. - L- a... 2... H35? 5.. 33 u . VAIv . . ..v : Apt . — .mgnmmkm‘umdj . IWJ,.3..3.3.3.32..\ .3. . . . . . 33 3. Fit» («£43 . £§¢hvlmuw .. , I: E I. :- ‘ .- .- [5 r =-=-=- 3.... .N L y. - r - 3 Maya 1 This is to certify that the dissertation entitled A COMPARISON OF THE EFFECTIVENESS OF BRAND DIFFERENTIATION AND INFORMATION-LEVEL STRATEGIES IN SOUTH KOREAN AND U.S. TELEVISION ADVERTISING presented by Charles Ray Taylor III has been accepted towards fulfillment of the requirements for PhD degree in Marke t in g (/400th [/3 '3,ng Major professor Date August 4, 1992 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State Universlty ~ ‘v-r. a. ‘ PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution A COMPARISON OF THE EFFECTIVENESS OF BRAND DIFFERENTIATION AND INFORMATION-LEVEL STRATEGIES IN SOUTH KOREAN AND U.S. TELEVISION ADVERTISING BY Charles Ray Taylor III A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Logistics 1992 {/7- {£57 ABSTRACT A COMPARISON OF THE EFFECTIVENESS OF BRAND DIFFERENTIATION AND INFORMATION-LEVEL STRATEGIES IN SOUTH KOREAN AND U.S. TELEVISION ADVERTISING BY Charles Ray Taylor III An experiment performed in both the United States and the Republic of Korea compared (a) the effectiveness of television commercials containing high and low information _content; and (b) the effectiveness of television commercials containing brand differentiating messages versus those that do not. In each country, 101 subjects viewed 20 television commercials for low involvement products selected from a content-analyzed database of over 2000 U.S. and South Korean. Several procedures to enhance equivalence of the data were incorporated in the study. The Elaboration Likelihood Model (ELM) of persuasion was used as a basis for hypOtheses pertaining to information level. For hypotheses Pertaining to brand differentiation it was predicted that Commercials with brand differentiating messages would be more favorably received than those without such messages. Results of the study indicate that (l) the U.S. SUbjects responded more favorably to high information level c0mmercials than did the Korean subjects; (2) the Korean SUbjects gave higher effectiveness ratings to commercials with brand differentiating messages than did the U.S. ‘subjects; (3) contrary to the predictions of the ELM, U.S. subjects consistently rated high information commercials higher than low information commercials; (4) U.S. subjects appeared indifferent to the presence of a brand differentiating message; and (5) Korean subjects strongly preferred commercials with brand differentiating messages versus those without them. Taken together, the results demonstrate differences in subject response patterns by country, suggesting that advertising strategies based information level and brand differentiation should not be standardized across the two countries. Specific implications for advertisers and marketers in the two countries and suggestions for further research are provided. Copyright by CHARLES RAY TAYLOR III 1992 To Hae—Kyong ACKNOWLEDGEMENTS There are many people to thank for their support and help during the performance of this research. First, I would like to express my sincere gratitude to my committee members. To Dr. R. Dale Wilson, my committee chairman, I am indebted not only for the high level of quality control he provided at each stage of the research process, but also for suggesting that I consider doing a PhD degree at Michigan State in the first place. MSU has been a very positive experience for me both professionally and personally, and I can sincerely state that I do not think I could have made a better choice. For his influence in my choice of programs and his help while I was here, I will always be grateful to Dr. Wilson. I would like to thank Dr. Gordon B. Miracle for helping me to develop an interest in international marketing and for his consistent generosity toward me throughout my time at MSU. Working with Dr. Miracle on his large database of Korean, Japanese and U.S. television commercials has been central to my professional development. I am also grateful to Dr. Miracle for the long hours we spent working together and for always having the time to meet with me in spite of his very busy schedule. The training I received from Dr. Thomas Page in performing marketing research was invaluable to my successful completion of this project. I am also very appreciative of his prompt, candid, and always astute Comments on earlier versions of this manuscript. I am vi grateful to Dr. Roger Calantone for agreeing to serve on this committee after a change in the policy of the College of Business mandated the addition of a member even though the research was well underway. That he was able to contribute to the success of this project through his knowledge of international research techniques and Korea itself is a tribute to his high level of scholarship and concern for doctoral students. This research would not have been possible without the financial support of the Center for International Business and Education Research and the Richard D. Irwin foundation. I would particularly like to express my gratitude to Dr. S. Tamer Cavusgil at CIBER for his encouragement and support, and also to Doris Scarlett for advising me as to how to successfully carry out the mechanics of the project. The research would also not have been possible without Dr. Miracle's permission to utilize his content analyzed database of commercials. The contribution of other members of Dr. Miracle's research team, especially Kyu Yeol Chang, Hiromi Maenaka, and Hae-Kyong Bang must also be acknowledged. The support of Dr. Robert Nason, Chairman of the Department of Marketing and Logistics, is also gratefully recognized. I must also thank Dr. Stanley C. Hollander for serving as a counselor to me throughout my PhD program. For his vital help with the Korean portion of the vii in 511 project, I am indebted to Dr. Doo Hee Lee of Korea University in Seoul. I must also thank Dr. Cha-Ock Lee of Sungshin Womens' University in Seoul for her aid in subject recruitment. The advice of Dr. Young-Sook Moon of the Lucky-Goldstar Advertising Agency was also valuable. I also appreciate my mother in law's (and family) efforts at making my stay in Seoul during the data collection process a very pleasant one. I must also thank Kyu-Yeol Chang, In-Young Noh and Ju-Hee Lee for their help with both the translation process and the use of Korean language software. Others who have given me much support throughout this process include my parents and family. Also, fellow PhD students, especially members of my prelim study group: Steve Kopp, Jay Laughlin, John Taylor, and Dan Wardlow. Finally, and most importantly, I must thank my wife, Hae-Kyong. Her expertise in advertising and mass media allowed her to provide me with valuable advice and help during the entire project. Her role in moderating the Korean subject groups was essential to the success of this research. More importantly, she has provided me with inspiration and emotional support during the trying times of the research process and has also been there to share many happy moments and times. Sarang hae. viii TABLE OF CONTENTS Page LIST OF EXHIBITS. ........ .. ..... .... ............. xi LIST OF TABLES......................... ..... ..... Xii I. INTRODUCTION....................... ........ ...... 1 Background........ ............................. 1 Overview Of The Experiment..................... 3 Contributions............. ...... ...... ...... ... 5 Advertising In The Republic Of Korea.... ...... . 10 Outline 0f Remainder 0f Dissertation. .......... 14 II. LITERATURE REVIEW.......QOOOOOOOOOOOO ....... .0... 15 Standardization Vs. Specialization..... ....... . 16 Standardization Advocates... ........... ..... l6 Customization Advocates................ ..... 20 Contingency Approaches.............. ...... .. 22 Critique of Prior Standardization Literature........................ ....... .. 24 Information Content In Advertising.... ......... 28 Measurement of Information Content in Prior Research.......................... 30 Prior Research on Informativeness of Magazine Advertising........... ... ...... 33 Prior Research on Informativeness of Television Advertising ................ .. 39 Summary and Discussion of Prior Findings....................... ...... . ..... 44 Brand Differentiating Messages .......... ....... 47 Theoretical Framework For Hypotheses.... ....... 50 Information Level ................. . ...... ... 50 Brand Differentiation......... .............. 56 Hypotheses. ......... ................. .......... 57 Main Effects................................ 58 Interactive Hypotheses ..................... . 63 III. METHODOLOGY.................... ............ . ..... 65 Database Of Commercials ........ ....... ..... .... 65 Data Collection.. ..... ............ .......... 65 Content Analysis........... ...... ........... 66 Operationalization of Test Variables. ......... . 68 Information Content............. ........ .... 68 Brand Differentiating Message. .......... .... 75 ix Research Design..... ........... . ........ . ...... Overview. ....... ..... ..... .................. Selection of Sample Commercials.. ........... Design of the Experiment .............. ...... Demand Artifacts ............. ............ Design of the Survey Instrument ............. Measurement....................... ...... ....... Attitude Toward the Ad....... .......... ..... Attitude Toward the Brand ....... ... ..... .... Purchase Intention................ ......... . Cognitive response.......................... Analysis..................... ..... ............. Reliability........... ...... ...... ...... .... Validity..................... ...... ......... Procedure........... ..... ..... ...... ........ IV. RESULTS AND DISCUSSION..... ...... .... ..... ....... Data Collection.... ............................ Subjects........ .......................... .. Analyses.... ........................... . ....... Demand Artifacts ............................ Prior Purchase and Exposure..... .......... .. Distributions of Dependent Variables... ..... Reliability of Dependent Variable Scales.... Manipulation Checks ................. . ..... .. Hypothesis Tests ...... . ................. . ...... Between Country Hypotheses ....... . ...... .... Within Country Hypotheses ..... . ............. Possible Covariation ....... ...... ........ ... V. CONCLUSION ...................................... Summary of Findings. ......... .................. Information content ..... . ................... Brand Differentiation ................... .... Limitations..... ............................ ... Suggestions For Future Research ................ APPENDIX 1: Description of Pre-tests and Final Survey Instrument... .......... APPENDIX 2: Cultural Differences Between the USA the ROK That May Influence Advertising Strategy ................ .... APPENDIX 3: Supplemental Tables ..................... REFERENCES... ....... . ..... . ................. ........ 105 116 118 123 126 143 166 174 177 177 182 187 191 194 205 218 232 Exhibit 1. 10. 11. 12. 13. LIST OF EXHIBITS Page Dimensions of Advertising that May or May not be Standardized Between Nations ......... 26 Information Cues Identified by Resnik and Stern (1977)............... ............. 31 Classification of High and Low Involvement Products Based on the Foote, Cone and Belding Planning Matrix..... ................ 58 Product Categorization Scheme.... ........... 67 Information Cues Included in Coding Scheme..... ................................. 70 Reliability of Information Cues. ............ 76 Cells in Experimental Design ................ 81 Some Key Studies Measuring A—Ad ............. 89 Some Key Studies Measuring A-Brand .......... 89 Some Key Studies Measuring Purchase Intention ......... . ....... ...... ............ 90 Operational Definitions of Cognitive Response Categories..... ..... . .............. 99 Commercials Utilized in Experiment — USA.... 103 Commercials Utilized in Experiment — ROK.... 104 xi LIST OF TABLES Table Page 1. Average Age of Subjects...... .......... ..... 106 2. Subjects by Major........................... 107 3. Percentage of Subjects Purchasing from Product Category Within the Past Year - united StateSOO.........OOOOOOOOOOOOOOO ..... 110 4. Percentage of Subjects Purchasing from Product Category Within the Past Year - Republic of Korea............. ........ ...... 111 5. Prior Exposure by Commercial - USA .......... 113 6. Prior Exposure by Commercial - ROK.......... 114 7. Reliability of Dependent Variables - USA.... 117 8. Reliability of Dependent Variables — ROK.... 119 9. Correlation of Dependent Variables.... ...... 125 10. ROK vs. USA Effectiveness by Level of Information........................ ......... 128 11. ROK vs. USA Effectiveness by Brand Differentiation...................... ....... 131 12. Number of Cognitive Responses by Country and Information Level........ ....... ... ..... 136 13. Number of Cognitive Responses by Country and Presence of Brand Differentiating Message............. ...... .................. 140 14. United States - Effectiveness by Level of Information......................... ..... 142 15. Republic of Korea - Effectiveness by Level of Information................... ...... ..... 144 16. United States - Effectiveness by Brand Differentiation...... ............ ........... 148 17. Republic of Korea - Effectiveness by 151 Brand Differentiation ..... ...... ............ xii Table 18. 19. 20. 21. 22. 23. Page United States - Cognitive Responses by Information Level................ ...... ..... 153 Republic of Korea - Cognitive Responses by Information Level...... ...... ... ..... ....... 155 United States - Cognitive Responses by Brand Differentiation.......... ........ ..... 157 Republic of Korea - Cognitive Responses by Brand Differentiation ....... . .......... ..... 158 United States - Effectiveness by Level of Information and Brand Differentiation.... 161 Republic of Korea — Effectiveness by Level of Information and Brand Differentiation. ...... ....... ....... . ....... 164 xiii CHAPTER 1 - INTRODUCTION BACKGROUND The topic of advertising effectiveness is one that has long been of interest to marketers and researchers. Research performed early in the 20th century by Walter Dill Scott (1903), Daniel Starch (1923), Percival White (1927), and many others started to build a base of knowledge on what executional devices lead to a consumer's ability to recall print advertising. More recent research by Twedt (1952); Assael, Kofron, and Burg (1967); Holbrook and Lehmann (1980); and many others has expanded this base of knowledge. Additionally, there have been many experiments performed in recent years investigating the effectiveness of various variables under the control of the advertiser for both broadcast and print advertising. Indeed, there has been much research on advertising effects in several areas, including: - The use of music in advertising (e.g., Gorn 1982; Haley, Richardson and Baldwin 1984; Park and Young 1986; Pitt and Abratt 1988; and Kellaris and Cox 1989). - The use of comparative advertising or particular types of comparative strategies (e.g., Shimp and Dyer 1978; Wilson and Muderrisoglu 1979; Belch 1981; Droge and Darmon 1987; Iyer 1988; and Droge 1989). - Number of exposures and strategies used in multiple exposures (e.g., Craig and Sternthal 1986; Machleit and Wilson 1988; Edell and Keller 1989; and Schumann, Petty and Clemens 1990). 1 - Type of appeal or mood created (e.g., Batra and Ray 1986; Holbrook and Batra 1987; Holmes and Crocker 1987; Edell and Burke 1987; and Batra and Stayman 1990). -- Characteristics or portrayal of presenters, including attractiveness, age, and sex (e.g., Petty, Cacioppo, and Schumann 1981; Kahle and Homer 1985; Whipple and Courtney 1985; Sewall and Sarel 1986; and Leigh, Rethans, and Whitney 1987). A few general comments about the collective impact and limitations of effects research should be made. First, while the above studies and many others have helped to improve our knowledge of what makes advertising effective, they have generally been performed in the United States. Clearly, in today's increasingly international marketplace, there is a need to extend effects research to other countries. A second limitation of prior effects research is that few large-scale studies dealing with the effectiveness of various executional techniques in television advertising have been published. Most academic studies on effects in television commercials have used only a few commercials. A notable exception is the Stewart and Furse (1986) study of 1,000 television commercials. Data on the effectiveness of 155 executional techniques were reported, utilizing three separate measures (related recall, key message comprehension, and persuasion). This study pointed out the need to do more in-depth effects research on several independent variables, some of which have not figured prominently in prior studies. In particular, the Stewart 3 and Furse study pointed out the need to devote more attention to do more effects research studying the impact of brand differentiating messages and information content. OVERVIEW OF THE EXPERIMENT The research extends effects research to an international context via the use of an experiment conducted in both the United States and the Republic of Korea. In both countries, approximately 100 subjects viewed twenty television commercials in which two key executional variables have been manipulated. The specific executional variables chosen for this study are: (a) the level of information content utilized in the commercial, and (b) the presence (or lack thereof) of a brand differentiating message. The latter variable was chosen for two primary reasons. First, in spite of being a very well known concept among practitioners (Ries and Trout 1981; Stewart and Furse 1986), academic study of the effectiveness of commercials containing brand differentiating messages has been very limited. Second, in the one large-scale study of effectiveness conducted in the United States which included the brand differentiating message as an independent variable, Stewart and Furse (1986, pps. 85-90) found it to be the single variable most strongly correlated with each of three measures of advertising effectiveness —- recall, message comprehension, and persuasiveness. 4 The level of information was also chosen in part because of the lack of prior effects research in the area. Another reason for the choice of this variable is its wide- ranging implications. For many years, advertising has been criticized by some for trying to persuade consumers to purchase products they do not need rather than attempting to help consumers make informed decisions. In fact, many academic studies (e.g., Resnik and Stern 1977, Aaker 1984) have focused on whether advertising generally tends to be informative. Very few, however, have asked whether more informative advertising tends to be more (or less) effective in general, or under certain circumstances. Since answers to this question have academic, practical, and societal implications, it warrants investigation. The following three dependent measures of effectiveness were used in both countries: (a) attitude toward the ad; (b) attitude toward the brand; and (c) purchase intention. In addition, cognitive response data were collected. Products from the same two product categories were included in the sample of commercials in each country in order to examine whether product category interacts with one of the independent variables to produce an effect on the dependent variables. 5 CONTRIBUTIONS The contributions of this dissertation are: 1) The author knows of no prior published research which has compared the effectiveness of the individual executional variables utilized in this study across two countries. Much controversy exists with regard to whether advertising or certain elements of advertising strategies can be standardized or transferred across countries. Determining whether the techniques investigated are associated with similar levels of effectiveness in the two countries has clear implications for managers considering standardizing advertising messages in the two countries. Additionally, data on the relative effectiveness of commercials containing given levels of information and/or brand differentiating messages should be highly useful to academicians attempting to build theories or identify contingency frameworks which help us to better understand when advertising can (or should) be standardized. 2) The study contributes to international research methodology by utilizing techniques designed to produce measures which are reliable and valid in an international context. It is notable that even many exclusively domestic studies fail to report measures of reliability of the scales utilized. While the lack of reporting of reliability in domestic studies is not excusable, it is probably even more 6 important to stress reliability and validity when designing international studies, where issues of data equivalence are paramount. The database from which the commercials were chosen made careful use of translations and back- translations in order to ensure the conceptual, functional, and linguistic equivalence of the coded variables across the two countries. Thus, we can be sure that when we refer to a commercial containing "high information," or a "brand differentiating message," the same variable is being measured in the United States and Korea. Indeed, the use of such translations in international research should be viewed as a prerequisite to obtaining any significant level of internal validity in an international context. It should also be noted that the measures of information level and brand differentiating message utilized in the database the commercials were drawn from showed high levels of intercoder reliability in both the United States and Korea. The translation and back-translation process will also be utilized for the purposes of choosing the measurement scales for A-brand, A-ad, and purchase intention. In this way, we can be certain we are measuring the same construct in each country. Thus, the internal validity of the measures across countries is enhanced. Commercials were also be chosen in such a way as to ensure that the treatment conditions (e.g., high information) do not systematically vary with some other executional variable (e.g., use of 7 humor, type of setting, type of appeal, use of humor, commercial format, etc.) The lack of systematic variance on other executional variables will also serve to enhance internal validity. The reliability of the chosen measurement scales was also measured and compared in each country, something which has been done infrequently in prior international research. One additional methodological consideration is to include an alternative measure of the informativeness of the advertisement on the survey instrument so that construct validity can be measured in each country and then compared. Measuring construct validity was accomplished by asking subjects to indicate whether they felt a commercial was informative and comparing these results to whether the objective scheme indicated that the commercial was informative. Prior international studies have not attempted to measure construct validity in spite of its importance in determining whether the construct is truly being captured by the measures being utilized. This study should help to demonstrate the importance of including construct validity measures in international research or, perhaps, point out some of the traps researchers can fall into when they do not attempt to measure validity. It should also be noted that the use of actual television commercials enhances the external validity of the results. An additional step to increase external validity I I 8 will be to have subjects view the commercials in as natural a setting as possible. 3) This study should also contribute a better understanding of the role of information in advertising. Numerous studies have measured the level of informativeness of advertising, both in the United States and in other countries, usually utilizing the Resnik and Stern (1977) framework. There have been at least two major problems with prior research on information in advertising that are addressed in this study. First, the Resnik and Stern coding scheme which has been used in most of the content analyses for the purpose of labelling advertising to be generally informative or generally uninformative ignores several important types of information cues that can be included in an advertisement. This difficulty is overcome by using an expanded and improved scheme for measuring information content. A second, and probably more important, problem with prior research on information in advertising has been that the focus, with only a very few exceptions, has been on whether the content of existing advertisements tends to be high or low. While the study of the general level of informativeness of advertising may be useful from a societal perspective, it is not particularly relevant from a managerial standpoint. Virtually no attention has been focused on the question pertaining to information content 9 that is the most directly relevant to managers -- whether there is a positive correlation between the amount of information contained in an advertisement and effectiveness. In this study, the Elaboration Likelihood Model (ELM) of persuasion will be used as a theoretical framework from which to study the impact of information on advertising effectiveness. It may be that under certain conditions, such as low involvement or within certain product categories, informative advertising tends to be ineffective. If this is the case, advertisers would be well advised to utilize commercials with relatively low information content in commercials for low involvement products. Conversely, if research showed a correlation between high information and effectiveness under some circumstance, it would be suggested that advertisers make use of more informative ads. Better knowledge of the correlation between informativeness and effectiveness under various conditions (e.g., high vs. low involvement) may well have strategic implications for advertisers. Thus, managers are very likely to be interested in the outcome of the experiment with regard to information. The Elaboration Likelihood Model will be tested in a new context in this study. One of the central tenets of the ELM is that consumers rely on peripheral cues in persuasive messages under conditions of low involvement. By measuring subject response to commercials with differing information 10 levels, this study will provide evidence as to whether low involvement does, indeed lead to a reliance on peripheral cues under conditions of low involvement. 4) As Stewart and Furse (1986) point out, greater understanding of brand differentiating messages is needed. Though they have long been an important component of the conventional wisdom of U.S. advertising practitioners with regard to what makes an advertisement effective, little academic research on the brand differentiating message has been performed. One exception is Stewart and Furse's (1986) study of 1,000 U.S. television commercials, which found the presenCe of a brand differentiating message to be the single executional variable most highly correlated with advertising effectiveness. This study will expand research on the brand differentiating message to another country, in an initial attempt to see if the Stewart and Furse finding holds up in another culture. Additionally, results will indicate whether the finding of relatively high effectiveness for brand differentiating messages is replicated within the United States. Further, the cognitive response component of the study will serve to investigate whether brand differentiating messages produce more cognitive responses than those that do not. Cognitive response may lead to higher memorability, and, in turn may lead to more effective advertising. 11 W1th these contributions in mind the research uestions investigated in this study are. 1) Are more informative television advertisements associated with higher levels of effectiveness? If so, what specific types of information cues seem to be associated with higher effectiveness? 2) Is the presence of a brand differentiating message in a television advertisement assoc1ated with higher effectiveness7 3) Are the findings for the level of effectiveness of each executional technique similar in the United States and the Republic of Korea? What do the findings suggest to international advertisers considering standardized advertising in the two countries or transferring advertising from one country to the other? What do the f1nd1ngs suggest to those developing theory on when advertising can be standardized? ADVERTISING IN THE REPUBLIC OF KOREA The country chosen for the purpose of comparison 15 the Republic of Korea (South Korea) Examining advertising in the ROK 15 particularly timely for a number of reasons. First the country has become an increasingly 1mportant player in the world economy over the past few decades. Park (1987) describes the country's economic development as being "among the most rapid and sustained in the contemporary world." Indeed, the figures are impressive Per capita income in the ROK rose from US$87 in 1962 to US$2 850 by 1987 (Hoare and Pares 1989), and it continues to rise. From 1962 to 1984, the South Korean economy grew at an annual rate of 8 56, indicative of the country's grow1ng competitiveness in world markets The ROK's competitiveness 12 is likely to play an increasingly important role in the world economy in the foreseeable future. By the year 2000, forecasts call for South Korea to sustain growth and reach a GNP level of 250-275 billion dollars, making it one of the 15 largest economies in the world (Whang 1987). Additionally, it is expected to become one of the world's ten largest exporters by the end of the century. Because of its role as a major player in the world economy, international researchers and marketers should attempt to gain a better understanding of Korean business practices. Since very little research of this type has been performed (Moon and Franke 1987), it is timely. The fact that South Korea possesses a culture which is unique and clearly different from the United States is also advantageous from a research standpoint. If we are to make generalizations that one executional technique is more effective than another in certain circumstances, it is important to examine whether such a finding holds up in cultures other than the United States. To ignore whether domestic research findings apply in international settings in effect denies the fact that there is now a high level economic interdependence among the nations of the world. It should also be noted that the Republic of Korea is located in a region which is becoming increasingly vital in the world's economy. The promising situation in the Pacific Basin in general with Newly Industrialized Countries (NIC's) such a the vc proni: shift Itlan the P is a busir espec volm ROK. Kore and rela true deca ROK advi frm Sam con It adv lin ma] 13 such as Taiwan and Singapore growing faster than the rest of the world coupled with Japan's economic strength and China's promising growth prospects has led to some predictions of a shift in the vortex of the world's economy from the North Atlantic region to the Pacific Rim (Park 1987). Because of the Pacific Rim's important role in the world economy, there is a need for greater understanding of Eastern culture and business practices. A final reason that study of Korean advertising is especially beneficial at the current time is the large volume of trade conducted between firms in the U.S. and the ROK. The U.S. actually has a higher volume of trade with Korea than it does with France, Italy, or Saudi Arabia (Moon and Franke 1987), and the importance of this trade relationship is expected to grow (Anand 1986). It is also true that advertising in Korea has boomed over the past decade and opportunities for U.S. firms to advertise in the ROK have increased. According to Shin (1990), total advertising outlay in the ROK expressed in U.S. dollars rose from $418 million in 1980 to $2.3 billion in 1989. Over the same time period, the number of color television sets in the country jumped from just a few thousand to over 6.5 million. At present, one of the largest difficulties in Korean advertising is a shortage of airtime available due to the limited number of television stations, indicating how the market has boomed. In late 1991, the government approved the 3 Gem supp ROK inve fin adv: ven' int bet of re) Sp. 11' ad 14 the creation of a fourth television network. Yet, growth in demand for commercial airtime has offset this additional supply. It must also be noted that in the last few years, the ROK has opened its advertising agencies to foreign investment (KOBACO 1989), creating opportunities for U.S. firms who advertise in the Korean market as well as U.S. advertising agencies wishing to participate in a joint venture. Thus, as a practical matter, it is in the interests of many U.S. firms and advertising agencies to better understand Korean advertising. OUTLINE OF THE REMAINDER OF THE DISSERTATION Chapter 2 is designed to place the study in the context of current international advertising research. First, a review and analysis of the standardization versus specialization controversy is provided. Next, prior literature and research findings oninformation content in dvertising and brand differentiating messages are reviewed. section on methodology then begins with a discussion of he research design. Data collection procedures, easurement issues are dealt with here. Next, results are iscussed. Finally, conclusions are drawn and the imitations of the research are discussed. rell eac the ele cor ove vii 1m in ea He ad CHAPTER 2 - LITERATURE REVIEW INTRODUCTION There are four primary bodies of literature that are relevant in view of the purpose of the proposed research, each of which will be discussed individually. First, since the research deals with the issue of whether certain elements of advertising strategy can be transferred across countries, the stream of literature dealing with the debate over whether elements of advertising can be standardized will be reviewed. Next, prior literature on each independent variable utilized in the study will be reviewed in order to provide the reader with relevant background in each of these areas, and to emphasize their importance. Hence, prior literature on information content in advertising will be discussed. A review of prior literature on brand differentiating messages will follow. Finally, prior literature on the Elaboration Likelihood Model will be reviewed since it serves as a theoretical basis for the experiment. 15 int: pm me acr iss dix ste id! e1 th 16 STANDARDIZATION VS. SPECIALIZATION To put this research in the context of the international advertising literature, it is necessary to provide a brief review of the controversy surrounding whether various elements of advertising can be standardized across multiple countries. The degree to which advertising can be standardized has been one of the most widely debated issues in marketing. In general, the literature can be divided into those taking three positions on the issue: (a) standardization advocates, (b) customization advocates, and (c) a "middle ground" position which attempts are to identify contingencies under which standardization of some elements of the advertising program is feasible. Each of these positions is now discussed individually. Standardization Advocates Several writers have advised that standardized advertising is beneficial and have even suggested that it is the "wave of the future." These writers agree that standardization strategies should be actively implemented in order to capitalize on narrowing national boundaries. Elinder (1961) is often credited with being among the first to bring the issue to the attention of modern day marketers. In observing advertising in Europe, he predicted that the same basic campaign could be used anywhere in Europe, allowing for substantial savings in both effort and dollar stat fea: ava of “in net acc FaI he' on 09 ca it 17 expenditures. Blinder argued that the time had come in Europe when similarities in various aspects of consumer behavior outweighed "national traits". He reasoned that standardization of advertising for the European market was feasible for this reason and also because of the increasing availability of international media and the greater mobility of European consumers. Further, Elinder suggested that an "international" approach to advertising was occurring naturally over time, and that efforts should be made to accelerate the transition to standardized advertising. Another proponent of standardization strategies was Fatt (1967). He noted that the U.S., itself is a heterogeneous market demographically, ethnically, culturally, and psychographically, and criticized those opposed to standardized advertising for focusing on what cannot be done. Fatt believed that improvements in international communications and transportation had made standardization increasingly possible. Additionally, when the opportunity for standardization did arise, he suggested that it should be exploited, since truly creative and effective promotional campaigns are a scarce commodity. While he acknowledged that there were some barriers to standardized advertising, Fatt predicted that there would, in fact, be a trend toward universal advertising themes due for the following reasons: (1) the lack of ability of political boundaries to circumscribe psychological or elot ml 1am new the the tel sta prc (1! an ge SP CC q: mi 18 emotional attitudes, (2) the worldwide spread of the marketing concept, (3) the increased speed of new product launches worldwide coupled with the growth of simultaneous new product launches in many countries, (4) a realization of the importance and scarcity of good advertising ideas, (5) the need to control costs, and (6) the advent of global television. By 1968, some observers were familiar enough with the standardization vs. specialization debate to begin to list pros and cons of either generalized approach. Buzzell (1968) was one such writer. While he admitted that there are some limitations based on local differences, he generally came out on the side of standardization. The specific benefits he pointed to were cost savings, consistency of dealing with customers, the exploitation of ideas with universal appeal, and improved planning and control. Thus, Buzzell reasoned, there were real potential gains to be made by standardizing various elements of the marketing mix, including advertising. After a period of relative neglect, in a very well written essay, Levitt (1983) revitalized the standardization vs. specialization debate by proclaiming that the dawn of the global village was upon us. Citing technology, communications, transport, and travel as forces working to reduce national boundaries, he maintained that the globalization of markets would lead marketers to use 19 standardization strategies more frequently, especially with regard to brands and products. In support of this assertion, Levitt cited the success of several firms who had successfully used standardization strategies, including Coca-Cola, McDonald's, Revlon, Sony, and Levi-Strauss. Peebles (1989) also advocated standardization strategies, primarily due to the need to create and maintain world brand images. He asserted that many opponents of standardization rely on examples where it has not worked in the past to infer that it can never work. Peebles admitted lthat the cost advantages of standardization had probably been overstated by past standardization advocates. The need for consistent images in an increasingly global market, however, was presented as a factor which should make managers strive to find standardized advertising approaches rather than simply assuming that they cannot be useful based on examples where it has been misapplied. It was stressed that world brand images have_to be very carefully managed, since advertising left to the judgment of individual countries will indeed produce different brand images in each country. In cases such as McDonalds and Rolex, Peebles pointed out how careful management of world brand images can allow firms to capitalize on the fact that human wants (e.g., value, quality, the latest technology at an affordable price, hunger, need to be loved and respected) are very similar if presented in a recognizable experience cont PTO] log. evi who loc Cu: ag. te na 20 context. It can be seen from the above discussion that the proponents of standardized advertising have relied mainly on logical and theoretical arguments rather than empirical evidence. As will be seen, the same can be said of those who believe that all advertising must be customized for local markets. Customization Advocates On the opposite side of the debate are those who argue against standardized advertising strategies in general terms. Emphasizing differences among residents of various nations, these writers view customization to local markets as being necessary in the vast majority of cases. A response to the standardization advocates of the early 1960's was provided by Weissman (1967), who emphasized differences of those living in different nations. He argued that there are deep-rooted dissimilarities among those living in different cultures which lead to the need for different marketing programs. Especially with regard to consumer goods, Weissman stressed that the differences among consumers of different nationalities outweigh the similarities. Ricks, Arpan, and Fu (1974) cited several examples where standardized advertising failed badly. Giving examples of actual blunders in international advertising, II the: bei: att The vi] rh: re] p1 ma ti 21 they stressed the need for advertisers to know the culture being operated in and argued that differences in customs, attitudes, and needs must always be taken into account. They concluded that domestic success of an advertising campaign does not serve as a guarantee that the campaign will succeed in any other market. Thus, an approach in which individuals from multiple cultures (at least representatives from headquarters and the host country) plan, execute, and evaluate advertising strategy for foreign markets was recommended. In a study comparing advertising campaigns with locally tailored Israeli advertisements for the same product Hornik (1980) found evidence supporting the use of specialized strategies in individual countries. Using 184 Israeli housewives, it was found that localized themes were preferred except in the small number of cases in which the advertisement was geared toward an international appeal, a worldwide corporate image, or a common international connotation. Hornik concluded that localization of advertising themes is advisable and that product universality does not necessarily imply global strategies. While limited to two countries, this study was significant in that it was among the first which used empirical evidence while focusing on the standardization vs. specialization debate. Noting that communications practices in Canada, Turkey, and Hit 599 a i adi st: ad 22 and the United Kingdom vary substantially, Kaynak and Mitchell (1981) also hailed the advantages of specialization. Because of these differences they expressed a belief that international marketers should recognize the advantages of tailoring their advertising and communications strategies to local markets. Harris (1984) argued strongly in favor of local adaptation on theoretical grounds. He challenged the four major arguments in favor of standardization: cost savings, increased similarity of markets based on demand characteristics, advantages of world brand images, and exploitation of marketing skills. Harris concluded that the range of products for which standardization can be applied is very limited. Contingency Approaches Many researchers have sought more neutral ground with regard to the standardization issue and have focused either on current practice and what can be learned from it or have attempted to identify contingencies which suggest when standardization is more/less appropriate. Several authors, including Miracle (1968), Dunn (1976), and Peebles, Ryans, and Vernon (1978), have suggested that when investigating an international market, process standardization is advisable. Thus, advertising managers must take a variety of features of a country's environment into account when making 23 advertising decisions. Similarly, several authors have suggested that a variety of contingencies must be considered in determining the degree to which advertising can be standardized (e.g., Quelch and Hoff 1986; Kreutzer 1988: Jain 1989). The level of compatibility of such environmental variables between countries determine the extent to which advertising can be standardized. Although much of the work involving standardization has not been empirical in nature, there have been some studies which have surveyed managers with regard to their use of standardized advertising. The results of these studies have tended to support the use of a contingency approach. Sorenson and Weichmann (1975), for instance, surveyed 37 consumer packaged goods managers of large multi-national corporation subsidiaries in Europe and concluded that while the benefits of standardizing marketing may be circumstantial, it is advisable to engage in process standardization. Boddewyn, Soehl, and Picard (1986) surveyed several European Economic Community firms, finding that while substantial barriers to standardized advertising existed, it was practiced by many managers. Specific variables linked to standardization success were the nature of the product (consumer vs. industrial), the extent of national differences (tastes, regulations, etc.), the level of competition, and economic conditions. 