it is important to discover if reviews of items with greater financial value result in more or less bias. Looking into the role of advertising in the financial media, the study Do Ads Influence Editors? Advertising and Bias in the Financial Media (2005) found that the personal finance publications in our sample are more likely to recommend the funds of their advertisers, even after controlling for the observable fund characteristics that their readers might value (Reuter and Zitzewitz, p. 1819). This study chose five of the top six recipients of m utual fund advertising dollars (p. 1), as test subjects in testing for bias. Proof of bias and agenda setting by these publishers is important to acknowledge, as the monetary risks of mutual funds can lead to gains and losses by investors often times equaling the price of many new cars. The
26 advertisers of these funds can make significant gains by consumer investments, just as automakers can make significant profit from consumer purchases. The resulting bias found in the study by Reuter and Zitzewitz is not surprising given Sutters (2002) conclusions in Advertising and Political Bias in the Media: The Market for Criticism of the Market Economy. As Sutter states, advertisers often choose to bundle their ads with the medias messages, thus paying for the production of news programs, because the associated news generates a larger audience than unbundl ed transmission ( p. 728). It is in the advertiser s best interest to ensure that their own product is being discussed in the media. Sutter seems to imply this phenomenon occurs by the advertiser paying for the production of new programs ( p. 728). Despite this, it must be acknowledged that the consumer ultimately decides what to make of the medias agenda independently. As Chiang and Knight (2011) discovered voters are sophisticated and attempt to filter out any bias in medi a coverage of politics ( p. 817). As noted earlier, this conclusion can be applied to consumers reading automotive reviews as well. In the study Revisiting the Clinton/Lewinsky Scandal: T he Convergence of Agenda Setting and Framing (2003) the authors find in their content analysis of the infamous Presidential scandal that Despite copious coverage of the scandal, and the public's concern over the president's alleged lying and obstruction of justice, Clinton's approval ratings were not negatively affected (Yioutas and Segvic, p. 571). The authors note that this is par ticularly interesting in regard to the fact that, Even in 1992, when the economy was a very salient topic, 67% of respondents
27 believed that during the presidential campaign too little attention had been paid to moral standards, specifically that the country's were not high enough ( p. 576). In a similar situation, cars praised by the press often do not motivate the consum ing public to purchase these vehicles, and cars that are panned by the press often do not discourage the consuming public from purchasing these vehicles. Specifically this can be seen with the 2012 Honda Civic, which merely a year after its introduction, r eceived significant changes after the press panned the subpar nature of the car. Despite this the 2012 Civic sold decidedly well, and as A utoblog notes, The Civic [was] selling near the t op of its class with 234,029 units [sold] through September of  (Paukert, 2012). This exemplifies that the agenda (particularl y at the secondlevel) set by reviews by automotive journalists does not necessarily correlate to consumer motivation. From the a bove literature related to the theory of secondlevel a genda setting, as well as literature on the theory of agenda setting at the first level, framing, gate keeping, and the general idea of advertising bias this study attempts to answer the following hypotheses : 1. As net advertising spending goes up for a manufacturer, the percentage of reviews containing no negative tone will go up. 2. As magazine advertising spending goes up for a manufacturer, the percentage of reviews containing no negative tone will go up. 3. As Internet advertising spending goes up for a manufacturer, the percentage of reviews containing no negative tone will go up. 4. As magazine and Internet advertising spending goes up for a manufacturer, the percentage of reviews containing no negative tone will go up.
28 5. As the number of manufacturer sponsored press drives increase, the percentage of reviews containing no negative tone will go up. Though the final hypothesis does not test for a monetary independent variable, it is an important variable to test for in its potential to bias automotive reviews. As Ji m Mateja, former automotive reviewer for the Chicago Tribune, told Frank Greve (2011), A few or more than a few [automotive journalists] would never say anything bad about a car, because they wanted to keep getting free cars to drive.
29 CHAPTER 3 RESEARCH METHODS This study relies on a content analysis to determine bias in the field of automotive journalism. As defined by Riffe, Lacy and Fico (2005), a content analysis is the systematic and replicable examination of symbols of communication, whic h have been assigned numeric values according to valid measurement rules and the analysis of relationships involving those values using statistical methods, to describe the communication, draw inferences about its meaning, or infer from the communication t o its context, both of production and consumption (p. 25). In laymans terms, a content analysis places numerical values on content in order to draw inferences about the meaning of the content. As noted above, the content analysis used in this study is th eory driven. The choice of a content analysis stems from previous studies in the field of agenda setting at the first and secondlevels, as well as studies relating to media bias. Of course, a content analysis is not without its negatives. As Shoemaker and Reese (1996) note, Reducing large amount of text to quantitative data does not provide a complete picture of meaning and contextual codes (p. 29). Yet, in the context of this study no better option is available to determine bias in automotive reviews than the content analysis. Shoemaker and Reese acknowledge this by concluding that most content analyses focus on the fact that media often singles out and highlights certain elements over others (p. 34) i.e. the media is prone to using framing and agenda setting. In regard to this study, no study was found that deals with the issue of bias in the automotive media, making this research particularly important to others studying the
30 field of agenda setting at the secondlevel. This study is also important to consumers looking for professional advice in the automotive publishing field before the purchase of a new vehicle. Studies were found that question the bias in other forms of review media, though. Looking at the study Does Advertising Spending Influence Media Coverage of the Advertiser? the authors used a content analysis to determine if the advertising spending of Italian fashion companies resulted in more coverage by a variety of fashion publications through Europe and the U.S. As the authors concl ude, With a sufficient advertising budget, the media may cover products without regard to their quality (Basuroy and Rinallo, 2009, p. 42). While these authors did not conclude an editorial bias in such coverage, their content analysis did conclude that it is possible for fashion companies to pay for editorial coverage. Such a conclusion is interesting to note in regard to the automotive publ ishing industry as well. In 2011, General Motors spent $3,055,700,000 in advertising, according to Advertising Age (Johnson, 2012) Such an advertising expenditure should result in General Motors receiving more coverage according to the conclusion of Basuroy and Rinallo (2009); however, this study seeks to find if General Motors large advertising expenditure results in biased reviews toward the products of General Motors. As Gentzkow and Shapiro (2005) note, media firms desire to build a reputation for accuracy ( p. 29). As such, it is unexpected that the automotive media will blatantly endorse a subpar product. Rather, it is expected that the automotive media will choose to use nonhostile diction and tone to avoid both endorsing a subpar product and putting down a major advertisers vehicle
31 Such dict ion and tone largely leads to the secondlevel of agenda setting in automotive reviews. Rather than simpl y providing its audience with what to think about, the reviewer can use tone to influence how to think about the object. Carroll and McCombs (2003) provided that study units in a content analysis can be arrayed along an affective dimension, typically defined as positive, negative, neutral, and mixed ( p. 3839), acknowledging the work of McCombs and Ghanem in noting that such an evaluative dimension at the secondlevel of agendasetting recognizes that news coverage conveys more than just facts, it also conv eys feelings and tone ( p. 39). With a focus on feeling and tone, this study analyzes editorial content from both Car and Driver and Motor Trend, noting each tested reviews positive, neutral or negative tone toward the vehicle being reviewed. Comparing the results of the affective dimension with the advertising expenditures, as well as the incentive factor of the number of manufacturer sponsored press drives, of the manufacturers that spent the most money on adv ertising in 2011 will reveal a relationship between reviews and manufacturer influence. As noted earlier, the content being analyzed was produced for the publishing year of January 2011 through December 2011. In order to compare the results of the content analyzed with manufacturer advertising expenditures, Advertising Ages June 25, 2012 article the 100 Leading National Advertisers, 2012 Edition is referenced. This article details the top 100 leading advertisers during the year of 2011. Specifically it notes the expenditures of the top 100 advertisers; however, additional spending information is provided and used in this study as well.
32 Diction and Definitions As suggested by Carroll and McCombs (2003) an affective dimension that consists of positive, ne gative and neutral is used to denote tone in the reviews of automotive products Using regression analys es, this study seeks to find a relationship of an exponent ially greater percentage, and number, of new car reviews with positive or not negative (the i nverse of negative) tone in automotive media for manufacturers with larger advertising expenditures. Another factor that is analyzed is the number of manufacturer press events put on by each manufacturer. These events are operationally defined as press drives. Press drives are manufacturer sponsored events that give journalists time behind the wheel of a new product in conditions generally controlled by the manufacturer. In these cases the manufacturer often supplies the vehicle, the route, and in many c ases perks, such as meals and hotel accommodations if the location of the press drive is located in a different city, state or country from the journalist. Alternatively, publications test vehicles, often supplied by the manufacturer, for multiple days (i n some cases longer) and conduct numerous tests on the product including performance testing. These tests are operationally defined as road tests and come in the form of a review of an individual car, a comparison between a handful of competing vehicles and longterm road tests, in which the publication uses the vehicle for an extended period of time ( Motor Trend holds onto longterm vehicles for a year, while Car and Driver generally holds onto longterm vehicles for 40,000 miles of driving) and proceeds to write a concluding review of their experience with the vehicle. The manufacturer generally supplies the vehicle(s) in road test scenarios.
