#13" mm, :11 rt’ 11“ ruangy-.. W515 5:551“! "355’“ L 1 - . 5 5:1:€'-.§I'7:555}H5 13$|l55k51fi ,1,- =5 $1: 5%“ RJ 51i‘:1'fi,1~£.5 ‘7: "' 55131“ fi:%w1 d}- 515’ w- M 3*? — 155 :' ' J ’ 5 :3 5 “4% ‘ I? k ."11 ‘ p I.) ‘ ‘ ..- L ~. , .121: 1.1.,5,‘ 1:: v . .5. éA::l. I ', :{!g; :).{2 ' gut: fig: 5.: ‘s ‘ ' vhf .: .7331}. .- ‘, :1? A .3 a; ' 3;“ 31K! ifig‘ -. :_ _ ~‘.-. .1231 ‘ t . ‘. .35: ‘ ' mg, 1 ‘Cflifififik I A: "‘5'“ 1 1. 0.13%“: 5:.» ,1: "1‘1: .2 , 1:1“ ‘ ‘ Q: 5 1155:: 1:515 '55"! ":5 ~ -. , . 1 (1 : . 1' 5:515 “ 531:;‘555155555155155”Law: _ g 17535; T ., 55!, to .5553} I 2525!": ‘i (@7113 15;;1'; ‘. . J: -‘ 1:51;. . ' “ ‘ ‘ 5 W15 5 ' 1555:“: ‘ w: :::11:Wfi:1%::w-ww«mfi111 ~. , v _ M 11- 1 ' * 1. : < ’* "a =:' #1:: 11:31 1: ’71” ‘ I“ :3: :11 :11: 1 1%: :1 , l '_ H4 :1? 2:. l f' £1. ‘. airfiflmu‘“ “TQM”! “1155055? 1 h! fafl. rum”, 'Ei' 1%)? f} 5355“ “Ht! k "4:" ‘ f . *‘1 ' : 8 51:5: 311-} 35;” 5 13 £535 l‘k- 31k?” ”1%3Ré‘fl5‘5’3 1:5” J53] UV :ng 3'. ' “ , 1 F5:- 5&4ng ‘5}1319'7351fl33‘:‘z‘J1,735C,’f§'71,31"l5:1: :‘WJi " "-l-' ' 1 5‘; 3511:: xii. 4:715:14? 50",): ”169333,:6‘1114 a"! kJ‘JJ‘ng’QVI,‘ 5?; £35. 1'” "’11745 a"; ‘43“, 1.25:; ng‘Lm , . 4‘13“", at”! “5125: ‘7 «in: l $111“ :75- 5’ “fl?“ a11:?:':: ”31‘ H :35 1% 1111-51,: E? {[3qu 1&6, F1 ragpknm h! :11 113:, 1’ ’: fl "f "75 1 =;;.~1::1;;: ”“1111:..1,1,:{;,:WWI: :11: 1:11. m '51: £914”; ----- ' 3 :1 v {111“‘5155'555’3551555' 55 .4' $1M) 5W55f5h J6? 13" 5; Iv “I ,. L {.393 ~. 1'71; 3353;11qu 49' ' "fit v ’5 1": ‘ {’- . .:.‘f'5" , :“' v‘ 2- '151J " .5 11mm 12:: :3 M51 LIBRARY ‘ Michigan State University This is to certify that the thesis entitled SCRIPTS, INVOLVEMENT, AND ADVERTISING EFFECTIVENESS: AN EMPIRICAL INVESTIGATION presented by Ingrid M. Martin has been accepted towards fulfillment of the requirements for M.A. . Advertising degree 1n | Major professor t Date 7!3’/87 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU LlBRARlES -:—. RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. JAN 1 3 2000 '0’} '7' 0 4. AUG 30 7 2004 ”PLEASE NOTE: In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark J . A O PPSPQPPN d .0 d d o 12. 13. 14. 15. 16. Glossy photographs or pages _ Colored illustrations, paper or print __ Photographs with dark background __ Illustrations are poor copy 4 Pages with black marks, not original copy ___‘/_ Print shows through as there is text on both sides of page lndistinct. broken or small print on several pages / Print exceeds margin requirements Tightly bound copy with print lost in spine Computer printout pages with indistinct print Page(s) lacking when material received, and not available from school or author. Page(s) 122 seem to be missing in numbering only as text follows. Two pages numbered . Text follows. Curling and wrinkled pages Dissertation contains pages with print at a slant. filmed as received Other U-M-I SCRIPTS, INVOLVEMENT, AND ADVERTISING EFFECTIVENESS: AN EMPIRICAL INVESTIGATION By Ingrid M. Martin A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Advertising 1987 ABSTRACT SCRIPTS, INVOLVEMENT, AND ADVERTISING EFFECTIVENESS: AN EMPIRICAL INVESTIGATION By Ingrid M. Martin This study was undertaken to investigate the potential of scripts as the theoretical framework to initiate and foster consumers’ advertisement recall and recognition in an attempt to create changes in their affective and cognitive structures. Scripts are proposed as an effective tool to increase consumer awareness, to develop a favorable attitude toward the brand and to increase purchase likelihood. The level of involvement is proposed as a variable that can affect the outcome of script applications by predicting the depth at which an individual processes the message. A two by two factorial design is employed to measure the level of recall and attitude that subjects display under high versus low involvement and exposure to scripted versus nonscripted advertisements. The results of the experiment found the script format to be an effective method to convey print advertisement. It is then proposed as an efficacious tool to develop public service announcements as well as to regulate the advertising of controversial products. To Wade for his undying patience and encouragement and to Sevgin for all the confidence and support without which the struggle would have been insurmountable. iii TITLE: I. II. III. IV. TABLE OF CONTENTS Chapter One Introduction Chapter Two A. Script Theory: A Theoretical Perspective Historical Development Fundamentals of Script Theory Script Theory: Some Empirical Applications Involvement A Process Orientation A State Orientation Central vs. Peripheral Routes Of Persuasion Chapter Three Methodology Research Definition and Hypotheses Dependent Measures Independent Measures The Medium The Product The Sample The Experimental Materials The Procedures Independent Variables Dependent Measures Coding Procedures Chapter Four Empirical Results Manipulation Checks Hypothesis One Hypothesis Two Hypothesis Three Hypothesis Four Hypothesis Five iv PAGE NUMBER: 10 18 29 30 36 42 51 51 52 54 56 59 60 61 65 66 68 70 71 71 72 72 73 75 76 VI. VII. Chapter Five Discussion and Conclusions Theoretical Implications Managerial Implications Public Policy Implications Limitations Future Research Appendices Appendix A: Questionnaire Appendix B: Booklet 1 Appendix C: Booklet 2 List of References 78 78 81 83 84 85 87 93 94 95 LIST OF TABLES TITLE: PAGE NUMBER: I. Script Format 64 II. Nonscript Format 64 III. Analysis of Variance Results for Hypothesis 3 74 IV. Analysis of Variance Results for Hypothesis 4. 