24 Hite and Fraser (1988) replicated the Boddewyn et al. (1986) study using a sample of successful multi-national U.S. firms. They found that only nine percent of the firms standardized for all foreign markets, with the majority relying on a mix of standardization and adaptation. They, too, developed a list of factors which managers felt were important determinants of the success or failure of standardization programs. Critique of Prior Literature on Standardization While theoretical arguments and research surveying managers may be useful in determining current business practices, they are inherently limited. Far too few studies focusing on the standardization debate have focused on consumer response to various marketing mix elements, including advertising, have been performed. This is not to say that there have not been useful cross-cultural studies of consumer behavior whose findings have implications for the debate. For example, Thorelli, Becker and Engledow (1975) compared similar psychographic groups ("information seekers") in the U.S. and Germany and found them to have a number of important differences including rank order of buying preferences, and attitude toward advertising in general. Green, Cunningham, and Cunningham (1978) also found differences among Americans and other groups with regard to what the most important attributes of convenience prc CO} en 25 products were, and Green and Langeard (1975) found several behavioral differences between French and American consumers. While these studies, taken collectively, do have implications for the debate, there have not been nearly enough of them. Additionally, results of such studies must be carefully analyzed before making conclusions regarding the standardization debate. There are simply many elements and sub-elements of the marketing mix which may or may not be standardized. Often, even discussions of the standardization debate do not clearly specify which element of strategy they are dealing with. Exhibit 1 (adapted from Miracle 1988b) is designed to help deal with the problem alluded to above. Many prior studies and discussions have referred to standardization without clearly defining what is being standardized. Exhibit 1 provides a framework which is useful for specifying which aspects of advertising are being studied in a standardization context. There are four primary dimensions of advertising between nations which may or may not be standardized: objectives, message strategies, media strategies, and budgets. Because all of these dimensions, or even subsets of them can be standardized, research directed at investigating the degree of standardization that is feasible should clearly speCify what dimensions are being dealt with. In this study, two message strategies —- the use of brand differentiating messages and the level of 26 EXHIBIT 1 Dimensions Of Advertising That May Or May Not Be Standardized Between Nations A. ADVERTISING OBJECTIVES - Clear communication tasks (many possibilities, e.g., hierarchy of effects) — Defined target audience (e.g., purchasers, users) - Specified periods of time B. ADVERTISING MESSAGE STRATEGIES - Themes - Appeals - Positive vs. negative approach - Much or little information - Image vs. product — Company vs. brand - Comparisons or no comparisons - Spokesperson vs. voice over — Color mix C. ADVERTISING MEDIA STRATEGIES Choice of media vehicles Size of ads or length of commercial Flighting - Pulsing D. ADVERTISING BUDGETS - Heavy, light, or medium weight - By product - By market area - By selected media - For selected creative or message approaches - Match or not match competitors — Contracyclical or not NOTE: Additionally, one may consider standardizing the methods or process by which the above decisions are made. SOURCE: Adapted from Miracle (1988b) 27 information in an advertisement -— are being explored in a standardization context. As indicated earlier, research testing the impact of specific variables related to standardization is needed. There have been a limited number of studies which have examined differences in the objectively measurable content (generally information content) of advertisements in different countries, including Dowling (1980), Madden, et. al (1986) and Hong et. al (1987). These studies have been useful in that they have demonstrated that differences in information content or type of appeal used in advertising in specific countries do exist. The solely descriptive nature of these studies, however, limits the conclusions that can be’drawn from them. It is important to investigate whether specific differences in the use of various executional techniques are due to their being more (or less) effective in different countries. Thus, by comparing the effectiveness of different types of television commercials in two countries, this study takes the next step in investigating whether two important elements of message strategy can be standardized in the United States and the ROK. Clearly, if a certain executional technique is found to be consistently less effective in a certain country over time and in numerous settings, it is very difficult to make a case for transferring commercials utilizing that technique to the country in question. The following sections provide a 28 background on the prior literature on information and brand differentiating messages. It should be kept in mind that research on these variables in the United States itself has not led to widely accepted conclusions. Thus, this study should be viewed as fitting into these research traditions as well as extending certain aspects of them to the international context. INFORMATION CONTENT IN ADVERTISING While there has been much attention paid to the issue of whether or not advertising is generally informative or generally persuasive, little prior research has directly investigated the relationship between informativeness and effectiveness. Thus, much prior research is an output of response to social critics of advertising. Ogilvy (1963), for example, noted that economists in particular were critical of advertising because of a belief that most of it simply attempted to get consumers to switch from one brand to another (e.g., "attempted to get the consumer to switch from Mogg's soap to Bogg's soap"). Ogilvy dismissed these charges primarily because he felt that more informative advertising was also more effective. He stated: "My own clinical experience would suggest that the kind of informative advertising which the dons endorse is more effective in terms of sales results than the 'combative' or 'persuasive' advertising which they condemn. Commercial self— interest and academic virtue march together." K . 29 While the societal debate over information in advertising has merit, it is unfortunate that no academic study to date has tested Ogilvy's hypothesis with regard to informativeness and effectiveness. Clearly, there are still some advertisements which appear to focus more on image than information, suggesting that not all advertisers agree with Ogilvy. It should be noted that most of Ogilvy's experience in advertising was with non-durable products (e.g., perfume, soap, food) -- those which seem to have more advertising which violates his assertion. At any rate, the basic question of whether more informative advertising tends to be more effective requires more research. Gail et al. (1984) recently noted that many critics of advertising continue to argue that advertisements are only useful to consumers if they are informative. While there have been a number of studies which have attempted to measure the general level of informativeness of advertising, they have not served to solve the controversy surrounding information vs. persuasion and advertising. Further, a study by Barksdale and Perrault (1980) indicated that consumers felt that advertisements were becoming less believable. This result suggests a need for a better understanding of why advertisers choose given levels of information and whether they are choosing levels which tend to be effective. It is worth investigating whether consumers find advertising which contains more information 30 to be more believable and, in turn, whether believable advertising tends to be more effective. Measurement of Information Content in Prior Research The pioneering work on measurement of information content in advertising was Resnik and Stern (1977). The authors developed a list of 14 evaluative criteria which could be used to aid the consumer in intelligent decision making. These evaluative criteria are referred to as "information cues," and are used to measure the level of informativeness of an ad. Exhibit 2 shows the information cues identified by the Resnik and Stern scheme. While the authors acknowledged that no coding scheme could ever hope to be a perfect instrument for measuring information content, since different consumers will find different stimuli to be informative, they stressed that their approach had the advantage of affording policy makers a reliable means of discriminating between informative and non- informative commercials. Under the Resnik and Stern definition, an advertisement was deemed informative if it contained at least one information cue. Hunt (1976) and Aaker (1984) elaborated further on the difficulty of devising an objective, accurate measuring , devise for level of informativeness. Hunt stressed that the quantity of information present in an advertisement should not be viewed independently of the quality or usefulness of 31 the measure. He concurred with the warning that the quantity of information in an ad is individual specific. Aaker (1984) adds that due to the gestalt-like nature of EXHIBIT 2 INFORMATION CUES IDENTIFIED BY RESNIK AND STERN (1977) 1. Price or Value 2. Quality 3. Performance 4. Components or Contents 5. Availability 6. Special Offers 7. Taste 8. Nutrition 9. Packaging or Shape 10. Guarantees or Warranties 11. Safety 12. Independent Research 13. Company-Sponsored Research 14. New Ideas advertisements, particularly television commercials, it is difficult to select appropriate cues. He also warns that the usefulness of an information cue depends not only on the audience member receiving it, but also on product category. Because of these limitations, Aaker suggested that consumer perceptions of informativeness, (generated by simply questioning the consumers as to whether they regard an ad to be informative), be measured in addition to objective content measures. He acknowledged that measures of consumer perceptions also have limitations because of their sensitivity to the scaling technique employed, the ambiguity 32 of the word "informative", and questions pertaining to whether all consumers have the capacity to process and/or understand some information cues. While it is acknowledged that there are limitations associated with developing an objective approach to measuring information content in advertising, it remains a worthwhile endeavor. An objective scale not only provides .policy makers with an objective baseline from which to work, it also facilitates comparisons across cultures and over time. Aaker's suggestion that consumer perceptions of informativeness be measured in addition to the objective measures is also a good one. The Resnik and Stern (1977) scheme will not be employed in this study. Past studies employing the Resnik and Stern scheme have assumed that an informative cue is one that, "permits a typical buyer to make an intelligent choice among alternatives." Because of the ambiguity in the word "intelligent" in this definition calls for subjective interpretation, this definition is not well suited to objective measurement of information content. Based largely on the ambiguity of the definition of information, Sepstrup (1985) and Stewart and Furse (1986) have pointed out that the Resnik and Stern framework describes a somewhat arbitrary group of information cues, and omits several important types of information seen fairly commonly in advertisements. The scheme utilized here, (discussed later 33 and shown as Exhibit 5) further refines the expanded schemes of Sepstrup and Stewart and Furse. For example, the inclusion of company information (cue # 30 in Exhibit 5) was based on the observation that it is not uncommon in Korean commercials for information about the company to be presented in an effort to convey relevant information about the company's image or reputation. The specifics of this measurement scheme and definition of information used will be discussed in greater detail in Chapter 3. Primarily due to the contentions of critics, several researchers have attempted to determine whether existing advertising is primarily informative or persuasive. Studies on the level of information present in U.S. advertisements have led to conflicting results. Various researchers have also measured the information content of advertisements in several countries. The results and implications of these studies, broken down by medium (television and magazine) will now be discussed. Prior Research on Informativeness of Magazine Advertising Marquez (1977) spurred much debate with his content analysis of 600 print ads. Specifically, he investigated whether ads were judged to be informative, persuasive, or intimidating based on dictionary definitions of the three terms. His findings were consistent with most prior consumer surveys, indicating that most consumers did not 34 find advertising to be informative. Of the print ads in Marquez's sample, 65.7% were found to have an appeal based primarily on persuasion. There was, however, a striking product category effect. The findings indicated that advertisements for products which were "inexpensive and nontechnical" used primarily persuasive appeals while products which were more expensive and/or technical tended to provide more information. This finding led Marquez to conclude that for product categories in which information is needed, the advertising is usually informative. In contrast to the Marquez finding, Laczniak (1979) found print ads to be generally informative. Using the objective Resnik and Stern (1977) classification scheme (as opposed to the Marquez framework), he found that 92% of ads were informative if "informative" meant they contained at least one information cue (It should be noted that this is the criterion Resnik and Stern applied to television ads in their sample; they, however, found less than half of the commercials to be informative on this basis.) Laczniak concurred with Marquez on one point, however, stating that the study of consumers' need for information on products which have been on the market for a long time is warranted. He explained: ‘ "For example, since almost everyone has tried Coke at one time or another, an ad for Coke with low information might be quite acceptable from a social standpoint." 35 Unfortunately, the question posed by Laczniak -- whether informative ads in certain product categories do not meet consumer needs -- has not been followed up. It would seem that an important step toward evaluating whether advertising in certain product categories which contains low levels of information is socially acceptable would be to determine whether informative advertising in that product category is effective. The possibility of a life cycle effect is also hinted at by Laczniak. It may be that more image-based advertising is shifted to at later life cycle stages because the consumer is already well informed about the product. Thus, ads which contain high levels of information may be perceived as boring, and, hence, may be ineffective. It is also possible that research could find that, in contrast to what is expected, informative ads can be effective in certain product categories and/or at particular life cycle stages. Researching the effectiveness of ads with varying levels of information in different product categories (as will be done in this study) or at different life cycles should lead to a better understanding of consumer response and, perhaps, add insight on advertisers' motives for including certain levels of information in their ads. Such insight should, in turn, help individuals and policy makers to make decisions pertaining to the societal debate on the information content of advertising. Laczniak's study was replicated by Stern, Krugman, and 36 Resnik (1981) using a wider variety magazines and product categories. Eighty—six percent of the advertisements in the total sample were found to be informative, with significant differences based on magazine type and product category. With regard to product category, the furniture/ durables/electronics categories were the most likely to be informative. The personal care, food/alcohol/tobacco, and household/lawn/garden were the product categories least likely to be characterized by informative advertising. The study also controlled for size of the ad, but found no significant differences on this basis. The impact of government regulation on the informativeness of magazine advertising was investigated Healy and Kassarjian (1983). Specifically, they examined the effect of the Federal Trade Commission's ad substantiation program on informativeness. Their findings indicated that those products for which substantiation was required showed stable levels of informativeness in the . 1971-1976 period. For industries not subjected to the substantiation program, the amount of information increased over the same time period. This finding suggests that more regulation does not necessarily lead to more information and implies that the reasons advertisers choose to include a given level of information in their ads warrant more intense investigation. Madden, Caballero, and Matsukubo (1986) performed the 37 first major cross-cultural study on information content in print advertising using the Resnik and Stern framework. The results of their analysis of 1,140 ads indicated that Japanese magazine ads were somewhat more informative than those in the U.S. (85% versus 75%). Like Stern et al. (1983), they found differences in informativeness levels based on both magazine type and product category. In the U.S., the food, personal care, and clothing categories were found to be low in informativeness while the household good, electronics, cars/car products, toys/recreational, and entertainment categories tended to contain more cues. In Japan, the food, electronics, and financial services categories were found to provide the least information, while the household goods, car/car related, and clothing categories provided the most. A particularly interesting finding of the Madden, et al. study was that Japanese advertisements were much more likely to contain a price cue than were their U.S. counterparts. In the Japanese sample, price was the most commonly recorded cue, while in the U.S. sample it rated only fifth. Results similar to those found by Madden et al. were recorded. in another study of Japanese magazine advertising by Hong, Muderrisoglu, and Zinkhan (1987). In their sample of 80 representative print ads from each country, they found the Japanese ads to be more informative than the U.S. ads. They also found that Japanese ads were much more likely to 38 contain a price cue than their American counterparts. The Resnik and Stern framework was extended to the analysis of Korean magazine ads by Moon and Franke (1987). In a content analysis of 573 Korean magazine ads, they found 80% of the ads to contain at least one information cue, with the average ad containing 1.8 cues. The authors noted that ads for international brands contained somewhat lower levels of information than those for domestic (Korean) brands. Like previous studies, they found some product category effects, but no differences in the number of cues based on ad size. Another country whose magazine advertising has been content analyzed using the Resnik and Stern scheme is the People's Republic of China. Rice and Lu (1988) applied the Resnik and Stern framework to 472 Chinese advertisements. They found that 100 percent of the ads contained at least one information one and that the average number of cues per ad was 2.26. They suggested two possible reasons for this surprisingly high total. First, they cited the limited experience of the Chinese with consumer products. This line of reasoning suggests that Chinese consumers may want more information so they can lower their purchase risk for products with which they are largely unfamiliar. The second explanation for the high level of informativeness found in Chinese advertisements suggested by Rice and Lu is the nature of regulation by the Chinese government. They noted 39 that the Chinese government stresses that advertising should educate the consumer and promote market socialism. While Rice and Lu do not directly discuss life cycle issues, it should be noted that virtually all of the "low experience“ consumer products they refer to were clearly in early life cycle stages in China in 1988. Thus, their findings are consistent with the notion that higher levels of informativeness are required at earlier stages of the product life cycle. The findings of a study of advertising in the People's Republic of China, Taiwan, and Hong Kong by Tse, Belk, and Zhou (1989) are also consistent with this hypothesis, since PRC ads, on the whole were found to be the most informative, followed by Taiwan, and then Hong Kong. Hong Kong is likely to have more products in the mature stage of the life cycle than Taiwan, which, in turn, is more likely to have mature product categories than is the PRC. Prior Research on Informativeness of Television Advertising The key finding in the Resnik and Stern (1977) study was that roughly one half (49.2%) of the television commercials analyzed were judged to be informative when an "informative" commercial was defined to contain at least one information cue. This finding created controversy, since it seemed to support critics' View that most advertising is uninformative, given the relatively liberal one cue criterion for judging a commercial to be informative. 4O Resnik and Stern did, however, find some variation in the level of informativeness for commercials in different product categories and those in different dayparts. Commercials containing less information were more common on weekday afternoons, weekend mornings, and in the personal care and food products categories. The first replication of the Resnik and Stern (1977) study in a country other than the U.S. was performed by Dowling (1980). He found 74% of a large sample of Australian commercials to be informative. A product category effeCt similar to what was found by Resnik and Stern in the U.S. was reported. Ads for food and personal care products tended to contain less information. The daypart effects, however appeared to be the opposite of that of the U.S., with Australian commercials which ran in the morning tending to be more informative. Dowling attributed the higher general level of informativeness in Australia to rigid government regulation and self-regulation of advertising in the country. Reid and Rotfeld (1981) focused on the informativeness of weekend morning television ads as a means of determining how much information is present in ads aimed at children. In their own sample of 324 weekend morning ads, they found a somewhat higher percentage (42.3%) of the commercials to be informative than did Resnik and Stern (34.9%) using the same criteria. The figure they arrived at was still less than 41 one half of the ads, however. In a study of Ecuadorian television commercials, Renforth and Raveed (1983) found that 82.4% met the one cue criterion. An interesting finding was the lack of finding of a product category effect. The authors asserted that the explanation for differences in the informativeness levels of advertisements in Ecuador, Australia, and the United States may be related to how long products have been on the market in each country. The Ecuadorian findings indicated that advertisements in the introduction and growth stages of the life cycle were significantly more informative than those ads for products in the mature stage. Since more Ecuadorian and Australian products are in early stages of the life cycle, Renforth and Raveed reasoned that ads in those two countries are more likely to be informative than those in the U.S. Since the regulation of advertising in Ecuador is minimal, they found the life cycle explanation to be more plausible than Dowling's hypothesis that informativeness is driven by regulatory levels. A replication of the Resnik and Stern study was performed seven years later on a sample of 348 U.S. television ads. In this study, Gail et al. (1984) found 54.8% of the ads to be informative, a figure similar to that found in the original study. The authors noted that there is much room for improvement by advertisers with regard to informing the consumer. 42 A content analysis of 472 Taiwanese commercials by Huang and Hou (1987) found 49.4% to contain at least one information cue. Their findings also indicated that, in Taiwan, restrictive regulation seemed to lessen the level of informativeness, especially for products for which it is easier to substantiate a claim. This hypothesis conflicts with the Dowling argument of a direct relationship between the level of regulation and the informativeness of a country's advertising. The relationship between regulatory levels and the degree of informativeness of a nation's advertising was also explored by Martenson (1987). Twenty-two U.S. Clio award winners and 102 Swedish cinema ads (television advertising is not permitted in Sweden) were content analyzed. The results indicated that 64% of the U.S. ads and 55% of the Swedish ads were informative. Martenson, in contrast to Dowling (1980), stated that the higher informativeness of [the U.S. commercials might be due to the fact that Swedish advertising is subject to more government restriction. She also suggested that product category might play a role, since ads for food and beverages comprised a larger share of the Swedish sample. Weinberger and Spotts (1989) used the Resnik and Stern scheme to analyze television ads from the U.S. and the United Kingdom. They found higher levels of informativeness in the U.S. ads than in the original Resnik and Stern 43 (1977), but reported that 34% of the commercials do not contain even a single information cue. Commercials in the United Kingdom were found to be even less informative, with only 55.8% containing at least one cue. The authors noted that in both countries commercials for high involvement products and those products that are purchased via rational motives were more informative than those for low involvement products and products in which emotion plays a role. Using an expanded coding scheme to measure informativeness which included 26 information cues, Stewart and Furse (1986) found U.S. television commercials to contain much higher levels of information than did previous studies. Content analysis of 1,000 U.S. commercials indicated that each commercial provided an average of four distinct categories of information. The figure for new products was even higher, at an average of five information cues. While the authors did not directly discuss why their findings varied so greatly from prior findings, it is clear that the expanded coding scheme used picked up some types of information that had been overlooked in prior content analyses. Additionally, the reliability figures cited by Stewart and Furse would appear to verify the value of their findings. One particularly interesting finding of this study was that high information content was negatively associated with recall and key message comprehension. This finding would seem to justify the use of less information 44 for some product categories, or perhaps for products at later life cycle stages. Further study comparing high and low information ads and their impact on effectiveness is clearly needed before such a conclusion is reached. Summary and Discussion of Prior Findings The above literature is insightful, but leaves much room for further research. First, the literature is split on whether the level of government regulation has an impact on level of informativeness of a country's advertising. Healy and Kassarjian (1983) did not find a significant relationship between the two variables in the United States. Dowling (1980) and Rice and Lu (1988) both suggested that higher regulation leads to higher informativeness, while Renforth and Raveed (1983) and Martenson (1988) suggest just the opposite - that more regulation leads to less information. In sum, the evidence to date seems to provide little support for the inference that more restrictive regulation leads to high levels of information or for the assertion that it leads to lower levels. Thus, it would probably be best to shift the focus of the study to advertising effectiveness, which is likely a key underlying factor leading to general levels of information. One consistent theme of much literature on information in advertising across cultures is that there seems to be less information in advertisements for products at late life 45 cycle stages. All studies that have studied general levels of advertising informativeness have found that advertisements in earlier life cycle stages contain more information cues than ads for products in later stages. No study however, has directly investigated whether this is related to a tendency for high information ads for products at late life cycle stages to be ineffective. Given the existence of theory on consumer information search (e.g., Assael 1984) which suggests that information search goes down over time, (as the product is on the market longer), this would seem to be a plausible, but unproven hypothesis. Another area where prior literature is consistent is in finding that there seems to be a relationship between product category and the level of information in advertising. Further exploration of these specific differences is needed, and research on the effectiveness of advertising of various information levels in individual product categories is an important place to start. Interestingly, there seems to be similarity in product category effects for some countries at similar stages of economic develOpment (e.g., the U.S. and Australia). At the same time, however, there appear to be some differences in informativeness by product category even among some countries at similar stages of economic development, as indicated by the Madden et al. (1986) finding that ads for electronic products in Japan contain a less than average 46 number of information cues. These results suggests a need to focus more on specific product categories to see if they have unique characteristics which lead to a need for more (or less) information in their ads in order to be effective. Involvement, ease of obtaining information on product features, and life cycle stage could all play an important role within certain categories across cultures. Cultural factors and country specific factors may also play a role, but there is clearly a need to shift to research on the effectiveness of advertisements and away from simply describing general levels of informativeness. While the primarily descriptive studies listed above contribute to the knowledge of what is, they cannot offer much insight on the motives advertisers have for including given levels of information in their commercials. The proposed research attempts to begin to look at possible motives by focusing on whether informative commercials in selected product categories tend to be more effective in the U.S. and South Korea than their uninformative counterparts. Results should be of use at both a practical and academic level. In surveying U.S. advertising managers, Perrien, Dussart, and Paul (1985) found that advertisers themselves tend to rate commercials with higher information content to be more effective than those with little or no information. The authors noted that this finding calls into question a 47 perception among some that informative advertisements have creative weaknesses. The idea that those in charge of creating advertising messages find informative messages more effective also raises the question of why advertising in general is not more informative. Clearly as several studies cited above indicate, and simple observation would suggest, not all television advertising is primarily designed to inform. Research which lends insight as to when informative advertising is associated with high (and low) levels of effectiveness is needed. BRAND DIFFERENTIATING MESSAGES Whether referred to as "brand differentiation", "product positioning", or creating a "unique selling proposition," the idea that it is desirable to develop an advertising message which clearly distinguishes a product from its competitors has been known to advertising practitioners for many years. Rosser Reeves (1961) was an early proponent of this strategy. From his experience at Ted Bates and Company in the 1940's, Reeves noticed that the most successful campaigns were those that made a claim about something unique offered by the product. Reeves thus concluded that advertisements should contain a unique selling proposition (USP) based on a benefit the customer will receive when they purchase this product. He stressed that the benefit must be one that is not offered by the 48 competition and that it must be strong enough to pull consumers to the product. Reeves cautioned that, "...slogans, slick phrases, strange pictures, (and) mere headlines..." do not comprise a unique selling proposition, asserting that the product benefit mentioned must be real. Borden (1942), in his classic academic study of the economic effects of advertising, also noticed the benefit of brand differentiation. Though he focused most closely on whether brand differentiation benefits society, it is clear that Borden understood the benefits of offering a unique product. Further, he pointed out that effective advertising should communicate the way in which the product is different from others. David Ogilvy (1963) shared the belief that memorable and effective advertising should clearly distinguish a product from its competitors. Ogilvy strongly recommended promising the consumer some tangible or intangible product benefit not offered by competitors. He stressed that factual information should be provided, stating: "The consumer isn't a moron; she is your wife. You insult her intelligence if you assume that a mere slogan and a few vapid adjectives will persuade her to buy anything. She wants all the information you can give her." . In this way, Ogilvy believed a unique brand image could be built and converted into sales. More recently, Ries and Trout (1972, 1981) have 49 popularized the idea of using advertising to position a product relative to its competitor. Citing advertising clutter and the finite ability of humans to take information, they focus on the importance of creating a memorable message. Thus, in order to be effective, advertising should strive to allow a product to occupy a unique position in the consumer's mind. Clearly, some type of message that differentiates the brand from its competitors is required to complete this task. In spite of the popular support for brand differentiating strategies, few academic studies have attempted to quantify their impact on the effectiveness of advertising. One study that did examine brand differentiating messages' impact on effectiveness was Stewart and Furse (1986), who found them to be, "...by far, the single most important executional factor for explaining both recall and persuasion of a product." Stewart and Koslow (1989) replicated this finding on a new a sample of commercials, again finding a positive relationship between the brand differentiating message and measures of effectiveness. Though no single variable was found to be highly correlated with effectiveness in these studies, the fact that the brand differentiating message placed first among the over 150 executional variables coded in these studies suggests the need for further study and understanding of the brand differentiating message. 