33 Each review was placed into a document using IBMs SPSS statistical software in order to compare the number and percentage of reviews with positive, negative and neutral tone given to different manufacturers by each publication. In order to make the analyses results consistent, only vehicles road tested or driven during a press event were analyzed as study units. F urthermore, vehicle reviews analyzed consist only of new (at the time of publication) production models. These vehicles will have been, or will soon be, made available to the public and were actually driven by the publication for editorial purposes. Meanwhile, when looking at road tests of longterm vehicles, only the conclusion of the publications longterm road tests were analyzed. As such, articles that introduce or update readers to the status of these longterm vehicles were not analyzed for content. Furthermore, feature stories that involve travel by the publication while using a vehicle outside of a press event setting were also ignored; unless the publication was able to retain the vehicle long enough to conduct their own performance testing. In so far as the definitive properties of positive, neutral and negative are concerned, positive is defined as a review by the publication that uses diction possessing a favorable tone. Similarly, negative is defined as a review by the publication pos sessing an unfavorable tone. And neutral is defined as a review by the publication that is neither favorable nor unfavorable in its tone. Such neutral tone is further defined in two ways for the purpose of this study. Neutral reviews are reviews by the publication that express neither favorable nor unfavorable tone to an object and its attributes Neutral reviews are reviews by the publication that express an evenhanded mix of favorable and unfavorable tone to an object and its attributes.
34 The inform ation units analyzed are the meaningful symbols grouped for comparative analysis (Riffe, Lacy & Fico, 2005, p. 70). As such, the information units in this content analysis are the publication, the manufacturer of the object being reviewed by the publication, the manufacturers total advertising spending in 2011, the manufacturers advertising spending toward magazines in 2011, the manufacturers advertising spending toward the Internet in 2011, the manufact urers combined spending toward magazines and the Internet in 2011, and the type of review conducted by the publication (i.e. road test or press drive). Another information unit used in this study is that of award. Award denotes that a vehicle was noted by Car and Driver as a 10Best car or by Motor Trend as a car, truck or SUV of the Year. Manufacturer total advertising spending in 2011, as well as manufacturer magazine advertising spending in 2011 and manufacturer Internet advertising spending in 2011 come from Advertising Ages 100 Leading Nati onal Advertisers, 2012 Edition. Manufacturer combined spending toward magazines and the Internet in 2011 is the total of the latter two information units manufacturer magazine advertising spending in 2011 and manufacturer Internet advertising spending i n 2011. Table 31 shows the automobile manufacturers included in Advertising Ages 100 Leading National Advertisers, 2012 Edition, as well as each manufacturers advertising expenditures in 2011, while Table 32 shows the expenditures of each manufacturer toward advertising in magazines and the Internet, as well as a combined total of the two expenditures. Reviews of vehicles produced from the ten manufacturers listed in Table 31 are the study units being analyzed in editorial reviews conducted by Car and Driver and Motor Trend. The decision to look at advertising spending by the
35 manufacturer for the year 2011 as a whole stems from both Breeds (1955) study, in which he noted the publisher plays a large role in social control of the newsroom, and the Wil lie Sutton Rule, which stresses the need for an individual to focus on activities that generate high returns (Investopedia). In accordance with Breeds results, if a publisher seeks to gain social control of the newsroom through agenda setting it is in his or her best interest as the publisher, and not the editor of the magazine, to encourage behavior that generates high returns. In regard to Car and Driver and Motor Trend, the publishers that own both magazines are involved in areas of the commercial m edia industry that extend beyond magazines ( Car and Driver is currently owned by Hearst after having been purchased in 2011 by Hachette Filipacchi Media, and Motor Trend is owned by Source Interlink Media). Furthermore, Internet branding for both Car and D river and Motor Trend extends the editorial functions of these publications past the supermarket newsstands of printed media. For definitive purposes the ten manufacturers that function as study units for this study are defined as such: General Motors Co. consists of the brands Chevrolet, GMC, Buick and Cadillac; Ford Motor Co. consists of the brands Ford and Lincoln; Toyota Motor Corp. consists of the brands Toyota, Lexus and Scion; Fiat/Chrysler Group consists of the brands Fiat, Ferrari, Maserati, Dodge, Jeep, Ram and Chrysler; Honda Motor Co. consists of the brands Honda and Acura; Nissan Motor Co. consists of the brands Nissan and Infiniti; Hyundai Motor Co. consists of Hyundai; Kia Motors Corp. consists of Kia; Volkswagen consists of Volkswagen, Audi, Lamborghini, Bugatti and Bentley; and Daimler consists of Mercedes Benz and Smart In order to complete
36 regression analysis, vehicles tested by both Car and Driver and Motor Trend are accompanied by the reported manufacturers advertising expenditure, as well as the manufacturers specific expenditures to magazine advertising and Internet advertising, and the combined spending of each manufacturer on magazine advertising and Internet advertising using data obtained from Advertising Ages 100 Leading National Advertisers, 2012 Edition . The publishing date span of January 2011 through December 2011 was chosen for a number of reasons. For one, this ensures that there are a large number of reviews to be analyzed. Another reason is the access available to the advertising expenditures each major manufacturer spent during the year of 2011, an important factor for testing this studys hypotheses and determining if there is a relationship between advertising expenditures and new car reviews. Analyzing the percentage of positive, neutral and negative reviews, as well as the number of reviews with observed net positive tone, for the automotive advertisers listed in Advertising Ages 100 Leading National Advertisers, 2012 Edition was done in order to account for the varied number of cases analyzed for each manufacturer. Calculating the percentage of reviews with positive, neutral and negative tone allows for an even weight when running regression analysis on these figures. Data This study tests the tone in automotiv e reviews for the automotive publishing industrys two biggest automotive magazine titles in 2011. Car and Driver recorded advertising rate card revenue of $148,622,418 and Motor Trend recorded advertising rate card revenue of $132,050,125 in the year 2011 (PIB, 2012). This study also uses
37 data from Advertising Ages 100 Leading National Advertisers, 2012 Edition, and the Advertising Age database. Reviews published in both Car and Driver and Motor Trend were obtained by purchasing reprints of both magazi nes 2011 publishing year (January 2011 December 2011), constituting 24 magazines in total. For every new car review that fit the operational definitions of road test or press drive by each magazine, the make, model and parent manufacturer were gathered, as well as the additional notation of whether the review possessed positive, neutral or negative tone. This resulted in a total number of 410 reviews, 192 of which were written by Car and Driver and 218 of which were written by Motor Trend. Tone was meas ured by looking at specific object attributes commonly noted in new car reviews. These attributes are outlined in the codebook attached in the appendix of this study. Each attribute was tested for the positive, neutral or negative tone in the review, or wa s marked N/A if the attribute was not mentioned. Two final factors, conclusion and other, were also tested for positive, neutral or negative tone. The conclusion of the review was tested for its tone, as were other attributes not listed in the codebook. Counting the number of positive, neutral and negative remarks, as well as using some personal judgment, aided in determining the overall tone of the vehicle being reviewed. Next, information was collected from Advertising Ages list of 100 Leading National Advertisers, 2012 Edition. Further information was collected from Advertising Ages database denoting each manufacturers measured spending in 2011 on magazine advertisements and Internet advertisements.
38 Makes not listed under the manufacturers (Table 32) listed in Advertising Ages 100 Leading National Advertisers, 2012 Edition were dropped from the study, and the total number of reviews analyzed was reduced to 296, with 131 sourced from Car and Driver and 165 sourced from Motor Trend. Of these 296 r eviews, each was tested again for positive, neutral or negative tone and marked accordingly all reviews were met with the same affective dimension score as originally found. Table 33 summarizes the independent and dependent variables that are used in th is study to test for regression. An observed seven of ten awards were given to a manufacturer included in this study by Car and Driver while an observed two of three awards were given to a manufacturer included in this study by Motor Trend. Tone by both publications was used as a variable, with positive, neutral and negative coded as one, zero and negative one, respectively. Car and Drivers average of .351, compared to Motor Trends average of .206, shows that less articles with positive tone are written about for vehicles tested by Motor Trend. Advertising expenditures in 2011 by each of the ten manufacturers in this study was also used as a variable. Finally the number of press drives and road tests conducted by each publication were used as study units to test for the effects that the manufacturer as a supplier of test vehicles and, in the case of press drives test locations, have on the positive, neutral or negative tone that enters product reviews written by these two publications. Table 34 shows the number and percentage of reviews given to each manufacturer by each publication with positive, neutral and negative tone.