75 V. Means for Each Experimental Cell on the Attitude Index 76 vi LIST OF FIGURES TITLE: PAGE NUMBER: I. A Continuum of Script Development 15 II. Four Level of Cognitive Activity 33 vii CHAPTER ONE INTRODUCTION Research concerning consumer response to advertising efforts years has become an important area of study in recent years. This field encompasses various topics ranging from measuring advertising effects on brand and product evaluation (Mitchell, and Olson 1981; Gardner 1983; Mitchell 1986) to determining advertising effectiveness through .consumer attitudes toward the advertising message (Lutz, MacKenzie and Belch 1982; Moore and Hutchinson 1982) and applying various consumer behavior models to measure consumer responses to advertising efforts (Wilkie and Pessemier 1973; Lutz 1975; Holbrook 1978; Shimp 1981; Mitchell 1982). While a massive literature exists in advertising effectiveness and consumer response to marketing communications, findings have been neither totally coherent nor complete (Fishbein and Ajzen 1981; Shimp 1981; Miniard and Cohen 1981). This current study addresses a portion of this controversy by presenting a new theoretical framework as a potential means to better understand and improve advertising effectiveness. Specifically, the script format is proposed as an effective framework to initiate and foster consumers’ advertisement recall and recognition in an attempt to create changes in consumers’ affective and 1 cognitive structures. The extensive schema and script literature is, therefore, reviewed to formulate the conceptual and empirical basis of this study. Scripts are routine, sequential sets of vignettes that an individual uses to guide his/her behavior and to reduce the amount of cognitive processing needed to perform a task. Their application in advertising effectiveness research is a fairly new development. One specific example of script applications to advertising is television media protrayals, which juxtapose people and plot elements in highly predictable ways so as to reinforce popular stereotypical beliefs. These portrayals are incorporated in widely accepted schemata and scripts which are familiar to their target audience. The result is a message which is readily processed, due to the familiar script and increased advertising effectiveness. Another example is the McDonald’s advertisement which shows the whole family going to McDonald’s selecting, ordering and paying for their food, and then eating their meal. Application of scripts to the print medium is a fairly new development. While this medium is very constrained in using pictures to present and execute a script mainly due to space limitations, it nevertheless promises to be a creative and potentially effective advertising technique in developing advertisements. Common to all of the above examples is the existance of an event-centered, coherently linked chain of episodic scripts. These scripts aid in conveying a message that is readily processed and is more likely to be recalled due partly to the scripted format of the advertisement. Some of the realistic goals of advertising include increasing the consumer’s memory for the advertising message, developing a favorable disposition towards the advertised brand, and consequently increasing purchase likelihood. Based on the theoretical and empirical literature in psychology and marketing, scripts are being proposed here as an effective tool to achieve these goals. By presenting a subject with a scripted advertisement, the advertiser may be able to invoke increased awareness and attention to the message. It is expected that such increased awareness and attention will then result in heightened memory for the event and the message, eventually culminating in a favorable action towards trying and adopting the advertised brand. Another implicit concern with advertising memorability is the effect that scripts can have on consumer attitudes towards brands as well as advertisements. Abelson (1981) postulates that concrete information is much more effective in forming attitudes towards an object than abstract information. This implies that scripted advertisements have the potential to favorably change and improve attitudes towards a message as well as a brand by attaching certain behaviors to the advertisement. A person’s attitude toward a product is the result of a series of '45 episodic scripts relating the experiences either in actual or vicarious terms to the attitude object. When the individual is queried about his/her attitude toward a certain product, s/he recalls the episodic scripts and responds accordingly. Due, perhaps in part, to the relatively recent development of script theory, research on its advertising applications is virtually nonexistent. The only exception is an exploratory study by Puto (1984) who found that script intrusions result in improved recall of advertising messages. Since, as noted by the author, the design was informal and the methodology inadequate, the study found no positive results for measures of attitude change and behavioral intentions. These findings have not only suggested the need for further research but also pointed to the necessity to include some other variables that could mediate the dynamics of script formats within the context of advertising effectiveness. A variable that is proposed to affect the outcome of script applications in advertising is the consumer’s level of involvement with the issue. Based on a review of the literature, a definition of involvement is presented as ”... the complexity of cognitive and behavioral processes characterizing the overall consumer decision making process" (Houston and Rothschild, 1977). This is further exemplified by Cohen (1982) when he states that we should discontinue the practice of referring to the consumer as in a state of low involvement or high involvement just because the individual’s attention is disproportionately going to something he or she may be interested in (i.e. aspects of the advertising execution such as the scripted format) rather than what the advertiser is interested in (i.e. the issue). Here, then, the concept of personal importance and/or relevance evoked by the variation in advertising format becomes the major measure of involvement. The key point in processing the advertisement is understanding it. The level of involvement will predict the depth at which an individual understands the message stimulus. By understanding, an individual finds the script that fits thestimulus (message) and then uses it to draw inferences about the claims made in the advertisement. This study is an attempt to demonstrate the interaction between scripted versus nonscripted advertisement format and the level of issue involvement in predicting advertising effectiveness. The study has three major objectives: (1) to develop a realistic operationalization