50 The proposed research will attempt to extend knowledge related to the brand differentiating message. One means of doing so is to examine whether the Stewart and Furse finding of relatively high effectiveness for commercials containing . a brand differentiating message holds up when different dependent measures or measurement scales are utilized. Investigating whether brand differentiating messages are associated with similar levels of effectiveness in the United States and South Korea will also expand knowledge on the topic. If brand differentiating messages are, indeed, related positively to effectiveness, it is important to investigate whether the phenomenon is uniquely to the U.S. or if it holds up across cultures. In either case, there are important implications for global advertisers. THEORETICAL FRAMEWORK FOR HYPOTHESES Information Level The hypotheses pertaining to the impact of information level on effectiveness will be based on the Elaboration Likelihood Model of persuasive communication (Petty and Cacioppo 1979, 1981, 1986). The Elaboration Likelihood Model posits that there are two basic processing routes any persuasive communication takes after being received - the "central route" and the "peripheral route." Individuals process via the central route when both their motivation and ability to process the stimulus are high. As Schumann, 51 Petty and Clemons (1986) state, "Under these conditions, people are expected to process the merits of the product as presented in the ad to make an informed purchase decision." In contrast, persuasion via the peripheral route is believed to occur when the receiver is either not motivated or unable to process the message. Under these circumstances, receivers are thought to rely on cues less related to the objective features of the product. Peripheral cues tested in prior studies include characteristics of the source (e.g., attractiveness, celebrity endorser, source credibility), the mere number of arguments, and cosmetic variations in print ads (e.g., layout, order, print type, wording of arguments). Several studies have provided support for the central tenets of the ELM model. Petty et al. (1983), Moore et al. (1986) and Kardes (1988) all found support for the idea that greater involvement leads to an increased likelihood of processing via the central route and that low involvement increases the probability of peripheral route processing. The relationship between involvement and processing route found in these studies supports the hypothesis of elaboration likelihood being positively related to motivation to process the stimuli. Additional support for the ELM has been found in studies of the impact of ability to process stimuli on elaboration likelihood. Findings have been related to 52 distraction, the presence of a quantitative message, and the use of a comparison. Petty et al. (1979) found subjects to be more likely to process via the central route under conditions of low distraction. Yalch and Elmore-Yalch (1984) found that the presence of quantitative messages reduce ability to process, and, hence, tend to lower elaboration likelihood. Finally, Droge (1989) found that the use of comparative ads enhances the ability to focus on message based features of an ad, leading to higher elaboration likelihood. In each case, greater ability to process the message was associated with higher elaboration likelihood. As Bettman (1986) notes, when both the motivation and ability conditions are met, the Elaboration Likelihood model predicts that certain types of cognitive response are likely. Specifically, processing via the central route will be more likely to elicit counterarguments and support arguments. In contrast, low elaboration likelihood should lead to cognitive responses such as source supports or source derogations. Clearly, the ELM has implications for the type of cognitive responses that should be expected in conditions of high and low motivation and/or ability. As shown in Exhibit 5, the types of cues which constitute information via the measurement scheme to be utilized are generally associated with either some characteristics of the product, or some benefit of 53 purchasing the product (whether tangible or intangible). Peripheral cues, such as attractiveness of source, presence of a celebrity endorser, pleasant commercial tone, use of certain colors or other cosmetic variations do not constitute information based on the operational definition of the construct employed in this study. This assumption is consistent with the procedure followed for classification of information cues in the database to be employed in this study. When content analyzing commercials, coders were instructed to count the presence of an information cue only if an advertiser clearly intended for the cue to be an informative component of the ad. Thus, level of information should be considered a construct which is clearly distinct from the number of arguments (another peripheral cue mentioned in prior ELM research). In other words, the information cues included in this study were designed to measure relevant product characteristics or images, and they generally do not reflect components of an ad that would serve as peripheral cues to those not motivated or able to process the stimuli in a commercial. The Elaboration Likelihood Model is not the only conceptual framework from which hypotheses pertaining to the level of effectiveness of advertising containing different levels of information can be formulated. It is, however, not inconsistent with any of the other available theoretical 54 bases. A classic hierarchy of effects model (Lavidge and Steiner 1961), for example, would predict a greater need to include information when the message receiver does not yet have high levels of experience with or knowledge of the product. It seems likely that under circumstances where a potential consumer has low familiarity with a product, motivation to process will be enhanced, leading to a relatively high level of motivation to process the stimulus. While the ELM focuses more specifically on motivation and ability to process, it is likely that, in most instances, its predictions would not be inconsistent with those of the hierarchy of effects model. Similar to the hierarchy of effects model, the product life cycle concept suggests that advertisements for products for which consumer experience is high (i.e., have been on the market a long time and are in the mature or decline stage of the life cycle) require less focus on information directly relevant to the product. Several writers have explicitly recommended including more information in commercials at early life cycle stages and then shifting to less information in later stages (e.g., Smallwood 1973; Sandage and Fryburger 1975; McCarthy and Perrault 1984; Kotler 1984; Rothschild 1987). Again, it is likely that products at late life cycle stages, at least in most instance, are those which consumers have much experience. It is also likely that experience with a product reduces 55 motivation to process, and thus, consistent with the ELM shift the focus of an effective communication away from objective product attributes and toward peripheral cues. Bettman (1986) and others (e.g., Kiel and Layton 1981; Punj and Staelin 1983; Reilly and Conover 1983; Brucks 1985) have investigated the effect of prior experience on information search in greater depth than the PLC framework provides. Findings from this research indicate that, generally, when prior knowledge is high, the motivation to search for information is low. Thus, products that have been on the market for long periods of time are likely to be associated with lower levels of involvement, suggesting a need to present less information in communications. Again, the logic is consistent with ELM in that low involvement leads to the peripheral route of persuasion where product related information is likely to be less important than under central route processing. Since it is not directly contradictory to these other theoretical bases, and provides more specifics regarding the processing of information as they relate to specific executions, the Elaboration Likelihood model of persuasive communication is judged to be the most appropriate theoretical base to draw hypotheses from. 56 Brand Differentiation Hypotheses for the effectiveness of commercials with brand differentiating messages will be based on the conventional wisdom that brand differentiation is an important component of all effective advertising. This notion is consistent with the thoughts of Borden, Reeves, Ogilvy, and Ries and Trout as well as the findings of Stewart and Furse (1986) and Stewart and Koslow (1989), in which the brand differentiating message was the single executional variable most positively correlated with effectiveness across product categories. While it may be possible to make a case for brand differentiating messages usually being tied closely enough to product features to enhance the probability of processing through the central route, prior operational definitions of the brand differentiating message and the one used in this study are not limited to objective product attributes. Thus, a reference to Coca Cola as "the real thing" may differentiate the brand without really focusing on product attributes. For this reason, it is not felt that the ELM in its current form is a suitable framework from which to draw hypotheses. Clearly, however, a greater understanding of the brand differentiating message and its role in effectiveness is needed. This study aims to provide information in a cross- cultural setting that enhances understanding of brand differentiating in advertising. 57 HYPOTHESES Main Effects a) Cross-cultural hypotheses Hypotheses related to differences in the effectiveness of the independent executional variables will not be directional in nature. Since no prior study has examined the effectiveness of commercials containing brand differentiating messages and/or high and low levels of information in Korea, hypotheses will predict no differences in mean levels of effectiveness in the two countries for the independent variables. If differences are found, they will be highlighted and attempts will be made to explain them. A central assumption in this paper is that the product categories included among the sample commercials (food/beverages and health and beauty aids) are low involvement in nature and are product categories the consumer is highly familiar with. For this reason, it is assumed that motivation to process the commercials (persuasive communications) is low, and ocCurs via the peripheral processing route. The classification of all products in the sample as low involvement will be consistent with the framework used by Foote, Cone and Belding as described by Weinberger and Spotts (1989). Exhibit 3 shows this classification. No product included in the sample will come from a product category classified as high involvement based on the FCB matrix. 58 EXHIBIT 3 Classification Of High And Low Involvement Products Based On The Foote, Cone, And Belding Planning Matrix A. High Involvement Product Categories — Cars, houses, home furnishings, appliances, cameras, insurance, traveler's checks, auto tire and batteries, securities, electric ranges, jewelry, fashion apparel, cosmetics, perfume, motorcycles, hair coloring, sunglasses, sports vehicles. B. Low Involvement Product Categories - Food, OTC drugs, household products, feminine hygiene products, coffee, tea, breakfast drinks, yogurt, motor oil, laundry products, personal care products, credit cards, banks, fast food, desserts, beer, Soft drinks, cigarettes, liquor, candy, long distance phone calling. SOURCE: Adapted from Weinberger and Spotts (1989) .— Since no published research suggests a systematic ariation in advertising effectiveness between U.S. and orean commercials based on level of information, the ollowing hypothesis is put forward: H1: Commercials containing similar levels of information in the U.S. and the ROK are associated with similar levels of effectiveness. 3his hypothesis implies that commercials containing high .nformation levels will be found to have similar mean levels 59 of effectiveness for each dependent measure utilized in the proposed experiment. Similarly, it is hypothesized that commercials containing low levels of information will be associated with similar levels of effectiveness, regardless of country, as indicated by each dependent measure. This hypothesis will serve to test whether subjects in each country tend to prefer commercials with similar levels of information. The Elaboration Likelihood Model would predict that low levels of information are associated with higher effectiveness in each country. This prediction will be tested within each country in hypothesis three. The hypothesis for the cross-cultural comparison of brand differentiating message effectiveness will also assume that it is not impossible to standardize this variable across the two countries. It should be noted, however, that rand differentiating messages were used much less requently in the ROK than they were in the United States, ased on the content analysis of the full database of ommercials. Thus: H2: The mean level of effectiveness for (a) commercials containing brand differentiating messages, and (b) commercials that do not contain brand differentiating messages, will not vary between the U.S. and the ROK. Hypotheses for cognitive response are also put forward. gain, it is assumed that cognitive responses do not utomatically vary between the two countries. The cognitive esPonse hypotheses are: 60 H3: The number of cognitive responses to commercials with a given level of information or brand differentiation (present or not present) does not vary between the U.S. and the ROK. H4: The frequency of individual types of cognitive responses elicited to commercials with given levels of the independent variables do not vary by country. b) Within Country Hypotheses In addition to the above hypotheses, others will be tested within the data set from each individual country. The first hypothesis serves to test the Elaboration Likelihood Model's projections in each country. The ELM posits that commercials for high involvement products are likely to be processed via the central route of persuasion (e.g. Petty et al 1983, Moore et al 1986, Kardes 1988). Conversely, commercials processed under conditions of low involvement are believed to be processed via the peripheral route. Under peripheral route processing, it is less likely that product related information will serve as an effective stimulus than is the case for central route processing (Petty and Cacioppo 1979, 1981, 1986; Schumann, Petty and Clemons 1986). Thus, we would expect effective commercials for low involvement products to contain lower levels of information than would be the case for high involvement products. Since only low involvement commercials are included in the sample, the following hypothesis can be put 61 Irward: H5: Commercials with low levels of information are associated with higher levels of effectiveness than are high information commercials in the USA. H6: Commercials with low levels of information are associated with higher levels of effectiveness than are high information commercials in the ROK. Within country hypotheses with regard to the brand differentiating message will assume that conventional U.S. wisdom (Borden 1942; Reeves 1961; Ogilvy 1963; Ries and Trout 1981; Stewart and Furse 1986) holds in both countries. Thus: H7: In the USA, commercials containing brand differentiating messages exhibit higher mean levels of effectiveness than those that do not. H8: In the ROK, commercials containing brand differentiating messages exhibit higher mean levels of effectiveness than those that do not. Within each country, there will also be hypotheses related to the specific types of cognitive response generated by high and low information commercials. As indicated by Bettman (1986), if the central route to persuasion is taken, counterarguments and support arguments are likely cognitive responses. Since there is more reliance on information and product characteristics under high involvement conditions, it makes sense to expect more arguments related to the product itself rather that ‘ responses related to the execution of the advertisement. 62 aversely, if the peripheral route of processing is taken, is likely that a higher number of source support and urea derogation responses will be elicited and that mmercials with lower levels of information will tend to 'oduce more source related responses. Thus,he following rpotheses are put forward: H9a: Commercials containing low levels of information in the United States are more likely to elicit source support and source derogation than are commercials containing high levels of information. H9b: In the USA, commercials containing high levels of information are likely to produce more support arguments and counterarguments than do commercials containing low levels of information. H10a: Commercials containing low levels of information in the ROK are more likely to elicit source support and source derogation than are commercials containing high levels of information. Hlob: In the ROK, commercials containing high levels of information are likely to produce more support arguments and counterarguments than do commercials containing low levels of information. These hypotheses are broken out separately between informative and uninformative commercials in each country, ‘so as to be able to test those commercials (containing high information) that do not conform to the pattern that would 1 e prescribed by ELM thinking. It is important to 'nvestigate whether even those commercials with the V uggested emphasis on the peripheral cues recommended for ow involvement products still induce source based rguments. 63 Additional hypotheses will relate the presence of a and differentiating message to c0gnitive response, in an .tempt to explore the underlying process via which brand .fferentiation seems to lead to effective advertising, at east in the United States. It is thought that brand ifferentiating messages are likely to make a commercials and perhaps a brand name) more memorable. This may lead to [reater effectiveness. It may be the case that the Increased recall is based on a higher level of cognitive response to commercials with brand differentiating messages. Thus: H11: U.S. commercials containing brand differentiating messages exhibit higher levels of cognitive response than those that do not. H12: Korean commercials containing brand differentiating messages exhibit higher levels of cognitive response than those that do not. Interactive Hypotheses Under conditions of low involvement, it has been .posited that commercials containing low levels of information are more effective than those containing high information levels. It has also been hypothesized that ,commercials which contain brand differentiating messages are ore effective than those that do not contain such a essage. Thus, commercials containing low information and a rand differentiating message should be the most effective 64 grouping. Hence: H13a: Commercials which contain a brand differentiating message and low levels of information are more effective than any other combination of the two independent variables in the USA. H13b: Commercials which contain high levels of information and no brand differentiating message are less effective than any other combination of the two independent variables in the USA. H14a: Commercials which contain a brand differentiating message and low levels of information are more effective than any other combination of the two independent variables in the ROK. H14b: Commercials which contain high levels of information and no brand differentiating message are less effective than any other combination of the two independent variables in the ROK. It should be noted that these interactive hypotheses cannot be regarded as a test of the Elaboration Likelihood Model, since ELM does not make assertions about the role of a brand differentiating message. For this reason, interactive hypotheses pertaining to the level and types of cognitive response will not be put forward. With these hypotheses stated, the discussion now turns to the methodology of the study. CHAPTER 3 - METHODOLOGY DATABASE OF COMMERCIALS The commercials shown to the subjects in the U.S. and the ROK were be drawn from a database of 2,095 commercials developed at Michigan State University under the supervision of Professor Gordon E. Miracle. Some background on this database is now provided. Data Collection Television commercials were taped off-the air in the U.S. and the Republic of Korea in the Fall, 1988 and Winter 1989. Carefully designed specifications were followed so that the sample collected in each country would be: (a) representative of television advertising in each country; (b) proportionate across nationally advertised product categories; and (C) comparable across the two countries Thus, taping was done during all major dayparts, covering Prime time, daytime, and fringe times. Additionally, taping Periods were selected to avoid holidays or other special events that might contain advertising which is not 65 66 representative of that run on a typical day. The taping was done on consecutive days until a sufficient number of commercials were obtained. Each sample excluded local advertising, political advocacy, government, and generic demand advertisements. Commercials of all lengths were included. Upon completion of the project, 1,228 U.S. and 876 Korean commercials were available for analysis. Content Analysis Once the commercials were taped, they were content analyzed on a wide variety of executional variables. Variables in the content analysis included: - Amount and type of information included - Commercial length - Use of visual devices - Use of auditory devices - Use of music and/or dancing - Type of promise, appeal or selling proposition - Commercial tone or atmosphere (e.g., warm, humorous) - Commercial structure (e.g., front end impact) - Commercial format (slice of life, demonstration, vignette, etc.) - Commercial ending - Positive vs. negative approach - Comparisons and type of comparison - Number and types of characters - Relationships among characters - Presence of brand differentiating message - Type of presentation (e.g., celebrity) Commercials were also classified by product category. The reakdown each sample by product category is shown in xhibit 4. In order to content analyze the samples in the U.S. and he ROK on a comparable basis, it was necessary to undergo a 67 EXHIBIT 4 Product Categorization Scheme U.S. ROK Type of Product Number Percent Number Percent Food/beverages 332 27.0 258 29.8 Alcoholic beverages 26 2.1 5 .6 Tobacco products 1 .1 1 .1 OTC drugs 123 10.0 114 13.2 Automobile/bike 85 6.9 9 1.0 Lawn & garden equipt. 0 0 2 .2 Auto-related supplies 25 2.0 10 1.2 Personal care 150 12.2 85 9.8 Detergents/cleaners 66 5.4 30 3.5 Household supplies 19 1.5 25 2.9 Clothing/shoes 24 2.0 99 11.4 Textiles/fabrics 1 .1 3 .3 Furniture 9 .7 27 3.1 Ceramics/glassware 0 0 4 .5 Electronic appliances 9 .7 16 1.8 Electronic prod. supplies 7 .6 24 2.8 Entertainment supplies 21 1.7 1 .1 Cameras/related equipt 6 .5 8 .9 Computers/related 2 .2 4 .4 Clocks, watches 2 .2 13 1.5 Communication equipt. 4 .3 3 .3 Toys, games 24 2.0 12 2.8 Sporting goods 9 .7 6 .7 Pets & related 21 1.7 0 0 Publications 18 1.5 16 1.8 Stationery 2 .2 13 1.5 Musical Instruments 0 0 4 .5 Building materials 3 .2 6 .7 Hotel, motel 6 .5 O 0 Restaurants, bars 42 3.5 2 .2 Movies, theater 13 1.1 10 1.2 Department stores 5 .4 13 1.5 Supermarkets 18 1.5 0 0 Other retailers 27 2.2 1 .1 Banking, finance 32 2-5 13 1-5 Transportation service 28 2.3 5 .6 Telecomm. service 19 1.5 l .1 School, education 1 .1 2 .2 Hospital, medical 2 .2 0 0 Other services 18 .5 5 .6 Other 28 2.3 15 1.7 68 translation and back-translation process. The data coding form and codebook of definitions of terms used for the type of information content and executions were first developed in English. They were then translated into Korean by one individual and back to English by another to ensure the coders in each country were capturing the same phenomenon. Once the final coding instrument was developed, native speakers of each language were carefully trained to perform the content analyses. Coding in each country was performed by college students with strong advertising backgrounds. The coders were instructed to watch each commercial as many times as was necessary to make a judgment about the presence of the variable being coded. The reliability of the coder judgments was measured on a subsample of the data, and was generally high in each country. Operationalization of Test variables Information content A major issue in the measurement of information is how it should be defined in theoretical terms. The Resnik and Stern (1977) study defined an information one as one that permits a buyer to, "make an intelligent choice among alternatives." Maenaka, Miracle and Chang (1991) criticized the Resnik and Stern definition on the grounds that the term "intelligent" is ambiguous. Another criticism of Resnikand Stern's scheme has been that it is an arbitrarily selected 69 group of several common types of information cues rather than an exhaustive set of the possible types of information that can be included in an advertisement (Sepstrup 1985; Stewart and Furse 1986). Both of the above criticisms of the Resnik and Stern system of measuring information suggest 'the need for an improved theoretical definition of information so that it can be operationalized, and measured, more efficiently. Maenaka, Miracle, and Chang (1991) reviewed prior theory on the role of information in persuasive communication. Drawing on the work of Shannon and Weaver (1949), Krippendorff (1975), and Ackoff (1958), they defined an information cue in a television commercial as a message by which the advertiser intends: (a) to reduce the listener‘s and/or viewer's uncertainty about the product being advertised (b) to provide the alternatives of choice, or (c) to effect a change in the mental state or behavior of the target audience about purchasing and using the product. Based on the above theoretical definition of information, the measurement of information content was measured via the use of a coding scheme containing 30 types of information that can be communicated in an advertisement 0 achieve one or more of the above purposes. The coding cheme and codebook definitions of each information cue are hown in Exhibit 5. In developing this scheme, refinements 7O EXHIBIT 5 Information Cues Included In Coding scheme 1) Price - Refers to the amount the consumer must pay for the product or service; this may be in absolute terms, like a suggested retail price, or relative terms, like a 10 percent off sale. 2) Variety of the product - Refers to claiming for or featuring more than one type of product. 3) Value - Refers to some combination of price and quality or quantity, as in more for the money, better quality at a low price, the best value for the dollar. 4) Quality - Refers to how good the product or service is; may refer to craftsmanship and/or attention during manufacture, use of quality (better, best) ingredients or components, length of time to produce or to create the product. May refer to or use symbol of quality standard, such as Good Housekeeping Seal or KS or Q in Korea. 5) Size - Refers to the physical size or capacity of the product, how long, tall, wide, heavy, capacity to do particular size tasks. 6) EconomyZSavings - Refers to saving money or time either in the original purchase or in the use of the product relative to other products in the category. 7) Su 1 antit available or limitation - Refers to how much or how many items are available and directly or indirectly the need to act before the supply is exhausted. 8) Method of payment - Information on preferred method to pay, for example by credit card over the telephone. 9) DependabilitngeliabilityzDurability - Information concerning how long the product will last without repair, service records and so on. 10) NutritiongHealth — Information concerning the nutritional or health-related characteristics of a product-- for example, "fortified with vitamin D," "the formula doctors recommend," "relieves iron— poor blood". 71 EXHIBIT 5 (Cont.) 11) Taste - Primarily for food, drink, or personal care products. 12) Sensory Information (other than taste) - (fragrance, touch, comfort, styling, sound): Information concerning a sensory experience, appearance, classic beauty, beautiful sound and so on of the product either when purchased or when prepared in final form; "smells April fresh," feels silky smooth," "luxurious comfort." 13) ComponentsLContentszIngredients - What went into the making or manufacture of the product-- for example, "contains lanolin," "made with pudding." These contents should be in the products purchased, not ingredients added by the consumer in preparation for use. Can also refer to several items served in a restaurant meal, or components of other services such as features in a life insurance policy. 14) Availability - Any information concerning the place(s) the consumer may purchase or otherwise obtain the product-for example, "available in supermarkets," "look for it in the dairy section." May also refer to places where the product is not available--for example, "not available in all areas." 15) Packaging or Shape - Information about the packaging of the product--for example, "look for the package with the red spoon," "look for the special two in one package," "the package is reusable," "in one convenient serving package." If no package, information about the shape, ease of handling, functions related to shape such as ease of storage, protected from damage. 16) GuaranteesZWarranty - Refers to any information concerning the presence of a guarantee or warranty, including but not restricted to money back offers, offers to repair or service the product in the event of problems, or offers to replace the product if the consumer or is dissatisfied or has a problem. 17) Safety - Information concerning the safety of the product-- for example, "has a built-in cut-off switch," "won't harm delicate hair," "nontoxic". 72 EXHIBIT 5 (Cont.) 18) Independent research results - Information offered about tests of the product or of product users that were carried out by an identified individual or organization other than the company manufacturing or distributing the product, such as Underwriter's Laboratory, a leading university, or the U.S. government. Such testing may concern objective product characteristics ("lasts twice as long") or may be related to user preferences ("preferred by two-thirds of the people surveyed"). 19) Company research results - Information about tests of the product or users of the product that were carried out by the company manufacturing or distributing the product--for example, the Pepsi challenge. 20) Research from unidentified source - Information about tests of the product or users of the product when the source of the test results is not identified. 21) New ideasl new uses - Refers to any information about a new way to use an established product-- for example, "use X brand paper cups for sorting and storing nuts and bolts," “use Y baking soda to deodorize refrigerator." 22) Performance, results of using (either tangible or intangible) - Any information concerning the outcomes associated with the use of a product. These outcomes may be in a positive form-- “gives hair bounce," “gives a smoother ride," "makes you feel healthier,"--or a negative form--"won't yellow floors." Performance deals with whether the product accomplishes a consumer purpose, e.g. clean dishes. Includes side benefits such as softer hands after using a detergent. 23) User's satisfactionlloyalty - Refers to any information concerning users' satisfaction, dedication, preference for the brand, or length of time consumer has used the advertised product-- for example, "I'd never give up my Tide," "I've always used...." 73 EXHIBIT 5 (Cont.) 24) Superiority claim - Information that claims the advertised product is better than competitive products or better than an older version of the advertised product in some particular ways. 25) Convenience in Use - Information concerning the ease which the product may be obtained, prepared, used, or disposed of. ‘26) Special offer or event - Information concerning special events such as sales, contests, two-for-one deals, premiums, or rebates to occur for a specified period of time. 27) New product or new and improved features - Refers to any information concerning a new product introduction, new components, ingredients, features, or characteristics of an existing product or improvement (qualitative or quantitative) in any feature, component, ingredient, or characteristic of an existing product--for example, "new and improved," "now with 50 percent less sugar," "new, milder....." "new, stronger....," "now with built in flash." 28) Use occasion - Information that clearly suggests an appropriate use occasion or situation for the product--for example, "buy film for the Christmas season," "enjoy Jello at a birthday party," "the beer for special occasions." 29) Characteristics or image of users - Refers to any information concerning the type(s) of individual(s) who might use the advertised product--for example, "the young at heart," "for the busy career woman." 74 EXHIBIT 5 (Cont.) 30) Company information - May refer to any information about the image or reputation of the company that manufactures or distributes the product--for example, "we've been in business longer than anyone else," "we try harder," "babies are our business." Indicates the company is competent, reputable, or trustworthy in certain ways. May refer to company age or size such as claiming accumulated business experience of the company throughout its history, or bigness in terms of sales or employees. May refer to claims relating to varieties of business that the company is involved in, for example a conglomerate. May refer to ’location of business such as claiming advantages because of the geographical location of the company's business, or its raw materials. to the widely utilized Resnik and Stern (1977) framework suggested by Sepstrup (1985) and Stewart and Furse (1986) were employed. Thus, several important types of information (e.g., convenience, new uses) which advertisers frequently attempt to communicate to their audience were included in the new measurement scheme. A few additional cues were added after reviewing many Korean commercials and the specific types of information they contained (e.g., company information). Upon viewing a commercial, the coder had to indicate ( a commercial was computed by summing the number of 1 individual cues present. It should be noted that each cue was counted only once. Thus, even if product quality was whether each information cue was present in the sound and/or picture of the ad. The overall level of informativeness for 75 mentioned, for example, 3 times in a commercial, this would only count as one cue. Nevertheless, the average number of information cues per commercial in each commercial was slightly over three in each country. Inter-coder reliability was measured for each information one by having all coders (12 in the ROK; 16 in the U.S.) view each of a subsample of 30 commercials and The percentage agreement by comparing their responses. country is shown in Exhibit 6. As can be seen, reliability for most individual information cues was very high. The average cue had an agreement rate of 93.0 percent in the United States and 93.7 in Korea. Brand Differentiating Message The presence of a brand differentiating message was analyzed as an individual item in the content analysis. following codebook definition was utilized: "Is the principal message of the commercial unique to the product being advertised, or could any product make this claim? The commercial must make it clear that the message is unique; that is, the commercial must explicitly indicate the uniqueness or difference of the product." The 76 EXHIBIT 6 Reliability Of Information Cues (Percent inter-coder Agreement by Country) USA ROK 1) Price 98.8 98.6 2) Variety of the Product 82.0 82.5 3) Value 97.3 97.8 4) Quality 85.0 89.7 5) Size 96.9 98.6 6) Economy or Savings 96.9 95.6 7) Supply, Quantity Available 100.0 99.4 8) Method of Payment 99.8 100.0 9) Dependability/Reliability/Durability 94.4 99.7 10) Nutrition or Health 93.3 97.2 11) Taste 76.9 86.4 12) Sensory Information 84.0 83.1 13) Components/Contents/Ingredients 81.3 84.2 14) Availability, Location 95.6 95.8 15) Packaging or Shape 96.7 85.3 16) Guarantees or Warranties, Return 97.0 98.6 17) Safety 100.0 98.3 18) Independent Research 98.3 100.0 19) Company Sponsored Research 99.4 100.0 20) Research from Unidentified Source 97.7 100.0 21) New Ideas, New Uses 95.4 98.6 22) Performance, Results of Using 83.3 86.1 23) User's Satisfaction or Loyalty 90.6 96.4 24) Superiority Claim 78.3 97.8 25) Convenience in Use 95.0 91.4 26) Special Offer or Event 99.4 97.5 27) New Product or New Features 89.2 88.3 28) Use Occasion . 92.1 84.2 29) Characteristics or Image of Users 93.1 84.4 30) Company Information 96.0 89.7 Reliability figures on the brand differentiating message varied substantially between the two countries. Percentage agreement in the U.S. was just under 80% (78.8%) while the comparable figure in the ROK was 92.2%. however, that the frequency of brand differentiating It should be noted, 77 messages in each country varied greatly. Nearly 30 percent of the U.S. commercials in the sample contained brand differentiating messages. In Korea, on the other hand, brand differentiating messages were present in only about 3 percent of the commercials coded. The relatively low inter-coder reliability figure for the brand differentiating message suggests a need to verify that the selected U.S. commercials chosen for the test sample in this study actually contain a brand differentiating message. This check was conducted during the pre-testing of the sample commercials (For more information on the nature of pre—tests, please see Appendix 1). The same type of manipulation check was performed on the Korean commercials coded as having a brand differentiating message. RESEARCH DESIGN Overview In both the United States and South Korea, 101 subjects each viewed 20 commercials in groups of 10. In each country, the 20 sample commercials were randomly ordered. Half of the groups were exposed to the commercials in the initial random order. The remaining groups viewed a rotated ersion of the randomly ordered commercials. Specifically, he commercials positioned eleventh to twentieth in the irst ordering became numbers one through ten in the rotated 78 grouping. The initial order and the rotated order were shown to alternating groups (i.e., first group saw initial tape, second group saw rotated tape...). The subjects were told that the researchers are interested in their response to each of a group of commercials which they will be shown. Each subject was then asked to provide thought listings and to carefully rate each commercial immediately after it was shown on three dependent measures (to be described below). In the Korean data collection stage, the researcher was be assisted by his wife, a native speaker of Korean and Professor Doo Hee Lee, a Korean citizen who holds a Ph.D. in Marketing from Michigan State University. Selection of Sample Commercials The test commercials were selected from the database described above. The sample chosen in each country included ads with high and low information content and with and without brand differentiating messages. The test commercials were drawn from two product categories: (a) food and beverages and (b)health and beauty aids in each country. These product categories were chosen due their ability to produce the desired number of test commercials and the fact that they are generally associated with low motivation to process. Efforts were made to ensure that the chosen sample 79 commercials containing one certain test variable level (e.g., high information) did not systematically covary with another important executional variable. Thus, data from the larger content analysis was used to show that advertisements in the sample which have brand differentiating messages do not systematically tend to use humor (or any other important executional variable) more often than other commercials in the sample. Pretests were run to ensure that the specific commercials chosen for inclusion in each sample indeed capture the desired level of the test variable (See Appendix 1). With regard to information, subjects in the pretest were asked to rate the commercial on a seven point semantic differential scale ranging from "informative" to "uninformative". This procedure served two purposes. First, it served as a manipulation check on the level of information of the sample commercials. Second, it allowed for testing of the construct validity of the information level variable. The pretests also asked subjects whether they believed a brand differentiating message was present. Design of the Experiment The basic experiment constitutes an A X (B x C X S) within subjects design. Here, A denotes the country of the commercial (U.S. or Korea), B denotes brand differentiation (present or not present), C denotes the information level 80 (low or high), and S indicates that each subject will view each commercial. Thus, there are four cells in the design for each of the two countries, since commercials containing brand differentiating messages were evenly split between high and low information levels. Based on the results of the pretesting, the breakdown of commercials to be included in the sample is shown in Exhibit 7. Advantages of the within subjects design are the need for fewer subjects in running the research, and greater sensitivity in detecting treatment effects when they exist (Keppel 1982). General disadvantages include some restrictive statistical assumptions which must be acknowledged and the possibility that subjects can change during continued service in all of the treatment conditions. This latter potential difficulty was considered in choosing the dependent measures. Given the dependent measures which were chosen, which are primarily attitude based, it is highly unlikely that learning or "practice effects" will contaminate the results. Even with an absence of practice effects, it is important to disguise the purpose of the primary study. It would be a major mistake to allow subjects to believe that there is a "right" answer being sought (e.g. commercials with high levels of information should be rated higher). Thus, every effort was made to disguise the purpose of the study. 