39 Using a frequency table, the number of positive, neutral and negative reviews tested for tone for each manufacturer by each publication was placed over the total number of reviews conducted for each manufacturer by each publication to calculate percentage totals. Since this study is using regression to map tone against a number of variables, creating equal weight between the reviews w ritten by both publications for the respective manufacturers is important so that the regression is not merely calculating frequency. Intercoder R eliability In order to ensure the reliability of the affective dimension used in this study, intercoder reliability was relied upon. Using the ten manufacturers listed in Advertising Ages 100 Leading National Advertisers, 2012 Edition, 296 reviews were analyzed for this study, with 131 sourced from Car and Driver and 165 sourced from Motor Trend. Intercoder rel iability requires that 10 percent of the material tested in the study be tested by an outside coder. This meant 30 (29.6) reviews needed to be coded by the individual serving as the coder. However, before vetting reviews of vehicles by manufacturers that were not listed in Advertising Ages 100 Leading National Advertisers, 2012 Edition, 410 reviews were analyzed. In the best interest of this study, and the reliability of the coding, it was decided that 10 percent of the original reviews would be coded f or reliability. This left the individual serving as the coder with 41 articles to code. All articles coded by this individual were reviews of vehicles produced by the ten manufacturers listed in Advertising Ages 100 Leading National Advertisers, 2012 Edi tion. Twenty of the articles coded for reliability were sourced from Car and Driver and 21 were sourced from Motor Trend. Due to the fact that Motor Trend published 34 more
40 reviews than Car and Driver that were analyzed for tone in this study, the coder coded an additional Motor Trend article. Using the intercoder reliability method of percent agreement, a method defined as the percentage of all coding decisions made by pairs of coders on which the coders agree (Lombard, Snyder Duch & Bracken, 2002, p. 590), this study sought a percentage agreement of .9 (or 90 percent) or greater between the 41 reviews coded by the independent coder. With 41 articles, this meant that at least 37 (.9 percentage agreement equals 36.9 agreed upon reviews) of the 41 review s needed to be in agreement with the original results of the affective dimension conducted during this study. Us ing the codebook (see appendix ) as a guide, the coder incorrectly coded three reviews of the 41 coded. As such, percent agreement was at 92.68 percent, supporting the reliability of the original coding. Even if the coder had only coded for reliability for 10 percent of the reviews used for regression in this study (n296), the three reviews that differed during intercoder reliability results in a percent agreement of 90 percent, or just at the threshold of what this study sought in percentage agreement when testing for intercoder reliability. To further ensure reliability of the coding in this study, Cohens Kappa was calculated by SPSS to be .878 with a p value of less than zero. Such a value is categorically outstanding according to the standards set by Landis and Koch (1977) that set the rule of thumb values of Kappa from 0.40 to 0.59 are considered moderate, 0.60 to 0.79 substantial, and 0.80 outstanding (TexaSoft, 2008). Empirical Framework Using the IBM statistical software SPSS, the empirical framework of this study is mapped out within the parameters of the program. Using this framework, data collected
41 for this study were run for linear regression in order to test for biased tone within the automotive publishing medium and the manufacturers that support it. An ANOVA provides a figure regarding the significance of the regression found in order to support whether the data input for this study confirm or deny bias by the automotive press, or whether such bias is significant enough to warrant further study on the subject in the future. Significance is defined as a figure below .05. Any figure above .05 was dismissed as statistically insignificant When mapping regression for the observed cases of positive, neutral and negative tone, the cases were coded to determine observed net positive tone in reviews. Positive, neutral and negative were coded as one, zero and negative one, respectively. The for mula to find observed net positive tone is as such, where X is the number of reviews recorded with positive, neutral or negative tone: X(1)+X(0)+X(1)= observed net positive tone When mapping regression for the percentage of cases deemed positive, neutral and negative, the inverse of percentage negative was scrutinized as well. This inverse, percentage not negative, is the culmination of the percentage of positive and neutral tone reviews over the independent variable. The formula to find percentage not negative is: Percentage positive + percentage neutral = percentage not negative
42 Table 31 Overall advertising expenditures in 2011 by manufacturer Manufacturer Adver tising dollars spent in 2011 ( millions) General Motors Co. $ 3,055.70 Ford Motor Co. $2,141.30 Fiat/Chrysler Group $1,770.90 Toyota Motor Corp. $1,727.60 Honda Motor Co. $1,139.90 Nissan Motor Co. $940.80 Volkswagen $729.60 Hyundai Motor Co. $559.20 Kia Motors Corp. $497.80 Daimler $388.30
43 Table 3 2 Advertising expenditures on magazine and Internet in 2011 by manufacturer Manufacturer Advertising dollar s spent in magazines in 2011 (thousands) Internet spending in 2011 (thousands) Total ma gazine and Internet spending ( thousands) General Motors Co. $214,926.00 $241,782.00 $456,708.00 Ford Motor Co. $136,179.00 $90,155.00 $226,334.00 Fiat/Chrysler Group $154,262.00 $84,823.00 $239,085.00 Toyota Motor Corp. $134,221.00 $137,931.00 $272,152.00 Honda Motor Co. $89, 359.00 $73,605.00 $162,694.00 Nissan Motor Co. $85,614.00 $38,877.00 $124,491.00 Volkswagen $29,799.00 $30,132.00 $59,931.00 Hyundai Motor Co. $39,529.00 $26,663.00 $66,192.00 Kia Motors Corp. $22,046.00 $24,554.00 $46,600.00 Daimler $14,259.00 $11,659.00 $25,918.00
44 Table 33 Variables tested Variable N Observed Average Std. Deviation Minimum Maximum Awards Car and Driver 10 7 0.7 0.823 0 2 Awards Motor Trend 10 2 0.2 0.632 0 2 Tone Car and Driver 131 131 0.351 0.6671 1 1 Tone Motor Trend 165 165 0.206 0.6198 1 1 Advertising dollars spent in 2011 (in millions) 10 296 $1,295.11 862.69196 $388.30 $3,055.70 Advertising dollars spent in 2011, magazines (in thousands) 10 296 $92,019.40 66940.19101 $14,259.00 $214,926.00 Advertising dollars spent in 2011, internet (in millions) 10 296 $76,018.10 70163.02263 $11,659.00 $241,782 Advertising dollars spent in 2011, Internet and magazines (in thousands) 10 296 $168,010.50 133975.9925 $25,918 $456,708 Press drives Car and Driver 10 32 3.2 2.348 0 7 Press drives Motor Trend 10 39 3.9 2.998 0 10 Road tests Car and Driver 10 99 9.9 4.383 3 19 Road tests Motor Trend 10 126 12.6 6.586 4 24
45 Table 34 Variables tested by manufactur er General Motors Co. Ford Motor Co. Fiat/Chrysler Group Toyota Motor Corp. Honda Motor Co. Nissan Motor Co. Volkswagen Hyundai Motor Co. Kia Motors Corp. Daimler Observed positive Car and Driver 6 9 7 3 7 1 15 2 1 9 Observed neutral Car and Driver 8 5 9 5 4 5 8 7 2 4 Observed negative Car and Driver 0 0 1 2 1 4 2 1 0 3 Observed positive Motor Trend 5 7 13 5 0 4 11 3 0 4 Observed neutral Motor Trend 13 11 15 12 6 5 15 7 3 8 Observed negative Motor Trend 3 3 2 1 3 1 1 3 1 0 Percent positive Car and Driver 42.90% 64.30% 41.20% 30% 58.30% 10% 60% 20% 33.30% 56.30% Percent neutral Car and Driver 57.10% 35.70% 52.90% 50% 33.30% 50% 32% 70% 66.70% 25% Percent negative Car and Driver 0% 0% 5.90% 20% 8.30% 40% 8% 10% 0% 18.80%