of the script advertisement concept that is consistent with its theoretical underpinnings, (2) to investigate the impact of script versus nonscript print advertisements on recall and attitudes, and (3) to explore the impact of issue involvement on the script versus nonscript print advertisement format. These issues will be examined to determine the impact of using scripted advertisements on advertising effectiveness and on selected consumer cognitive response variables. In the next chapter, the theoretical foundation of script theory is examined in detail along with some of the empirical applications of script theory in psychology and marketing. Further, the literature on involvement and specifically product involvement, is presented. Chapter Three discusses the research problem and the corresponding hypotheses. Also presented here is the rationale for the product, medium, and sample chosen as well as relevant definitions of measures and their operationalizations. This chapter includes the results of the pretest conducted to verify the existence of a story in the scripted version of the advertisements. Last, the experimental design and execution are covered along with an overview of the questionnaire included in the Appendix. In Chapter Four, the empirical results, along with the statistical tables, are discussed. Finally, Chapter Five presents the conclusions based on the empirical results presented in Chapter Four. The conclusions are discussed in the context of theoretical, managerial, and public policy implications. Finally, the limitations of the experiment as well as propositions for future research are presented. CHAPTER TWO A. SCRIPT THEORY: A THEORETICAL PRESENTATION Part of our knowledge is organized around stereotyped activities such as going to a restaurant, shopping at a grocery store and visiting a doctor. Either by direct or vicarious experiences, we acquire numerous cultural stereotypes along with some personal, idiosyncratic variations on these stereotypes. These routine, sequential sets of vignettes are termed scripts by Schank and Abelson (1976) and are the basis for script theory. In the last ten years, this field has provided an important approach to research in human memory by psychologists. This aspect of memorability is what has made script theory a potential new field of research in advertising. The effectiveness of an advertisement is governed by its memorability and script theory provides a promising theoretical perspective of improving consumers’ memories for advertisements. This study is developed to determine if there is a measurable difference of effectiveness between script and nonscript ads. The major theoretical proposition espoused here is that scripts provide a useful framework for understanding and assessing the outcomes of various advertising activities. The first part of this discussion will provide the historical background of scripts and their theoretical foundations in the form of a literature review. This is followed by an explanation of the fundamentals of script theory concentrating on Abelson’s views. Historical Development The notion of scripts and script-related theories provide the conceptual framework used to develop the theoretical and empirical basis proposed in this study. Consequently, it is essential to explain the concept in some depth. Script theory has its theoretical foundation in 'the concept of schemata which are proposed to aid individuals in information processing and recall. To fully understand the theory on which scripts are founded, a brief discussion of modern schema theory is needed. In 1932, J.C. Bartlett developed the concept of schemata which was then further developed by Minsky (1975) with frame theory, and finally, Schank and Abelson (1977) with script theory. In his book Rgmgmbggigg, Bartlett (1932) defines a schema as an active set of past reactions and experiences which are organized so as to operate in any well-adapted organic response. He proposes that the characteristic properties of schemata are that they are organized, composed of old knowledge and generative, in the sense that they can deal with an indefinitely large number of new instances. His major application of schema was its role in aiding the reconstruction process individuals use in recall. For many years Bartlett’s conceptualization of schemata lay dormant until Minsky (1975) published his research on frame theory. This theory builds on the concept of schemata by giving it a narrower and empirically testable definition through applications to artificial intelligence. Minsky (1975) defines a frame or schema as a knowledge structure which provides general information expected in a situation in an ordered sequence. The next development in schema theory came with Abelson’s (1976) work in script theory. Basically, this theory posits scripts as schemata held in memory which contain knowledge of sequences of events or behavior that are appropriate for a given stereotyped situation [Abelson 1976 and Schank & Abelson 1977]. More specifically Abelson (1981) describes scripts as one type of schema which encompasses most of the conceptual issues raised by other types of schemata "... yet simple enough and well structured enough to allow more focused analysis and representation." (Abelson, p. 715). Information processing via the use of scripts begins when an individual encounters a new stimulus event. S/he then draws upon its representation of the event and uses it to complete the attributes of the stimulus and to generate predictions about other attributes and ensuing events. The 10 basic structure of the script is a hierarchical one based on levels of abstractions and a temporal ordering directed by imputed goals (Taylor & Crocker, 1980). The theory predicts that the memory trace which embodies a highly typical and frequent event that occurs in a particular episode will be forgotten or omitted from the trace. In other words, scripts aid the encoding, storage and retrieval of information in episodic memory. Certain redundant information need not be stored since it can be extracted from the prototypical script. One needs to only remember that a scripted event occurred to recall highly