81 EXHIBIT 7 Cells In Experimental Design A. USA i) food/beverages 1) high information/ (3) brand differentiating message 2) high information/ (3) no brand differentiating message 3) low information/ (3) brand differentiating message 4) low information/ (3) brand differentiating message ii) health and beauty aids 1) high information/ (2) brand differentiating message 2) high information/ (2) no brand differentiating message 3) low information/ (2) brand differentiating message 4) low information/ (2) brand differentiating message B. South Korea i) food/beverages 1) high information/ (3) brand differentiating message 2) high information/ (3) no brand differentiating message 3) low information/ (3) brand differentiating message 4) low information/ (3) brand differentiating message ii) health and beauty aids 1) high information/ (2) brand differentiating message 2) high information/ (2) no brand differentiating message 3) low information/ (2) brand differentiating message 4) low information/ (2) brand differentiating message 82 Demand Artifacts Procedures to reduce or eliminate contamination based on demand artifacts were built into the study. It was decided that subjects would be told that the researcher is interested in their response to a group of advertisements which they will be shown. Upon completion of the response form for the last commercial, each subject was be asked to express their belief as to what the purpose of the study was. This step was designed to ensure that demand characteristics (or artifacts) did not taint the results. Demand artifacts occur when subjects believe that a certain outcome is expected and then adjust their responses, thereby biasing results. As Sawyer (1975) points out, it is particularly important to consider whether demand artifacts are creating bias in within-subjects designs. For this reason, the subjects' perceptions of the purpose of the study were analyzed, as suggested by Hastak and Olson (1989), as a means of assessing whether demand characteristics were present. Design ofsurvey instrument Before proceeding to discuss specific measures, it is necessary to comment on the importance of the translation and back-translation process in the design of the survey instrument. As pointed out by Brislin (1970), Douglas and Craig (1985), and Miracle (1988), among others, translations 83 and back-translations can be used to help overcome some of the problems that arise in doing international research. In order to aid in overcoming problems of equivalence, two bilingual speakers were employed in the back-translation process of the response instrument to be utilized in the experiment. The first translated the response instrument from its original form in English to Korean. The second then independently translated the new Korean version of the response instrument back to English. As Brislin (1970) points out, this process allows for the two English versions to be compared and for problems of equivalence to be identified. Generally, if the original and the back- translated version match exactly, it is suggested that the middle version'(in Korean) is equivalent to the original English version. To enhance the formation of the original response instrument, the following rules suggested by Werner and Campbell (1970) to allow for better translations were observed: 1) Simple sentences and phrases were utilized. 2) The use of pronouns was avoided. When necessary, nouns were repeated. 3) Metaphors and colloquialisms were avoided. 4) The use of the English passive tense was avoided. 5) Hypothetical phrases or subjunctive moods_were avoided. 84 In specific terms, the back-translations were used to help overcome problems of equivalence and to ensure that only etic generalizations are being made. Etic generalizations are those that are valid in more than one culture. Emic generalizations, on the other hand, are those that are valid in only one culture. Emic concepts or behaviors may be understood in the context of only one culture. Thus, it is possible that some terminology or methodology utilized in the United States can not be found or understood in certain other nations and cultures. Three types of equivalence problems that can create difficulty in doing international research are conceptual, linguistic, and functional equivalence. Problems of conceptual equivalence stem from situations in which the same concept is simply not present in two cultures. If a concept is present in only one of the cultures being studied, clearly, it makes little sense to investigate it in the other(s). Thus, conceptual equivalence is concerned with the interpretations which individuals make with regard to objects, stimuli, or behavior, and whether these exist or are expressed in the same way in different cultures (Douglas and Craig 1985). An example of a conceptual difference among cultures was identified by Triandis and Vassilou (1972) who discussed the Greek concept of "philotimo", which refers to behaving in a manner that is expected by members on an ingroup, such as the family, friends, or guests. In 85‘ Greece, there are clear expectations of ingroup members with regard to meeting obligations and sacrificing one's self for the good of the group. Apparently, philotimo is a concept which is unique to Greece. Other conceptual equivalence issues may not be as striking. For example, the connotation of certain colors may be different in various cultures. Thus, red may be associated with danger in the USA, but not in other cultures. What is humorous in one nation may not be at all humorous in another. Also, the concepts of "ethnic" and "foreign" may be different in the USA, but virtually identical in Korea. These types of differences can have important implications for advertising objectives and strategies (Miracle, Bang, and Chang 1992). Linguistic equivalence stems from the ability to express certain concepts or behaviors easily and clearly. According to Miracle, Bang and Chang (1992), "Linguistic differences and similarities relate to the translatability of functions from one language into another." Eskimos, for example, may have a greater number of words to describe snow than are available to speakers of English. Similarly, speakers of English, appear to have more words available to them to distinguish between the probability of performing a certain behavior (see discussion of purchase intention scales) than do Koreans. There may also be certain idioms or metaphors that simply do not translate to another language. Another possible problem related to linguistic 86 equivalence is that it a certain term or expression is translatable, but requires a much more detailed explanation in the target language. The above similarities and differences can all influence the choice of advertising strategies and objectives. Ultimately, a researcher must make certain the concept being investigated is communicated effectively in each language when doing cross-cultural research. Functional equivalence refers to whether the same concept or behavior (or a product) serves the same function in two or more cultures. From the viewpoint of marketers, the relevant question is whether or not purchasing and/or usage behavior are similar in multiple cultures. For example, Douglas and Craig (1985) observe that in the Netherlands and several other countries, bicycles are used as a basic mode of transportation for most consumers as opposed to recreational use, as is the case in the United States. These types of differences can also have important implications for advertising objectives and strategies. If not, clearly the researcher must make adjustments by changing As indicated above, the research will make use of translations and back-translations to see that only emic generalizations are made and to identify and deal with problems of equivalence. An additional measure to overcome problems of equivalence was having a bilingual assistant (fluent in both Korean and English) present during the 87 instruction process in the data collection of both countries in order be certain that similar instructions were given. MEASUREMENT The independent variables utilize the measures employed in the content analysis. In order to validate that the treatment groups (high vs. low information and presence or lack of presence of brand differentiation), pretests were conducted to verify that the inclusion of a commercial in a given treatment group was appropriate. This procedure was particularly important in the case of information content, since, as pointed out by Aaker and Norris (1982), it is possible that consumer perceptions may deviate from objective coding schemes. Thus, the pretests were used to measure consumer perception of the commercial's informativeness on a seven point semantic differential scale ranging from informative to uninformative. Prior effects research has made use of many types of dependent measures. These include: - Cognitive measures such as source derogations, counterargumentation, and message discrepancy (e.g., Belch 1981; Nelson, Duncan and Frontczak 1985; Edell and Keller 1986; Droge and Darmon 1987; Batra and Stayman 1990) - Recall (e.g., Belch 1981; Kahle and Homer 1985; Stewart 1986; Stewart and Furse 1986; Sewall and Sarel 1986; Kirmani 1990; Kent and Machleit 1990; Biel and Bridgewater 1990; Schumann, Petty, and Clemons 1990) 88 - Recognition (e.g., Krugman 1986; Zinkhan and Gelb 1986; Singh and Churchill 1987; Singh, Rothschild, and Churchill 1988; Kent and Machleit 1990) ’ - Choice Behavior (e.g., Gorn 1982; Kellaris and Cox 1989) - Advertiser Credibility - (e.g., Belch 1981; Appel 1987; Droge and Darmon 1987; Goldberg and Hartwick 1990) - Leyel of Attention - (e.g., Park and Young 1986; Stewart and Pechmann 1990) - Physiological measures - Attitude toward the ad LAéadL (see Exhibit 7) Attitude toward the brand (A-brand) (Exhibit 8) - Purchase intention (See Exhibit 9) The dependent measures of advertising effectiveness employed were: (a) Attitude toward the ad, (b) Attitude toward the brand and (c) Purchase intent. While it is clearly difficult to identify the most appropriate measure of effectiveness under given circumstances, a review of literature on advertising effects encompassing over 50 articles led the author to favor measures more closely tied to persuasion rather than earlier portions of the learning hierarchy. While it is not argued that recall or recognition measures are largely without merit, as some believe (e.g., Ross 1982; Gibson 1983), all advertisers should be concerned with whether their advertising is ersuading the consumer. Cognitive measures and advertiser redibility measures were not chosen because it was believed 89 EXHIBIT 8 Some Key Studies Measuring A-AD Mitchell and Olson (1981) Shimp (1981) Edell and Burke (1986) Batra (1986) Batra and Ray (1986) Cacioppo, Losch and Kim (1986) MacKenzie, Lutz and Belch (1986)- Droge and Darmon (1987) Holmes and Crocker (1987) Leigh, Rethans and Whitney (1987) Wilson and Machleit (1988) Iyer (1988) Edell and Keller (1989) Biel and Bridgewater (1990) Kirmani (1990) EXHIBIT 9 Some Key Studies Measuring A-Brand Belch (1981) Olson, Toy and Dover (1982) Kahle and Homer (1985) Zinkhan and Fornell (1985) Zinkhan and Gelb (1986) Park and Young (1986) Edell and Burke (1986) Holmes and Crocker (1987) Leigh, Rethans and Whitney (1987) Holbrook and Batra (1987) Muehling and Laczniak (1988) Iyer (1988) Machleit and Wilson (1988) Kardes (1988) Edell and Keller (1989) Kirmani (1990) Yi (1990) Schumann, Petty and Cacioppo (1990) Batra and Stayman (1990) Biel and Batra (1990) 90 EXHIBIT 10 Some Key Studies Measuring Purchasing Intention - Belch (1981) - Zinkhan and Fornell (1985) - Kahle and Homer (1985) - Zinkhan and Gelb (1986) - Park and Young (1986) - Leigh, Rethans and Whitney (1987) - Cox and Locander (1987) - Edell and Keller (1989) - Stewart and Pechmann (1990) - Yi (1990) that they are of more value to studies focusing more directly on routes to persuasion rather than persuasion itself. Choice behavior and physiological measures were not well suited to the proposed experiment at a practical level. Choice behavior would be difficult, if not impossible to track, and it is unlikely that the equipment used to perform physiological tests is widely available in South Korea. It is felt that it is important to include multiple measures of effectiveness, since different executional techniques may manifest their effects on different measures. The decision as to which specific measurement scales to employ will be made by giving preference to those which have shown demonstrated reliability and validity in past studies. Preliminary choices and measurement procedures for each chosen dependent variable are now be discussed individually. For a discussion of the final choice of scales based on the results of pretesting, please see Appendix 1. 91 Attitude toward the ad Lutz (1985) and Muehling and Laczniak (1988) defined attitude toward the ad in the following manner, "a predisposition to respond in a consistently favorable or unfavorable manner to a particular advertising stimulus during a particular exposure occasion." The scales chosen to operationalize A-ad in this study reflect this definition. An additional criterion for choice of individual scales is that the words chosen translate well into Korean. If the dependent measures do not reflect equivalent concepts, the study's internal validity is harmed, if not destroyed. Thus, a series of previously used scales were translated and back-translated before an individual scale was chosen for pretesting. Many scales have been utilized in the measurement of A- ad in the marketing, advertising, and consumer psychology literature. Exhibit 8 shows some high quality studies that have used A-Ad as a dependent measure. Clearly, there is no consensus among these studies as to what scales should be utilized. In most instances, seven-point semantic differential scales are employed. The specific bipolar adjectives chosen and the number of scales employed has varied greatly from study to study. In most cases, the reliability of the scales has not been reported in the studies. For the purpose of the selection of specific scales, 92 those that had previously demonstrated reliability were favored as candidates to go through the back-translation process. Semantic differential scales were utilized, and the number of points was determined through the pretesting process. As Barbara Mueller (cf. Miracle, et al. 1990) points out, there has been some question of the appropriateness of the use of seven-point scales in Japan (Fields 1980) and China (Ricks 1974), since individuals in each country have been found to systematically respond in the middle of such scales. This finding is in contrast to Americans, who seem to favor extremes on such scales. Results of the pretest indicated that seven-point scales were appropriate for both the Korean and U.S. subjects. In choosing the number of scales to utilize there is a tradeoff between possible respondent fatigue and the ability to establish reliability and validity. Zinkhan and Fornell (1985) noted that in spite of difficulties related to reliability and validity, measurement of A-brand had been "shown to be satisfactory in other studies" (e.g. Petty, Cacioppo and Schumann 1983). The approach chosen here insisted that reliability be measurable, but took the possibility of fatigue into account. Thus, it was decided that a scale of three to four items would be employed. Candidates included: 93 - Yi (1990) -- good/bad; like/dislike; favorable/unfavorable (Chronbach's alpha = .90); - Cox and Locander (1987) -- good/bad; pleasant/unpleasant; very likable/not very likable (alpha = .90); - Leigh, Rethans, and Whitney (1987) -- very interesting/not very interesting; very good/very bad; dislike very much/like very much; very irritating/not very irritating (alpha reported to be higher than standard of .80); - Droge and Darmon (1987) -- good/bad; interesting/not interesting; like/dislike; offensive/nonoffensive; irritating/not irritating (alpha = .93); and - Machleit and Wilson (1988) unfavorable/favorable; good/bad; enjoyable/unenjoyable; not fond of/fond of; dislike very much/like very much; irritating/not irritating; well made/ poorly made; insulting/not insulting (alpha = .95) It should be noted that some authors (e.g., Shimp 1981; Batra 1986) have suggested that A-ad be split into cognitive and affective (or emotional) components. Such a distinction may be considered in this study. Machleit and Wilson (1988), however, found there to be no discriminant validity on cognitive and affective dimension in their study. Additionally, it seems that the studies that have suggested such a distinction exists have employed very large numbers of scales, some items of which may stray from the theoretical definition of A-ad employed in this study. 94 Attitude toward the brand Exhibit 9 shows a number of studies that have used various measures of A-brand. Again, scales which have been previously shown to be reliable were given preference in selecting candidate scales. Scales falling into this category included: - Cox and Locander (1987) -- good/bad: pleasant/unpleasant; very likable/not very likable (Chronbach's alpha =.94); - Droge and Darmon (1987) -— bad/good; pleasant/unpleasant; favorable/unfavorable (alpha = .93); ‘ - Machleit and Wilson (1988) -- dislike very much/like very much; useful/useless; valuable/worthless; important/unimportant; beneficial/not beneficial; not fond _ of/ fond of; enjoyable/not enjoyable (alpha = .94); - Edell and Keller (1989) -- very bad/very good; very dislikable/very likable (correlation = .97); and - Muehling and Laczniak (1988) -- bad/good; uggavorable/favorable; negative/positive (alpha = As with A-ad, a limited number of semantic differential scales were utilized to avoid respondent fatigue and pretests were utilized to determine the appropriate number of points on the scales. An additional consideration was the need to choose different endpoints on the scales used for A-ad and A-brand so that any potential of artifactual relationships being caused by scale similarity was reduced (Droge and Darmon 1987; Edell and Burke 1984; Madden, Dillon and Twible 1984). 95 Separate data on brand attitudes was collected from a different group of subjects before the experiment was run. This was be done so that a baseline measure of brand preferences is available without sensitizing the subjects viewing the commercials to the purpose of the study. Purchase intention Exhibit 10 shows previous studies that have measured purchase intention in a variety of ways. Two types of scales were chosen to measure purchase intention in the pretests. The first was a probability scale developed by Juster (1964) and since utilized by Smith (1967), Zinkhan and Fornell (1985), Zinkhan and Gelb (1986) and others. The scale appears as follows: Taking everything into account, what are the chances that you will buy this brand if you buy any brand of the advertised product class sometime during the next year. , Circle the number that corresponds to the one best answer: - practically certain (99 in 100) - almost sure (9 in 10) - very probably (8 in 10) — probably (7 in 10) - good possibility (6 in 10) fairly good possibility (5 in 10) - fair possibility (4 in 10) - some possibility (3 in 10) - slight possibility (2 in 10) - very slight possibility (1 in 10) - almost no chance (0 in 10) or-Hvuasmoxqooxoo I 96 As with all other scales selected, the use of this probability scale was subject to modifications based on the outcome of back-translations and the pretesting process. As a result of the back-translation process, it was determined that the scale could not be utilized in its general form. It was discovered that there was a problem of linguistic equivalence associated with the use of the Juster scale. Specifically, during the back-translation process, it became apparent that the Korean language simply does not contain as many words to indicate intention to perform an action at a future time. Thus, it was determined that the use of probabilities (without written descriptors) alone would be a more appropriate in this study. Some prior studies which have measured purchase intention have dealt with only one or a few products. This has enabled the use of pictures for forced brand choices sometime after the subject is exposed to the commercials. The large number of commercials does not make this feasible here. It was felt here that multiple measures would allow for a check on construct validity, which could provide some level of assurance that the desired construct is indeed being measured. Thus, a second set of scales were employed. In several prior studies, only one semantic differential scale has been used to measure purchase intention (e.g., a five- point scale anchored by "likely to buy" and "unlikely to buy"). For purposes of measuring 97 reliability, this study employed more than one scale. Candidate scales included: - Yi (1990) -- Asked, "What are the chances you will buy the brand next time you make a purchase in this product category," likely/unlikely; possible/impossible; probable/improbable (Chronbach's alpha = .92); - Edell and Keller (1989) -- extremely likely/ extremely unlikely; - Cox and Locander (1987) -- likely/ not very likely In all cases, back-translation results were considered. Cognitive Response Measurement In addition to the above three measures, cognitive responses were recorded. For each commercial, subjects were asked to list any thoughts they have about the commercials, subject to a one minute time limit. These responses were then coded into a cognitive scheme based on a slight modification of a scheme devised by Wright (1973) by native speakers in each language. Wright's categories are counterargument, support argument, source derogation, and curiosity thoughts. Curiosity statements were dropped from the scheme due to their generally low frequency across studies in which they are measured and the apparent random nature of their occurrence (i.e. they do not appear to be directly related to executional elements of the stimulus). One additional category, source support, will be added, as 98 suggested by Bettman (1986). Operational definitions of the four cognitive response categories are shown in Exhibit 11. The coding process followed the established guidelines relating to objectivity, systemization, sampling, and reliability of Kassarjian (1977) and Kolbe and Burnett (1991). Objectivity was enhanced by providing written rules and procedures for judges, training the judges, analyzing pre-test data, using judges independent of the author, and having the two judges in each country code the responses independently. In cases of disagreement, the disagreement was discussed with both judges and resolved by the researcher. Systemization was achieved by stating clear hypotheses regarding the cognitive response categories. As previously discussed, the samples were selected via a process which incorporated randomization and are substantially larger than those of most previously reported television studies (Kolbe and Burnett 1991). Finally, coefficients of agreement for each measure are reported as measures of reliability in each country. Chance agreement is not a major problem in this case since there are four distinct categories which are treated as mutually exclusive. If the categories were not mutually exclusive, calculation of Cohen's kappa would have been advisable. 99 EXHIBIT 11 Operational Definitions Of Cognitive Response Categories a) Source support statements - Statements which express trust in or support for the advertiser or advertisement itself or statements expressing a liking for the overall means used by the advertiser in the presentation. b) Source derogation statements - Statements which express distrust for the advertiser or advertisement itself or statements eXpressing a dislike for the overall means used by the advertiser in the presentation. c) Support arguments - Statements which are directed in favor of use of the product and which: i) state a specific favorable consequence of using the product, ii) state a specific desirable use of the product, iii) suggest an undesirable consequence of not using the product, or iv) reaffirm the accuracy or validity of an argument presented in the advertisement. Simple statements of liking the product or positive emotional reactions unaccompanied by any statement are not included in this category. d) Counterarguments - Statements which are directed against the idea or use of the products in the advertising communication and which i) state a specific unfavorable result of using the product, ii) state a specific undesirable attribute of the product, iii) suggest an alternative method for handling one of the problems cited in the advertising message, iv) state a specific favorable or desirable consequence or attribute of an alternative product, or v) challenge the accuracy or validity of a specific argument contained in the advertising message. Simple statements of dislike for the product idea or emotional reactions not accompanied by any of the above types of statements are not included in this category. 100 ANALYSIS For the previously described design, a multiple analysis of variance (MANOVA) was be run to test for differences in the mean level of effectiveness between the U.S. and South Korea. Since there are 6 sources of variance in this design (A, B, AXB, S, AXS, BXS, and AXBXS), it was also necessary to run individual ANOVA's both between and within (e.g., high vs. low information) countries. MANCOVA and ANCOVA procedures were also run in an attempt to adjust within subjects scores as appropriate, based on prior exposure or by sex of respondent. Analysis of whether the dependent measures are correlated was also be conducted. The SPSS PC+ statistical package was used to perform the statistical analyses. Reliability Reliability figures are reported in the next chapter for both level of coder agreement on the independent test variables and on the dependent scales measuring effectiveness. The lack of reporting of reliability has been a flaw present on much of the prior research in research on both advertising effects and international advertising. It is particularly important in international research, where cultural differences complicate the research design to provide assurances that the research is replicable. For this reason, the ability to measure 101 reliability has been given strong preference in the experimental design. Validity Throughout the formulation of the design, efforts to enhance both the internal and external validity of the data were taken. To enhance internal validity, every effort was made to show that the commercials included in the sample did not systematically covary with some other important advertising executional variable (e.g., type of appeal; use of music or use of humor). Additionally, measures of A- brand were taken on a different sample of subjects in order to be able to analyze whether previous brand perceptions of the test population are associated with A-ad, A-brand and purchase intention without sensitizing the population to the purpose of the study. The purpose of the study was also disguised as fully as possible in a further effort to enhance internal validity. The use of a sample of actual television commercials enhances the study's external validity. Though possible effects of prior exposure must be considered in the analysis, the sample will better simulate "real world" ads than would those designed explicitly for a laboratory setting. An additional measure taken in order to aid external validity was to have subjects view the commercials in as natural a setting as is possible within the parameters 102 of the experiment. Construct validity is also measured for the operationalization of the information variable in the two countries studied. As indicated by Peter (1981), Calder, Phillips, and Tybout, (1982) and others, it is important to determine whether chosen measures actually reflect the theoretical construct being investigated. This is particularly important in areas where theory development is in the early stages, as is the case in both advertising effects research and international marketing research. Construct validity was also be considered for the purchase intention variable. Procedure It was determined that subjects would be exposed to the commercials in small groups. The settings were matched as closely as possible. The order of commercials was selected randomly, and a rotated version of the tape was preferred in order to avoid order effects. After each commercial was shown, the subjects were asked to fill out a survey instrument pertaining to that commercial. The survey instrument asked whether the subject had seen the commercial before and requested a response to the multiple scale items for the three dependent measures of effectiveness (to be‘ discussed below). Additionally, subjects were asked to provide thought listings. 103 Commercials Included in the Sample Exhibits 12 and 13 show the specific commercials chosen for inclusion in the samples based on the procedures describe above. level for both of the independent variables. EXHIBIT 12 The Exhibits also indicates the treatment COMMERCIALS UTILIZED IN THE STUDY - USA Brand Information Differentiating *3 Product Level Message 1 Roman Bread high yes 2 Charmin high yes 3 Pillsbury (potatoes) low no 4 Sugar Twin low yes 5 Clusters (cereal) high yes 6 Close-Up (toothpaste) low no 7 Tone (soap) high no 8 Swanson (pot pies) low no 9 Classico (pasta sauce) low no 10 Studio Line (hair care) low no 11 Caress (soap) high no 12 Snickers 19W yes 13 Libby's (canned meat) high yes 14 Hershey's Kisses hlgh no 15 Peppridge Farms (cookies) low yes 16 Maybelline (eye pencil) low ' yes 17 White Rain (hair care) lOW yes 18 Diffrinse (cold cream) hlgh yes 19 Creamettes (pasta) h19h no 20 I Can't Believe its h1gh no Not Butter 104 EXHIBIT 13 COMMERCIALS UTILIZED IN THE STUDY - ROK ' v w ‘ 'I '- uv ‘ '- v' w vv ‘v ' Brand Information Differentiating Product Leygl Message Lotte (toothcare) high no Hae-tae Milk low no Mona Lisa (tissue) high yes Fresh Brown (bread) low no Touch Me Lotion low no Post (cereal) high no Kellogg's (cereal) high no Pepsi low yes Coca Cola high yes Rhumba (ice cream bar) high yes Amore (eye jel) high no Andersen (candy) high no Penaten (lotion) low yes ‘ Alchon Mandu low no 1 Samboomil (honey) low yes Q.T. (skin care) high yes 1:1 (cookies) high yes Cardin (parasol) low no Julia Cosmetics low yes ' Cider (soft drink) 10W yes CHAPTER 4 - RESULTS AND DISCUSSION This chapter will begin with a brief description of the ata collection process and the subjects utilized. Then, reliminary analyses conducted prior to hypothesis testing ill be discussed. These analyses include investigation of he possible impact of demand artifacts on the study and escriptions of data collected on subjects' prior purchase rom the product categdry and prior exposure to the ommercials. Additionally, dependent variable istributions, reliability of the dependent variable scales, 1d manipulation checks are discussed. After these reliminary tests, results of the hypothesis tests are aported and analyzed. Finally, data collected on potential >urces of covariation are discussed. DATA COLLECTION lbjects A total of 101 usable questionnaires were collected in bth the USA and the ROK, for a total of 202 response Lstruments. In the ROK, the data were collected in ptember, 1991 at the Korea University in Seoul. Since male subjects were needed and the business school at Korea iversity has a predominately male student population (over %), subjects were recruited from nearby Sungshin Womens' 105 106 liversity. A total of nine groups were run. The largest soup was composed of 14 students. Average group size was L. In the USA, 11 experimental groups were run at Michigan :ate University between November, 1991 and January, 1992. 1e average group size was 9. Table 1 shows the frequency distribution for age of the arean subjects. The range was from 18 to 26 years old. 1e mean age was 21.1 years and the modal age was 21. In 1e USA, the range was from 19 to 39 years old, the mean was 1.5 years, and the mode was 21. TABLE 1 Average Age of Subjects Frequengy Percent Mean 2 2.0 19 18.8 51 50.2 21 21.0 3 3.0 2 2.0 2 2.0 1 1.0 101 100.0 21.5 1 1.0 11 10.9 19 18.8 31 30.7 28 27.7 7 6.9 3 3.0 1 1.0 101 100.0 21.1 107 Table 2 shows the breakdown of the subjects by major. 1 Korea, a majority of the subjects were business majors, :counting for almost 90% of the sample. While only 44.6 arcent of the U.S. subjects were business majors, most of me remaining subjects majored in business related fields e.g., advertising, merchandise management). TABLE 2 Subjects by Major p Freguency_ Percent 3A Merchandise 41 40.6 Management Business 45 44.6 Communications/ 11 10.9 Advertising *Other 4 4.0 tal 101 100.0 public of Korea Business 90 89.1 English 2 2.0 Home Economics 11 7.9 Education 1 1.0 tal 101 100.0 (1 each from Political Science, Nursing, Horticulture and Engineering Arts). 108 ANALYSES Demand Artifacts Upon finishing their response to the final commercial, each subject was given a separate sheet of paper to indicate what they felt the purpose of the study was. Analysis of responses in each country indicates that the subjects were not aware of which variables were being manipulated in the experiment. No subject in either country mentioned that the researcher was interested in reactions to commercials with varying levels of information content or with/without a brand differentiating message. A few subjects in each country did make references to studying "how advertising informs the consumer", but they seemed to be referring to the basic role of advertising described in consumer behavior texts (i.e., hierarchy of effects) rather than considering how different levels of information content might be used as a strategic device. These results indicate that subjects ere unfamiliar with the purpose of the study. Hence, it an be concluded that demand characteristics did not ontaminate the data set. rior Purchase and Exposure ) Prior purchase Table 3 shows the product category for each U.S. ommercial and the average percentage of subjects who eported purchasing from the product category within the 109 previous year. Generally, findings indicate that the subjects are part of the target market for the advertised products. As can be seen, across commercials an average of nearly 80 percent of the subjects reported purchase from the product category. There were only two product categories in the U.S. sample for which less than fifty percent reported purchase. These were canned meats (Commercial #4) and artificial sweetener (Commercial #13). The lower figures for these categories are not surprising, since both appeal to a relatively narrow target market compared to other product categories in the sample. Thus, it appears that the product categories used in the U.S. sample were generally appropriate for the analyses to be performed, since it appears that the group of subjects chosen is part of the target audience for the products. As Table 4 shows, across the ROK commercials, the verage percentage of subjects reporting purchase from the roduct category within the past year was 68.6%, a figure omewhat lower than in the United States. Four product ategories showed figures of under fifty percent. However, 't is contended that Korean advertisers would regard the ubjects as a relevant group. In the case of dental floss (Commercial #1), only 8 ercent of the subjects reported purchasing the product ithin the previous year. While this was a bit surprising iven results of the pretest which suggested somewhat higher 110 usage among those Koreans living in the United States, the result seems to be explained by the fact that dental floss has not traditionally been used in Korea and, in fact, was not even sold there until the recent liberalization of import laws. Thus, the product is in the introductory stage of its life cycle. It is very likely that advertisers of dental floss would regard 18 to 26 year old females as an important part of the target market. TABLE 3 Percentage of Subjects Purchasing From Product Category within The Past Year - United States Commercial Number Percentage Product Category 1 88.1 Bread 2 92.1 Toilet Paper 3 62.4 Scalloped Potatoes 4 32.7 Artificial Sweetener 5 84.1 Cereal 6 93.1 Toothpaste 7 97.0 Soap 8 64.4 Chicken Pot Pies 9 75.3 Pasta Sauce 10 86.1 Hair Spray/Hair Care 11 97.0 Soap 12 96.0 Candy Bars 13 30.0 Canned Meat 14 93,0 Candy (Chocolate Kisses) 15 91.1 Cookies 16 92.1 Eyeliner 17 88.1 Hair Spray 18 59.4 Cold Cream 19 87.1 Pasta 20 82.0 Margarine Average 79.9 111 TABLE 4 Percentage of Subjects Purchasing From Product Category Within the Past Year - Republic of Korea Commercial flpmber Percentage Product Category 1 8.0 Dental Floss 2 95.0 Milk 3 97.0 Toilet Paper 4 93.1 Bread 5 80.0 Lotion 6 56.4 Cereal 7 57.4 Cereal 8 97.0 Soda Pop 9 98.0 Soda Pop 10 92.1 Ice Cream Bar 11 15.8 Eye Repair Jel 12 76.2 Cookies 13 52.4 Baby Lotion 14 89.1 Mandu (Food) 15 11.9 Honey 16 6.0 Baby Powder 17 78.2 Rice Cookies 18 80.2 Parasol sun) 19 90.1 Cosmetics 20 98.0 Soda Pop Average 68.6 Two of the other product categories -- eye repair jel 1nd baby powder -- are categories from which it would seem .ikely that the subjects would begin purchasing from within a limited number of years. With regard to eye repair jel Commercial #11), it is customary for Korean females to -egin using the product, designed to prevent crow's feet, oon after leaving college- The baby lotion commercial #16) encouraged women to use the product themselves, 112 (similar to the marketing of baby shampoo in the United States). Though it is clear that this appeal had not yet taken hold among the subjects, it is still likely that marketers of baby lotion would find the subject group of interest, since most Korean females are likely to get married after college, and relatively few (in comparison with the United States) are likely to become "career women" due to traditional customs. In the case of commercial #15 (which was for honey), the likely reason for the low figure of 11.9% of subjects purchasing the product within the past year is that honey is regarded as a "health food" in Korea and, again, is not a traditional product. Thus, honey is a product in Korea which is designed to appeal to a limited target audience (the health conscious) rather than the mass market. It seems likely that the females in the sample would be considered part of the target audience for this product category. Overall, it is apparent that advertisers of the products included in each sample would regard the subject pOpulation to be part of the target market in the majority, if not all, cases. 113 b) Prior exposure Tables 5 and 6 show the percentage of subjects in each country reporting prior exposure to each commercial. Not surprisingly, due to factors such as age of commercial, its TABLE 5 Prior Exposure by Commercial - USA Number Reporting Commercial Prior Exposure Percentage 1 18 17.8% 2 24 23.7 3 36 35.6 4 74 73.3 5 77 76.2 6 81 80.2 7 85 84.2 8 9 8.9 9 26 2.6 10 92 91.1 11 35 34.7 12 56 55.4 13 2 1.9 14 73 77.2 15 15 34.7 15 59 58.4 17 53 67.3 13 42 41.6 19 31 30.7 20 76 75.2 Potal 985 48.7 114 frequency of appearance, etc., there is a wide range of values. In the United States, prior exposure ranged from 1.9% of the subjects for commercial #13 (Libby's canned meat) to 91.1% for commercial #10 (L'Oreal). Average prior exposure across commercials was just under 50 percent. In Korea, the range was from 34.7% for commercial #15 (a Korean brand of honey) to 100% for commercial #20 (Coca Cola). TABLE 6 Prior Exposure by Commercial - Korea Number reporting Commercial prior exposure Percentage 1 54 53.5% 2 92 91.1 3 78 77.3 4 34 33.7 5 41 40.6 6 72 71.3 7 72 71.3 8 96 95.0 9 83 82.2 10 73 72.3 11 54 63.4 12 42 41.2 13 42 41.2 14 100 99.0 15 35 34.7 15 37 86.1 17 57 56.4 13 70 69.3 19 42 41.6 20 101 100.0 Potal 685 55-9 115 Average prior exposure across commercials was slightly over two thirds (66.9%). The higher overall mean value for prior exposure in Korea is not surprising, given that there are just four commercial television networks in Korea (there are no non-network independents or cable networks) and that television programming does not air on weekdays in the midday or late night dayparts. These facts lead to less clutter in Korean television advertising. Additionally, television time must be purchased through the government agency KOBACO well in advance of its airing. In practice, this has led to large firms tending to buy up most available time and, in turn, frequent repetition of many of the same commercials. Distributions of Dependent Variables Frequency distributions for all dependent variables (A- ad measures, a-brand measures, purchase intention measures, cognitive measures) were examined, as were measures of skewness and kurtosis. In both countries, A-ad and A-brand measures tended to be slightly skewed left from the scale midpoint of four. In general, it was determined that assumptions of normality for the purposes of hypothesis testing were justified. 