46 Table 34 Cont inued General Motors Co. Ford Motor Co. Fiat/Chrysler Group Toyota Motor Corp. Honda Motor Co. Nissan Motor Co. Volkswagen Hyundai Motor Co. Kia Motors Corp. Daimler Percent positive Motor Trend 23.80% 33.30% 43.30% 27.80% 0.00% 40% 40.70% 23.10% 0% 33.30% Percent neutral Motor Trend 61.90% 52.40% 50% 66.70% 66.70% 50% 55.60% 53.80% 75% 66.70% Percent negative Motor Trend 14.30% 14.30% 6.70% 5.60% 33.30% 10% 3.70% 23.10% 25% 0% Press drives Car and Driver 2 2 5 5 2 2 6 1 0 7 Road tests Car and Driver 12 12 12 5 10 8 19 9 3 9 Press drives Motor Trend 3 4 6 3 1 3 10 2 0 7 Road tests Motor Trend 18 17 24 15 8 7 17 11 4 5
47 CHAPTER 4 RESULTS Empirical Results Testing for Bias in Relation to A wards This section attempts to test whether Car and Driver a nd Motor Trend favor larger advertisers when awarding vehicles. Car and Driver is known for its 10Best award, an award given to the ten vehicles on the market the magazine views as the best. In 2011 these cars were: the Hyundai Sonata, Volkswagen Golf/ GTI, Honda Accord and Honda Fit, Ford Mustang GT, Cadillac CTS V, Chevrolet Volt, BMW 3 Series/M3, Mazda MX 5 Miata and Porsche Boxster/Cayman. Likewise, three of the Year articles were published in 2011 by Motor Trend. Chevrolet took 2011 Car of the Year and 2011 Truck of the Year with the Volt and Silverado HD, while Land Rover took 2012 SUV of the Year with the Range Rover Evoque (though not included in this study, the 2011 SUV of the Year, which was published in 2010, was the Porsche Cayenne). As such, seven of Car and Drivers 10Best cars are produced by the ten manufacturers observed as a unit of study due to their inclusion in Advertising Ages 100 Leading National Advertisers, 2012 Edition, while two of the three awards given by Mot or Trend were occupied by a top automotive advertiser. More interesting is that the two awards given to a major automotive advertiser from Motor Trend were awarded to vehicles produced by General Motors, the highest spending advertiser among automotive manufacturers in 2011, and the second highest spending advertiser overall,
48 with only Procter & Gamble, Co. spending more on advertising in 2011 according to Advertising Ages 100 Leading National Advertisers, 2012 Edition (Johnson, 2012). Correlation, however, is not causation, and with only one years worth of awards collected for use in this study, it would be irresponsible to draw conclusions from these numbers. Yet, such statistics are hard to ignore and in order to determine if, to paraphrase Shakespeare, there is something rotten with the state of automotive awards, further research should be conducted comparing advertising spending with nominal awards over a larger period of time. Testing for Bias in Relation to Advertising Expenditures With the advertising expenditures of ten automotive manufacturers serving as the independent variable, and the percent of reviews with positive, neutral and negative tone given to each manufacturer by both Car and Driver and Motor Trend serving as the dependent variable, this study used regression analysis to test for bias by one or both of these publications for manufacturers with greater advertising expenditures. The regression test relied upon overall advertising spending in 2011 by each manufacturer as the independent variable; however, three subsequent tests were performed to see if expenditures restricted to only advertising spending in magazines, the Internet, and both magazines and the Internet correlated to bias in reviews as well. The decision to test for r egression against each manufacturers spending in 2011 on magazine advertising and Internet advertising was made due to the fact that both Car and Driver and Motor Trend operate as print and online publications. With each manufacturers individual spending on magazine and Internet advertising extracted from Advertising Ages database, it was decided to test the combined sums of the two spending units by each manufacturer as both Car and Driver and Motor Trend receive
49 funding from magazine and online adverti sing. For each regression test performed, figures related to Beta, R2 and significance were taken. These figures are listed together in Table 41. Manufacturers advertising expenses 2011 In 2011 General Motors Co. spent $3,055,700,000 on advertising, Ford Motor Co. spent $2,141,300,000, Fiat/Chrysler Group spent $1,770,900,000, Toyota Motor Corp. spent $1,727,600,000, Honda Motor Co. spent $1,139,900,000, Nissan Motor Co. spent $940,800,000, Volkswagen spent $729,600,000, Hyundai Motor Co. spent $559,200,000, Kia Motors Corp. spent $497,800,000 and Daimler spent $388,300,000 (Johnson, 2012) Table 42 shows that with a Beta of .166 and an R2 of .028, Car and Drivers regression of the percentage of reviews with positive tone against manufacturer adverti sing spending is statistically insignificant (pvalue .646). The regression of the percentage of reviews with negative tone has a Beta at .341, an R2 of .117, but is also statistically insignificant (p value .334). Likewise, regression reveals a statistic ally insignificant (p value .577) Beta of .202 and an R2 of .041 for the number of reviews published by Car and Driver with observed net positive tone against manufacturer advertising spending. Table 42 also shows that in testing for bias in reviews against manufacturer advertising expenditures in Motor Trend, similar results to Car and Driver are found. Beta is .151 and R2 measures at .023 for regression of the percentage of reviews with positive tone, but is statistically insignificant (pvalue .676). An R2 of .003 for regression of the percentage of reviews with negative tone and a Beta of .059 is also statistically insignificant (p value .872). The regression for observed net reviews with positive tone
50 by Motor Trend has a Beta of .139 and an R2 of .019, but is statistically insignificant (pvalue .701). Manufacturer s m agazine advertising expenses 2011 This section uses the same dependent variables to test a new monetary independent variable: advertising expenses by each automotive manufacturer towar d magazine advertising in the year 2011. Given that both Car and Driver and Motor Trend continue to be, and historically have been, magazines, this figure reported by Advertising Age is relevant to both publications. In fact Car and Driver recorded print a dvertising rate card revenue of $148,622,418 and Motor Trend recorded print advertising rate card revenue of $132,050,125 in 2011, the two largest rate card revenues recorded in the automotive media that year, according to the Publishers Information Bureau (2012). In 2011 General Motors Co. spent $214,926,000 on magazine advertising, Ford Motor Co. spent $136,179,000, Fiat/Chrysler Group spent $154,262,000, Toyota Motor Corp. spent $134,221,000, Honda Motor Co. spent $89,359,000, Nissan Motor Co. spent $85,614,000, Volkswagen spent $29,799,000, Hyundai Motor Co. spent $39,529,000, Kia Motors Corp. spent $22,046,000 and Daimler spent $14,259,000 (Johnson, 2012). Table 43 shows that with Beta at .021, an R2 at less than .001 regression of the percentage of reviews by Car and Driver with positive tone is statistically insignificant (p value .954). The regression of the percentage of reviews by Car and Driver with negative tone has a Beta of .207 and an R2 of .043, but is also statistically insignificant (p value .566). Likewise, the number of reviews with observed net positive tone has a Beta of .056 and an R2 of .003, but is statistically insignificant (pvalue .879).