probable segments. The result of this representational array then is the efficient storage of information in longterm memory (LTM). Fundamentals of Script Theory In his seminal work on scripts, Abelson (1976) provides the theoretical foundation where he departs from the ‘”classical” view of cognitive and social psychology in information processing. He attempts to organize the application of a set of available ideas to areas where they have been little used. This is seen in the previously mentioned concept of the relationship between scripts and episodic memory. The dominant view in cognitive psychology is that knowledge is represented through propositional networks such as John Anderson’s ACT model (Howard, pp. 204-21). In 11 his ACT model, Anderson developed a theory of long term memory which explains the structures and processes underlying memory as well as inferential reasoning, language acquisition and comprehension. These types of network models are based on the idea that understanding involves some form of match between inputs and known propositions. Abelson attacks this view not in the abstract theoretical but in the practical application of this concept. The experimental approach used in cognitive psychology uses what he terms "silly sentences" exclusively with little or no concern paid to pragmatics, that is, "... to the effects of available episodic or catergorical vignettes or scripts on the interpretation of language." (Abelson, p. 36) The basic characteristic of a script is in the vignette which represents the "... raw constituents of remembered episodes in the person’s experience" (Abelson, p.34). A vignette encompasses an image of a perceived event and a conceptual representation of that event. It is important to remember that the representation is not a verbal- acoustic code but it would be such that verbal inputs and outputs could be attached to it (Abelson, 1976). A convenient conceptualization of vignettes is then thought of as a picture plus a caption. A realistic conceptualization of vignettes could be the idea of networks of nodes containing vignettes which are linked together by the propositions that contain the 12 temporal ordering characteristics of scripts. The node contains an abstract, hypothetical vignette which is linked to other hypothetical vignettes by the variations that can occur in a script. These variations result from distractions, interferences, or free behaviors. This may be the pragmatic theory that Abelson (1976) alludes to in his research. Scripts then become a coherently linked chain of vignettes with the simplest version consisting of two panels. The first panel sets up the situation and the second resolves it. Abelson posits three processes that aid in the development of vignettes in long term memory. First, vignettes can be stored as single experiences resulting in storage in; episodic memory. Second, " similarity groupings can then build up categorical vignettes instantiated by many single experiences in a given type of situation" (Abelson, p. 35). Finally, a third level of processing occurs in which individual features are processed instead of the specific vignettes from which they came. These three levels of processing move from episodic memory to semantic memory as can be applied to a restaurant script in the following example. (1) At the episodic level, the individual would process this ad in the following postulated manner. "I remember the day I went to Bennigan’s and met this great guy, Brad. He asked me for a cigarette and the next thing you know we were deep in conversation getting to know each other.” (2) At the categorical level, the individual would store the script in the following way. This would be a generic 16 vignette of going to a restaurant with friends which is linked with the generic vignette of different ways to meet a great guy. These two generic vignettes would form a script as seen in the ad in Figure l. (3) At the highest level of abstraction, this set of vignettes would be stored in hypothetical form. Episodic memory is no longer directly involved but instead semantic memory is processed using a set of conditional and inferential concepts and referring to possible categorical scripts. For example one way to meet a great guy is in a restaurant where you can strike up a conversation by asking for a cigarette. If he doesn’t smoke, you can ask him to light your cigarette. These two hypothetical situations are all dependent on the individual situation. In a later article, Abelson (1981), explores the idea of a weak script versus a strong script in addressing the differences in sequencing properties across events. He defines a weak script as similar to other types of cognitive schemata that do not use sequence information and which organize expectations about certain objects or events. This is in contrast to strong scripts which involve expectations about the temporal ordering as well as the hierarchical ordering of events. Evidence for these properties are found in a series of experiments conducted by Bower, Black and Turner (1979) discussed later. Abelson (1981) contends that scripts play a double role in cognitive psychology. One can "behave" a script, that is take on the role such as a customer in a restaurant. To do this, one must understand that this possibility exists as well as to be able to perform the role. The second role of scripts is understanding a situation which implies the cognitive retrieval of previous situations to which the present situation is similar. 