116 Reliability of Dependent Variable Scales The multiple-item scales used to measure attitude toward the ad, attitude toward the brand and purchase intention were found to be highly reliable in both countries. Table 7 shows Chronbach's alpha for each multiple scale item used on the U.S. questionnaire. For attitude toward the ad, Chronbach's alpha for the four scales (which were anchored by good/bad; like/dislike; positive/negative; and well-made/poorly made) was .95. Deletion of the well made/poorly made item would have resulted in a slightly higher alpha of .98. For attitude toward the brand, the five scales (anchored by favorable/unfavorable; pleasant/unpleasant; positive/negative; beneficial/not beneficial; and good/bad) was .96. The beneficial/not beneficial item was the only one which would have led to a slightly higher alpha if it was omitted. Finally, for purchase intention, the three scales (anchored by likely/unlikely; probable/improbable; and possible/impossible was .98. These results indicate that the each set of multiple item scales used to measure attitude toward the ad, attitude toward the brand, and purchase intention can be viewed as measuring a single construct. 117 TABLE 7 Reliability Of Dependent Variables - United States Chronbach's Scale Mean Variance Alpha ALPHA If Dg1eted If Deleted If Deleted AAD AADl 9.22 20.72 .926 AADZ 9.16 20.32 .923 AAD3 9.28 21.58 .931 AAD4 9.26 24.14 .967 TOTAL .953 ABRAND ABRANDl 11.59 32.52 .948 ABRANDZ 11.60 33.63 .947 ABRAND3 11.61 33.75 .945 ABRAND4 11.29 35.90 .969 ABRANDS 11.53 34.36 .950 TOTAL .961 El PIl 7.81 17.93 .962 PIZ 7.84 18.06 .959 P13 8.16 19.06 .986 TOTAL .979 Table 8 shows data on the reliability of the multiple item scales used on the Korean questionnaire. As in the United States, all three multiple scales demonstrated a high evel of reliability, illustrating the effectiveness of the ranslation/back-translation process in producing comparable cales. For attitude toward the ad, Chronbach's alpha for 118 the four item scale was .96. For the five item scale used to measure attitude toward the brand, alpha was .93. For purchase intention, the scales' alpha was .97. As in the USA, these figures suggest that the multiple item scales effectively measured the same construct. For the cognitive response measures, percentage agreement in the ROK was .86 compared to .91 in the USA. Both figures are in excess of the .85 standard (Kassarjian 1977) which has been widely applied in marketing and advertising studies. When the coders disagreed, differences were resolved in consultation with the researcher. Manipulation Checks a) Brand differentiating message As a manipulation check for the presence of a brand differentiating message, subjects were asked to indicate whether the commercial had a brand differentiating message based on the operational definition used in this study. The mean value reported by U.S. subjects for commercials coded as having a brand differentiating message was .54. The mean for those commercials without a brand differentiating message was .41. One commercial coded as having a brand differentiating message (#16) was clearly not viewed by the subjects as having one. Omitting this commercial would result in a mean of just over .59 for the remaining nine commercials coded as having a brand differentiating message. 119 TABLE 8 Reliability of Dependent Variable Scales - Republic of Korea Chronbach's Scale Mean Variance Alpha ALPHA If Deleted If Deleted If Deleted AAD AADl 10.37 21.83 .942 AAD2 10.28 21.68 .943 AAD3 10.39 22.21 .947 AAD4 10.25 22.95 .962 TOTAL .961 ABRAND ABRANDl 12.82 23.98 .911 ABRANDZ 12.84 27.01 .910 ABRAND3 12.92 25.23 .899 ABRAND4 12.80 27.61 .926 ABRANDS 12.91 27.23 .911 TOTAL .928 21 P11 7.23 14.28 .949 PI2 7.12 15.24 .950 PI3 7.50 14.86 .962 TOTAL .969 In the Korean sample, the commercials coded as having a rand differentiating message were given a mean rating of .50 compared to a mean of .36 for the commercials coded as ot having a brand differentiating message. verage value on this variable was .42 for the Korean ample, compared to .47 for the U.S. sample. The overall 120 While the differences between the means of subject perception of whether the manipulation was present in each country is not large, they are in the correct direction in each case. This is a positive result for several reasons. First, the brand differentiating message concept itself is somewhat elusive and difficult to achieve. U.S. advertising professionals have been known to compare attempting to differentiate the message in competitive product categories to dancing on the head of a pin. Coming up with truly unique messages is particularly difficult in low involvement product categories. Another reason the result is encouraging is that the subjects were only able to view each commercial just one time. Moreover, brand differentiation was the last item on the questionnaire, meaning some time had passed since subjects saw the commercial. When the commercials were coded in the content analysis, coders were allowed to view the commercial repeatedly, closely scrutinizing the operational definition of the construct as well as information contained in the sound and picture of the commercial. This latter process led to reliable classification of the variable. It is strongly suspected that if the subjects had the opportunity to focus only on whether the commercial had a differentiating message and/or were allowed to watch the commercial repeatedly to make this assessment, the difference in the subjects' reported means would be wider. However, the apparent gap between objective 121 coding of the brand differentiating message construct and subjects' perceptions should be conSidered in evaluating the generalizability of the results. b) Information content In the United States, commercials identified as having high levels of information content based on the objectively measured 30 information cue scheme shown in Exhibit 5 were given a mean rating of 3.1 on a seven point scale anchored by informative (1) and uninformative (7). The subjects gave an average rating of 3.9 to those commercials identified as having low levels of information content via the 30 cue scheme. In the Republic of Korea, the subjects' mean for high information commercials was 3.2 on a seven point scale anchored by the Korean language equivalents of informative/uninformative, while the mean for low information commercials was 3.6. The overall mean level of informativeness for all commercials was 3.4 in the Korean sample. This figure is identical to the mean given to the twenty U.S. commercials by the U.S. subjects. Again, in each country, the mean value for the high and low information commercials are classified in correct direction relative to each other. While the gaps between the means are not huge, they are statistically significant. Also, it should be noted that the endpoints of this scale 122 were not widely utilized in either country. Subjects in both countries were, apparently, reluctant to classify a commercial as being either wholly informative or wholly uninformative. In fact, in the United States, no single commercial was given a mean rating of less than 2 (lowest was 2.31) or more than 5 (highest was 4.81). Generally, it appears that the U.S. subjects were very reluctant to label a commercial as being highly uninformative or, to a lesser extent, as being highly informative. Similarly, the Korean subjects, on average did not rate commercials as being very informative or very uninformative. The commercial rated most informative had a mean value of 2.16 (second was 2.83) while the commercial ranked least informative had a mean value of 5.26 (second was 4.56). While the fact that the subjects' mean perceptions of informativeness were in the correct direction is encouraging, and the avoidance of scale endpoints reduces concern about the difference in the means to some extent, it must be noted that subject perception of informativeness may be a construct which is separate from objective measures of information. Indeed, Aaker and Norris (1982) have suggested that subjects' perceptions of informativeness should be measured as a separate construct. Clearly, what is perceived as "informative" varies from individual to individual based on prior knowledge, involvement, etc. While it is likely that subject perception of 123 informativeness is a construct somewhat distinct from objective measurement schemes, it is encouraging that there is a tendency toward convergence rather than away from it. Further, for managerial purposes, objective schemes such as the one employed in this study have the advantage of being easily measurable. Additionally, no prior study has demonstrated that subjects' perceptions of informativeness can be measured reliably, as has been demonstrated by the objective scheme utilized here. For these reasons, the data will be analyzed based on the objective classification of level of information content derived prior to the design of this study. HYPOTHESIS TESTS Each hypothesis test will be conducted both for the combination of the three scaled measures of effectiveness (A-ad, A-brand, and purchased intention) using multivariate analysis of variance and also for each individual construct. The use of a MANOVA procedure using all three dependent variables is justified, since the three measures were positively correlated with each other in both countries (see Table 9), indicating all three dependent variables can be viewed as measures of advertising effectiveness for the purpose of analysis. Differences between the two countries in the correlations between individual combination of the dependent variables were not large. The correlation between 124 A-ad and A-brand was identical in the two countries at .60. The A-ad/purchase intention correlation was a bit higher in Korea (.51) than in the USA (.40). A-brand and purchase intention were correlated at levels of .56 in Korea and .65 in the USA. When dependent measures are correlated, as is the case here, the use of individual t-tests alone is insufficient for determining differences between means because correlations are ignored and, hence, not all available information for assessing differences in group means (Hair, Anderson, and Tatham 1987) is utilized. Additionally, a series of t-tests does not allow for effective control of the Type One error rate. For these reasons, the results of multivariate tests are reported prior to reporting the results of individual ANOVA's and/or t-tests when testing the hypotheses. Individual t-tests are performed because, in some instances, they allow for a clearer picture of the specific managerial implications of the data. In all cases where MANOVA was utilized, tests of the assumptions of normality of the multivariate distribution and the determinant of the variance/covariance matrix were conducted. In general, the multivariate distributions did not violate the normality assumption. Additionally, Cochran's C and the Bartlett-Box F tests were used to assess homogeneity of individual variables and Box M's multivariate 125 TABLE 9 Correlation Of Dependent Variables A-Ad A-Brand Purch. Int. Overall A-Ad 1.00 .61 .44 A-Brand .61 1.00 .60 Purchase Int. .44 .60 1.00 USA A-Ad 1.00 .60 .40 A-Brand .60 1.00 .65 Purchase Int. .40 .65 1.00 ROK A-Ad 1.00 .60 .51 A-Brand .60 1.00 .56 Purchase Int. .51 .56 1.00 tests were performed. Homogeneity of variance assumptions did not appear to be violated. Appendix 3 provides data on the means and standard deviation for the global means of A-ad, A-brand, and purchase intention, as well as providing tables on individual scale items (for the aforementioned three measures) and the mean number of each cognitive response. Individual scales for A-ad, A-brand, and purchase intention (e.g., attitude toward the ad good/bad) will not be analyzed in the context of the hypothesis tests due to the high reliability of the multiple scales in both countries. 126 Between Country Hypotheses Hypothesis One Hypothesis 1 suggested that there would be no difference between the mean level of effectiveness of commercials containing similar levels of information in the Republic of Korea and the United States. As shown in Table 10, the null hypothesis is rejected at the .05 level of significance. The MANOVA indicates a significant country effect, F(3,4027) = 81.4, p < .001; a significant effect for level of information, F(3,4027) = 9.4, p < .001; and a significant country X level of information interaction effect F(3,4027) = 20.8, p < .001. Thus, it is highly likely that the differences in group mean vectors tested in this hypothesis are not the result of sampling error. In other words, Korean and U.S. subjects react to commercials in the two treatment conditions (high vs. low information) in different ways. Indeed, visual inspection of the means shown in part b of Table 10 indicates that U.S. group gave substantially higher ratings (A-ad = 2.81; A-brand = 2.67; P.I. = 3.64) to high information commercials than did the Korean group (A-ad = 3.47; A-brand 3.23; P.I. = 3.77). The t-tests shown in Table 10 shed additional light on differences between the USA and the ROK in the effectiveness of commercials with high and low levels of information. In the high information condition, the differences in means were statistically significant for attitude toward the ad (t ‘3 '2'— 127 = -9.56, p < .001) and attitude toward the brand (t = —9.37, p < .001), but not for purchase intention (t = -1.50, p = .134). The lack of a statistically significant difference for purchase intention in the high information treatment seems to be related to a large drop-off between purchase intention and A-brand ratings by the U.S. subjects. It would appear that Koreans are more likely to indicate intention to purchase the product if they have a positive attitude toward the brand. This finding is consistent with observations (e.g., Lee 1991) that Koreans tend to be very brand loyal. In the low information condition, purchase intention (t = 8.94, p < .001) shows a highly significant difference in means between the two countries, while A-brand shows a moderately significant difference (t = -1.83, p = .067) and A-ad does not show a statistically significant difference at even the .10 level (t = -1.11, p = .269). It is clear from Table 10 that the U.S. subjects had a striking preference for commercials containing high levels of information, while the Korean subjects gave slightly better ratings to commercials with lower levels of information. These results suggest that U.S. and Korean consumers systematically react differently to television commercials containing different levels of information. 128 TABLE 10 ROK VS. USA - Effectiveness By Level Of Information F- F- Value_ Statistic si . F a) Multivariate Test (MANOVA) Country X Information Hotelling's .20 20.8 <.001 Wilks' .98 20.8 <.001 Country Hotelling's .06 81.4 <.001 Wilks' .94 81.4 <.001 Information Hotelling's .01 9.4 <.001 Wilks' .99 9.4 <.001 b) T-tests Mean Mean USA ROK d.f. t-val. sig. t A-Ad High 2.81 3.47 2018 -9.56 <.001 Low 3.34 3.41 2018 -1.11 .269 A-Brand 1 High 2.67 3.23 2016 -9.37 <.001 Low 3.09 3.20 2016 -1.83 .067 Purch. Int. High 3.64 3.77 2018 -1.50 .134 Low 4.29 3.50 2018 8.94 <.001 129 While more exhaustive explanations of preferences relating to the level of information in each country will be attempted in the within country hypotheses (including the Elaboration Likelihood model), it seems likely that cultural differences play a central role (See Appendix 1 for some cultural differences between the USA and the ROK). Differences in the contextual level of communication in each country, for example, may help to explain why Koreans seem to prefer a somewhat more subtle approach to advertising in general. Clearly, the differences found here cannot be explained by the product life cycle concept or a "stage of economic development" arguments often utilized in comparative content analyses, sinCe they would predict a preference for more information at earlier life cycle stages or during earlier stages of development. In reality, the majority of the product categories in each country have been on the market a long time in each country and tend to be mature products. Even if it was argued that the ROK was in an earlier stage of economic development, the findings would be in the opposite direction of "stage of development" predictions, since U.S. subjects preferred higher levels of information than the Koreans. 130 Hypothesis Two The second hypothesis suggested that there would be no differences between the means of the with and without brand differentiation treatments between the U.S. and Korean samples. As indicated by Table 11, the null hypothesis of no differences cannot be accepted. There is a significant effect not only by country itself F(3,4027) = 81.3, p < .001; but also for the country X brand differentiating 23.3; p < .001. Thus, it message interaction (F 3,4027) can be concluded that there is a difference in the way the subjects in the two countries reacted in at least one of the brand differentiating message conditions. In fact, the results of individual ANOVA's on each individual dependent variable indicates a significant country effect for each (p < .001 in each case). The means and t—tests shown in Table 11 show more about the nature of the differences in reactions between subjects in the two samples. For attitude toward the ad, while U.S. and Korean subjects responded in a virtually identical manner to commercials containing a brand differentiating message (t = .10; p = .921), Koreans responded much less favorably to commercials without brand differentiating messages. While U.S. subjects gave an average rating of 131 TABLE 11 ROK VS. USA - Effectiveness By Presence Of Brand Differentiating Message F- F- Value Statigtic sig. F a) Multivariate Test (MANOVA) Country X Brand Differentiating Message Hotelling's .17 23.3 <.001 Wilks' .98 23.3 <.001 Country Hotelling's .06 81.3 <.001 Wilks' .94 81.3 <.001 Brand Differentiating Message Hotelling's .02 27.0 <.001 Wilks' .98 27.0 <.001 b) T-tests Mean Mean USA ROK d.f. t-val. sig. t ~Ad Yes 3.13 3.12 2018 .10 .921 No 3.03 3.76 2018 -10.65 <.001 ~Brand Yes 3.03 3.17 2017 -2.26 .024 No 2.73 3.26 2015 —8.94 <.001 rch. Int. Yes 4.20 3.50 2016 7.62 <.001 No 3.73 3.79 2017 -0.60 .551 132 3.03 to commercials without a brand differentiating message, the Korean mean was just 3.76. Apparently, Koreans tend to have less favorable opinions of advertisements that do not explicitly offer anything unique in comparison to the Americans. Interestingly, several cognitive responses from the Korean sample reinforce this notion. Comments such as, "this advertisement is too similar to the commercial for a competitive product" or "I did not find any difference between this ad and that of competitive product X" were frequently made by the Korean subjects. It may be the case that the presence of a brand differentiating message allows the Korean subject to view the advertisement itself as being unique, and, in turn, leads to higher A-ad rankings. The fact that the Korean advertising industry is relatively new (even marketing departments within large corporations did not become common until the late 70's/early 80's) may also contribute to the less positive reaction to commercials without a unique message. Korean advertising executives indicate that "copycat" or at least less innovative approaches to advertising are used frequently in the ROK. This lack of innovation may allow fresh approaches to succeed more easily in the ROK, while commercials similar to others might be regarded as, "tired and worn out." For attitude toward the brand, statistically significant differences between means for the two countries were found in both treatment conditions. For commerclals 133 with brand differentiating messages, however, the raw difference in means was relatively small (USA 3.03, ROK 3.17; t = -2.26; p = .024). Perhaps of more interest is the large gap in reaction to commercials that do not provide a brand differentiating message (USA 2.73 ROK 3.26; t = -8.94; p < .001). Koreans reacted considerably less favorably to those brands in commercials without differentiating messages than did the U.S. group. Thus, at least in low involvement product categories, it would appear that Koreans are less favorably disposed to commercials which do not contain brand differentiating messages than are Americans. _Given the findings for A-ad and A-brand, it is interesting that there is no significant differences between the U.S. and Korean means for the no brand differentiating condition (t = -.060; p = .551). The reason for this finding is the dramatic drop-off between A-brand and purchase intention in the U.S. sample. U.S. subjects seemed less willing to indicate that they plan to purchase a brand even if respond favorably toward the commercial or indicate a general liking for the brand than do Koreans. This phenomenon may be associated with high levels of brand loyalty among Koreans. The same pattern is apparent in the brand differentiation condition (t = 7.62; p < .001). Koreans who indicate a liking for the brand are more likely to also indicate that they intend to purchase the product. In summary, the Koreans subjects responded less 134 favorably in terms of attitude toward the ad and attitude toward the brand for commercials without brand differentiating messages than did U.S. subjects. When looking at commercials with brand differentiating messages, the Korean subjects showed higher purchase intention. As a whole, these results would suggest caution be taken when considering the use of brand differentiating messages in a standardized advertising campaign for a low involvement product in the two cultures. It appears that brand differentiation in Korea led to higher effectiveness ratings in general, while no substantial difference based on brand differentiation was found in the USA. While specific reasons for these finding will be speculated on in more depth when within country hypotheses are analyzed, it would appear that brand differentiation in these categories is a more effective strategy (or at least more possible to execute) in the ROK than it is in the USA. Hypothesis Three The third hypothesis predicted that the number of cognitive responses in the four primary treatment conditions would not differ for the Korean and U.S. samples. The overall mean number of cognitive responses was 1.52 per subject in the United States compared to 1.63 for the Korean subjects. In general, the four conditions did not produce departures from these means. As indicated by Tables 12 and 135 13, the largest difference between number of cognitive responses was in the low information content condition, in which the Koreans averaged 1.63 cognitive responses compared to 1.45 for the U.S. group. While this difference is statistically significant (t = -5.23; p < .001), it is doubtful as to whether the fact that Koreans gave 11 percent more responses per subject is practically meaningful, especially given the fact that the Koreans averaged 7 percent more cognitive responses across all conditions. It would, however, appear that Koreans are slightly more prone to offering cognitive thoughts when the level of information is low. Koreans were about 10 percent more likely to offer cognitive responses for commercials with brand differentiating messages (t = -4.37; p < .001). In both cases, the difference in not large enough to clearly suggest strategic implications. While the no brand differentiating message category (t = -1.96; p = .05 ) also produced statistically significant results, the means are very similar to those for the overall samples (1.54 vs. 1.52 in the USA and 1.61 vs. 1.63 in the USA). In the high information condition, the difference in means was not statistically significant (t = -1.13; p < .259). While the null hypothesis of no differences in means could not be accepted in three of the four cases above, the results do not suggest that there are substantial differences between the overall number of cognitive 136 responses in the USA based on the level of information and presence (or lack of presence) of a brand differentiating message. TABLE 12 Number of Cognitive Responses by Country and Information Level a) High Information Mean Mean U__SA ____ROK __d . f - tie; £1911: Number of: Cognitive 1.59 1.63 2018 -1.13 .259 responses Source supports 0.85 0.62 2014 6.51 <.001 Source derogations 0.27 0.53 2018 -8.96 <.001 Support arguments 0.27 0.21 2017 2.85 .004 Counterarguments 0.08 0.14 2018 -4.13 ‘ <.001 b) Low Information Mean Mean USA ROK gyfy t-val. sig. t Number of: Cognitive 1.45 1.63 2018 -5.23 <.001 responses Source supports 0.56 0.71 2018 -4.43 <.001 Source derogations 0.35 0.49 2018 -4.55 <.001 Support arguments 0.27 0.16 2018 4.70 <.001 Counterarguments 0.11 0.15 2018 -2.26 .024 137 Hypothesis Pour a) information content The fourth hypothesis predicted that there would not be differences in the mean number (per subject) of the four specific types of cognitive responses coded for each treatment condition. Table 12 reports the results of t- tests comparing the U.S. and Korean means for both the high and low information conditions. Interestingly, the U.S. subjects were more likely to provide source support statements for commercials containing high information levels than were their Korean counterparts (t = 6.51; p < .001), while the reverse was true in the low information condition (t = -4.43; pi< .001). Apparently, there is a positive relationship between information level and source support arguments in the USA, while the relationship is negative in the ROK. U.S. subjects gave about 37% more source support statements to commercials with high information condition than they did to those with low levels, while Koreans gave almost 27% more source supports to the commercials with low information levels. It may be that while U.S. viewers tend to view information as being useful (if not always persuasive), Koreans might sometimes find high levels of information to be associated with a "hard sell" type of approach, which is inconsistent with normal high context communication patterns in their culture. Alternatively, higher levels of trust in large companies 138 based on collectivistic values may lead Koreans to treat commercials that provide high levels of information with a certain level of skepticism. In any case, if maximizing the number of source supports (or, more generally, positive cognitive responses about the advertiser or advertisement) was the primary goal of the advertiser, these results would suggest that different levels of information be utilized in advertising in the two countries. The most striking feature of the results for the two information categories with regard to source derogations is that Koreans were more likely to give this type of response in both the high (t = -8.96; p < .001) and low conditions (t = -4.55; p < .001). For high information commercials, Koreans were twice as likely to give source derogations (.53 vs. .27 per subject) than were the U.S. subjects, while the gap was not quite as wide for low information commercials (.49 vs. .35). Hence, it can be concluded that the Korean subjects are more likely to give source derogations in general in comparison to U.S. subjects, but are even more prone to do so for commercials with high levels of information. Again, it seems that information level and positive cognitive response are positively related in the United States, but not in the Republic of Korea. For support arguments and counterarguments, which focus on reaction to specific claims or implications of a commercial, the U.S. subjects were somewhat more likely to 139 provide support arguments in the high information condition (.27 vs. .21 per subject; t = 2.85; p = .004) while Koreans were slightly more likely to give counterarguments (.14 vs. .08 per subject; t = -4.13; p < .001). The same relationships held in the low information condition, with U.S. subjects providing more support arguments per subject (.27 vs. .16; t = 4.70; p < .001), but less counterarguments per subject (.15 vs. .11; t = -2.26; p = .024). Overall, these differences, while statistically significant, are not large. The basic finding is that U.S. subjects gave more support arguments under either condition, while the Koreans gave more counterarguments. b) brand differentiation Table 13 shows the results of the t-tests comparing the mean number of each type of cognitive response for the U.S. and Korean samples in the brand differentiating message present and not present categories. When a brand differentiating message was present, Koreans were more likely to offer source supports (t = -3.53; p < .001), source derogations (t = -5.40; p < .001), and counterarguments (t = -1.18; p = .238), but less likely to offer support arguments than were their U.S. counterparts (t = 5.48; p < .001). When a brand differentiating message was not present, Koreans were more likely to offer source derogations (t = -8.00; p < .001) and counterarguments (t (t = 5.95; p < .001), 0031.). 140 -5.39; p < .001), but less likely to offer source supports or support arguments (t = 2.16; p = TABLE 13 Number of Cognitive Responses by Country and Brand Differentiation a) Brand Differentiating Message Present Number of: Cognitive responses Source supports Source derogations Support arguments Counterarguments b) Brand Differentiating Message not Present Number of: Cognitive responses Source supports Source derogations Support arguments Counterarguments Mean USA 1.49 0.66 0.29 0.28 0.12 Mean USA 0.75 0.34 0.26 0.07 Mean ROK 1.65 0.78 0.43 0.15 0.14 Mean 395_ 0.54 0.58 0.21 0.15 d.f. 2018 2016 2018 2017 2018 d.f. 2018 2016 _2018 2018 2018 t-val. si . t -4.37 <.001 -3.53 <.001 -5.40 <.001 5.48 <.001 -1.18 .238 t-val. sig. t -1.96 .050 5.95 <.001 -8.00 <.001 2.16 <.031 ~5.39 <.001 141 The most interesting result here is the greater tendency of the Koreans to offer source support for those commercials with brand differentiating messages while being less likely than U.S. subjects to provide source support for commercials without brand differentiation. In fact, Koreans gave 18% more source supports to commercials with brand differentiating messages than did the U.S. group. At the same time, Koreans gave 39% less source supports to commercials without brand differentiation than did the U.S. group. Thus, it appears that the presence of a brand differentiating message has a more favorable impact on the number of source support arguments in the ROK than it does in the USA. For source derogations, support arguments and counterarguments, the results do not provide such a sharp contrast. Koreans were, however, somewhat less likely to offer source derogations when there is a brand differentiating message in the commercial than when there was not, reinforcing the idea that more positive response is generated among Koreans by including a brand differentiating message. Compared to the U.S. subjects, Koreans were more likely to give source derogations in both the present and not present conditions. Similarly, U.S. subjects were generally more likely to give source support regardless of treatment, and Koreans averaged slightly more counterarguments under both conditions. 142 TABLE 14 United States - Effectiveness by Level of Information F— Value Stat1§tic sigi F a) Multivariate Test (MANOVA) Information Hotelling's .04 26.2 <.001 Wilks' .96 26.2 <.001 b) T-tests Mean Mean High Low d.f. t-val. s1 . t A-Ad 2.81 3.34 2018 7.78 <.001 A-Brand 2.67 3.09 2014 6.49 <.001 Purchase 3.64 4.29 2015 6.90 <.001 Intention Taken together, the results for the average number of individual types of cognitive responses indicate that there are differences between the ROK and the USA based on type of cognitive response. Koreans appear to have somewhat more favorable opinions of the source under the low information and brand differentiation conditions than do U.S. viewers. For the most part, from a managerial standpoint, the magnitude of the differences in means was not overwhelmingly large. Yet these differences are large enough to suggest that extreme caution be taken in considering standardizing 143 advertising between the two countries. It is apparent that similar strategies can lead to different types of cognitive response in the two countries. This finding is consistent with the findings across the two countries for the A-ad, A- brand, and purchase intention measures. within Country Hypotheses Hypothesis Five Hypothesis five predicted that the U.S. commercials containing low levels of information would be more effective than commercials containing high levels of information. As can be seen from Table 14, this hypothesis cannot be accepted. While the results of the MANOVA are statistically significant (Wilks' lambda = .96; f < .001) the means for all three of the dependent variables are in the opposite direction of what was predicted by the elaboration likelihood model. The U.S. subjects gave substantially higher mean rankings to the commercials containing high levels of information content. For attitude toward the ad (t = 7.78; p < .001), the difference was over one-half of a scale point, with the high information condition having a mean of 2.81 compared to 3.34 for the low information condition. For attitude toward the brand, the gap was a bit narrower, but still substantial (High 2.67; Low 3.09; t = 6.49; p < .001). The largest gap was in purchase intention I where the commercials with higher information content were 144 rated 3.64 compared to 4.29 for the low information commercials (t = 6.90; p < .001). TABLE 15 Korea - Effectiveness by Level of Information F- Value Statistic 519. F a) Multivariate Test (MANOVA) Information Hotelling's .01 4.5 .004 Wilks' .99 4.5 .004 b) T-tests Mean Mean High Low d.f. t-val. sig. t A-Ad 3.47 3.41 2018 -0.77 .441 A-Brand 3.23 3.20 2018 -0.48 .635 Purchase 3.78 3.50 2018 -3.26 .001 Intention These results suggest that U.S. viewers tend to like high levels of information in commercials for low involvement products. This finding seriously calls into question the ELM's prediction that consumers will rely on peripheral cues in advertisements for low involvement products. While it is possible that heightened attention was given to the commercials by the subjects in the 145 experimental situation, thereby enhancing favoritism for informative commercials, it is doubtful that such an effect could account for the wide gap in the ratings. At a minimum, the results suggest that if U.S. viewers pay attention to the commercials they watch they will tend to prefer those that contain more information cues. It should also be noted that this result is consistent with what would be predicted by the cultural variables listed in Appendix 2. If U.S. residents live in a low context culture in which individualism and directness are highly valued, it is not surprising that advertising in which the benefits of the product are clearly articulated and specific, logical reasons for purchase are offered would be relatively effective. Advertising containing several information cues is likely to have these characteristics. Hypothesis Six Table 15 shows the results of the tests based on level of information for the Korean sample. Based on the MANOVA results, hypothesis six, which predicted greater effectiveness for low information commercials is accepted (Wilks' lambda = .99; p = .004). However, as the t—test results show, the differences in means for each of the three dependent measures are generally not large. In fact, for both attitude toward the ad (t = -0.77; p = .441) and attitude toward the brand (t = -0.48; p = .635), the 146 differences in means, while in the predicted direction, did not produce statistically significant results. For purchase intent, low information commercials received a mean value of 3.50 compared to 3.78 for those with high information. This was a statistically significant difference (t = -3.26; p = .001), and is consistent with the prediction of the elaboration likelihood model. However, when this result is compared with the U.S. results, it seems likely that some other explanation for the greater receptiveness of Koreans to commercials with lower information content may be more plausible. The ELM has not been presented as a culture- bound model. Indeed, if there is a cultural bias, it is toward a U.S. perspective. Thus, especially given the relatively small differences in means, it does not seem appropriate to attribute the Korean response pattern to the behavior predicted by the Elaboration Likelihood Model. The cultural dimensions listed in Appendix 2 may offer some insight into these results. Since Korea is a higher context culture in which collectivism and consensus are valued, it makes sense that less hold, more subtle approaches to advertising might be favored. Since commercials with less information tend to usually involve more subtle approaches, it makes sense that Koreans would prefer approaches with less information,-or, at least, be more receptive to them than U.S. viewers. 