51 Table 43 shows that Motor Trends results were similar to Car and Driver when observing regression in the tone of automotive reviews against each manufacturers advertising expenditures in 2011 for magazines. Beta measures .178 with an R2 of .032 for regression of the percentage of reviews with positive tone, and Beta measures .047 with an R2 of .002 for regression of the percentage of reviews with negative tone. Regression for both proved statistically insignificant (pvalues of .622 and .897, respectively). Regression for the number of reviews with observed net positive tone has a Beta of 151 and an R2 of .023, but is statistically insignificant (pvalue .678). Manufacturer s Internet advertising expenses 2011 This section uses the same dependent variables to test a new monetary independent variable: advertising expenses by each automotive manufacturer toward Internet advertising in the year 2011. In 2011 General Motors Co. spent $241,782,000 on Internet advertising, Ford Motor Co. spent $90,155,000, Fiat/Chrysler Group spent $84,823,000, Toyota Motor Corp. spent $137,931,000, Honda Motor Co. spent $73,605,000, Nissan Motor Co. spent $38,877,000, Volkswagen spent $30,132,000, Hyundai Motor Co. spent $26,663,000, Kia Motors Corp. spent $24,554,000 and Daimler spent $11,659,000 (Johnson, 2012). Table 44 shows that with Beta at .056 and R2 a t .003 the regression of the percentage of reviews by Car and Driver with positive tone is statistically insignificant (p value .877). Meanwhile, the regression of the percentage of reviews by Car and Driver with negative tone has a Beta of .291 and R2 of .085, but is statistically insignificant (pvalue .414). Likewise, the regression of the number of reviews by Car and Driver with
52 observed net positive tone has a Beta of .085 and R2 of .007, but is statistically insignificant (p value .815). Table 44 a lso shows that Motor Trend offers similar results to Car and Driver when running regression against Internet advertising expenditures of the ten manufacturers observed in this study. R2 measures less than .001 for regression of both the percentage of reviews with positive tone (Beta .012), and the percentage of reviews with negative tone (Beta .010). Both are statistically insignificant (pvalue of .974 and .977, respectively). Regression for the number of reviews by Motor Trend with observed net positive tone is much the same, with Beta at .013 and R2 at less than .001, regress is statistically insignificant (p value .971). Manufacturer s Internet and magazine advertising e xpenses 2011 This section uses the same dependent variables to test a new monetar y independent variable: advertising expenses by each automotive manufacturer toward Internet and magazine advertising in the year 2011. This figure combines the reported figures by Advertising Age each manufacturer spent on advertising on the Internet and in magazines in 2011. In 2011 General Motors Co. spent $456,708,000 on Internet and magazine advertising, Ford Motor Co. spent $226,334,000, Fiat/Chrysler Group spent $239,085,000, Toyota Motor Corp. spent $272,152,000, Honda Motor Co. spent $162,694,000, Nissan Motor Co. spent $124,491,000, Volkswagen spent $59,931,000, Hyundai Motor Co. spent $66,192,000, Kia Motors Corp. spent $46,600,000 and Daimler spent $25,918,000 (Johnson, 2012). Table 45 shows that by comparing the percentage of reviews by Car and Driver with positive tone against the combined advertising spending by the ten observed
53 manufacturers, regression has a Beta of .040 and R2 measures .002, though this data is statistically insignificant (p value .913). Regression of the percentage of rev iews with negative tone has a Beta of .256 and R2 of .066, but is also statistically insignificant (pvalue .475). Likewise, the regression of the number of reviews by Car and Driver with observed net positive tone has a Beta of .072 and R2 of .005, but i s statistically insignificant (p value .843). Table 45 shows that Motor Trend follows a similar pattern, with the regression of the percentage of reviews with positive tone and the regression of the percentage of reviews with negative tone lacking any st atistical significance. R2 measures .007 with Beta at .083 for regression of the percentage of reviews with positive tone, and R2 measures .001 with Beta at .029 for regression of the percentage of reviews with negative tone. Both are statistically insignificant (p value measured at .819 and .936, respectively). Regression for the number of reviews by Motor Trend with observed net positive tone is no more telling, with Beta at .069 and R2 at .005, it is statistically insignificant as well (p value .850). T esting for Bias in Relation to Manufacturer sponsored E vents Of the 296 cases observed in this study, 71 of the reviews were operationally defined as press drives. These are manufacturer sponsored events that give journalists time behind the wheel of a new product in conditions generally controlled by the manufacturer. On top of this, perks are often times provided to the reviewer during a press drive. The fifth hypothesis of this study questions the role these events have on the type of tone present in aut omotive reviews. The number of press drives each manufacturer offered to both Car and Driver and Motor Trend during the publishing year of January 2011 to December 2011 served
54 as the independent variable. As such the number of press drives offered by each manufacturer to each publication was compared against the percentage of reviews each publication gave each manufacturer with positive, neutral and negative tone, as well as the observed net positive tone each publication gave each manufacturer. Specific fi gures for Beta, R2 and significance for each regression test can be found in Table 46. Testing Car and Driver for bias from manufacturer sponsored events In 2011 Car and Driver reviewed two vehicles by General Motors Co. in a press drive setting, two vehi cles by Ford Motor Co., five vehicles by Fiat/Chrysler Group, five vehicles by Toyota Motor Corp., two vehicles by Honda Motor Co., two vehicles by Nissan Motor Co., six vehicles by Volkswagen, one vehicle by Hyundai Motor Co., zero vehicles by Kia Motors Corp. and seven vehicles by Daimler. Table 46 shows that when running regression for the percentage of reviews by Car and Driver with positive tone against the number of press drives offered by each of the ten observed manufacturers produces an R2 of .148 and a Beta of .385, while the percentage of reviews by Car and Driver with negative tone produces an R2 of .053 and a Beta of .230. Both regression calculations were statistically insignificant (pvalue measured at .272 and .522, respectively). Likewise, regression revealed that the number of reviews by Car and Driver with observed net positive tone has a Beta of .456 and an R2 of.208, but is statistically insignificant (pvalue .185). Testing Motor Trend for bias from manufacturer sponsored e vents In 2011 Motor Trend reviewed three vehicles by General Motors Co. in a press drive setting, four vehicles by Ford Motor Co., six vehicles by Fiat/Chrysler Group, three vehicles by Toyota Motor Corp., one vehicle by Honda Motor Co., three vehicles by
55 Nissan Motor Co., ten vehicles by Volkswagen, two vehicles by Hyundai Motor Co., no vehicles by Kia Motors Corp. and seven vehicles by Daimler. Table 46 shows that when measured for regression against the number of press drives provided by each of the ten observed manufacturers in this study, the percentage of reviews by Motor Trend with both positive and negative tone have Betas of .760 and .791, respectively, and R2 at .578 and .625. P value measured at .011 and .006 show a statistically significant relationship between the percentage of reviews with positive, negative and not negative (the inverse of negative) tone that Motor Trend gives to manufacturers that offer a greater number press drives. Likewise, the number of reviews by Motor Trend with observed net posi tive tone also reveals a statistically significant relationship between the two variables, with a Beta of .849, R2 at .720 and the pvalue measured at .002. To further examine the significance of the data, Spearmans rank correlation was applied in order to constrict variance to the independent variable by applying rank order to the figures used. Each case remained statistically significant, with the level for the percentage of reviews with positive tone and negative tone measuring .001. The net observed n umber of reviews with positive tone was met with level of significance of less than .001.
56 Table 41 Results by independent variable Advertising expenditures Magazine advertising expenditures Internet advertising expenditures Magazine and Internet ad vertising expenditures Manufacturer events Percent positive Car and Driver Beta 0.166 0.021 0.056 0.04 0.385 R 2 0.028 0 0.003 0.002 0.148 P value 0.646 0.954 0.877 0.913 0.272 Percent negative Car and Driver Beta 0.341 0.207 0.291 0.256 0.23 R 2 0.117 0.043 0.085 0.066 0.053 P value 0.334 0.566 0.414 0.475 0.522 Observed net Car and Driver Beta 0.202 0.056 0.085 0.072 0.456
57 Table 41 Cont inued Advertising expenditures Magazine advertising expenditures Internet advertising expenditures Magazine and Internet advertising expenditures Manufacturer events P value 0.577 0.879 0.815 0.843 0.185 Percent positive Motor Trend Beta 0.151 0.178 0.012 0.083 0.76 R 2 0.023 0.032 0 0.007 0.578 P value 0.676 0.622 0.974 0.819 0.011 Percent negative Motor Trend Beta 0.059 0.047 0.01 0.029 0.791 R 2 0.003 0.002 0 0.001 0.625 P value 0.872 0.897 0.977 0.936 0.006 Observed net Motor Trend Beta 0.139 0.151 0.013 0.069 0.849 R 2 0.019 0.023 0 0.005 0.72 P value 0.701 0.678 0.971 0.85 0.002