14 Three conditions then become necessary for scripted behavior to occur. First, one must have a concrete representation of the script. Second, a situation eliciting the script must be experienced. Third, the individual must enter the script. An example of this could be how a person experiences the restaurant script. First, she must know how to be a customer in a restaurant; second, there needs to be a stimulus which makes going to the restaurant attractive (i.e., satisfy her hunger). Finally, a decision must be implemented to enter the restaurant and begin the scripted behavior. Performance of the script results in behavior which is appropriate for a given situation. It is the result of some conscious or unconscious processing of environmental cues by the individual. Conscious, active information processing can occur in situations where the scripted behavior is relatively new to the individual (i.e., going to an expensive restaurant requires some new variations on the restaurant script if one is used to only going to fast food places). This is in contrast to unconscious processing which is described as ”mindless", essentially automatic behavior. For example, when someone eats at a Burger King on a daily basis, this behavior can be described as more or less automatic or mindless. It can be generally accepted that people do think about what they are doing but the intensity and depth of the processing varies according to the specific situation. L5 This is explained in Figure 1 where the progression toward automatic script processing is represented as a continuum of script development. FIGURE 1 A CONTINUUH 0E SCRIPT DEVELOPMENT Novel Infrequent Conventional Partially Stereotypical Situation Situation Situation Stereotyped Situation Situ];ion Unecripted Script Conacioue Heat hoto- Strong Photoecript Behavior and Development Performance of Script Modified Performed Mindleeely Action Cued Expected Behavior and Performed LOU HIGH 3; 4 , Typicality of the Situation (Adapted from "Scripts in Organizational Behavior”. (198‘), Acadggz of Management Sciencea. p. 653.) Abelson (1981) addresses several variations that can be predicted in a particular enactment of a certain script. These unexpected sources of variations include interference, distraction, and free behavior. Interference can be in two forms, as obstacles and as errors. An N obstacle is ... something that removes a precondition for a given event ...” (p. 726) such as the great guy sitting next to you can’t give you a cigarette since he doesn’t smoke. An error is the incorrect completion of a certain script such as instead of giving you a cigarette this great guy hands you a cigar. The second type of variation is distraction which is 1.6 defined to be an event of sufficient importance to interrupt the script action, though it does not interfere with any uniquely specific event. (Abelson, 1981) An example would be the unexpected event of a fire in the restaurant where you are having lunch with this great guy. The third variation is free behavior which involves those activities that may plausibly intermix with the ongoing script such as conversation at the next table. This conversation may be unattended by the subject and not affect the ongoing script actions or something in the conversation may catch the attention of the subject. This would then become a distraction much like Cherry’s cocktail party phenomena where semantic processing does occur outside the subject’s consciousness. Relating scripts to text comprehension, den Uyl and Oostendorp (1980) address another important property related to scripts. They contend that scripts make it possible to expect certain events which is a critical notion in advertising. Expectations about product attributes or a situation can facilitate understanding the advertisement because it changes the task of understanding to one of elementary recognition which in turn makes the search of long term memory much less problematic. Expectations also provide the groundwork for making inferences about the future outcome of a script. This processing can be either active as a top-down process or passive as a bottom-up process. The level of specificity 17 of one’s expectations can determine the future outcome of a script. This is dependent on the amount of detail in which future events are represented and in turn can aid in greater comprehension of scripts. A high level of specificity in advertising would imply a decrease in misconceptions about product/brand attributes. To test the difference between the effectiveness of a scripted and a nonscripted ad, it would be possible to affect different levels of specificity. These different levels of specificity determine if there is greater comprehension resulting in greater ad recall. B. SCRIPT THEORY: SOME EMPIRICAL APPLICATIONS The most widely referenced research work in script theory is the set of seven experiments undertaken by Bower, Black and Turner (1979), (herein referred to as BBT). The findings in these experiments have been replicated and extended by den Uyl and Oostendorp (1980), Leigh & Rethans (1984), Whitney & John (1984), and Puto (1985). These experiments were all designed to investigate people’s knowledge of routine activities and how this knowledge is used to organize, interpret, and remember narrative texts. These researchers use scripts as a means to identify how subjects use knowledge to expand upon what they are reading. The set of seven experiments conducted by BBT provided the foundation for empirically testing script theory. All the other studies done in pscyhology and marketing are either replications or applications of these seven experiments. A discussion about each of the experiments is therefore necessary at this point. The first experiment was designed to generate a script for a routine, automatic behavior. BBT found that there was a high degree of agreement in the "basic action” language used by respondents to describe a continuum of events. For example, in the restaurant script of the 730 actions mentioned in total, only four were completely unique (given by a single person). The second experiment explored the concept that a I8 19 script is not an undifferentiated linear chain but instead it is an organization of chunks composed of subsequences of actions. To determine the boundaries of these chunks, subjects were required to segment lower level action sequences in a script. An example is deciding what subsequence of actions made up a scene (i.e.: ordering food) in the RESTAURANT script. In the third experiment, BBT explored the premise that subjects in recalling text will use the underlying script 'to fill in gaps of intervening actions not explicitly mentioned in that text. As their short term memory of the text fades, they rely upon the fully completed (underlying) script, which results in unstated actions intruding into recall. This includes