147 Hypothesis Seven Hypothesis seven predicted that U.S. subjects would rate commercials with brand differentiating messages higher than those without them. Table 16 shows the results of the MANOVA and t—tests. While the MANOVA produced statistically significant results, the means are in the opposite direction from what was predicted. For all three dependent variables, the subjects gave commercials without brand differentiating messages higher ratings. The difference was not statistically significant for attitude toward the ad (t = - 1.62; p = -1.62), but was for attitude toward the brand (t = 6.49; p < .001) and purchase intention (t = —4.96; p < .001). As Table 16 shows, the gap in means based on brand differentiation for A-ad and A-brand is not large. However, for purchase intention the margin by which commercials without a brand differentiating message was preferred was fairly substantial: they received a mean ranking of 3.73 compared to a mean of 4.20 for the commercials that had brand differentiating messages. It may be the case that this unexpected result is in part due to the fact that only low involvement products were included in the sample. It is particularly difficult to brand differentiate low involvement products since competing products tend to have somewhat similar features. The competitive nature of the U.S. advertising industry probably serves to compound the problem, as rival 148 TABLE 16 United States - Effectiveness by Brand Differentiation F— Value1 Statigtic _sig. F a) Multivariate Test (MANOVA) Information Hotelling's 5.55 3713.2 <.001 Wilks' .15 3713.2 <.001 b) T-tests Mean Mean with w/out BDM BDM d.f. t-Val. Si . t A-Ad 3.13 3.02 2018 -1.62 .106 A-Brand 3.03 2.73 2014 6.49 <.001 Purchase 4.20 3.73 2015 -4.96 <.001 Intention products/firms/ad agencies can act quickly in response to a new product feature or advertising campaign. The measurement of brand differentiation, performed in an academic setting, may also play a role in this result. The results of the manipulation check indicated that while subjects perceived some level of difference in those commercials coded as having and not having brand differentiating messages, they do not a1yay§ perceive brand differentiation in exactly the way that trained coders do. This is probably, at least in part, due to the fact that the subjects in this study saw each commercial only once, while 149 those who coded the commercials could review the commercial multiple times to see if the operational definition of "brand differentiating message" was met. It may be that it is simply difficult to pick out brand differentiation in some cases when exposed to a commercial only once, particularly for low involvement products. Because of the unexpected nature of the finding, the possible effects of purchasing from the product category and prior exposure on reaction to brand differentiation were tested (See Table K in Appendix 3). Generally, the results of these analyses suggest that neither variable substantially alters the results. While the gap in means narrowed a bit for purchase intention if only those who have purchased from the product category were considered, and widened a bit when only those previously exposed were considered, the direction of the results did not change. In each case, attitude toward the ad itself was almost identical, while a statistically significant difference existed for A-ad and purchase intention. Overall, the magnitude of the differences does not suggest recommending against the use of brand differentiating strategies in U.S. advertisements for low involvement products. However, the findings do suggest that brand differentiation alone, even if it can be achieved, does not guarantee success and that other strategic variables, including information content, should be closely examined when considering what makes 150 advertising effective in low involvement product categories. Hypothesis Eight Table 17 shows that, as predicted, Korean commercials which contain brand differentiating messages were judged to be more effective than those that did not, as predicted. MANOVA results were highly significant (Wilks' lambda = .12; f < .001), and the means for all three of the individual measures varied in the predicted direction. The difference was particularly large for attitude toward the ad (t = 9.31; p < .001), with commercials containing a brand differentiating message receiving an average rating of 3.12 compared to 3.75 without. The difference in attitude toward the brand measures was not statistically significant at the .05 level (t = 1.67; p = .094). For purchase intention, a moderate (3.50 vs. 3.78) and statistically significant gap (t = 3.41: p = .001) was found. These results suggest that, at the present time, including a brand differentiating message in a Korean television commercial for a low involvement product will enhance its chance of being effective. Based on conversations with Korean advertising executives familiar with U.S. advertising practices, it is easier to differentiate a low involvement product through advertising in the ROK than is the case in the USA. Though Korean 151 TABLE 17 Korea - Effectiveness by Brand Differentiation F- F- Valuep spgpiptic :gig. F a) Multivariate Test (MANOVA) Information Hotelling's 7.29 4899.2 <.001 Wilks' .12 4899.2 <.001 b) T-tests Mean Mean with w/out BDM BDM d.f. t-val. si . t A-Ad 3.12 3.75 2018 9.31 <.001 A-Brand 3.17 3.26 2018 1.67 .094 Purchase 3.50 3.78 2018 3.41 .001 Intention corporations have been well established now for over 30 years in many cases, separate marketing and advertising departments are relatively new. As was once the case in the USA, some Koreans in the advertising industry complain that advertising has traditionally been viewed as an expense. As a result, until very recently, the advertising industry has lagged in terms of creative execution and competitive vigor. The industry has usually been characterized by imitation rather than innovation, and the concept of brand management has not been widely implemented. Under these conditions, it 152 is probably easier to successfully devise a brand differentiating message than is the case in the United States. Further, brand differentiating strategies may be viewed as novel, and a relief from typical, often similar approaches which have traditionally been used by Korean advertisers. Apparently, those advertisers who do brand differentiate in Korea tend to be rewarded with more favorable consumer response. The results reported here clearly suggest that including a brand differentiating message in Korean television commercials for low involvement products probably enhances the chances of producing an effective ad. Hypothesis Nine The ninth hypothesis suggested that in the USA low information levels would be associated with higher levels of source-based responses, while high information levels would be associated with relatively more product based arguments. Table 18 shows the results of the t-tests which tested these assertions. For the source based arguments, the high information commercials produced considerably more source support arguments (t = -8.47; p < .001), but less source derogations (t = 3.33; p = .001). While these results lead to a the rejection of hypothesis 9a, they may provide insight into the apparent greater effectiveness of high 153 information commercials in the U.S. sample. That commercials with high information levels generated more positive responses about the source but fewer negative statements about the source suggests that informing the viewer tends to enhance the chance of favorable cognition on the part of the viewer. Favorable cognition, in turn, is likely to be associated with higher effectiveness ratings. TABLE 18 United States - Cognitive Response by Information Level Mean Mean High Low Info Info d.f. t-value 1§ig. t Source supports 0.85 0.56 2016 -8.47 <.001 Source derogations 0.27 0.35 2018 3.33 .001 Total source 1.12 0.92 2016 -5.88 <.001 supports and source derogations Support arguments 0.27 0.27 2017 -0.24 .808 Counterarguments 0.08 0.11 2018 2.30 .021 Total support 0.35 0.38 2017 0.98 .328 arguments and counterarguments 154 The cognitive responses which are more closely related to the product itself show little difference based on information level. The number of support arguments was virtually identical for the high and low categories at .27 per commercial (t = -0.24; p = .808). The number of counterarguments were also very similar at .08 for the high category compared to .11 for the low category (t = 2.30; p = .021). Thus, hypothesis 9b is also rejected. In summary, it would appear that in the USA commercials containing high levels of information produce more positive statements about the source than commercials with low information, but that there is very little difference in product based responses based on information level. Hypothesis Ten Table 19 shows the mean frequency of cognitive responses by information level for the Korean sample. Interestingly, the results for the source based responses are in the opposite direction of those in the U.S. sample. Consistent with what was predicted by hypothesis 10a, low information commercials produced more source supports than high information commercials by a margin of .71 to .62 per commercial (t = 2.50; p = .010). High information commercials, however, produced more source derogations by a slim and not statistically significant margin (t = -1.25; p = .210). Thus, overall, hypothesis 10a must be rejected. 155 TABLE 19 Korea - Cognitive Responses by Information Level Mean Mean High Low Info Info d.f. t-value_1§iq. t Source supports 0.62 0.71 2016 2.50 .010 Source derogations 0.53 0.49 2018 -1.25 .210 Total source 1.15 1.20 2016 1.30 .194 supports and source derogations Support arguments 0.21 0.16 2018 -2.34 .019 Counterarguments 0.14 0.15 2018 0.59 .557 Total support 0.35 0.31 2018 -1.42 .154 arguments and counterarguments In contrast to the U.S. subjects, the Koreans gave more positive and fewer negative source based responses to commercials with low information levels. Apparently, Koreans are more likely to think positive thoughts about a source when a lower-keyed approach is used, especially in comparison to U.S. viewers. This result is consistent with the differences in the contextual level of the two cultures and the favoring of less direct communication approaches by Koreans. As in the USA, there was little difference in the number of product related arguments based on the level of information provided in the ROK sample. For both support 156 arguments (t = —2.34; p = .019) and counterarguments (t = 0.59; p = .557), the raw difference in means was quite small. Thus, hypothesis 10b cannot be accepted. Hypothesis Eleven As shown in Table 20, hypothesis 11, which predicted that the presence of a brand differentiating message would produce more total cognitive responses in the U.S. sample is rejected at the .05 level of significance (t = 1.91; p = .056), since commercials without brand differentiating messages actually produced slightly more cognitive responses. Raw differences in means based on brand differentiation were small, although brand differentiation showed slightly lower levels of source based statements and slightly higher levels of product related arguments. As with overall levels of effectiveness, this finding may be best explained by the difficulty of getting the consumer to perceive brand differentiation in a short commercial given a highly competitive environment. Alternatively, it may be that brand differentiation simply does not produce more cognition when included in commercials for low involvement products. 157 TABLE 20 United States -cognitive Responses by Brand Differentiation No BDM BDM d.f. t-value sig. t Source supports 0.66 0.75 2016 2.51 .012 Source derogations 0.29 0.34 2018. 2.03 .043 Support arguments 0.28 0.26 2017 -0.63 .527 Counterarguments 0.12 0.07 2018 -3.53 <.001 Total 1.34 1.41 2016 1.91 .056 Hypothesis Twelve As in the USA, overall levels of cognitive response did not vary significantly (t = -0.72; p = .474) based on the presence of a brand differentiating message. Thus, hypothesis 12 is rejected. However, a somewhat different response pattern was present in the ROK. Commercials with brand differentiating messages generate considerably more source support arguments (.79 vs. .54 for no brand differentiation) and considerably fewer source derogations (.43 vs. .58). This finding suggests that in including a brand differentiating message in a Korean television commercial can often improve effectiveness, since they tend to induce more positive source-based cognitive responses and fewer negative source-based responses. As was also 158 evidenced by the scaled measures of effectiveness, it appears that there are some advantages to using brand differentiating strategies in Korean television commercials for low involvement products. TABLE 21 Korea - Cognitive Responses by Brand Differentiation No BDM BDM d.f. t-value sig. t Source supports 0.79 0.54 2016 -6.91 <.001 Source derogations 0.43 0.58 2018 2.03 <.001 Support arguments 0.16 0.21 2018 2.75 .006 Counterarguments 0.14 0.15 2018 0.47 .638 Total 1.51 1.48 2016 -0.72 .474 Hypothesis Thirteen Hypothesis 13 predicted that commercials with a brand differentiating message and low information content would get the highest mean ratings while commercials with no brand differentiating message and high information would be ranked the lowest in the U.S. sample. Table 22 shows that neither of the predictions are verified. The low information/brand differentiating message present cell, in fact, got the 159 lowest ratings, ranking fourth (out of the four information X brand differentiating message cells) on both attitude toward the ad and attitude toward the brand and third on purchase intention. Meanwhile, the high information/no brand differentiating message cell was highly rated: first on A-ad and second on A-brand and purchase intention. The results of the MANOVA and individual univariate tests for differences between the four cell means provide more insight into this finding. The MANOVA shows strong statistical significance for the main effects of each dependent variable (information Wilks' lambda = .96; brand differentiating message Wilks' lambda = .98; p < .001 in each case) as well an interactive brand differentiating message X information effect (Wilks' = .95; p < .001). As discussed earlier, the U.S. subjects showed a preference for commercials with high levels of information. This effect is strong enough that the high information cells are consistently ranked number 1 and 2 for all three dependent measures. Additionally, the univariate tests shown in Table 22 indicate significant effects based on information for each dependent measure (p < .001 in all cases) for each measure. In spite of the statistical significance, the effects of having a brand differentiating message are not as striking. Cells with brand differentiation ranked first and fourth for A-ad, second and fourth for A-brand, and second 160 and third for purchase intention. In the univariate tests, statistically significant results were obtained for A-brand (f = 21.2 p < .001) and purchase intention (f = 26.2; p < .001), but not for A-ad (f = 2.3; p = .118). For all variables, however, the presence of a brand differentiating message alone did not produce differences in means large enough to suggest managerial relevance. The exact nature of the information X brand differentiation interaction appears to vary by dependent measure. For attitude toward the ad, the combination of a brand differentiating message and high information appears to have a synergistic effect, producing the highest ratings, while low information/no brand differentiation produces the lowest rating. For attitude toward the brand, the univariate test shows no interactive effect (f = 0.02; p = .909). The information effect alone is very strong for A- brand, with high information cells ranking first and second. The substantially higher mean rating (3.13) in the high information/no brand differentiating message cell compared to the others appears to account for the interactive effect 161 ‘ TABLE 22 United States - Effectiveness by Information Level and Brand Differentiation a) Multivariate F- Test (MANOVA) Value Statistic sig. F Information Hotelling's .04 26.2 <.001 Wilks' .96 26.2 <.001 Brand Diff. Message Hotelling's .02 10.4 <.001 Wilks' .98 10.4 <.001 Information X Brand Diff. Message Hotelling's .04 28.8 <.001 Wilks' .95 28.8 <.001 b) Cell Mean Mean Mean Means A-Ad Rank A-Brand Rank P.Intent Rank High info/ 2.73 1 2.81 2 4.16 2 BDM High info/ 2.90 2 2.53 1 3.13 1 No BDM Low info/ 3.52 4 3.23 4 4.25 3 BDM Low info/ 3.15 3 2.94 3 4.33 4 No BDM c) Univariate tests F- d.f. Statistic sig. f Information A-Ad 2009 61.6 <.001 A-Brand 2009 42.4 <.001 Purch. Intent 2009 47.9 <.001 Brand Diff. Message A-Ad 2009 2.3 .118 A-Brand 2009 21.2 <.001 Purch. Intent 2009 26.2 <.001 Information X Brand leferentiating Message A-Ad 2009 15.5 <.001 A-Brand 2009 0.0 .909 Purch. Intent 2009 35.3 <.001 162 for purchase intention. The means of the other three cells are actually quite closely bunched, ranging only from 4.16 to 4.33. This finding for purchase intention is surprising, since the cell predicted to rate last actually was rated first. As a whole, the findings for the U.S. sample suggest that the elaboration likelihood model itself and conventional wisdom regarding the brand differentiating message do not adequately predict consumer response to television commercials for low involvement products. The subjects clearly preferred the commercials with more information, and appeared to be indifferent to the presence of a brand differentiating message. In comparison to Korea, the finding on information level makes sense -- more direct forms of communication tend to be preferred in the USA. However, more investigation of the factors driving this finding is warranted. Advertising has long been criticized by some for focusing on persuading rather than informing. This finding suggests the advertisers are generally better off trying to inform. For the brand differentiating message results, it would appear that advertisers may be being forced from the head of the pin on which they try to dance in attempting to differentiate low involvement products. With relatively few Objective differences in product features present in these categories and high levels of competition, brand 163 differentiation through advertising strategy may be becoming less and less possible. Hypothesis Fourteen Hypothesis 14 predicted that commercials with a brand differentiating message and low information content would receive the highest mean ratings from Korean subjects while high information/ no brand differentiating message commercials would be ranked the lowest. Table 23 shows that the high information/no brand differentiating message was indeed the cell with the lowest rating on A-brand and purchase intention and was second to last on A-ad. The low information/brand differentiating message present cell, predicted to be rated the highest, was ranked first for A- ad, but only third for A-brand and purchase intention. Thus, the predictions were accurate in some instances, but not in others. The MANOVA shows significant main effects for information (Wilks' lambda = .99; p = .004) and brand differentiation (Wilks' lambda = .95; and an interaction between the two (Wilks' lambda = .97; p < .001). Clearly, the presence of a brand differentiating message was the dominant effect. For A-ad, the brand differentiating message present cells ranked first and second, and for A- brand and purchase intention, first and third. The presence of a brand differentiating message was associated with 164 TABLE 23 Republic of Korea - Effectiveness by Information Level and Brand Differentiation a) Multivariate F- Test (MANOVA) Value Statistic sig. F Information Hotelling's .01 4.5 .004 Wilks' .99 4.5 .004 Brand Diff. Message Hotelling's .06 37.7 <.001 Wilks' .95 37.7 <.001 Information X Brand Differentiating Message Hotelling's .04 23.6 <.001 Wilks' .97 23.6 <.001 b) Cell Mean Mean Mean Means A-Ad Rank A-Brand Rank P.Intent Rank High info/ 3.18 2 3.05 l 3.40 1 BDM High info/ 3.76 3 3.41 4 4.15 4 No BDM Low info/ 3.07 1 3.29 3 3.59 3 No BDM Low info/ 3.76 4 3.14 2 3.42 2 BDM c) Univariate tests F- d.f. Statistic sig. F Information A-Ad 2009 0.6 .432 A-Brand 2009 0.2 .629 Purch. Intent 2009 10.9 .001 Brand Diff. Message A-Ad 2016 86.6 <.001 A-Brand 2016 2.8 .093 Purch. Intent 2016 11.9 .001 Information X Brand Differentiating Message A-Ad 2016 0.8 .348 A-Brand 2016 22.6 <.001 Purch. Intent 2016 30.3 <.001 165 higher rankings for all three measures, at statistically significant levels for A-ad (f = 86.6; p < .001) and purchase intention (f = 11.9; p = .001), but not for A-brand (f= 2.8; p = .093). While the low information cells tended to have slightly higher levels of mean effectiveness than those with high information, the difference was much smaller than for brand differentiation. In fact, the univariate tests indicate a statistically significant effect only for purchase intention (f = 10.9; p = .001). i The information X brand differentiation interaction was statistically significant for A-brand (f = 22.6; p < .001) and purchase intention (f = 30.3; p < 4001): bUt not A-ad (f = 0.8; p = .348). In the case of A-brand and purchase intention the combination of high information and a brand differentiating interacted to lead to higher effectiveness, while high information and no brand differentiation was clearly the least effective cell. These results indicate differences in the way the Korean and U.S. subjects reacted to the independent variables studied. For low involvement products, brand differentiation appears to be an effective technique in the ROK. Additionally, Koreans seem to be more likely than U.S. viewers to react positively to commercials which contain low levels of information. While this latter result seems consistent with the elaboration likelihood model's 166 prediction, the effect within the ROK sample for information does not seem large enough to claim support for it. Further, since ELM is a U.S. based theory, it would be illogical to suggest that it applies in the ROK but not in the USA. A more plausible explanation for the differences found with respect to the information variable would be the cultural differences based on contextual level, and the relative value places on collectivistic and/or harmony seeking behaviors. For the brand differentiating message result in the ROK, the best explanation may be the greater plausibility of successfully executing this strategy given Korean market conditions and the newness or novelty associated with the approach. Possible Covariation Variables considered as possible covariates prior to conducting the study were age, major, purchase from product category, prior exposure, and prior attitude toward the brand. The effects of each of these variables will now be discussed. Age and Major As shown in Table 1, there was very little variance in the age of the subjects included in both samples. In the USA, 97% of participants were aged 19 through 23, and 96% of the Korean sample also fell into this range. For this 167 reason, there is no theoretical basis for treating age as a covariate for the purposes of analysis. There is simply no reason to expect that women at the high and low ends of the age range will systematically differ in response to the advertisements. For this reason, results of MANOVA's in which age was treated as a covariate (no significant differences were found) are not reported. As with age, the type of majors students were drawn from did not dramatically differ. In the ROK, over 89% of the sample consisted of business majors. In the USA, 96% of the sample majored in business of a business related field (this includes merchandise management majors, who primarily study retailing). Again, there is no sound basis for expecting covariation given the mix of majors present and, indeed, no differences in the analyses based on major were found. Prior purchase from product category Table K in Appendix 3 shows the mean rankings of U.S. subjects who indicated that they had purchased from the advertised product category within the past year. Generally, prior purchase led to slightly better rankings on each dependent variable. However, the differences in means based on the independent variable cells were similar to that of the overall sample. For brand differentiation, overall, the subjects rated 168 commercials with brand differentiating messages slightly lower than those without. When only those who purchase from the product category are considered, the same relationship holds, although the margin by which no brand differentiation was preferred was even narrower than in the overall sample. Thus, the general conclusion drawn -- that brand differentiation was not associated with a substantial difference in means -- does not appear to have been affected based on prior purchase from the product category. As in the overall sample, high information commercials were clearly preferred by subjects who purchase from the product category. Indeed, as shown in Table K, the differences in the high and low information categories are very similar in magnitude to that of the whole sample. T- tests indicate that all of the differences between the high and low categories are significant at p < .001. As in the USA, prior purchase does not appear to have biased results in the ROK. As can be seen from Table L (See Appendix 3), brand differentiating messages were preferred by a substantial margin by both those who purchase form the category in terms of attitude toward the ad and attitude toward the brand. In both cases, attitude toward the brand appears to have been unaffected by the use of a brand differentiating strategy. For information level, as found in the overall sample, subjects who purchased from the product category expressed a 169 slight preference for commercials with low levels of information in terms of attitude toward the ad. For a- brand, information level appears to have not made a difference among either purchasers or the overall sample. One difference that is noticeable is that low information commercials had a more positive impact on the overall sample (mean = 3.50 vs. 3.78 for high information category) among those who purchase from the category (3.08 for low vs 2.97 for high information). Apparently, the level of information present made little difference to the Korean subjects who purchased from the product category in terms of purchase intention, and less information tended to be received favorably from those who had not purchased from the product category. This result is not particularly surprising, since both experience with the product and intensity of brand loyalty likely play major roles in Koreans' intention to purchase. However, the finding does reinforce the earlier caution given about making a blanket generalization about differences in information level leading to differences in effectiveness. Prior exposure As with prior purchase, prior viewing of the commercial generally tended to be associated with slightly higher effectiveness ratings in the U.S. sample. However, the differences in means associated with information level and 170 brand differentiation categories generally were very small. The presence of a brand differentiating message was associated with slightly lower ratings for A-ad and A-brand among those who had seen the commercial previously. For purchase intention, where the gap was widest, those previously exposed favored the commercials without a brand differentiating message by an even wider margin (4.16 - 3.47 .69) than was the case in the overall sample (4.20 - 3.73 = .47). Apparently, in terms of purchase intention only, those who have not seen a commercial previously were more likely to be swayed by brand differentiation. The means by information category for those who had seen the commercial previously do not show any substantial differences from the overall sample. Again, high information commercials were rated higher in every category at statistically significant levels. In the ROK, brand differentiating messages were associated with higher ratings on A-ad and purchase intention, but they made little difference in A-brand ratings. This finding echoes that of the overall sample. The same general statement can be made about information level, although on purchase intention, the gap in favor of low information was somewhat wider (.59) than in the overall sample. Generally, the examination of the mean ratings based on a) subjects who purchase form the advertised product 171 category and b) subjects who had seen the commercial previously does not suggest that these variables skewed the results of the overall samples, at least not on a large scale. While a few differences were slightly intensified with regard to purchase intention, there do not appear to be major differences in consumer reaction based on purchase from the category or prior exposure in either country. Prior attitude toward the brand To get some idea of general attitudes toward the brands utilized in the study, in both countries a questionnaire was given to a group of subjects separate from those who participated in the experiment. Those subjects who participated in the main study were not surveyed in this way due to fear of sensitizing them to the stimuli in advance of the experiment. In the USA, the results of the survey strongly suggest that differences in prior attitudes toward the brand did not bias the results. The subjects rated the brands included in the high information cells at an average of 3.23. This figure is virtually identical to the 3.27 given to the brands included in low information cells. Commercials that had no brand differentiating messages were rated at 3.16 compared to 3.34 for those that had a brand differentiating message. It is very doubtful that this gap plays much of a role in explaining the overall results. In general, brand 172 differentiation did not produce substantial differences in mean ratings of effectiveness. Since commercials without brand differentiating messages received slightly higher ratings in the overall study, this slight gap seems to reinforce the idea that the U.S. subjects seemed indifferent to brand differentiating messages. Larger differences in prior brand attitudes were present in the ROK data. The brands in high information commercials received a mean rating of 3.51 compared to a mean of 3.69 for those for low information commercials. This finding suggests the possibility of the high information category commercials being rated slightly higher based on prior brand attitudes. If this were the case, commercials with low information levels would be rated higher relative to those in the high category. However, the magnitude of the difference in prior brand attitudes does not suggest that this had a major impact on the results. For brand differentiation, brands included in the brand differentiation present cell averaged a rating 3.44 compared to 3.75 for those without. This result suggests a possible bias towards commercials which contained a brand differentiating message. However, it is very unlikely that this difference alone wholly accounts for the dramatic differences found based on brand differentiation. Additionally, it could well be that it is prior exposure to the brand differentiating messages serves to enhance A-brand 173 ratings. In summary, it seems very unlikely that the results of the U.S. study were contaminated by differences in prior brand attitudes, since large differences simply were not present. In the ROK, the differences, while not overly large, should be considered in interpreting results. It could be that if prior brand attitudes were constant, low information commercials would have gotten slightly higher means. Given the magnitude of the differences in prior brand attitudes, this effect might be noticeable, but in all likelihood, not large. For brand differentiation in Korea, there may have been some bias towards those commercials with brand differentiation based on better prior brand attitudes. Yet, it is likely that this possible bias was a secondary, rather than primary contributor to the overall effects. CHAPTER 5 - CONCLUSION This chapter will begin with a discussion of the conclusions that can be drawn from the research questions which were posed in Chapter 1. First, implications for researchers and managers interested in the standardization controversy will be discussed. Second, conclusions relating to information content in advertising will be discussed. This will be followed by a summary of the findings pertaining to brand differentiating messages. Then, the focus will shift to the limitations of the study. Finally, suggestions for future research will be made. Implications for Standardization/Specialization Argument The results of this study suggest that advertising strategies based on information content and brand differentiating messages should not be standardized across the USA and the Republic of Korea. The results showed that the U.S. subjects preferred commercials with higher levels of information than their Korean counterparts. Meanwhile, brand differentiation seemed to have a positive impact on advertising effectiveness in the ROK, but not in the USA. These findings indicate that subjects in the two countries 174 175 responded differently to strategies based on information level and brand differentiation. Thus, at a micro level, the research suggests a need for advertisers operating in both the Korean or U.S. markets to tailor strategy relating to information and brand differentiation based on national preferences. More broadly, this research demonstrates the need to focus on consumer response in international research on standardized advertising. If appropriate procedures are followed, consumer response can be compared and specific strategic recommendations can be made. Content analysis alone does not allow for sufficient insight into why differences in advertising in two countries are present. For example, the content analysis performed on the database from which commercials were drawn for this study showed that only 3% of commercials in the ROK sample contained brand differentiating messages, compared to 30% in the USA. On first glance, this result might lead one to suspect that brand differentiation is not an effective strategy in Korea. Results of the experiment conducted here show otherwise. The content analysis showed that approximately the same number of information cues are present in advertising in the two countries. The quick conclusion would, thus, be that U.S. and Korean consumers react similarly to commercials with similar levels of information. The findings here refute this, indicating that Koreans are more likely to 176 respond favorably to commercials containing lower levels of; information in the product categories studied than are U.S. viewers. These results show what can be gained by moving beyond content analysis (which is often useful and necessary in preliminary hypothesis development) and shifting the focus of research to what consumers respond to. The need to follow procedures to enhance equivalence in international advertising research on consumer response cannot be stressed enough. In this study, the database of commercials was initially compiled via a procedure designed to ensure comparability. Then, the commercials were content analyzed via a process in which: coding schemes were carefully translated and backtranslated; coders were native speakers; coders were extensively trained; and reliability was measured. The fact that the commercials were content analyzed on a wide variety of variables provided for a unique opportunity to utilize a matching process which allowed for control of several important elements of executional strategy. When dependent measures were chosen, again a translation/back-translation process was followed, reliability was considered in pretests, and measures were taken to ensure the equivalence of experimental procedures. It is only through the use of these types of procedures that valid conclusions regarding executional strategy in two countries can be drawn. With regard to the differences found in consumer 177 response, it would appear that full explanations for their existence are complex. Fuller explanations for differences based on brand differentiation and information level will be attempted in the next section. However, the findings do support Lee's (1991) statement that U.S. theories of consumer behavior probably have limited usefulness in international research. It is very likely that cultural differences such as those listed in Appendix 2 influence communication patterns within a society. Thus, theories based on U.S. communication patterns alone may contain a cultural bias. The findings of this study support the need to consider cultural and other environmental factors (e.g., economic, political, social, and regulatory forces) when developing theory pertaining to consumer response to advertising or persuasive communication in international settings. SUMMARY OF FINDINGS Information Content a) USA The hypothesis tests relating to information content in each country provide some fascinating results. First, counter to the prediction of the Elaboration Likelihood Model, it is apparent that U.S. subjects favored commercials with high levels of information in the product categories studied. For all three scaled dependent measures, the 178 subjects preferred high information commercials at statistically significant levels. The content analyses of Resnik and Stern'(1977) and Dowling (1980) found that commercials for food/beverages and personal care products contain lower levels of information than those for other product categories. Thus, it is somewhat surprising that U.S. viewers appear to prefer commercials with more information in these categories. However, the preference for higher information is consistent with Ogilvy's (1963) contention that advertising which contains more information is generally more effective. It is also consistent with the Perrien, et al. (1985) finding that when surveyed, managers believed commercials with higher levels of information content were more effective. The obvious recommendation of the findings pertaining to information content to U.S. advertisers is that strategies incorporating high levels of information should be more strongly considered. If commercials with more information tend to be more effective, it makes sense to utilize them. A shift to commercials with more information might have the additional impact of allowing for a better response to critics who believe most advertising is designed to persuade rather than inform. Further, more information might help deal in part with the issue of consumer perception of the credibility of advertising identified by Barksdale and Perrault (1980). 179 Also notable is the finding that commercials with higher levels of information content received substantially more source support arguments. U.S. subjects tended to express more positive thoughts about the advertiser or execution of the commercial if there was more information in the commercial. Conversely, the U.S. subjects were more likely to give source derogations to commercials containing less information. These findings reinforce the suggestion that managers consider using level of information as a strategic tool. A few caveats regarding the above finding must be mentioned. First, results are based on objective measurement of information content rather than consumer perceptions of informativeness. While the manipulation checks indicated some tendency toward convergence of the two constructs, there are clearly differences between what consumers regard as information and what academic measures indicate is information. This difference suggests a need to treat consumer perception of informativeness as a separate construct in future studies. A second caveat with regard to the information findings is that only low involvement product categories were tested. It may be that high information levels are even more preferred in commercials for high involvement product categories. For this reason, some caution must be taken in dismissing the Elaboration Likelihood Model's predictions 180 altogether. However, the findings cast serious doubts on the ELM's ability to describe consumer response to persuasive communications in low involvement product categories. A final caution with respect to including more information in commercials relates to the Stewart and Furse (1986) finding that commercials with more information are negatively correlated with recall measures. While more information may enhance consumer perception of attitude toward the ad and brand and purchase intention, it may not enhance memorability. Thus, the goals of the advertiser must be taken into account in considering level of information from a strategic viewpoint. In spite of these cautions, the findings do suggest that more consideration be given to using high information levels in commercials for low involvement products. Additionally, as will be discussed below, the findings indicate that cultural differences may play a role in tendencies for consumers in the two countries to prefer high or low levels of information in persuasive communications. Clearly, the ELM, would suggest that when motivation to process a communication is low, peripheral route processing takes place and peripheral route cues are relied upon. The findings here suggest that information is appreciated by U.S. consumers even when motivation to process is relatively low. If the theory's predictions are not verified in the 181 USA, where it was developed, there is little reason to suspect that it would hold in another culture. b) Republic of Korea While information level did not produce the large differences in mean ratings for each dependent variable found in the U.S. sample, it appears that Koreans have a greater tolerance for low information levels in commercials in the product categories studied. Interestingly, the Koreans, on average, gave more source support statements and fewer source derogations to commercials containing low levels of information. The Korean subjects also tended to give more source derogations to the high information commercials. Both of these findings are in sharp contrast to those in the USA. Coupled with the clearly lower ratings Koreans gave to commercials with high information in comparison to the U.S. subjects, it is clear that consumer response based on the level of information in the two countries varies. It would appear that cultural explanations based on the contextual level of the culture, individualistic vs. collectivistic behaviors, and harmony seeking behaviors help to explain the results. Koreans gave slightly higher ratings for low information commercials on A-ad and A-brand and a larger preference on purchase intention. In contrast, U.S. subjects rated high information commercials higher on 182 all three measures. These findings make sense, given that Korean communication patterns have traditionally been less direct and more value is placed on low-keyed communication aimed at preserving harmony and strengthening ties within groups. Based on these cultural dimensions (discussed at more length in Appendix 2), one would expect Koreans to respond more favorably to commercials containing less information, since higher levels of information are associated with directness and serving the needs of individual preferences. Meanwhile, U.S. viewers would be expected to respond more favorably to high information commercials since they are more direct, aggressive, and aimed at individual preferences. Brand Differentiation a) United States The relative indifference of the U.S. subjects to the brand differentiation variable was the most unexpected finding in this study. To some extent this finding conflicts with the Stewart and Furse (1986) finding that brand differentiation is more correlated with advertising effectiveness than other executional variables. However, the Stewart and Furse study differed from this one on two important dimensions. First, two of Stewart and Furse's three dependent measures (related recall and key message comprehension) are recall based. No recall based measures 183 were utilized in this study. Thus, different dependent variables were utilized. If recall based dependent measures had been employed in this study, it is possible that the results may have been different. Second, the Stewart and Furse sample of U.S. commercials contained products from a wide variety of categories, including many for high involvement products. It is quite possible that brand differentiation is an effective strategy in commercials for high involvement product categories, or even in others that were not studied here. These differences from the Stewart and Furse study do not fully explain the findings. Conventional wisdom has suggested that brand differentiation is unambiguously a good strategy. Perhaps the strongest explanation for the results is product category or involvement based. For foods and health and beauty aids, the are relatively few objective differences in product features. Thus, objective means of brand differentiation in advertising (e.g., "Contains an ingredient that no other product has") may often not be available. Further, differentiation based on intangibles may be relatively easy for savvy competitors to duplicate, or at least offset. Thus, it may simply be difficult to successfully brand differentiate in commercials for low involvement products. It should also be remembered that the product categories used in this experiment were in late life cycle stages. It is