58 Table 42 Results by advertising expenditures in 2011 Beta R 2 P value Percent positive Car and Driver 0.166 0.028 0.646 Percent negative Car and Driver 0.341 0.117 0.334 Observed net Car and Driver 0.202 0.041 0.577 Percent positive Motor Trend 0.151 0.023 0.676 Percent negative Motor Trend 0.059 0.003 0.872 Observed net Motor Trend 0.139 0.019 0.701
59 Table 43 Results by magazine advertising expenditures in 2011 Beta R 2 P value Percent positive Car and Driver 0.021 0 0.954 Percent negative Car and Driver 0.207 0.043 0.566 Observed net Car and Driver 0.056 0.003 0.879 Percent positive Motor Trend 0.178 0.032 0.622 Percent negative Motor Trend 0.047 0.002 0.897 Obser ved net Motor Trend 0.151 0.023 0.678
60 Table 44 Results by Internet advertising expenditures in 2011 Beta R 2 P value Percent positive Car and Driver 0.056 0.003 0.877 Percent negative Car and Driver 0.291 0.085 0.414 Observed net Car and Driver 0 .085 0.007 0.815 Percent positive Motor Trend 0.012 0 0.974 Percent negative Motor Trend 0.01 0 0.977 Observed net Motor Trend 0.013 0 0.971
61 Table 45 Results by magazine and Internet advertising expenditures in 2011 Beta R 2 P value Percent po sitive Car and Driver 0.04 0.002 0.913 Percent negative Car and Driver 0.256 0.066 0.475 Observed net Car and Driver 0.072 0.005 0.843 Percent positive Motor Trend 0.083 0.007 0.819 Percent negative Motor Trend 0.029 0.001 0.936 Observed net Motor T rend 0.069 0.005 0.85
62 Table 46 Results by number of press drives Beta R 2 P value Percent positive Car and Driver 0.385 0.148 0.272 Percent negative Car and Driver 0.23 0.053 0.522 Observed net Car and Driver 0.456 0.208 0.185 Percent positive Mo tor Trend 0.76 0.578 0.011 Percent negative Motor Trend 0.791 0.625 0.006 Observed net Motor Trend 0.849 0.72 0.002
63 CHAPTER 5 DISCUSSION Conclusion The purpose of this study was to see if manufacturers are able to influence agenda setting at the sec ondlevel in reviews written by the automotive press by studying the tone used in automotive editorial content. The two publications Motor Trend and Car and Driver were chosen as the publication study units for this work, while the ten automobile manufacturers, General Motors Co., Ford Motor Co., Fiat/Chrysler Group, Toyota Motor Corp., Honda Motor Co., Nissan Motor Co., Volkswagen, Hyundai Motor Co., Kia Motors Corp. and Daimler were chosen as manufacturer study units due to their large advertising expendi tures in 2011 based off of information from Advertising Ages 100 Leading National Advertisers, 2012 Edition. The dependent variables studied were the observed net positive tone of reviews given by each publication to vehicles reviewed from each manufact urer, as well as the percentage of reviews with positive, neutral and negative tone given by each publication to vehicles reviewed from each manufacturer. Independent variables consisted of the reported net advertising spending in 2011 by the ten manufact urers in this study, reported magazine advertising spending in 2011 by the ten manufacturers in this study, reported Internet advertising spending in 2011 by the ten manufacturers in this study, reported Internet and magazine advertising spending in 2011 by the ten manufacturers in this study, and the number of press drives each publication recorded for each of the ten manufacturers included in this study. Overall this study found very little consistent evidence of bias in nearly every area tested. Neither Car and Driver nor Motor Trend showed any statistically significant
64 tendency to bias the tone in the reviews that each magazine published in accordance with the advertising expenses of the manufacturers; be it net advertising expenses, magazine advertising expenses, Internet advertising expenses, or both Internet and magazine advertising expenses. Furthermore, Car and Driver showed no statistically significant tendency to increase the number of reviews with positive tone, or decrease the number of reviews with negative tone, for manufacturers that offered the magazine more time behind the wheel of its products in the setting of a manufacturer sponsored press drive. The same cannot be said of Motor Trend, though, as the magazine showed significance levels that were statistically significant for regression when review tone was regressed against the number of press drives by each of the manufacturers studied. Motor Trend appears to give a greater percentage of reviews with positive tone, a lower percentage of reviews with negative tone, a greater percentage of reviews without negative tone, and a larger number of reviews with net observed positive tone to manufacturers that invite the magazine to a greater number of press drives. As noted earlier, no conclusi ons can be drawn regarding manufacturer influence on the awards given out by either publication, but observed statistics from the awards given out by both manufacturers in 2011 do warrant further study of manufacturer influence on awards. The results of t he content analyses led to the rejection of all five hypotheses for Car and Driver revealing no statistical significance by the magazine in favoring its reviews towards manufacturers with higher advertising spending or a greater number of press drive events.
65 Hypotheses one through four were also rejected in the case of Motor Trend; however regression analysis revealed that Motor Trend gives a greater number and percentage of reviews with positive and not negative tone to manufacturers that offer the magaz ine press drives. Thus, hypothesis five fails to be rejected. The lack of biased tone in reviews of automotive products gives consumers reason to trust that neither Car and Driver nor Motor Trend are allowing monetary influence to bias their reviews. Despi te the fact that it does not appear that either publication slants its reviews in favor of manufacturers that spend more on advertising, the fact that more than half of the reviews used in this study (152) were analyzed as having neutral tone (i.e. either no discernible tone, or an approximately even amount of positive and negative tone) makes it possible that both publications attempt to offer a better match betw een advertisers and readers ( Gal or, Geylani & Yildirim 2010, p. 26), by providing reviews t hat use a combined effort of positive and negative tones to both keep the manufacturer at ease by complimenting the product while also alerting consumers to problem areas in the product. Given the information discovered in this study, it appears the publis her does not have undo influence in the newsrooms of these two publications. The same cannot be said about the information found in this study regarding the tone in reviews and the perks manufacturers offer automotive journalists during press drives. While reviews by Car and Driver did not show any statistical significance in providing additional positive, or not negative, tone in reviews to manufacturers that offer more press drives, Motor Trend showed statistically significant regression, which reveals that as the publication attends more press drives for a given manufacturer a
66 higher percentage of reviews with positive tone, a lower percentage of reviews with negative tone, a higher percentage of reviews with not negative tone, as well as a greater number of reviews with net observed positive tone, is observed in the publishing year of 2011. Such a policy may explain why Motor Trend had seven additional reviews that took place in a press drive setting compared to Car and Driver (39 versus 32) The reasons for the relationship between tone and press drives are not fully known, but from what John Phillips (2011) and Frank Greve (2011) wrote, it appears it is in Motor Trends best interest to write positively, or not negatively, about manufacturers who are mor e likely to hold press drives so that the writers can both go where the cars, as well as the executives and engineers who build the cars (Phillips) are, while simultaneously continuing to enjoy the perks of free business class flights and expense paid s tays (Greve) at lavish places around the world. These findings counter Warren Breeds (1955) findings, as it is not the publisher but the writers who seem to be influenced by the media in the setting of a press drive. Press drives are arguably more benefi cial to the writer than the publisher as the individual reviewing the vehicle indulges in the perks associated with these events. As automotive publicity manager Fred Heiler told Frank Greve (2011), t he reason behind such lavish trips is, not to promote good journalism, but simply because reviews have proven more effective and cheaper than ads at motivating customers to buy. By inviting automotive journalists who often live and socialize in the same rulingclass circles as the people they cover to writ e reviews of a manufacturers product in such an environment, it is possible that those reviewing the product are inclined to give
67 [their peers product] the benefit of certain doubts ( Friedersdorf 2013). It is also possible that the settings, which oft en times are grand and exotic, may serve as a distractor to the reviewer, or that the settings are even purposely chosen by the manufacturer to make the most of a vehicles abilities (i.e. a manufacturer is unlikely to choose a setting with a curvy road for a vehicle not designed for handling such conditions). In essence, the events and environment of the press drive may distract the reviewer from giving an evenhanded review of the product. Such a case can be exemplified by the fact that Car and Driver wro te the review of the Nissan Versa in a neutral tone during a press drive, but later wrote in a negative tone when reviewing the car in a road test setting. No matter the reasons, press drives are less likely to be reviewed with a negative tone and more lik ely to contain positive tone. Of the 71 press drives in this study, three were negative (4 percent). Meanwhile of the 225 road tests in this study, 29 were negative (13 percent). Conversely nearly 30 percent of road tests were reviewed with a positive tone (66), while almost 65 percent of press drives (46) were reviewed with a positive tone. These numbers show that, statistically, a review is more likely to skew positive or not negative in tone to vehicles tested in a press drive setting. As a manufacturer offers more press drives, the percentage of reviews with not negative tone is likely to skew in the manufacturers favor. Thus, while the results do show a statistical significance for Motor Trend skewing the tone it gives to manufacturers based off of the number of press drives offered, the reality seems to be that as more press drives are offered by any manufacturer the observed percentage of reviews with positive tone, as