generic scripts for similar actions such as VISITING the DOCTOR and VISITING the DENTIST scripts. Some primary findings include the following. First, stated script actions are recalled more than unstated script actions. Second, the percentage of unstated script intrusions increased when other story versions of the same script were read by the subjects. For example, the DOCTOR and DENTIST scripts have many overlapping actions which was found to increase the probability that the unmentioned analogous action can intrude in recalling the related story of VISITING the CHIROPRACTOR. The fourth experiment was an attempt to replicate the previous experiment using recognition measures rather than 20 recall measures. The important point here is the extent that "... studying a sentence describing an action activates associated actions in the underlying script and ." will later be recognized as having that these actions been read. (BBT,1979,p.193) BBT attribute this to a failure to discriminate the source of activation, whether it is due to an explicit or implicit presentation. Their results confirm that subjects have a fairly high rate of false recognition for actions in an underlying Script. In both experiments three and four, it was found that intrusions and false alarms increase if subjects read more than one text that instances the same script (i.e.: Jim visits his doctor or his dentist). The issue here is does advertising which uses scripted behavior have the potential to evoke a script by showing just small pieces of that script. By examining the advertisement in Figure l, is it possible that this two vignette script is capable of instantiating a favorable memory of a similar situation? In the fifth experiment, BBT tested the concept that some scripts are strongly ordered actions. If their order is arbitrarily changed, recall should prove difficult because the cannonical order in memory should serve as a source of proactive interference (PI) against learning of this new version. Proactive interference is the confusion that an individual experiences when attempting to process new information due to information learned at an earlier 21 point in time. The results found that "... in ordered scripts, misordered actions were recalled less accurately than their (equivalent) yoked actions (in unordered scripts), but ordered actions were recalled more accurately than their yoked actions." (BBT,1979,p.205) The cannonical order of a script aids people in learning the order of any new text as well as improving recall of script actions. The script order stored in memory is a source for guessing when the presented order has faded from memory. Relating events and stories in'a manner 'other than reciting. their temporal order is common. An example is sports reporting where the final score is related first then significant events are recounted in order of their importance. This tells us that even though scripts are defined in terms of temporal ordering this is not a necessary condition in recalling a text. The sixth experiment tested the idea that temporal distance between events can be used as an index to determine the number of alternative actions that a person can access in the underlying script. Their results found there to be little relationship between temporal distance between actions and their distance in long term memory. An example given by BBT is based on the point that scripts often have some causal or contingent connections that link large temporal gaps between events (i.e.: the size of a tip given to a waiter depends on the service given earlier in 22 the restaurant script). In the final experiment, BBT investigated how well people recall different types of interruptions in comparison to the script actions. They test three types of interruptions termed obstacles, errors and distractions. (These terms were defined earlier in the paper.) The hypothesis is that interruptions will be more memorable than script actions and that irrelevant actions will be the least memorable. The results of the seventh experiment found that obstacles and distractions were better remembered than errors. This alone has important implications for its use in advertising strategy. BBT examines memory for occasional events by inserting them into script-based stories to determine the level of deviation from the "smooth running" script. A scripted advertisement can use a distraction to increase consumers attention level hence the memorableness of the advertisement is heightened. The empirical applications of script theory in marketing include the Leigh & Rethans (1984) exploratory study. This experiment was designed to address issues related to script elicitation and organization within a consumer decision making context. They developed four experiments to investigate the existence of script norms in the context of an automobile purchase. The first experiment established the existence of script norms. The second experiment investigated the possibility that oral 23 elicitation was not as effective in determining script norms as paper and pencil elicitation. The last two experiments manipulated instructions to subjects in an effort to determine response sensitivity to task delineation. The objective of all four experiments is to explore the conceptual issue of order of events in scripts in a consumer behavior context. The results of the first experiment found a high degree of agreement among subjects on a basic set of actions used to describe the decision making process. BBT call these basic action language. This sequence of events was determined to fit a five step conceptual model of consumer decision making developed by Olshavsky and, Granbois (1979,1980). The steps include problem recognition, search, alternative evaluation, choice and outcomes. In experiment two the possible effect of oral elicitation on the context of the proscribed script was determined. They found that the script seemed to overlap on some of the most salient features of the paper and pencil script used in the first experiment. The results were that the written elicitation involved a more thoughtful process with more detailed reporting of actions. The third experiment verified the concept