probably easier to brand 184 differentiate in earlier stages of the life cycle, when competitors (and their advertising strategies) are not so firmly entrenched. Subject perceptions of brand differentiation must also be taken into account. Results of the manipulation checks suggest that one exposure to a commercial may not be enough for a viewer to perceive brand differentiation. While trained coders did not have difficulty identifying a brand differentiating message reliably, the subjects did not systematically agree with the coders. However, this phenomenon may be related to the lower motivation associated with processing commercials for low involvement products or the sheer difficulty of differentiating in these categories. For the above reasons, some caution should be taken in generalizing the findings for brand differentiating messages. It is not advisable to conclude that brand differentiation is generally an ineffective tool in U.S. advertising. The findings do, however suggest that it is very difficult to successfully brand differentiate in low involvement product categories, perhaps especially so when a product is at the mature stage of the life cycle. They also cast some doubt on the effectiveness of brand differentiation in commercials for low involvement products. Thus, the best advice for U.S. advertisers is be cautious in assessing what can be accomplished through brand differentiation in commercials for low involvement products, 185 particularly in those categories in which the product has few objective differences from its competitors. b) Republic of Korea The results of this study showed that the Korean subjects had a clear preference for commercials with brand differentiating messages. In fact, the effect was large enough to suggest that Korean advertisers strongly consider utilizing brand differentiation strategies in television commercials. It is important to speculate on why the effect of brand differentiation was so much more powerful in Korea than in the United States. One reason that has been suggested for the greater impact of brand differentiation on effectiveness in the ROK is the feasibility of implementing such a strategy. Brand differentiation is still new to Korean advertising, and, in large part, to Korean viewers. Korean firms were built primarily on large-scale manufacturing efficiency, and production was highly emphasized throughout the 60's and 70's. During this time, Korean firms were able to produce many products at a cost advantage, and hence were able to compete effectively on price or value in world markets. Marketing, while not wholly ignored, was not stressed as much as manufacturing efficiency. It was only recently that the Korean chaebols began to establish formal marketing departments, partly as a response to more competitive world 186 markets. U.S. educated Korean advertising professionals suggest that the Korean advertising industry has lagged behind the U.S. in terms of creative strategy and level of competition as a result. In fact, many Korean firms have been prone to copying the advertising strategies of competitors. It is very likely that this type of copying and a general lack of innovative approaches would leave more opportunity to differentiate products (even low involvement products) due to boredom on the part of consumers. Thus, it may well be easier to brand differentiate in Korean commercials than in U.S. commercials. In addition to greater ability to clearly differentiate the brand, a few other practical considerations may help explain the effectiveness of brand differentiating messages in the Korean sample. Related to the idea of a lack of innovation, is the fact that Korean consumers are exposed to few commercials with brand differentiating messages and find it to be a novel approach. The content analysis performed on over 800 Korean commercials indicated that only 3% contained a brand differentiating message. Thus, Korean viewers might find brand differentiation to be an innovative approach and, in turn, react positively to commercials that use this strategy. Conversations with Korean advertising professionals also suggest that brand management is a concept which has not been widely implemented in Korea. As a result, few 187 advertisers have run campaigns for given brands. Traditionally, the focus of Korean firms has been on the company's reputation rather than stressing the merits of a particular brand. The results of this study seem to suggest that a change toward more emphasis on brands in Korean advertising may be an effective strategy. On the whole, the results suggest that brand differentiation strategies should be more strongly considered by Korean advertisers. Clearly, the young female subjects utilized in this experiment responded significantly more favorably to commercials that contained a brand differentiating message. It should also be noted that conditions in the ROK's advertising environment are changing rapidly, and, thus, it may be advisable to attempt brand differentiation before it becomes an industry trend. If the entire industry begins to use the strategy, as seems to be the case in the USA, it becomes more difficult to effectively implement, especially in low involvement product categories. LIMITATIONS One limitation of this study is that it is virtually impossible to control for all of the possible variables that can have an impact on advertising effectiveness. The research does have the advantage of controlling for several key advertising executional variables (e.g., music, commercials length, comparisons, creation of a mood, type of 188 principal character) through a matching process incorporated in the experimental design. Clearly, this provides a greater measure of control than present in most previous advertising effects studies. An additional limitation of the study is based on the fact that it employs 20 commercials in each country -- a larger number than has been employed in most previous studies. Using this number of commercials provided the advantage of being able to calculate aggregate means in each cell in the design, allowing for a greater level of confidence that quirks in individual commercials were not biasing the results. However, with such a large sample and a large number of subjects, there are simply limits on how naturalistic of a viewing setting was feasible. It was not feasible to embed the ads in regular programming. Additionally, subjects are likely to pay more attention to the ads in the laboratory setting than they would in their own living rooms. Thus, some limitations on external validity based on the experimental setting must be acknowledged. The use of student subjects also places limits on the generalizability of the findings. It may be that people in different age groups or of different educational backgrounds may respond differently from the subjects surveyed here. At the same time, it should be noted that the subjects in each country would be regarded as representing an important 189 portion of the target audience for the product categories which were investigated. The use of a relatively large number of "real" commercials (taped off-the-air) also enhances external validity, since they more closely simulate what viewers see on a daily basis than artificially produced commercials would. The experimental setting also suggests a need for caution in terms of evaluating the Elaboration Likelihood Model's predictions for the effectiveness of advertising under varying levels of information. (It is possible that subjects were more motivated to complete the experimental task than if they were watching the commercials in a normal situation in their homes. If motivation to process was higher than normal, it could argued that the subjects did not process the stimuli via the peripheral processing route. It is probable, however, that the large number of commercials evaluated in this study (the survey instrument for each subject was 62 pages long) offset high motivation levels to some degree. Subjects are more likely to maintain high motivation levels if just a few commercials are employed. A few limitations of the measurement techniques must be acknowledged. The 30 cue coding scheme for measuring information is not culturally biased and allows for objective measurement of the level of information contained in each commercial. However, as indicated by Hunt (1976) 190 and Aaker (1980), this type of scheme cannot measure the "quality" or usefulness of individual pieces of information to consumers, since perception of what is informative varies by individual. Additionally, based on the content analysis, where an average commercial in each country contained over 3 information cues, the logical cut-off for "low information" was 2 or less cues (high was defined as 4 or more). It is possible that subjects could have found useful information in the "low information" commercials in some cases. However, the measurement technique represents a clear improvement over the widely utilized Resnik and Stern scheme, and is still judged to be the best suited scheme for performing the task conducted here. Another measurement issue relates to advertising effectiveness. It is not yet clear that any measure of advertising effectiveness closely approximates sales response. Since increasing sales is the ultimate goal of most advertising, it is one which researchers should be concerned with. Particularly with regard to purchase intention, it seems that consumers in laboratory settings indicate much higher intention than is actually the case. Clearly, efforts should be made to develop measures that approximate sales response as closely as possible. The use of aggregate response creates another limitation of the study. While focus on aggregates does approximate what many advertisers attempt to accomplish, it 191 is also limited in what it can tell us about the individual and his/her thought processes. A final limitation relates to the fact that commercials in two product categories in two countries were tested. Ideally, one would like to test for differences over all commercial lengths, product categories, and countries. Since such a comprehensive study was not done, the generalizability of the findings is somewhat limited. SUGGESTIONS FOR FUTURE RESEARCH The findings of this study suggest a need for more research in several areas. The fact that the cultural dimensions outlined in Appendix 2 seem to help explain differences in the effectiveness associated with varying levels of information suggests that more efforts need to be made to define and measure important aspects of culture. While Hofstede's (1984a) work in this area is seminal, marketing and advertising researchers should focus more narrowly on cultural values that may have an impact on communication patterns. There is clearly a need to identify and quantify cultural differences in studying differences in advertising between countries. Further research based on the dimensions of context, individualism/collectivism, and confrontation/harmony seeking values would be a good starting point. Such research should be performed in more countries, and in different regions of the world. 192 This study also points out the need to incorporate aspects of culture into theories of consumer behavior when studying international marketing. While a few attempts have been made to adapt theories (e.g. Fishbein's model by Lee 1991) to other cultures, little attempt to teSt such theories has been made. Theories developed and tested only in the USA are clearly limited in their usefulness in understanding world markets. Further research on the relationship between the information content of advertising and advertising effectiveness is also necessary. Results of the manipulation checks in this study suggest a need to study subject perception of what constitutes "informative advertising" more closely. It may be that certain types of information are regarded as more useful in general, or in certain product categories. More investigation of the Stewart and Furse (1986) finding that more information is associated with lower memorability is also warranted. The findings here suggest U.S. viewers tend to prefer high levels of information (as measures by A-ad, A-brand, and purchase intention) in commercials for low involvement products. There may be a trade off between this positive response and memorability. The findings based on the presence of a brand differentiation clearly indicate that more research is needed. In particular, investigation of how brand 193 differentiation can be achieved is badly needed. There may be different ways of accomplishing brand differentiation (e.g. tangible vs. intangible methods), each of which may be more feasible for products in different categories, at different levels of involvement, or at different life cycle stages. There is also a compelling need to explore whether means of achieving brand differentiation vary by culture. The insights gained from this research suggest a need to extend the research to other areas. First, the impact of information level and brand differentiation on effectiveness should be extended to other product categories. In particular, high involvement products need to be studied. Second, the research should be extended to other media vehicles, to see if the findings can be generalized or if they are applicable only to television advertising. Finally, research in this area should be extended to other countries. Thus, much work remains to be done in this line of research. However, it is hoped that this study has helped to demonstrate the need to perform more experimental research in international ad effects research. While content analyses may be an important first step in many instances, insight based on consumer response is badly needed. APPENDICES APPENDIX 1 DESCRIPTION OF PRE-TESTS 194 Appendix 1 Description of Pro-tests Pre-tests of candidate commercials for inclusion in the final sample were run for both countries. While all pre- tests were conducted in the United States, Korean commercials were tested only on citizens of the ROK who currently resided in the USA. 10 subjects were included in the Korean pretests and 21 in the U.S. pretest. The format of the pre-test resembled that of the overall experiment. Subjects were told that the researchers were interested in their reaction to a number of commercials and will be asked to indicate their response on a variety of scales (to be described below). The commercials included in pre-testing will be those being considered for inclusion in the larger experiment. The goals of the pre-test were as follows: a) To screen the usability of selected commercials in the final experiment. This was accompllshed by observing responses to the commerc1al 1n general and by conducting manipulation checks on the independent commercials. Respondents were asked. to indicate whether they believe the commer01al.1s informative on a seven point semantic differential scale anchored by "informative" and "uninformative" and also to indicate whether they believe each commercial contains a brand differentiating message. Subjects were each given a sheet which clearly deflned the term brand differentiating message, so they could refer to 1t when considering this question. 195 b) To assess the usability of selected measurement scales in the final experiment. The usability of both the words chosen as endpoints for the semantic differential scales and the number of points will be considered. Thus, variance on each scale was measured and reliability of the overall set of measures of each dependent variable was considered. Additionally, the variance of responses on each scale was compared between countries as an aid in determining the appropriateness of a seven point scale. c) To assess the overall viability of the process, particularly with respect to respondent understanding of the task at hand and the possibility of fatigue playing a role in the testing process. A specific assessment of how many commercials cognitive responses can be requested in the final experiment was made based on a judgement of how subjects responded to the task. Additionally, the process carefully watched for potential equivalence problems so that any inconsistency in administration of the experiment in the two languages could be anticipated and corrected. Hae—Kyong Bang, a native speaker of Korean who is knowledgeable in advertising and mass media research techniques was be present at both pre-tests to aid in this purpose. In each pre-test, subjects were assembled in a room in groups small enough to ensure that each participant has a good View of the video screen on which the commercials will be shown. Once the group was assembled, subjects were read the following instructions, "Thank you for participating in this study. We are interested in your reaction to the twenty commercials which we are about to show to you. We will show each commercial individually and allow you sufficient time to fill out the response sheet before moving on to the next commercial. Please watch each commercial carefully, since it will be shown only once. Then, respond 196 to each commercial on the given scales, which ask you to indicate your attitude toward the advertisement, your attitude toward the brand, and whether you intend to purchase the advertised product in the study." In the pre- test of Korean commercials, these instructions were read in Korean. Before the commercials were shown, the text of the response instrument was read and discussed, and subjects were allowed to ask questions. The English version of the response instrument appeared as follows: 197 1) Please describe your response to this advertisement on the following scales: good bad 1 2 3 4 5 6 7 dislike ' like 1 2 3 4 5 6 7 positive negative 1 2 3 4 5 6 7 poorly made well made 1 2 3 4 5 6 7 uninformative informative 1 2 3 4 5 6 7 2) Please indicate your attitude toward the brand in the commercial on the following scales: favorable unfavorable 1 2 3 4 5 6 7 unpleasant pleasant 1 2 3 4 5 6 7 positive negative 1 2 3 4 5 6 7 not beneficial beneficial 1 2 3 4 5 6 7 bad good 1 2 3 4 5 6 7 3) Do you ever purchase items from this product category? Yes No 198 4) Please indicate whether you intend to purchase the advertised product within the next year on the following scales: likely unlikely 1 2 3 4 5 6 7 improbable ~ probable 1 2 3 4 5 6 7 possible impossible 1 2 3 4 5 6 7 If you had to assign a probability to the likelihood of your purchasing the product in the next year, which figure would be most appropriate (check one only): 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 4) In your opinion, does this commercial contain a message that clearly differentiates the brand from its competitors? In other words, is the principal message of the commercial unique to the product being advertised, or could any product make this claim? Yes No 5) Please list any thoughts you had while viewing this ad. Include any thoughts you had about the advertisement or advertiser while viewing the commercial. 199 Results of Protest Results of the pretests in each country indicated that the commercials chosen for inclusion in each sample suggested that each was suitable for inclusion in the final test. Blatant disagreement with the coder classifications was not present in either country. The pretest generally indicated that the chosen scales could be utilized. The scales showed reliability (albeit for a small sample size) and generally were viewed by the subjects as being straightforward. However, it was found that including the cognitive response portion of the questionnaire as the last item created difficulty for the subjects, who had some difficulty remembering the content of the commercial by the time they had completed all of the multiple item scales. Collecting the cognitive responses subject to a one minute time limit did not appear to cause unacceptable fatigue. For this reason, it was decided that cognitive responses would be collected for all commercials included in the final study. The Korean pre-test subjects did not experience difficulty with the use of a seven point scale. In fact, all points (1-7) were very frequently utilized. There was no general tendency to check points toward the middle of the scale or the end of the scale. In addition to moving the location of the cognitive responses to the front of the questionnaire, a few changes 200 in wording were made on the Korean instrument, based on the comments of pretest subjects. In particular, words to describe the concept of a "product category“ and "attitude toward the brand" were modified. Two other changes were to ask whether the respondent had seen the commercial previously and to omit the probability scale for purchase intention, since it did not produce results substantially different from the scaled items. After these modifications were made, the final instruments appeared as follows: 201 Student Number Age Major 202 1) Have you seen this commercial before? Yes No 2) Please list any thoughts you had while viewing this ad. Include any thoughts you had about the advertisement or advertiser while viewing the commercial. 7 7——— ’*—'—Ww_=fik —-.:=~=-——.__=—s-_,—-—-_ - , < -",—-—--— - 203 3) Please describe your attitude toward this advertisement on the following scales: good bad 1 2 3 4 5 6 7 dislike like 1 2 3 4 5 6 7 positive negative 1 2 3 4 5 6 7 poorly made well made 1 2 3 4 5 6 7 uninformative informative 1 2 3 4 5 6 7 4) Please indicate your attitude toward the brand in the commercial on the following scales: favorable unfavorable 1 2 3 4 5 6 7 unpleasant pleasant 1 2 3 4 5 6 7 positive 4 negative 1 2 3 4 5 6 7 not . . beneficial benef1c1al 1 2 3 4 5 6 7 bad 900d | 1 2 3 4 5 6 7 5) Do you ever purchase items from this product category? Yes No 204 6) Please indicate whether you intend to purchase the advertised product within the next year on the following scales: likely unlikely 1 2 3 4 5 6 7 improbable probable 1 2 3 4 5 6 7 possible impossible 1 2 3 4 5 6 7 7) In your opinion, does this commercial contain a message that clearly differentiates the brand from its competitors? In other words, is the principal message of the commercial unique to the product being advertised, or could any product make this claim? Yes No APPENDIX 2 CULTURAL DIFFERENCES BETWEEN THE USA AND THE ROK THAT MAY INFLUENCE ADVERTISING STRATEGY 205 Appendix 2 Cultural Differences Between the USA and the ROK That May Influence Advertising Strategy Korea and the U.S. provide useful examples in investigating differences and similarities in international television advertising (Miracle, Chang, and Taylor 1992). On one hand, both countries have well-developed mass media and relatively high levels of advertising. The U.S. is an economically advanced nation and, as indicated in Chapter 1, the ROK is well on its way to such status. On the other hand, there are differences in the market systems of the two countries and, more broadly, in the cultural, economic, legal, political, and social conditions that are present. While this dissertation does not attempt to deal with all of these variables, some important cultural variables which are likely to influence differences in U.S. and Korean advertising will be identified. The use of cultural differences in explaining differences in advertising between two cultures is consistent with the often expressed belief that advertising processes are directly influenced by cultural norms (Hallowell 1972; Mueller 1987). It is held that an understanding of cultural variables is important in analyzing and comparing advertising in two or more countries. In particular, it is important to focus on those cultural characteristics which are likely to influence 206 communication patterns in a society. To this end, an extensive literature review was undertaken. The literature review suggested three cultural dimensions which may lead to differences in advertising in the two cultures. The dimensions are the degree to which the cultures: 1) are high or low context 2) are individualistic or group oriented, and 3) value direct/confrontational or indirect/harmony seeking behaviors. Context Language is a central part of a society's culture. Klopf (1981) notes that as cultures differ from each other, the communication practices exercised in each culture will also differ. Edward T. Hall (1966, 1976, 1983) has suggested that different languages exhibit contextual variations. Hall (1976, p.79) distinguishes between high context and low context languages in the following way: A high context (HC) communication or message is one in which most of the information is already in the person, while very little is in the coded, explicit, transmitted part of the message. A low context (LC) communication is just the opposite; i.e. the mass of the information is vested in the explicit code. Twins who have grown up together can and do communicate more economically (HC) than do two lawyers in a courtroom during a trial (LC), a mathematician programming a computer, two politicians drafting a regulation, or two administrators writing a regulation. Cultures which are high context are intuitive and 207 contemplative, and tend to utilize indirect and ambiguous messages. In contrast, low context cultures are analytical and action oriented, and tend to use explicit, clearly articulated messages. Wells (1986) and Mueller (1991) have suggested that global advertisers must consider whether contextual variation takes place in planning strategy. Takada and Jain (1991) suggest that there are even more widespread marketing implications if contextual variation is present, focusing on differences in diffusion rates and processes. Thus, if two cultures differ in their contextual level, advertisers and marketers should consider the possible consequences. Prior literature suggests that there is a difference in the contextual level of the U.S. and Korean cultures. In fact, Hall (1976, 1987) explicitly describes the United States and some Western European countries as being low context, while listing South Korea, Japan, and Taiwan as being high context literature. The ROK's classification as a high context culture is supported by numerous writers and experts on Korean culture. Yum (1987), for example, notes that there is an abundance of implicitness and indirectness in the Korean language and culture rooted in Confucian and Buddhist traditions. He stresses that a traditional de- emphasis of oral (vs. written) communication has led Koreans to focus on nonverbal cues in communication and has contributed to making implicitness a central part of Korean 208 communication. Kim (1985) also notes that U.S. consistently rely more on verbal elements, starting from a very early age, than do Koreans. He discusses the concept of "nunch'i", which involves the interpretation of others' facial expressions during conversation plus a "mysterious alpha hidden in one's heart". Nunch'i is an essential element of communication in Korea's high context environment. Kang (1988) and Gudykunst, et. al. (1987) are among others who have discussed the relatively high context nature of the Korean culture, finding that Koreans tend to prefer indirect and non-obvious communication, in contrast to the relatively direct and obvious nature of much communication in the U.S. Given the high context nature of Korea culture in comparison with the relatively low context USA, one would expect styles of communication used in television advertising to differ among the two countries. In particular, one would expect Koreans to rely more on the contextual elements (e.g. mood, tone) in a commercial and less on direct, explicit claims than would U.S. consumers. Thus, it is expected that Korean viewers are more likely to respond positively to commercials with low levels of information content than are U.S. viewers. Conversely, it is expected that Americans will react more positively to commercials with high levels of information than will Koreans. 209 With regard to brand differentiation, the impact of contextual variation is a bit more difficult to assess because brand differentiation was defined in each culture based whether the message was unique to the product being advertised. On first glance, one would be tempted to predict that Koreans are less prone to react favorably to commercials containing brand differentiating messages, since they must "explicitly indicate the difference or uniqueness of the product". Indeed, the low frequency of brand differentiating messages in Korean advertising in general would seem to support this claim. However, it must be remembered that it was Korean coders who made the judgement as to what constitutes uniqueness. It is possible that "uniqueness" for a Korean, in some instances, might be tied to elements of the context of the message as opposed to being linked only to explicit pieces of information about the features or benefits of the product, as is typically the case in the USA. For this reason, it is not suggested here that the contextual level of a culture is unambiguously related to the effectiveness of brand differentiation in advertising. Individualism/Collectivism Hofstede (1980, 1984a, 1984b), in his comprehensive study of work-related values, defines individualism as "a preference for a loosely knit social structure in which individuals take care of themselves and their immediate families only" and collectivism as, "a tightly knit social organization in which individuals can expect other in-group persons to look after them". Hofstede found individualism to be one of four underlying dimensions of work-related values, and found that it varied substantially from country Of the 50 countries in the Hofstede data set, to country. the United States ranked first on this dimension with an index score of 91, while South Korea was 39th, with an index of 18, indicating a wide variance between the two countries on the value placed on this cultural dimension. Kim (1985, p.39) agrees that Americans and Koreans place differential value on individualistic and collectivistic behaviors, stating: In America the self, or individuality, has been based on an individual-centered mode which views the individual as the center, with the rest of the world around him. In Korea, as in Japan, the self is conceived as part of the larger social context surrounding the individual, in which the individual learns to subordinate himself to a larger social system whose members are linked by close emotional bonds. To maintain the solidarity of the members constituting the group, the individual learns to see how his fellow members are feeling rather than how to assert his individuality. other literature on Korean culture supports the notion 211 that collectivistic behaviors are more highly valued than is the case in the United States. Hoare and Pares (1988) indicate that Koreans value dependency highly, as evidenced by their willingness to rely on extended families for various forms of emotional, social, and economic support. Paik (1968), Kim (1984), Kim (1985), also note that the individual has not been the unit of social life in traditional Korean culture. Instead, a spirit of mutual assistance among groups prevails, manifesting itself (among other ways) in cooperation among rural and urban dwellers in harvesting crops and dealing with droughts. Indeed, the ROK's economic development process required much hard work and sacrifice on the part of the people. In this way, the South Korean people seem generally more willing to "pitch in for the good of their country" to a stronger degree than is the case in the United States. Baum (1987), Thomas (1986), and Kearney (1991) have noted that the emphasis on collectivism has allowed Korean firms to gain great employee loyalty, with the latter indicating that familialism may be the dominant principle guiding Korean corporations. With regard to the information content of advertising, a high value on collectivistic behaviors may lead to a need for company's to develop some type of personal relationship with the audience. Miracle (1987) reported that the goal of advertising in Japan, which, like South Korea, places a high value on collectivism, is often to make friends with the 212 consumers and to get them to depend on the seller. Thus, Japanese advertisements often begin by telling a story or entertaining the audience. Miracle (1987) describes the sequence as follows: a) Make friends with the audience. b) Prove that you understand their feelings. c) Show that you are nice. d) Consumers will then want to buy from you because they feel familiar with you and trust you. In contrast, the logic behind Western advertising is essentially the opposite. The audience is often told how the product is different and why it is preferable to other brands so that consumers are given a clear justification for purchase. Then, if consumers are satisfied with the purchase, they will begin to develop trust in the company and prefer its products. While it certainly should not be assumed that Korean advertising parallels that of Japan, since the two countries have distinct cultures and some dissimilarities, it is seems quite possible that the collectivistic values of each country could tend to have similar impacts on advertising. With the economy having been dominated by large conglomerates (chaebol) which produce a very wide variety of products by U.S. standards, it is also clear that consumers in Korea have a strong attachment to those companies with which they associate quality and, in turn, trust. For this 213 reason, it seems probable that the Koreans, like the Japanese, may be well served to concentrate on building a mood, creating imagery, or telling a story. For this reason, it is expected that commercials containing low levels of information will tend to be more effective in Korea than in the United States. Conversely, it would be expected that commercials containing high levels of information will tend to be more successful in the United States than in the ROK. Again, for brand differentiating messages, the effect of the value placed on individualistic/collectivistic behaviors on advertising success seems somewhat ambiguous. If differentiation could only be achieved through stressing of a tangible product benefit, one would expect brand differentiating strategies to be relatively ineffective in Korea. However, it may be that Korean advertisers differentiate based on imagery or feelings of trust associated doing business with the company. Thus, no direction is predicted for the brand differentiating message's impact on advertising effectiveness in the ROK. Confrontation/Harmony Seeking Behaviors Some writers have noted that Koreans tend to avoid confrontation in communication. Yum (1987) characterizes Korean communication patterns as being "accommodation oriented" as opposed to confrontation oriented. Klopf 214 (1981) states that acquiescence and conformity are highly valued in Korea, and play a role in communication in the country. Paik (1968) and others have also observed that the traditional face-consciousness of Koreans make confrontations in social interaction generally unpleasant. Hofstede's (1980, 1984a, 1984b) research findings included a dimension titled "Uncertainty Avoidance" which measures the degree to which members of society feel uncomfortable with uncertainty and ambiguity, which leads them to support beliefs promisingcertainty and to maintain institutions protecting conformity." In this category, which clearly includes some cultural values not directly related to communication as well as values related to context harmony seeking, Korea ranked 34th (higher value on uncertainty avoidance while the United States ranked 11th (less value on uncertainty avoidance). This finding supports the idea that Koreans tend to place a relatively value on behaviors associated with conformity or building harmony. Lee (1991) has discussed motivations toward group conformity within Confucian cultures in more specific terms. In particular, he notes that the desire to conform to the norms of various groups, including the extended family, neighbors, friends, co-workers, and even general members of society influences consumer behavior in Korea (Lee 1983). Often, this can translate into people purchasing a brand 215 which has been established as a group norm. A good example of this type of behavior is the tendency of Koreans who live in the same area, such as an apartment building, to use the same brand of detergent (Lee 1991). Choosing a different brand might be seen as upsetting the social norm, and therefore be viewed as confrontational, and thus undesirable. Face-saving behavior is a related aspect of culture that is often important in Korean consumer behavior. The fact that Koreans are motivated not to lose face leads to a tendency to choose a product whose price, brand, and packaging are consistent with their social position (Lee 1983). Since face saving is based on conformity to social expectations in Korea, the purchase decision is, thus, sometimes influenced by the desire not to violate norms or upset harmony within the society. Before speculating on the potential impact on advertising strategy, a few words of caution relating to the confrontation avoidance dimension in Korea should be given. First, it has been noted by observers of East Asia (e.g., Park, Klopf, and Kambra 1978; Graham, et al. 1988) that Koreans tend to be more comfortable with confrontation in business situations than either their Japanese or Chinese counterparts. Kim, D.K. (1985, p.4) noted that foreigners who are unaware of the importance of personal relationships in business in Korea are often surprised to find that 216 Koreans who are extremely hospitable towards acquaintances can actually be quite hostile and blunt to those who they do not know. Jang (1985) concurs, noting that some modern day Koreans have developed some behaviors that are not necessarily consistent with traditional Confucianism, including the ability to be very tough, even ruthless negotiator, in spite of a calm and gentle manner. Jang's observation relates to the second caution which is, again, that Korea is not Japan. While confrontation avoidance and its influences on Japanese advertising have been well documented in research on Japanese practice (e.g., Wagenaar 1978; Madden et al. 1986; Becker 1986) and many authors have been tempted to group "Asians“ together, no prior research known to the author has investigated whether the same relationship holds true in Korea. That Korea, a land historically surrounded by and often invaded by powerful neighbors (Japan, China, and Russia) would develop some tolerance for confrontational behavior is in spite of entrenched Confucian values is really not particularly surprising. It is likely that harmony seeking behavior in Korea is valued more highly than in the USA, but not as highly as it is in Japan. Further, the particular communication situation investigated here, which utilizes mass media, does not constitute a personal relationship, and could possibly even be a situation where harmony-seeking values do not play a key role in determining communication. 