68 well as the number of observed net positive reviews, allows tone to skew in the manufacturers favor. Of course other factors may be at play, as well. It is possible that manufacturers that make superior vehicles are more inclined to invite the press to drive their vehicles in a press drive event; however, it is curious that Car and Dri ver did not show a statistical relationship between its tone in reviews based off of the number press drives offered by a manufacturer. Admittedly it is possible that this difference may be an attempt by one or both publications to increase polarization i n order to alleviatecompetition (Gal or, Geylani & Yildirim, 2010, p. 26). Another factor possibly contributing to Motor Trends relationship of press drives and the percentage of reviews written without negative tone may be due to personal bias on the part of the author. Ignoring personal relationships, the author may have a personal preference to one brand over another, and for reasons that may or may not be intentional this bias results in the author writing about different vehicles he or she reviews in a tone that reflects his or her personal brand preference. Future Study The majority of the hypotheses raised by this study were rejected; however, in the case of Motor Trend, one hypothesis was statistically significant and was thus not rejected as the number of press drives offered by a manufacturer increases, Motor Trend is statistically more likely to review said manufacturers products with more positive and less negative tone in its writing. This study establishes that such a case existed for Mo tor Trend during the publishing year of January 2011 through December 2011; however, future studies should seek to find if such a pattern continues through other publishing years, through other automotive publications and even through other
69 industry review publications that rely upon manufacturer sponsored events to write reviews. Time also limited the independent variables related to advertising dollars in this study. Future study of this subject should seek to expand the role of advertising budgets to a micro level of advertising spending by month or by week for each manufacturer tested against publishing dates for reviews. Such a study may reveal if manufacturers invest more heavily in advertising before or after the release of an award or a positive published review. This study is thus limited in its ability to study the role advertising budgets play preceding or following such a situation and must assume that each manufacturer a relatively equal amount on advertising on a monthto month basis in the ye ar 2011. Another limitation of this study happens to be the different calendar years used by publishers and advertisers (i.e. a January 2011 magazine is usually completed at least one month before the written publication date). In order to accommodate for this limitation, the dependent variables of this study were run against data provided by Advertising Ages 100 Leading National Advertisers, 2011 Edition for the ten manufacturers tested in this study for the independent variables of overall advertising spending, advertising spending in magazines, advertising spending on the Internet, and advertising spending on magazines and the Internet. The data found was statistically insignificant as well. Furthermore, it was deemed that the low number of cases in this study involving each publications award winners in 2011 made making a conclusion on the impact of manufacturer advertising spending on awards received unobtainable. Future studies,
70 however, could and should look into the effects of awards in the automotive and other review journalism fields by looking at the correlation between manufacturer advertising expenditures for multiple award years. Though no conclusion can be made, this study noted that exactly 70 percent of the awards given by Car and Driv er went to manufacturers with advertising expenditures large enough to be listed in the Advertising Ages 100 Leading National Advertisers, 2012 Edition, and that twothirds of the awards given by Motor Trend in 2011 went to vehicles produced by the secondhighest advertiser overall in 2011 according to Advertising Age, General Motors (Johnson, 2012). It is conceivable that due to pride in their products achievements that manufacturers invested more in advertising; however, in the case of General Motors, Advertising Age reports that in 2010 the company was the third largest advertiser with $2,746,300,000 spent and the company received no awards from the Motor Trend and only one from Car and Driver. Such a case also supports that General Motors vehicles, a s well as the vehicles of other manufacturers given awards by both publications, may have simply been superior in 2011 (Johnson, 2012). In any case, future study in this area might put to rest the c onsiderable suspicion that the Car of the Year on the cov er of an auto magazine is the result of an automaker's extensive purchase of advertising space in that magazine (Wilcox & Cameron, 2012, p. 88). An area of study future research can also investigate is the role of tone in rank order comparison tests. For example, in Motor Trends comparison test of compact sedans the Mazda 3, a vehicle not included in this study due to Mazdas small advertising budget, was written in an operationally defined negative tone despite the fact it finished in third place in an eight car comparison test, while vehicles ranked lower
71 in the test were written in an operationally defined neutral tone. Such a study will need to introduce a formula to make up for the fact that different comparison tests use a different number of vehicl es. Future studies can also look into the role such tone has on consumer purchase decisions. Such a study can compare the role of tone in reviews on a specific vehicle model against said vehicle models sales figures. Similarly, this can be carried out on other reviewed products such as phones, tablets or computers. Another aspect that future studies may want to consider is the role of the author. Comparing the tone used by individual authors over various product reviews may reveal if an author has a pers onal tone preference in writing his or her reviews, or if the authors of a given publication generally share a standardized writing tone a finding that may reveal if writers learn by osmosis (Breed, 1955, p. 328) to write in a specific tone as a res ult of a newsroom policy enacted by the publisher. Following in the vein of Golan, Kiousis and McDaniel (2007), future studies may also want to look at the relationship reviews play in advertising of automotive or other reviewed products. Manufacturers of ten quote from media reviews in their advertisements. Future research can look at how manufacturers use these quotes to influence the agenda of their advertisement at both the first and secondlevel, and if using more media endorsements translates into more interest in the vehicle being advertised. Nevertheless, the information found in this study demonstrates that Motor Trend provided manufacturers that offered more manufacturer sponsored press drives a higher percentage, and net number, of positive and not negative reviews in 2011. This
72 study concludes that the relationship between magazine and manufacturer can potentially influence the tone, and influence the agenda of the press at the secondlevel, in automobile reviews if a manufacturer offers more pr ess drives to automotive journalists.
73 APPENDIX CODEBOOK Make Model Publication Award/Road Test/Press Drive (Check one) Positive Neutral Negative N/A Vehicle acceleration and engine performance Transmission performance Ride (comfort) Han dling (performance) Brake performance Steering performance Styling Interior comfort and space Interior ergonomics and materials Seat comfort Cargo space Features and amenities Conclusion Other Overall
74 LI ST OF REFERENCES Basuroy, S., & Rinallo, D. (2009). Does advertising spending influence media coverage of the advertiser?. Journal Of Marketing, 73(6), 33 46. doi:10.1509/jmkg.73.6.33. Retrieved from: http://uh7qf6fd4h.search.serialssolutions.com/?ctx_ver =Z39.882004&ctx_enc=info%3Aofi%2Fenc%3AUTF 8&rfr_id=info:sid/summon.serialssolutions.com&rft_val_fmt=info:ofi/fmt:kev:mtx:jo urnal&rft.genre=article&rft.atitle=Does+advertising+spending+influence+media+c overage+of+the+advertiser%3F&rft.jtitle=Journal+of+Marketing&rft.au=Rinallo%2 C+Diego&rft.au=Basuroy%2C+Suman&rft.date=200911 01&rft.pub=American+Marketing+Association&rft.issn=00222429&rft.volume=73&rft.issue=6&rft.spage=33&rft.externalDBID=n%2Fa&rft.exte rnalDocID=212230673 Breed, W. (1955). Social control in the newsroom: A functional analysis. Social Forces 33(4): 326335. Brown, K., & Cavazos, R. (2005). Why is this show so dumb? Advertising revenue and program content of network television. Review of Industrial Organization, 27(1), 1734. doi: http://dx.doi.org/10.1007/s1115100548366 Retrieved from : http://www.sciencedirect.com/science/article/pii/S0167268109001917 Cammisa, J. (2012). Old vs new Scirocco shootout! head 2 head episode 15. [Video/DVD] YouTube: Motor Trend. Retrieved from : http://www.youtube.com/watch?v=SJtcTgBdLF8 Car and Driver. ( 2012). About us. Car and Driver media kit. Retrieved from: http://www.caranddrivermediakit.com/r5/showkiosk.asp?listing_id=4170499 Carroll, C. E., & McCombs, M. (2003). Agenda setting effects of business news on the public's images and opinions about majo r corporations. Corporate Reputation Review, 6 (1), 36 46. Retrieved from : http://search.proquest.com/docview/231576825?accountid=10920 Champlin, D., & Knoedler, J. (2002). Operating in the public interest or in pursuit of private profits? News in the age of media consolidation. Journal of Economic Issues 36(2), 459468. Retrieved from : http://www.jstor.org/stable/pdfplus/4227797.pdf?acceptTC=true Chiang, C., & Knight, B. (2011). Media bias and influence: Evidence from newspaper endorsements. The Review of Economic Studies, 78(3), 795820. doi: 10.1093/restud/rdq037. Retrieved from : http://restud.oxfordjournals.org/content/78/3/795 DataCenter. (2013). 100 leading national adver tisers. Advertising Age. Retrieved from: http://adage.com/datacenter/datapopup.php?article_id=235491
75 Davis, D. E. (1968, April). Turn your hymnals to 2002. Car and Driver Retrie ved from : http://media.caranddriver.com/files/1968bmw 2002car anddriverdavis 2002.pdf Ellman, M., & Germano, F. (2009), What do the papers sell? A model of advertising and media bias. The Economic Journal 119, 680704. doi: 10.1111/j.14680297.2009.02218.x. Retrieved from : http://onlinelibrary.wiley.com/doi/10.1111/j.1468 0297.2009.02218.x/full Entman, R. M. (2007). Framing bias: Media in the distribution of power. Journal of Communication, 57(1), 163173. doi: 10.1111/j.14602466.2006.00336.x. Retrieved from : http://uh7qf6fd4h.search.serialssolutions.com/?ctx_ver=Z39.882004&ctx_enc=info%3Aofi%2Fenc %3AUTF 8&rfr_id=info:sid/summon.serialssolutions.com&rft_val_fmt=info:ofi/fmt:kev:mtx:jo urnal&rft.genre=article&rft.atitle=Framing+Bias%3A+Media+in+the+Distribution+ of+Power&rft.jtitle=Journal+of+Communication&rft.au=Robert+M+Entman&rft.dat e=200703 01&rft .pub=Blackwell+Publishing+Ltd&rft.issn=00219916&rft.volume=57&rft.issue=1&rft.spage=163&rft.externalDocID=1294143071 Esrock, S., & Leichty, G. (1998). Social responsibility and corporate web pages: Self presentation or agendasetting?. Public Relations Review, 24 (3), 305319. doi: 10.1016/S03638111(99)801428 Friedersdorf, C. (2013). Why does the American media get big stories wrong. The Atlantic. Retrieved from: http://www.theatlantic.com/politics/archive/2013/06/why does the americanmediaget big stories wrong/276454/ Gal or, E., Geylani, T., & Yildirim, T. P. (2010). The impact of advertising on media bias. 1 41. Ret rieved from : http://econ.hunter.cuny.edu/mediaeconomics workshop/conferencepapers/Gal Or%20Geylani%20Yildiri m%20Ads%20and%20Media%20Bias.pdf Gentzkow, M., & Shapiro, J. (2005). Media bias and reputation. National Bureau of Economic Research, 1 43. Retrieved from : http://www.nber.org/papers/w116 64.pdf?new_window=1 Greve, F. (2011). Taking readers for a ride. American Journalism Review. Retrieved from: http://www.ajr.org/Article.asp?id=5141 Golan, G. J., Kiousis, S. K., & McDaniel, M. L. (2007). Second level agenda setting and political adverti sing. Journalism Studies, 8 (3), 432443. doi: 10.1080/14616700701276190. Retrieved from : http://www.tandfonline.com/doi/pdf/10.1080/14616700701276190
76 Harris, C. (2011). How Ferr ari spins. Jalopnik Retrieved from : http://jalopnik.com/5760248/how ferrari spins Investopedia. Definition of Willie Sutton rule. Investopedia. Retrieved from: http://www.investopedia.com/terms/w/willie suttonrule.asp Johnson, B. (2012). 100 leading national advertisers, 2012 edition. Advertising Age. Retrieved from: http://adage.com Linkov, J. (2011). Are Volkswa gens media vehicles the same as what you can buy?. Consumer Reports Retrieved from : http://news.consumerreports.org/cars/2011/11/are volkswagens mediavehicles thesameas what youcanbuy.html Lombard, M., Snyder Duch, J., &Bracken, C. (2002). Content analysis in mass communication assessment and reporting of intercoder reliability. Human Communication Research. Retrieved from: http://atar.mscc.huji.ac.il/~mb/analyzing_CA.pdf Marilyn, R., Wanta, W., & TzongHorng, D. (2002). Agenda setting and issue salience online. Communication Research, 29(4), 452465. doi: 10. 1177/0093650202029004004 McCombs, M. (2005). A look at agenda setting: Past, present and future. Journalism Studies, 6(4), 543557. McCombs, M.E., & Shaw, D.L. (1972). The agendasetting functions of mass media. Public Opinion Quarterly 34(2), 176 187. McCombs, M., & Shaw D. (1993). The evolution of agenda setting research: Twenty five years in the marketplace of ideas. Journal of Communications 58 67. Retrieved from: http://www4.ncsu.edu/~amgutsch/MccombsShawnew.pdf McElroy, J. (2009). Thus spake David E.. Autoline Retrieved from : http://www.autoline.tv/daily/?p=5606 Motor Trend. (2013). Motor Trend 2013 Media Kit. Retrieved from : http://ads.simautomotive.com/CONSUMER/_PDF/_MEDIA_KITS/MTD GenMediaKit.pdf Paukert, C. (2012). Honda warns dealers to clear out 2012 civic stock, 2013 emergency refresh on sale Nov. 29. Autoblog. Retrieved from : http://www.autoblog.com/2012/10/23/honda warns dealers to clear out 2012civic stock2013emergenc/
77 Phillips, J. (2011). Yes, a car manufacturer once sent me money. Car and Driver. Retrieved from: http:// www.caranddriver.com/columns/johnphillips yes a car manufacturer oncesent me money PIB. (2012). January December 2011 v 2010. Publishers Information Bureau. Retrieved from: http://www.magazine.org/insights resources/pib/magazinetitles dataytd/january december 2011vs2010 Reuter, J. (2009). Does advertising bias product reviews? An analysis of wine ratings. Retrieved from: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1480293&http://www.google .com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=0CEwQFjAC&url=http %3A%2F%2Fpapers.ssrn.com%2Fsol3%2FDelivery.cfm%3Fabstractid%3D1480 293&ei=TZmAT_unBIi6tweEtu2iBg&usg=AFQjCNHRFlqvZXY60wG_7yQ Id9DHjt VDBA Reuter, J., & Zitzewitz, E. (2005). Do ads influence editors? Advertising and bias in the financial media. Quarterly Journal of Economics 1 28. Retrieved from: https://www2.bc.edu/jonathanreuter/research/ads.pdf Riffe, D., Lacy, S., & Fico, F. (2005). Analyzing media messages: Using quantitative content analysis in research. Lawrence Erlbaum Associates, Inc. Retrieved from: http://books.google.com/books?id=t6ezVKmiCVcC&pri ntsec=frontcover#v=onepa ge&q&f=false Socalspeedzone. (2010). Jim Wangers and the Car and Driver ringer GTO. YouTube. Retrieved from: http://www.youtube.com/watch?v=Qy6Zn4TIOr4 Scheufele, D. ( 1999). Framing as a theory of media effects. Journal of Communication, 49(1), 103122. doi: 10.1111/j.14602466.1999.tb02784.x Retrieved from : http://onlinelibr ary.wiley.com/doi/10.1111/j.14602466.1999.tb02784.x/abstract Shoemaker, P., & Reese, S. (1996). Mediating the message: Theories of influences on mass media content. Longman Publishers USA. Retrieved from: http://shoemaker.syr.edu/docs/mediatingthemess age2ndedition1996shoemaker reese.pdf Sutter, D. (2002). Advertising and political bias in the media: The market for criticism of the market economy. American Journal of Economics and Sociology, 61(3), 725745. doi: 10.1111/15367150.00187 Tankard, J (2001). The empirical approach to the study of media framing in Framing Public Life edited by Reese, Gandy and Grant (Erlbaum).
78 TexaSoft. (2008). Interrater reliability (kappa) using SPSS. SPSS Tutorials for Statistical Data Analysis. Retrieved from: http://www.stattutorials.com/SPSS/TUTORIAL SPSSInterraterReliability Kappa.htm Weaver, D. H. (2007). Thoughts on agenda setting, framing, and priming. Journal of Communication, 57(1), 142147. doi: 10.1111/j.14602466.2006.00333.x. Retrieved from : http://onlinelibrary.wiley.com/store/10.1111/j.14602466.2006.00333.x/asset/j.14602466.2006.00333.x.pdf?v=1&t=h7mdvs38&s=54e0b1c60b2fbdf72fe9d6a76a086 a5f0d34dd57 Wert, R. (2011). This is the only way Chris Harris is allowed to drive a Ferrari. Jalopnik, Retrieved from : http://jalopnik.com/5856492/this is theonly way chris harris is allowed to test drive a ferrari White, D.M. (1950). The gate keeper: A case study in the selection of news. Journalism Quarterly 27(4), 383390. Wilcox, D., & Cameron, G. (2012). Public relations strategies and tactics. Pearson, Tenth edition, 88. Yioutas, J., & Segvic, I. (2003). Revisiti ng the Clinton/Lewinsky scandal: The convergence of agenda setting and framing. Journalism and Mass Communication Quarterly, 80(3), 567582. Retrieved from : http://search.proquest.com/docview/216929626?accountid=10920; http://uh7qf6fd4h.search.serialssolutions.com/?ctx_ver=Z39.882004&ctx_enc=info:ofi/enc:UTF8&rfr_id=info:sid/ProQ%3Aabiglobal&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.g enre=article&rft.jtitle=Journalism+and+M ass+Communication+Quarterly&rft.atitle =REVISITING+THE+CLINTON%2FLEWINSKY+SCANDAL%3A+THE+CONVER GENCE+OF+AGENDA+SETTING+AND+FRAMING&rft.au=Yioutas%2C+Julie% 3BSegvic%2C+Ivana&rft.aulast=Y ioutas&rft.aufirst=Julie&rft.date=2003 1001&rft.volume=80&rft.issue=3&rft.spage=567&rft.isbn=&rft.btitle=&rft.title=Journa lism+and+Mass+Communication+Quarterly&rft.issn=10776990
79 BIOGRAPHICAL SKETCH Gregory Fink earned his Bachelor of Science degree i n Journalism from the University of Florida in December of 2011. He joined the University of Floridas Master of Arts in Mass Communication program in January of 2012 and received his degree in August of 2013.