217 If harmony-seeking behaviors are indeed more highly valued in the ROK, one would not expect to aggressive forms of advertising to be effective there. Thus, low-keyed (or "soft-sell") approaches would tend to be more favored. This would likely mean that commercials with low levels of information content would be expected to be relatively effective. In contrast, in the USA, where bold, assertive forms of communication are not taboo, one would expect commercials with higher information content to be more effective relative to the Korean case. For the brand differentiating message, it is likely that U.S. consumers, more tolerant of direct approaches, would prefer messages which clearly stress a unique element of the product. In the ROK, a less direct approach might be preferable on this basis, although, again, it is not completely clear as to whether Koreans view brand differentiation in exactly the same manner as U.S. viewers. APPENDIX 3 SUPPLEMENTAL TABLES 218 TABLE A Means For Attitude Toward The Ad Measures By Commercial ' United States A-AD1 A-AD2 A-AD3 A-AD4 Mean Dev. Mean Dev. Mean Dey. Mean Deyy COMM 1 3.04 1.39 3.09 1.36 2.68 1.12 2.97 1.18 2 2.19 1.25 2.24 1.32 2.13 1.25 2.43 1.19 3 2.50 1.14 2.63 1.15 2.47 1.14 2.62 1.08 4 3.59 1.50 3.69 1.42 3.51 1.37 3.50 1.40 5 1.71 1.22 1.73 1.28 1.80 1.15 1.62 0.82 6 3.32 1.90 3.34 1.91 3.31 1.83 3.38 1.51 7 2.59 1.36 2.51 1.21 2.47 1.25 2.48 1.26 8 3.67 1.75 3.74 1.84 3 61 1.74 3.94 1.63 9 3.23 1.74 3.43 1.76 3.14 1.61 3.15 1.55 10 2.86 1.52 2.91 1.56 2.90 1.49 2.93 1.37 11 3.63 1.65 3.62 1.72 3.50 1.73 3.11 1.23 12 3.34 1.38 3.48 1.41 3.17 1.32 3.53 1.28 13 4.21 1.60 4.43 1.63 4.13 1.59 3.98 1.58 14 1.82 1.06 1.94 1.13 1.95 1.10 2.19 1.17 15 3.69 1.79 3.67 1.73 3.41 1.42 3.64 1.69 16 2.93 1.42 2.94 1.48 3.00 1.42 2.90 1.25 17 4.18 1.61 4.39 1.69 3.99 1.63 3.92 1.33 18 2.65 1.61 2.64 1.63 2.66 1.53 2.45 1.42 19 2.21 1.19 2.23 1.24 2.27 1.29 2.39 1.27 20 4.35 2.16 4.31 2.26 4.37 2.06 3.81 1.69 TOTAL 3.08 1.77 3.15 1.74 3.02 1.63 3.04 1.50 219 TABLE B Means For Attitude Toward The Ad Measures By Commercial Republic Of Korea A-ADl A-AD2 A-AD3 A-AD4 Mean Dev. Mean Deyep Meen Dev. Mean Deyy COMM # 1 3.99 1.41 4.41 1.35 3.86 1.46 4.02 1.33 2 2.45 1.28 2.38 1.20 2.36 1.20 2.95 1.19 3 3.81 1.60 3.84 1.53 3.74 1.57 3.81 1.52 4 3.71 1.46 3.82 1.38 3.77 1.43 3.92 1.38 5 3.72 1.31 3.82 1.27 3.57 1.24 3.94 1.26 6 3.36 1.32 3.37 1 21 3 41 1 32 3.43 1.20 7 2.50 1.33 2.61 1.38 2.51 1.38 2.46 1.39 i 8 2.91 1.57 2.91 1.52 2.94 1.53 2.59 1.33 9 2.20 1.16 2.07 1.11 2.11 1.05 2.15 1.04 10 3.42 1.63 3.57 1.68 3.50 1.59 3.86 1.50 11 4.16 1.44 4.27 1.47 4.00 1.43 4.14 1.36 12 4.59 1.43 4.74 1.48 4.54 1.47 4.66 1.34 13 4.05 1.27 4.07 1.12 3.95 1.30 4.38 1.19 14 2.58 1.00 2.78 0.97 2.70 1.06 2.95 1.04 15 3.90 1.37 4.13 1.40 3.87 1.41 3.93 1.22 16 1.93 0.74 1.99 1.40 2.00 0.84 2.10 0.93 17 4.43 1.32 4.46 1.39 4.23 1.33 4.45 1.23 18 5.82 1.00 6.01 1.02 5.84 1.10 6.15 0.90 19 2.85 1.43 2.89 1.38 2.83 1.41 2.99 1.28 20 1.50 0.87 1.48 0.86 1.69 1.04 1.45 1.28 TOTAL 3.39 1.65 3.48 1.68 3.37 1.63 3.52 1.62 220 TABLE C Means For Attitude Toward The Brand Measures By Commercial United States ABRANDl ABRANDZ ABRAND3 ABRAND4 Mean Dey. Mean Dev. .Mean Dev. pMean Dev. COMM # 1 2.64 1.17 2.53 1.02 2.57 1.08 3.15 1.26 2 1.63 0.80 1.76 0.78 1.69 0.81 1.95 1.05 3 2.29 1.34 2.33 1 21 2.28 1.23 2.66 1.20 4 4.09 1.46 3.98 1.30 3.91 1.21 3.81 1.53 5 2.50 1.44 2.43 1.41 2.50 1.36 2.82 1.25 6 3.02 1.77 2.90 1.65 3.00 1.60 3.15 1.70 7 2.51 1.29 2.52 1.13 2 49 1 24 2.56 1.33 8 3.41 1.64 3.25 1.52 3.34 1.49 3.62 1.39 9 3.00 1.43 3.02 1.46 2.95 1.44 3.32 1.28 10 2.78 1.72 2.72 1.52 2.74 1.50 2.94 1.45 11 2.59 1.37 2.60 1.33 2.68 1.41 2.88 1.31 12 2.24 1.48 2.30 1.36 2.35 1.48 3.42 1.74 13 4.30 1.83 (4.40 1.81 4.15 1.74 4.12 1.70 14 1.49 0.99 1.52 0.90 1.60 0.89 2.73 1.54 15 2.30 1.30 2.33 1.30 2.39 1.26 3.17 1.34 16 2.69 1.47 2.70 1.38 2.65 1.36 3.17 1.52 17 4.57 1.71 4.40 1.67 4.28 1.60 4.35 1.69 18 2.75 1.52 2.81 1.41 2.78 1.37 2.84 1.42 19 2.25 1.33 2.28 1.18 2.28 1.13 2.61 1.30 20 3.34 1.87 3.27 1.73 3.28 1.75 3.03 1.58 TOTAL 2.82 1.65 2.80 1.56 2.80 1.53 3.12 1.54 221 TABLE C (Continued) Means For Attitude Toward The Brand Measures By Commercial United States ABRAND5 Mean ygey. COMM # l 2.65 1.07 2 1.86 0.87 3 2.45 1.21 4 3.91 1.35 5 2.46 1.32 6 3.06 1.61 7 2.62 1.30 8 3.42 1.37 9 3.18 1.42 10 2.73 1.42 11 2.77 1.37 12 2.51 1.55 13 4.24 1.68 14 1.80 1.13 15 2.44 1.24 16 2.77 1.40 17 4.26 1.68 18 2.83 1.34 19 2.39 1.17 20 3.14 1.55 TOTAL 2.87 1.51 222 TABLE D Means For Attitude Toward The Brand Measures By Commercial Republic Of Korea ABRANDl ABRAND2 ABRAND3 ABRAND4 Mean. Dev. Mean Dev. Mean Dev. Mean Deyp COMM # 1 4.08 1.81 4.01 1.35 3.44 1.48 2.92 1.24 2 2.70 1.39 2.58 1.17 2.50 1.15 2.50 1.18 3 3.38 1.70 3 21 1.35 3 14 1.38 3.07 1.33 4 3.31 1.60 3.29 1 19 3.26 1.39 3.18 1.25 5 2.83 1.41 3 18 1.06 2.85 1.17 2.67 1.11 6 3.01 1.34 3 20 1.12 3.16 1.41 3.16 1.38 7 2.60 1.42 2 74 1 31 2.66 1.40 2.96 1.38 8 3.44 1.60 3.24 1 37 3.44 1.49 3.93 1.42 9 2.22 1.28 2.49 1.44 2.50 1.42 3.49 1.51 10 3.60 1.73 3.44 1.40 3.55 1.49 4.03 1.12 11 3.57 1.66 3.61 1.15 3.52 1.43 3.25 1.31 12 4.33 1.52 3.85 1.08 4.07 1.43 4.40 1.24 13 3.53 1.42 3.52 1.04 3.43 1.33 3.37 1.35 14 2.44 1.10 2.66 0.98 2.50 1.04 2.90 1.04 15 4.02 1.79 3.81 1.21 3.73 1.54 3.20 1.45 16 2.49 1.29 2.46 1.03 2.43 .1.16 2.63 1.26 17 3.23 1.58 3.34 1.19 3.09 1.39 3.07 1.24 18 4.77 1.64 4.35 1.35 4.46 1.47 3.88 1.58 19 3.34 1.45 3.30 1.06 3.12 1.13 3.43 1.04 20 2.11 1.41 2.29 1.39 2.29 1.34 3.44 1.63 TOTAL 3.25 1.66 3.23 1.33 3.16 1.46 3.27 1.39 223 TABLE D (Continued) Means For Attitude Toward The Brand Measures By Commercial Republic of Korea ABRANDS Mean Dev. COMM # 1 3.28 1.23 2 2.45 1.05 3 3.07 1.34 4 3.18 1.23 5 2.83 1.12 6 3.14 1.31 7 2.74 1.34 8 3.54 1.28 9 3.06 1.48 10 3.67 1.18 11 3.41 1.17 12 4.06 1.14 13 3.28 1.18 14 2.64 1.01 15 3.40 1.36 16 2.45 1.07 17 3.08 1.25 18 3.96 1.45 19 3.21 1.08 20 2.80 1.60 TOTAL 3.16 1.32 224 TABLE E Means For Purchase Intention Measures By Commercial United States P11 P12 P13 (Mean Dev. Mean Dev. Mean Dey. COMM # 1 3.89 1.95 3.86 1.93 3.34 1.85 2 2.43 1.70 2.43 1.70 2.06 1.85 3 4.33 1.92 4.33 1.86 3.88 1.83 4 5.77 1.75 5.72 1.78 5.27 1.75 5 4.03 2.05 3.94 2.00 3.63 1.95 6 4.26 2.24 4.22 2.18 3.84 2.12 7 3.79 1.86 3.73 1.88 3.51 1.87 8 5.17 1.69 5.11 1.68 4.71 1.73 9 4.53 1.99 4.55 2.00 4.18 1.96 g 10 4.06 2.20 4.04 2.22 3.77 2.13 11 3.47 2.11 3.50 2.11 3.18 1.97 12 2.45 1.98 2.46 1.98 2.36 1.97 13 6.18 1.43 6.09 1.46 5.84 1.55 14 1.90 1.45 1.87 1.40 1.76 1.34 15 3.80 1.98 3.80 1.94 3.54 1.91 16 4.24 2.13 4.17 2.12 4.01 2.02 17 5.53 1.98 5.53 1.94 5.10 1.96 18 5.05 2.02 4.97 2.04 4.57 2.13 19 2.96 1.78 2.98 1.70 2.74 1.69 20 4.09 2.21 3.98 2.15 3.64 2.09 TOTAL 4.09 2.21 4.06 2.18 3.75 2.13 225 TABLE F Means For Purchase Intention Measures By Commercial Republic Of Korea PIl PI2 PI3 Mean Dev. Mean Dev. Mean Dev. COMM # 1 5.31 1.87 5.19 1.83 4.92 1.92 2 2.41 1.63 2.74 1.57 2.23 1.52 3 2.97 1.68 3.11 1.56 2.64 1.51 4 3.74 1.90 3.92 1.75 3.42 1.85 5 3.50 1.63 3.69 1.43 3.14 1.58 6 3.58 1.89 3.76 1.66 3.36 1.79 7 2.92 1.81 3.16 1.65 2.74 1.71 8 3.10 1.71 3.26 1.69 2.91 1.57 9 1.69 1.05 1.88 1.20 1.64 1.01 10 3.77 1.88 3.89 1.73 3.58 1.73 11 4.94 1.75 4.99 1.61 4.41 1.91 12 4.42 1.82 4.57 1.62 4.06 1.86 13 4.96 1.64 4.92 1.55 4.51 1.79 14 2.48 1.15 2.68 1.07 2.25 1.11 15 5.25 1.62 5.36 1.43 5.07 1.67 16 5.13 2.03 5.18 1.93 4.97 2.12 17 3.60 1.83 3.66 1.69 3.34 1.85 18 5.15 1.61 5.11 1.47 4.76 1.74 19 3.46 1.57 3.61 1.40 3.24 1.56 20 1.49 1.04 1.42 0.82 1.30 0.64 TOTAL 3.69 2.03 3.81 1.90 3.43 1.97 226 TABLE G Means For Global A-Ad A-Brand And PI Measures By Commercial United States A-Ad A-Brand PI .Mean Dev. Mean Dev. .Mean Dev. COMM # 1 2.95 1.14 2.71 1.01 3.70 1.84 2 2.25 1.17 1.78 0.77 2.30 1.56 3 2.55 1.01 2.40 1.11 4.18 1.81 4 3.57 1.28 3.94 1.24 5.59 1.68 5 1.72 1.01 2.54 1.23 3.87 1.95 6 3.33 1.65 3.03 1.58 4.11 2.14 7 2.51 1.16 2.54 1.16 3.68 1.83 8 3.74 1.64 3.39 1.39 5.00 1.65 9 3.24 1.58 3.09 1.33 4.42 1.94 10 2.90 1.38 2.78 1.41 3.96 2.12 11 3.47 1.46 2.70 1.24 3.38 2.00 12 3.38 1.23 2.56 1.36 2.42 1.95 13 4.19 1.47 4.23 1.63 6.04 1.43 14 1.97 1.03 1.82 0.91 1.84 1.37 15 3.60 1.53 2.52 1.16 3.72 1.91 16 2.94 1.27 2.80 1.29 4.14 2.06 17 4.19 1.44 4.37 1.59 5.39 1.91 18 2.60 1.43 2.80 1.35 4.86 2.01 19 2.27 1.17 2.36 1.12 2.89 1.67 20 4.21 1.94 3.21 1.58 3.89 2.07 TOTAL 3.07 1.54 2.88 1.58 3.97 2.13 227 TABLE H Means For Global A-Ad A-Brand And PI Measures By Commercial Republic Of Korea A-Ad A-Brand PI Mean Dev. Mean Dev. Mean Dev. COMM # l 4.07 1.20 3.54 1.16 5.14 1.79 2 2.53 1.10 2.55 1.04 2.48 1.51 3 3.80 1.46 3.17 1.26 2.91 1.49 4 3.81 1.33 3.24 1.21 3.69 1.76 5 3.76 1.13 2.87 1.07 3.44 1.47 6 3.39 1.14 3.14 1.17 3.57 1.70 7 2.52 1.27 2.74 1.25 2.94 1.67 8 2.83 1.40 3.52 1.23 3.09 1.59 9 2.13 1.00 2.75 1.21 1.74 1.03 10 3.59 1.51 3.66 1.20 3.75 1.72 11 4.14 1.29 3.48 1.19 4.78 1.65 12 4.63 1.32 4.14 1.13 4.35 1.71 13 4.12 1.09 3.43 1.11 4.80 1.58 14 2.75 0.91 2.63 0.86 2.47 1.00 15 3.96 1.24 3.63 1.34 5.22 1.52 16 2.01 0.74 2.49 1.00 5.09 1.97 17 4.39 1.19 3.16 1.20 3.53 1.75 18 5.96 0.88 4.28 1.36 5.01 1.53 19 2.89 1.29 3.28 1.04 3.44 1.45 20 1.53 0.80 2.58 1.31 1.40 0.75 TOTAL 3.44 1.56 3.21 1.27 3.64 1.91 228 TABLE I Mean Number Of Cognitive Responses By Commercial United States Source Source Support Total Support Derogation Argument Mean Dev. Mean Dev. Mean Dev. Mean Deyy COMM # 1 1.59 0.80 0.85 0.74 0.23 0.49 0.35 0.59 2 1.79 0.86 1.19 0.84 0.18 0.41 0.28 0.55 3 1.80 0.79 0.91 0.74 0.14 0.38 0.48 0.72 4 1.28 0.63 0.43 0.57 0.36 0.59 0.14 0.38 5 1.91 0.80 1.48 0.80 0.07 0.35 0.25 0.54 6 1.44 0.70 0.63 0.77 0.51 0.69 0.10 0.33 7 1.60 0.84 0.96 0.84 0.17 0.43 0.20 0.50 8 1.48 0.87 0.48 0.70 0.42 0.64 0.25 0.61 1 9 1.46 0.82 0.61 0.79 0.45 0.69 0.22 0.59 10 1.58 0.80 0.82 0.83 0.22 0.48 0.34 0.68 11 1.64 0.79 0.56 0.75 0.55 0.74 0.23 0.61 12 1.32 0.76 0.54 0.59 0.20 0.47 0.33 0.55 13 1.41 0.83 0.25 0.52 0.37 0.56 0.30 0.64 14 1.50 0.77 1.08 0.76 0.10 0.33 0.24 0.51 15 1.37 0.78 0.43 0.61 0.48 0.73 0.25 0.48 16 1.39 0.88 0.48 0.68 0.24 0.47 0.42 0.71 17 1.38 0.79 0.26 0.56 0.54 0.71 0.16 0.46 18 1.48 0.77 0.68 0.63 0.20 0.49 0.32 0.65 19 1.42 0.83 0.85 0.74 0.09 0.29 0.41 0.64 20 1.53 0.76 0.54 0.76 0.73 0.79 0.16 0.42 i TOTAL 1.52 0.81 0.70 0.78 0.31 0.58 0.26 0.57 Mean Number Of Cognitive Responses By Commercial 229 TABLE I (CONT.) United States Counter Argument Mean Dey. COMM # 1 0.02 0.14 2 0.02 0.14 3 0.05 0.22 4 0.12 0.33 5 0.03 0.17 6 0.07 0.26 7 0.10 0.30 8 0.19 0.44 9 0.07 0.43 10 0.10 0.30 11 0.05 0.22 12 0.09 0.32 13 0.34 0.61 14 0.00 0.00 15 0.03 0.17 16 0.13 0.39 17 0.27 0.56 18 0.16 0.37 19 0.01 0.10 20 0.04 0.20 TOTAL 0.93 0.33 Mean Number Of Cognitive Responses By Commercial 230 TABLE J Republic Of Korea Source Source Support Total Support Derogation Argument Mean Dev. Mean Dev. Mean Dev. Mean Deyy COMM # 1 1.67 0.84 0.42 0.67 0.72 0.78 0.20 0.42 2 1.72 0.81 1.04 0.80 0.35 0.59 0.21 0.43 3 1.70 0.83 0.59 0.71 0.58 0.79 0.23 0.47 4 1.54 0.78 0.45 0.68 0.45 0.62 0.22 0.46 5 1.38 0.73 0.44 0.54 0.47 0.63 0.15 0.36 6 1.43 0.82 0.46 0.72 0.31 0.50 0.24 0.49 7 1.66 0.77 0.95 0.83 0.25 0.56 0.34 0.60 8 0.54 0.69 0.85 0.71 0.34 0.53 0.11 0.37 9 1.51 0.76 0.96 0.80 0.15 0.43 0.21 0.48 10 1.62 0.68 0.78 0.66 0.57 0.68 0.08 0.27 11 1.70 0.84 0.35 0.56 0.87 0.81 0.14 0.38 12 1.58 0.87 0.32 0.53 0.82 0.85 0.20 0.45 13 1.65 0.77 0.32 0.58 0.69 0.74 0.15 0.41 14 1.77 0.80 0.98 0.86 0.17 0.40 0.42 0.64 15 1.62 0.76 0.35 0.50 0.53 0.69 0.21 0.48 16 1.68 0.77 1.19 0.88 0.10 0.33 0.30 0.52 17 1.63 0.76 0.18 0.38 0.88 0.77 0.15 0.36 18 1.68 0.77 0.06 0.24 1.41 0.72 0.01 0.10 19 1.72 0.83 1.08 0.86 0.39 0.60 0.06 0.24 20 1.79 0.91 1.53 0.88 0.08 0.31 0.09 0.32 TOTAL 1.63 0.80 0.66 0.78 0.51 0.71 0.18 0.44 231 TABLE J (Cont.) Mean Number Of Cognitive Responses By Commercial Republic Of Korea Counter Argument Mean Dev. COMM # l 0.17 0.38 2 0.08 0.27 3 0.15 0.38 4 0.27 0.51 5 0.13 0.37 6 0.20 0.42 7 0.08 0.27 8 0.08 0.27 9 0.04 0.20 10 0.06 0.24 11 0.18 0.41 12 0.12 0.33 13 0.28 0.47 14 0.11 0.34 15 0.31 0.54 16 0.08 0.27 17 0.31 0.60 18 0.14 0.35 19 0.02 0.14 20 0.07 0.29 TOTAL 0.14 0.38 232 TABLE X Composite Means By Independent Variable Categories - USA I. All SUBJECTS A-Ad A-Brand PI Mean Deya Mean Dev. aMean Dey. BDIFF 0 3.03 1.57 2.73 1.37 3.73 2.04 1 3.13 1.50 3.03 1.52 4.20 2.19 INFORMATION Low 3.34 1.47 3.09 1.48 4.29 2.10 High 2.81 1.56 2.67 1.39 3.64 2.11 II. SUBJECTS WHO PURCHASE FROM PRODUCT CATEGORY ONLY (n = 1612) A-Ad A-Brand PI Mean Dev. Mean Dev. pMean Dev.- BDIFF 0 2.95 1.57 2.63 1.34 3.39 1.95 1 2.99 1.46 2.76 1.39 3.62 2.09 INFORMATION ’ Low 3.27 1.48 2.91 1.46 3.83 2.06 High 2.69 1.50 2.47 1.24 3.19 1.93 III. SUBJECTS WHO HAVE SEEN COMMERCIAL PREVIOUSLY ONLY (n = 985) A-Ad A-Brand PI Mean Std. D. Mean Std. D. Mean Std. D. BDIFF 0 2.87 1.56 2.62 1.56 3.47 2.02 1 2.89 1.37 3.03 1.37 4.16 2.21 INFORMATION Low 3.15 1.37 3.05 1.45 4.13 2.16 High 2.59 1.54 2.53 1.29 3.38 2.03 233 TABLE L Composite Means By Independent Variable Categories - ROK I. ALL SUBJECTS A-Ad A-Brand PI Mean Dey. .Mean Dev. Mean Dey: BDIFF 0 3.75 1.53 3.26 1.53 3.78 1.84 1 3.12 1.52 3.17 1.52 3.50 1.96 INFORMATION Low 3.41 1.60 3.20 1.27 3.50 1.85 High 3.47 1.52 3.23 1.26 3.78 1.95 II. SUBJECTS WHO PURCHASE FROM PRODUCT CATEGORY ONLY (n = 1386) A-ad A~Brand PI .Mean Dev. Mean Day! Mean Dev. BDIFF 0 3.66 1.61 3.09 1.28 3.27 1.69 1 3.03 1.53 3.11 1.24 2.81 1.64 INFORMATION Low 3.23 1.63 3.09 1.69 3.08 1.68 High 3.45 1.55 3.11 1.25 2.97 1.66 III. SUBJECTS WHO HAVE SEEN COMMERCIAL PREVIOUSLY ONLY (n = 1335) A-Ad A-Brand PI Mean Dev. Mean Dev. Mean Dev. BDIFF O 3.52 1.55 3.08 1.26 3.56 1.86 1 2.81 1.45 3.04 1.24 3.16 1.91 INFORMATION Low 3.09 1.61 3.04 1.26 3.05 1.76 .High 3.21 1.47 3.08 1.24 3.64 1.98 REFERENCES Ackoff, Russell L. (1958), "Toward A Behavioral Theory of Communication," Management Science, 4 (April), 218-235. Aaker, David (1984), "Measuring the Information Content of Television Advertising," Current Issues and Research in Advertising, 93-108. Aaker, David A. and Donald Norris (1982), "Characteristics of TV Commercials Perceived as Informative," Journal of Advertising Research, 22 (2), 61-70. Ajzen, Icek and Martin Fishbein (1980), Understanding Attitudes and Predicting Social Behavior, Englewood Cliffs, NJ: Prentice Hall. Appel, valentine (1987), "Editorial Environment and Advertising Effectiveness," Journal of Advertising Research, (Aug/Sept) 11-16. Assael, Henry, J.H. Kofron, and W. Burg (1967), "Advertising Performance as a Function of Print Ad Characteristics," Journal of Advertieinq Reeearch, 7, 20-26. Batra, Rajeev and Michael L. Ray (1986), "Affective Responses Mediating Acceptance of Advertising," Journal 13 (September), 234-249. of Consumer Research, "The Role of Batra, Rajeev and Douglas M. Stayman (1990), Mood in Advertising Effectiveness," Journal of Consumer Research, 17 (September), 203-214. "Korea's Newest Export: Management Baum, Laurie (1987), (January 19), 66. Style," Business Week, Becker, Carl B. (1986), "Reasons for Lack of Argumentation and Debate in the Far East," International Journal of Intercultural Relations, 10, 75-92. ,Belch, George (1981), "An Examination of Comparative and Noncomparative Television Commercials: The Effects of Claim variation and Repetition on Cognitive Response and Message Acceptance," Journal of Marketing Research, 18 (August), 333-349. "Consumer Psychology," Annual Bettman, James R. (1986), 37, 257-289. B§Yiew of Peygholoqy, Biel, Alexander L. and Carol A. Bridgewater (1990), "Attributes of Likable Television Commercials," Journal of Advertising Research, (June/July), 38-44. 234 235 Boddewyn, J.J., Robin Soehl and Jacques Picard (1986), "Standardization of International Marketing: is Ted Levitt In Fact Right?" Business Horizons, (Nov/Dec), 69-75. Borden, Neil (1942), The Economic Effects of Advertising. Homewood, IL: Richard D. Irwin. Brislin, Richard W. (1970), "Back-Translation For Cross- Cultural Research," Journal of Cross-Cultural Psychology, 1 (September), 185-216. Burke, Marian C. and Julie A. Edell (1986), "Ad Reactions Over Time: Capturing Changes in the Real World," Journal of Consumer Research, 13 (June), 114-118. Brucks, Merrie (1985), "The Effects of Product Class Knowledge on Information Search Behavior," Journal of Consumer Research, 12, 1-16. Buzzell, Robert D. (1968), "Can You Standardize Multinational Marketing?" Harvard Business Review., (Nov/Dec), 102-113. Cacioppo, John T., Mary E. Losch, and Hai Sook Kim (1986), "Social Psychological Procedures for Cognitive Response Assessment: The Thought Listing Technique," in Cognitive Assessment Thomas V. Merluzzi, Carol L. Glass, and Myles Genest (eds.). New York: Guilford Press, 309-342. Calder, Bobby, Lynn W. Phillips and Alice Tybout (1982), "The Concept of External Validity," Journal of Consumer Research, 9 (December), 240-244. Cox, Dena Saliagas and William B. Locander (1987), "Product Novelty: Does it Moderate the Relationship Between Ad Attitudes and Brand Attitudes?" Journal of Advertisin , 16 (3), 39-44. Craig, C. Samuel and Brian Sternthal (1986), Repetition Effects Over the Years: An Anthology of Classic Articles, New York: Garland. Cutler, Bob D. (1989), "International Advertising: Some Propositions for Determining When Local Adaptation is Appropriate," in Marketing: Positioning for the 1990's, Robert L. King (ed.), Charleston, SC: The Citadel, 343- 346. Douglas, Susan P. and C. Samuel Craig (1983), International Marketing Research. Englewood Cliffs, NJ: Prentice Hall. 236 Droge, Cornelia (1989), "Shaping the Route to Attitude Change: Central versus Peripheral Processing through Comparative versus Noncomparative Advertising," Journal of MarketingaResearch, 26 (May), 193-204. Droge, Cornelia and Rene Y. Darmon (1987), "Associative Positioning Strategies Through Comparative Advertising: Attribute versus Overall Similarity Approaches," Journal of Marketing Research, 24 (November), 377-388. Dowling, Graham E. (1980), "Information Content in U.S. and Australian Television Advertising," Journal of Marketing, 44 (4), 34-37. Dunn, S. Watson (1976), "Effect of National Identity on MUltinational Promotion in Europe," Journal of Marketing, (October), 50-57. Edell, Julie A. and Marian C. Burke (1987), "The Power of Feelings in Understanding Advertising Effects," Journal of Consumer Research, 14 (December), 421-433. Edell, Julie A. and Kevin Lane Keller (1989), "The Information Processing of Coordinated Media Campaigns," Journal of Marketing Research, 26 (May) 149-163. "How International Can International Blinder, Erik (1961), Advertising Be?," International Advertiser, (December), 12-16. International Fatt, Arthur C. (1967), "The Danger of 'Local' Advertising," Journal of Marketing, (January), "Advertising Strategy in Japan," Fields, George (1980), Dentsu's Japan MarketingZAdvertising, (Fall/Winter), 52-56. Fields, George (1983), From Bonsai to Leviie: When West Meete East: An Insider'e Surprising Account of How Japanese Live. New York: MacMillan Publishing Company. "Why Are the Japanese not Fields, George (1984), 'Westernized'?: But Why Ask at All," Japan MarketinggAdvertising (Summer), 25-29. Gail, Tom, Steven Calvert, Rita Goolkatsian and Arlene Gumsteg (1984), "An Analysis of Information Content in Advertising: An Update," Current Research and Issues in Advertieing, 158-165. 60-62. Gardner,.Meryl Paula (1985), "Does Attitude Toward the Ad Affect Brand Attitude Under A Brand Evaluation Set?“ Journal of Marketing Research, 22 (May), 192-198. 237 Gibson, Lawrence D. (1983), "Not Recall," Journal of Advertising Research, 23, 39-46. Gorn, Gerald R. (1982), "The Effects of MUsic in Advertising on Choice Behavior: A Classical Conditioning Approach," Journal of Marketing, 46 (Winter), 94-100. Graham, John L., Dong-Ki Kim, Chi-Yuan Lin, and Michael Robinson (1988), "Buyer-Seller Negotiations Around the Pacific Rim: Differences in Fundamental Exchange Processes," Journal of Consumer Research, 15 (June), 48-54. Green, Robert T., W.H. Cunningham, and I.C.M. Cunningham (1975),"The Effectiveness of Standardized Global Advertising," Journal of Advertising, (4 (3), 25-30. Gudykunst, William B., Young-Chul Yoon, and Tsukasa Nishadsa (1987), "The Influence of Individualism-Collectivism on Perceptions of Ingroup and Outgroup Relationships," Communication Monographs, 54 (September), 295-306. Haley, Russell I., Jack Richardson, and Beth Baldwin (1984), "The Effects of Nonverbal Communication in Television Advertising," Journal of Advertising Research, 24 (4), 11-18. Hall, Edward T. (1966), The Silent Language. New York: Doubleday. Hall Edward T. (1976), Beyond Culture. Garden City, NY: Anchor Press/Doubleday. Hall, Edward T. and Mildred Reed Hall (1972), Hidden Differences: Doing Busineee With the Japaneee. Garden City, NY: Anchor Press/Doubleday. Hallowell, A.F. (1972), "Cultural Factors in the Structuralization of Perception," in Intercultural Communication (L.Samovar and R. Porter eds.), Belmont CA: Wadsworth Publishing Co. Hair, Joseph F., Jr., Rolph E.Anderson, and Ronald L. Tatham (1987), MUltivariate Data Analysis. New York: MacMillan Publishing Co., 145-164. Harris, Greg (1984), "The Globalization of Advertising," International Journal of Advertising, 223-234. Hastak, Manoj and Jerry C. Olson (1981), "Assessing the Role of Brand-Related Cognitive Responses as Mediators of Communication Effects on Cognitive Structure,“ Journal of Consumer Research, 15 (March), 444—456. 238 Hite, Robert E. and Cynthia Fraser (1988), "International Advertising Strategies of Multinational Corporations," Journal of Advertising Research, (Aug/Sept), 9-17. Hoare, James and Susan Pares (1989), Korea: An Introduction. New York: Kegan Paul International. Hofstede, Geert (1980), "Motivation, Leadership, and Organization: Do American Theories Apply Abroad?" Organizational Dynamics, (Summer), 42-63. Hofstede, Geert (1983), "Dimensions of National Cultures in Fifty Countries and Three Regions," In J. Deregowski, S. Dziurawiec, and R. Annis (eds) Explorations in Cross-Cultural Psychology, Lisse Netherlands: Swets and Zeitlinger, 335-355. Hofstede, Geert (1984a), Culture's Consegpences: International Differences in Work-Related Values, Beverly Hills: Sage. Hofstede, Geert (1984b), "Cultural Dimensions in Management and Planning," Asia Pacific Journal of Management, (January), 335-354. Holbrook, Morris B. and Rajeev Batra (1987), "Assessing the Role of Emotions as Mediators of Consumer Responses to Advertising," Journal of Consumer Research, 14 (December), 404-420. Holbrook, Morris and Donald R. Lehmann (1980), "Form vs. Content in Predicting Starch Scores,"Journal of Advertising Research, 20, 53-62. Holmes, John H. and Kenneth E. Crocker (1987), "Predispositions and the Comparative Effectiveness of Rational, Emotional and Discrepant Appeals for Both High and Low Involvement Products," Journal of the Academy of Marketing Science, 15 (Spring), 27-35. Hong, Jae W., Aydin Muderrisoglu and George M. Zinkhan (1987), "Cultural Differences in Advertising Expression: A Comparative Content Analysis of Japanese and U.S. Magazine Advertising," Journal of Advertisin , 16, (1), 55—62. Hornik, Jacob (1980), "Comparative Evaluation of International vs. National Advertising Strategies," Columbia Journal of World Business, (Spring), 36-45. Huang, Jen-Hung and Juei-Fu Hou (1987), "The Effects of Regulation on the Level of Information Content of Television Advertising -- A Comparison of Three Nations," 1987 AMA Summer Educators Conference Proceedings, 93-96. 239 Hunt, Shelby D. (1976), "Informational vs. Persuasive Advertising: An Appraisal," Journal of Advertisin , 5 (Summer), 5-8. Iyer, Easwar S. (1988), "The Influence of Verbal Content and Relative Newness on the Effectiveness of Comparative Advertising," Journal of Advertisin , 17 (3), 15-21. Jain, Subhash (1989), "Standardization of International Marketing Strategy: Some Research Hypotheses," Journal of Marketing, (January), 70-79. Jang, Song H. (1985), "Negotiating in Korea," Business Korea, (March), 36-37. Johnson, Eric J. and J. Edward Russo (1984), "Product Familiarity and Learning New Information," Journal of Consumer Research, 11 (June) 542-550. Juster, Thomas (1964), Anticipations of Purchase: An Analysis of Consumer Behavior. Princeton, NJ: National Bureau of Economic Research. Kahle, Lynn R. and Pamela M. Homer (1985), "Physical Attractiveness of the Celebrity Endorser: A Social Adaptation Perspective," Journal of Consumer Research, 11 (March), 954-961. Kang, Shin Pyo (1988), "Korean Culture, The Seoul Olympics, and World Order," Korea and World Affairs, 12 (Summer), 347-362. Kardes, Frank (1988), "Spontaneous Inference Processes in Advertising: The Effects of Conclusion Omission and Involvement on Persuasion, Journal of Consumer Research 15, (September), 225-233. Kaynak, Erdener and Lionel A. Mitchell (1981), "Analysis of Marketing Strategies Used in Diverse Cultures," Journal of Advertising Research, 21 (June), 25-32. Kassarjian, Harold H. (1977), "Content Analysis in Consumer Research," Journal of Consumer Research, 4 (June), 8- 18. Kearney, Robert P. (1991), "Managing Mr. Kim," Across the Board, (April), 40-46. Kellaris, James J. and Anthony D. Cox (1989), "The Effects of Background Music in Advertising: A Reassessment,“ Journal of Consumer Research, 16 (June), 113-118. 240 Kent, Robert J. and Karen A. Machleit (1990), "The Differential Effects of Within-Brand and Between-Brand Processing on the Recall and Recognition of Television Commercials," Journal of Advertising, 19 (2), 4-14. Keppel, Geoffrey (1982), Design and Analysis: A Researcher's Handbook. Englewood Cliffs, NJ: Prentice Hall. Kiel, G.C. and Roger A. Layton (1981), "Dimensions of Consumer Information Seeking Behavior," Journal of Marketing Research, 18, 133-239. Killough, James (1978), "Improved Payoffs from Transnational Advertising," Harvard Business Review, (July/Aug), 92- 102. Dong-Ki (1985), "The Impact of the Traditional Korean Kim, Values on Korean Patterns of Management," paper presented at the Pan-Pacific Conference, Seoul, Korea. Kim, Kyong-Dong (1984), Man and Society in Korea's Economic Growth. Seoul: Seoul National University Press. Kim, Edward H. (1985a), Korea: Be ond the Hills, Seoul Korea: Seoul and Samhwa Printing, Ltd. "Expression of Emotions by Koreans and Kim Ki-Hong (1985b), 9, 38-56. Americans," Korean Studiee, "The Effect of Perceived Advertising Kirmani, Anna (1990), Costs on Brand Perceptions," Journal of Consumer Research, 17 (September), 160-171. Klopf, Donald W. (1981), "Communicating Across Cultures: Understanding the Basic Elements," Korea Journal, 21 (10), 11-24. Kotler, Philip (1984), Marketing Management: Analysis, Planning, and Control. Englewood Cliffs, NJ: Prentice Hall. Kreutzer, Ralf Thomas (1988), "Marketing Mix Standardization: an Integrated Approach in Global (10), 19-30. Marketing," European Journal of Marketing, Krippendorff, Klaus (1975), "Information Theory," in Communication and Behavior, Gerhard Hanneman and William McEwen (eds.), ReadingzMA: Addison-Wesley, 351- 389. "Low Recall and High Recognition Krugman, Herbert E. (1986), of Advertising," Journal of Advertising Research, (Feb/Mar), 79-86. 241 Laczniak, Gene (1979), "Information Content in Print Advertising," Journalism Quarterly, 56, 324-327, and 345. Lavidge, Robert and Gary A. Steiner (1961), "A Model for Predictive Measurement of Advertising Effectiveness," Journal of Marketing, 25 (October), 61-65. Lee, Chol (1991), "Modifying an American Consumer Behavior Model for Consumers in a Confucian Culture: The Case of the Fishbein Behavior Intention Model," Journal of International Consumer Marketing, 3 (1), 27-50. Lee, K.T. (1983), Pattern of Korean Way of Thinking I. Seoul: Sinwon Publishing Company. Leigh, Thomas W., Arno J. Rethans, and Tamatha Whitney (1987), "Role Portrayals of Women in Advertising: Cognitive Responses and Advertising Effectiveness," Journal of Advertising Research, (Oct/Nov), 54-63. Lutz, Richard J. (1985), "Affective and Cognitive Antecedents of Attitude Toward the Ad: A Conceptual Framework," in Psychological Processes and Advertising Effects: Theory. Research, and Application, Linda F. Alwit and Andrew A. Mitchell, eds. Hillsdale, NJ: Lawrence Erlbaum Associates Publishers, 45-63. Machleit, Karen A. and R. Dale Wilson (1988), "Emotional Feelings and Attitude Toward the Advertisement: The Roles of Brand Familiarity and Repetition," Journal of Advertieing, 17 (3), 27-35. Madden, Charles 8., Marjorie J. Caballero, and Shinya Matsukubo (1986),"Analysis of Information Content in U.S. and Japanese Advertising," Journal of Advertising, 15 (Spring), 38-45. Madden, Thomas J., Chris T. Allen and Jacquelyn L. Twible (1988), "Attitude Toward the Ad: An Assessment of Diverse Measurement Indices Under Different Processing 'Sets'," Journal of Marketing Reeearch, 25 (August), 242-252. MacKenzie, Scott B., Richard J. Lutz and George E. Belch (1986), "The Role of Attitude Toward the Ad as a ‘Mediator of Advertising Effectiveness," Journal of Marketing Research, 23 (May), 130-143. Maenaka, Hiromi, Gordon E. Miracle and Kyu Yeol Chang (1991), "An Analysis of Information in Japanese and US Television Advertising," Paper Presented at the Conference on Communication in Japan and the United States at Fullerton, CA in March, 1991. 242 Marquez, F.T. (1977), "Advertising Content: Persuasion, Information, or Intimidation,“ Journalism Quarterly. 54, 482-491. Martenson, Rita (1987), "Is Standardization of Marketing Feasible in Culture-Bound Industries: A European Case Study?" International Marketing Review, (Autumn), 7-17. Miniard, Paul W., Sunil Bhatla, and Randall L. Rose (1990), "On the Formation and Relationship of Ad and Brand Attitudes: An Experimental and Causal Analysis," Journal of Marketing Research, 28 (August), 290-303. Miracle, Gordon E. (1968), "International Advertising Principles and Strategies," MSU Business Topics, (Autumn), 29-36. Miracle, Gordon E. (1987), "Feel—Do-Learn: An Alternative Sequence Underlying Japanese Consumer Response to Television Commercials," The Proceedings of the 1987 Conference of the American Academy of Advertising (F. Feasley, ed.), Columbia: Univ. of South Carolina. Miracle, Gordon E. (1988), "Management Implications of Advertising Regulation in Japan and the USA," Developments in Pacific-Asian Business: Education and Research, Proceedings of the First Annual International Bu51ness Symposium on Pacific—Asian Business, Univers1ty of Hawaii, 1: 70-71. Miracle, Gordon E. (1988), "An Empirical Study of the Usefulness of the Back-Translation Technique for International Advertising Messages in Print Media," Proceedings of the 1988 Conference of the American Academy of Advertising, John D. Leckenby (ed.), Austin, TX: The University of Texas, RC51-RC56. Miracle, Gordon E., Kyu Yeol Chang and Charles R. Taylor (1989), "Relationships Between Cultural Variation and Differences in U.S. and Japanese Advertising," in Marketing: Positioning for the 1990's, Robert L. King (ed.), Charleston SC: The Citadel, 347-351. Miracle, Gordon E., Roberto Friedman, Charles S. Madden, Barbara Mueller, Russel W. Belk, Wei-Na Lee, and David K. Tse (1990) "Research Methods for International Advertising," 1990 AAA Conference Proceedings. Miracle, Gordon E., Hae—Kyong Bang, and Kyu Yeol Chang (1992), "Achieving Reliable and Valid Cross-Cultural Research Results," Paper presented at the National Conference of the American Academy of Advertising in San Antonio, TX. 243 Miracle, Gordon E., Kyu Yeol Chang, and Charles R. Taylor (1992), "Culture and Advertising Executions: A Comparison of Selected Characteristics of Korean and U.S. Television Commercials," International Marketing Review, forthcoming. Mitchell, Andrew A. and Jerry C. Olson (1981), "Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand Attitudes?" Journal of Marketing Research, 18 (August), 318-332. Moon, Young Sook and George E. Franke (1987), "The Information Content of Domestic and International Advertising: An Analysis of Korean Magazine Ads," 1987 American Marketing Association Summer Educators' Conference Proceedings, 98-101. Moore, Danny L. and J. Wesley Hutchinson (1983) "The Effects of Ad Affect on Advertising Effectiveness,"in Advances in Consumer Research, Vol.10, Richard P. Bagozzi and Alice M. Tybout (eds.), Ann Arbor, MI: Association for Consumer Research, 526-531. Muehling, Darrel D. and Russell N. Laczniak (1988), "Advertising's Immediate and Delayed Influence on Brand Attitudes: Considerations Across Message-Involvement Levels," Journal of Advertisin , 17 (4), 23-34. L Mueller, Barbara (1987), "Reflections of Culture: An Analysis of Japanese and American Advertising Appeals," Journal of Advertieing Research, (June/July), 51-59. Mueller, Barbara (1991), "An Analysis of Information Content in Standardized vs. Specialized Multinational Advertisements," Journal of International Business Studies, 1, 23-39. Nelson, James E., Calvin P. Duncan, and Nancy T. Frontczak (1985), "The Distraction Hypothesis and Radio Advertising," Journal of Marketin , 49 (Winter), 60-71. Ogilvy, David (1964), Confessions of an Advertising Man. New York: Atheneum. Paik, Hyun-ki (1968), "The Korean Social Structure and its Implications for Education," The Korea Journal, 8 (March), 11-15. Park, C. Whan and 8. Mark Young (1986), "Consumer Response to Television Commercials: The Impact of Involvement and Background on Brand Attitude Formation," Journal of Marketing Research, 23 (February), 11-24. 244 Park, Myung-Seok, Donald Klopf, and Ronald Cambra (1978), "A Comparison of Communication Apprehension in Korea, America, Australia and Japan," Korea Journal, 18 (11), 33-38. Park, Ungsuh Kenneth (1987), "A Bird's Eye View of Korean Economic Development," Korea and World Affairs, 11 (Spring), 137-148. Pechmann, Cornelia and David W. Stewart (1990), "The Effects of Comparative Advertising on Attention, Memory, and Purchase Intentions," Journal of Consumer Research, 17 (September), 180-191. Peebles, Dean M. (1989), "Don't Write Off Global Advertising: A Commentary," International Marketing Review, (1), 73-78. Peebles, Dean M., John K. Ryans and Ivan R. Vernon (1977), "A New Perspective on Advertising Standardization," European Journal of Marketing, (8), 569-576. Perrien, Jean, Christian Dussart, and Francoise Paul (1985), "Advertisers and the Factual Content of Advertising," Journal of Advertising, 14 (1), 30-35. Peter, J. Paul (1981), "Construct Validity: A Review of Basic Issues and Marketing Practices," Journal of Marketing Research, 18 (May), 133-145. Petty, Richard E. and John T. Cacioppo (1979), "Issue Involvement Can Increase or Decrease Persuasion by Enhancing Message Relevant Cognitive Responses," Journal of Personality and Social Sciences, 37 (10), 1915-1926. Petty, Richard E. and John T. Cacioppo (1984), Attitudes and Persuasion: Classic and Contemporary Approaches. Dubuque IA: Brown. Petty, Richard E. and John T.Cacioppo (1986), "The Elaboration Likelihood Model of Persuasion," in Advances in Social Psychology, Leonard Berkowitz (ed.), 123-205. Petty, Richard E., John T. Cacioppo and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Reeearch, 10 (September), 135-146. Pitt, Leyland and Russell Abratt (1988), "Music in Advertisements for Unmentionable Products--A Classical Conditioning Approach," International Journal of Advertising, 7 (1), 130-137. 245 Punj, G.N. and R. Staelin (1983), "A Model of Consumer Information Search for New Automobiles," Journal of Consumer Research, 9, 366-380. Quelch, John A. and Edward J. Hoff (1986), "Customizing Global Marketing," Harvard Business Review, (May/June), 59-68. Rau, Pradeep and John F. Preble (1987), "Standardization of Marketing Strategies by Multinationals," International Marketing Review, (Autumn), 18-28. Reeves, Rosser (1961), Reality in Advertising. New York: Alfred A. Knopf. Reid, L.N. and J.J. Rotfeld (1981), "How Informative are Ads on Children's Shows?" Journalism Quarterly, 58, 108-111. Renforth, William and Simon Raveed (1983), "Consumer Information Cues in Television Advertising: A Cross Country Analysis," Journal of the Academy of Marketing Science, 3 (11), 216-225. Resnik, Alan and Bruce L. Stern (1977), "An Analysis of Information Content in Advertising," Journal of Marketing, 41 (January), 50-53. Rice, Marshall D. and Zaiming Lu (1988), "A Content Analysis of Chinese Magazine Advertisements," Journal of Advertising, 17 (4), 43-48. Ricks, David A., Jeffery Arpan and Marilyn Y. Fu (1974), "Pitfalls in Advertising Overseas," Journal of Advertising Research, (6), 47-50. Ricks, David A., Marilyn Y. Fu and Jeffery S. Arpan (1974), International Business Blunders. Columbus, OH: Grid. Ries, Al and Jack Trout (1972), "Positioning Cuts Through the Clutter," Advertising Age, 43 (May 1), 51-54. Ries, Al and Jack Trout (1981), Positioning: The Battle for Your Mind. New York: McGraw Hill. Ross, Harold L., Jr. (1982), "Recall vs. Persuasion: An Answer," Journal of Advertising Research, 22, 13-16. Rothschild, Michael I. (1987), Advertising: From Fundamentals to Strategies. Lexington, MA: D.C. Heath and Company. Sandage, C.H. and Vernon Fryburger (1975), Advertising Theory and Practice. Homewood, IL: Richard D. Irwin. 246 Sawyer, Alan G. Experiments in Consumer Research," Journal of Consumer (1975), "Demand Artifacts in Laboratory Research, 1 (March), 20-30. Schumann, David W., Richard E. Petty, and D. Scott Clemons (1990), "Predicting the Effectiveness of Different Strategies of Advertising Variation: A Test of the Repetition-Variation Hypothesis," Journal of Consumer Research, 17 (September), 192-202. Scott, Linda M. (1990), "Understanding Jingles and Needledrop: A Rhetorical Approach to Music in Advertising," Journal of Consumer Research, 17 (September), 223-236. Scott, Walter Dill (1903), The Theory of Advertising. Boston: Small, Maynard and Company. Sewall, Murphy A. and Dan Sarel (1986), "Characteristics of Radio Commercials and Their Recall Effectiveness," Journal of Marketing, 50 (January, 52-60. Shannon, Claude E. and Warren Weaver (1949), The Mathematical Theory of Communications. Urbana, IL: University of Illinois Press. Shimp, Terence A. (1981), "Attitude Toward the Ad as a Mediator of Consumer Brand Choice," Journal of Advertising, 10 (2), 9-15. Shimp, Terence and David C. Dyer (1978), "The Effects of Comparative Advertising Mediated by Market Position of Sponsoring Brand," Journal of Advertising, 7 (Summer), 13-19. Shin, In Sup (1989), Advertising in Korea. Seoul: Si-sa- yong-o-sa, Inc. Shin, In Sup (1990), The Changing Korean Advertising Scene. Seoul: The Korean Audit Bureau of Circulations. Singh, Surendra N. and Gilbert A. Churchill, Jr. (1987), "Response Free-Bias-Free Recognition Tests to Measure Advertising Effects," Journal of Advertising Research, (June/July), 23-36. Singh, Surendra M., Michael L. Rothschild and Gilbert A. Churchill, Jr. (1988), "Recognition vs. Recall as Measures of Television Commercial Forgetting," Journal of Marketing Research, 25 (February), 72-80. Smallwood, John E. (1973), "The Product Life Cycle: A Key to Strategic.Marketing Planning," MSU Business To ics, (Winter), 29-35. 247 Soley, Lawrence C. and Leonard N. Reid (1983), "Is the Perception of Informativeness Determined by the Quantity or the Type of Information in Advertising?" Current Issues and Research in Advertising, 241-251. Sorenson, Ralph Z. and Ulrich E. Weichmann (1975), "How Multinationals View Marketing Standardization," Harvard Business Review, (May/June), 38-167. Starch, Daniel (1923), Principles of Advertising. Chicago: A.W. Shaw Company. Stern, Bruce L., Dean M. Krugman, and Alan Resnik (1981), "Magazine Advertising: An Analysis of its Information Content," Journal of Advertisin Research, 43. 21 (2), 39— Stern, Bruce L., Alan J. Resnik and Edward L. Grubb (1977), "Information Content in Television Advertising: A Further Analysis," in Contemporary Marketing Thought, B.A. Greenberg and D. M. Bellenger (eds.). Chicago: American Marketing Association. Stewart, David W. and David H. Furse (1986), Effective Television Advertising: A Study of 1000 Commercials, Lexington: Lexington Books, D.C. Heath and Company. Stewart David W. and Scott Koslow (1989), "Executional Factors and Advertising Effectiveness: A Replication," Journal of Advertising, 18 (3), 21-32. Takada, Hirokazu and Dipak Jain (1991), "Cross—National Analysis of Diffusion of Consumer Durable Goods in Pacific Rim Countries,“ Journal of Marketing, 55 (April), 48—54. Tellis, Gerard J. (1988), "Advertising Exposure, Loyalty, and Brand Purchase: A Two Stage Model of Choice," Journal of Marketing Research, 25 (May), 134-144. Thomas, Hester (1986), "Growth Prospects Attract Business to South Korea," Accountancy (November), 97-98. Thorelli, Hans B., Helmut Becker, and Jack Engledow (1975), Information Seekers: An International Study of ConsUmer Information and Advertising Image. Cambridge, MA: Ballinger Publishing Co. Triandis, Harry (1972), The Analysis of Subjective Culture. New York: John Wiley. Triandis, Harry and V. Vassilou (1972), "A Comparative Analysis of Subjective Culture,“ in The Analysis of Subjective Culture. Harry Triandis (ed.). _r—w——v 248 Tse, David K., Russell W. Belk, and Nan Zhou (1989), "Becoming a Consumer Society: A Longitudinal and Cross- Cultural Content Analysis of Print Ads from Hong Kong, the People's Republic of China, and Taiwan," Journal of Consumer Research, 15 (March), 457-472. Twedt, Dik Warren (1952), "A Multiple Factor Analysis of Advertising Readership," Journal of Applied Psychology, 37, 207-215. Wagenaar, Jan Dirk (1978), "Creativity in Ads: Hard on Mood, Soft on Conflict," in "Marketing in Japan," Advertising Age: Supplement, 49 (January 16), n.p. Walters, Peter G.P. (1986), "International Marketing Policy: A Discussion of the Standardization Construct and its Relevance for Corporate Policy," Journal of International Business Studies, (Summer), 55-69. Weissman, George (1967), "International Expansion," in PlottingaMarketing_Strategyi A New Orientation, Lee J. Adler (ed.), New York: Simon and Schuster, 197. Wells, William (1986), "Global Advertisers Should Pay Heed to Contextual Variations, Marketin News, (February 13), 18. Werner, O. and D.T. Campbell (1970), "Translating, Working Through Interpreters, and the Problem of Decentering," in A Handbook of Cultural Anthropology. R. Naroll and R. Cohen (eds.), New York: American Museum of Natural History. Whang, In-Joung (1987), Century," Korea Journal, "Korea Toward the Twenty-First 27 (April), 21-32. "Female Whipple, Thomas W. and Alice E. Courtney (1985), Role Portrayals in Advertising and Communication 4-8. Effectiveness," Journal of Advertising, 14 (3), White, Percival (1927), Advertising Research. New York: D. Appleton and Co. Wilson, R. Dale and Aydin Muderrisoglu (1979), Claims in Comparative Advertising: A Study of "Strength of Competitive Vigor," in Educators' Conference Proceedings, Series 44, N. Beckwith, et. Chicago: American Marketing Association. al., eds. "The Cognitive Responses Mediating Wright, Peter (1973) Acceptance of Advertising," Journal of Marketing Research, 10 (February), 53-67. 249 (1984), "The Effect of Yalch, R.F. and Elmore-Yalch, R. Numbers on the Route to Persuasion," Journal of gonsumer Research, 11, 522-527. Yi, Youjae (1990), "Cognitive and Affective Priming Effects of the Context for Print Advertisements," Jourmal of Advertising, 19 (2), 40-48. Y1, Youjae (1990), "The Effects of Contextual Priming in Print Advertisements, Journal of Consumer Research, 17 (September), 215-222. Yum, June-Ock (1987), "Korean Philosophy and Communication," in Communication Theory From Eastern and Western Perspectives, (D.L. Kincaid ed.), New York: Academic Press, 71-86. Zinkhan, George M. and Claes Fornell (1985), "A Test of Two Consumer Response Scales in Advertising," Journal of Marketing Research, 22 (November), 447-452. Zinkhan, George M. and Betsy D. Gelb (1986), "What Starch Scores Predict," Journal of Advertisin Research, (Aug/Sept), 46-50. IllWWIWill!HI"NHIHIHIIIIHIllHHllJlHlJHlHHHI