that scripts are hierachically ordered and that the entering conditions define the level of appropriate abstraction. For example, the subjects were asked to report the scripted actions of a salesperson rather than their own. The data showed that 24 the subjects had little difficulty reporting what they thought should be the scripted behavior of a salesperson. The final experiment asked subjects to list all actions that both the consumer and the salesperson would be involved in during a sales transaction. The result was a summary script that reflects not only the temporal ordering but the interactive nature of the script. This interactive procedure allowed more specific role decomposition than the previous experiments but still maintained the same script ,structure. In summary, the authors proposed the script concept as a useful tool in describing and explaining consumer decision making processes. This is due to the fact that there seems to be a basic action language used to describe certain behavior and that the structure of scripts is found to be consistent across different behavior patterns. The important point is the consistent finding that scripts are hierarchically organized with varying levels of abstraction. In the Whitney and John’s (1984) study, they suggest that sometimes a stimulus may vary from the stored script, dominating the subsequent recall. As short term memory fades, this reliance on scripted information will intrude upon the unstated script actions in recall. They set up their study by first identifying two scripts which were relevant to the subjects in the study. The two scripts were for a McDonald’s purchase episode and the purchase of a stereo from a large, local discount 25 retailer. The goal of the study was to show that consumers will intrude unstated script actions when recalling events. They used two types of manipulations to develop their narratives. In one version, a main event was not mentioned in its usual ordering but instead was moved to another position in the sequence of events. If subjects were truly guided by stored scripted actions, then there should not have been any difference between either the correct version or the changed version. In other words, Whitney and John hypothesized that memory for the event is identical regardless of where it was explicitly mentioned in the narrative. The second manipulation was the insertion of actions in the narrative that was not present in the baseline scripts. The authors did not expect unstated script intrusions because no script—based expectations are triggered regarding these events. Instead, they expected differences between control and experimental groups. The importance of this second manipulation is that it establishes the conditions under which expectations elicited from scripts do not affect memory. The results of Whitney and John’s experiments point toward a script-theoretic mechanism due to the fact that intrusion did occur when recalling script-based narratives. They also found that those events that are not part of the ordered script are not subject to intrusion errors. 26 Another finding was the idea that salient events in a script are similar to non-salient events in that they are not affected by intrusion errors. An important implication of Whitney and John’s study is its extension to advertising of products. If there are scripts for advertising, it is possible to expect intrusion errors to exist in memory across brands in a product category. This means that advertisers should design their advertisements to conform to script-based expectations so as to increase advertising recall. The Puto (1985) study proposed scripts as an effective way to improve advertising memorability. He positions scripts as the ideal mechanism to aid consumers in monitoring and perceiving multiple incoming stimuli without having to devote full attentional capacity to each stimulus. Advertisers seek to increase memory for these multiple stimuli and to have consumers retain a favorable attitude toward the brands and products. Scripts are therefore proposed as a method to better understand and increase the memorability of advertising messages. The major finding of Puto’s study was that script interruptions result in better recall than standard script actions. This suggests to advertisers that improving memorability for key points in advertisements can be done by inserting them as interruptions. The study found no positive results for attitude and behavioral intention measures due to the informal design and the inadequacy of 27 the measures. In another exploratory marketing study, Bozinoff and Roth (1984) tested some important properties of scripts by applying them to consumer low involvement decision making behavior. Consumers who are in the process of making a routine, inexpensive purchase will use the automatic nature of this behavior to activate a script which will guide their behavior. The results of their experiments provided support for the automatic nature of scripts. The implications for advertisers are threefold. First, because of the unconscious nature of scripts, advertising effectiveness can be improved immensely by creating interruptions which make the unconscious script conscious. Second, to create behavior the advertiser wishes, advertisements should be related to uncommon activities to make it memorable. Third, it is not wise to depend on consumers verbalized intentions because ultimately it is the unconscious script that guides the routine behavior. Some of the unresolved issues about script theory are discussed by Bower, Black, & Turner (1979). The first, and most important is the inability to measure the tacit, nonintrospective knowledge people have about stereotyped procedures and activities. Individuals are only able to report script knowledge that is accessible to conscious introspection. This makes the application of script theory to advertising research problematic in the sense that much 28 of the exposure consumers have to advertising is difficult to measure due to a cluttered environment. This means it is difficult to determine what is truly unattended, what is perceived by the receiver, and what is unattended yet subconsciously "perceived" by the receiver. In the next section, the literature on involvement will be reviewed in terms of its applications to script theory and measuring advertising effectiveness. C. INVOLVEMENT The literature provides different conceptualizations of the involvement construct. These concepts include ego involvement, commitment, product or purchase importance, communication involvement, and issue involvement, just to name the most frequently used terms. The common dimensions underlying the involvement measure have yet to be identified. This literature review divides involvement into three basic approaches or orientations. The first is process orientation which defines involvement as the result of the level of cognitive processing by the individual. The second approach is state orientation which views involvement as the determinant of the direction and intensity of the information processing performed by the individual. The third approach is based primarily on Petty and Cacioppo (1979, 1983) conceptualization of involvement as a moderating variable in the attitude formation process. Each approach is reviewed here based on the most relevant studies in the literature. To avoid the confusion that exists in the literature, only the involvement studies with advertising and communication applications are discussed. Once the three orientations are developed then five involvement typologies are identified and defined as the most commonly used in the advertising and marketing literature. Finally, the review ends with the identification of issue involvement as the pertinent 29 .30 construct with its application to advertising recall. A Process Orientation The first approach which views involvement as a process oriented construct is founded in Krugman’s (1965,1966) seminal work on attitude change due to an individual’s exposure to advertising messages. Krugman postulated that the individual’s level of involvement was the result of the depth of the cognitive information processing and decision making employed. This will in turn determine whether an individual will develop or change an attitude toward a persuasive message and/or brand. Krugman goes on to suggest that under high involvement, attitude development is more likely to occur due to exposure to a persuasive message. Attitude is also mediated by the type of medium used to present the message. In other words, the print medium is considered to be a high involvement medium because it is passive, requiring an active consumer to receive information. The broadcast media, in turn, are low involvement because they take the active role allowing the consumer to passively receive information. This generalization concerning the role of media as a predictor of involvement level may be disputed. For example, how active is an individual’s cognitive processing activity when s/he flips through a magazine in a grocery store line or in a doctor’s reception area? Therefore, this claim made by Krugman must be used within 31 the situational context to which it applies. Another implication underlying Krugman’s conceptualization of this construct is the two different models of the learning hierarchy, which explain the different levels of involvement. Under conditions of low involvement, affect or attitude does not necessarily precede behavior whereas under high involvement the traditional hierarchy of effects still holds. The common thread throughout the literature of process oriented conceptualizations is the assumption that cognitive activity determines the level of involvement. Some other researchers who have conceptualized the involvement construct as process related are Batra and Ray (1982), Antil (1984), Greenwald and Leavitt (1984), Rothschild and Houston (1977), Zinkhan and Mudderisoglu (1982), and Sherrill and Shimp (1982). A selected review of some of the empirical studies which define involvement as a process oriented construct will be discussed. The first paper to be reviewed is Zinkhan and Mudderisoglu (1982) (herein, 2&M) study on how the constructs of involvement, familiarity and cognitive differentiation affect advertising recall. They define the three constructs so as to avoid the semantic confusion that has resulted in the literature. They point out that although all three constructs will have similar effects on recall, they are not identical. The purpose of this study is then twofold. First, Z&M attempt to clarify the 3_2 measurement confusion that exists between the constructs. Secondly, they investigate the relationship of these three measures to advertising recall. The empirical application of the paper examines the validity of the measurement process while testing the strength of the causal relations. The results of the proposed model, though, show a weak relationship between the three predictors and advertising recall. The strength of this study lies in its progress in isolating measures of these three constructs which can pass stringent tests of validity. In other words, this is an attempt to show involvement, as well as familiarity and cognitive differentiation, as the result of information processing and decision making employed by an individual. The next paper is a study by Greenwald and Leavitt (1984) (herein, G&L) who define involvement at four levels of audience attention. This is an attempt to conceptualize involvement within a cognitive framework as a result of information processing on the part of the subject. G&L use Krugman’s theory of involvement as the foundation to develop a model to differentiate across various levels of involvement. The authors state that by combining increases in involvement with qualitatively distinct levels of cognitive activity which require greater amounts of attentional capacity, the result should be increasingly durable effects on longterm memory. They present four distinct levels of involvement which 33 are associated with these qualitatively different levels of cognitive activity. The difference or distinction between the four levels is the capacity required to affect involvement. The four levels are presented in the diagram in Figure 2. FIGURE 2 Four Levels of Cognitive Activity Pre- Focal Compre- Elabora- Attention_-P Attention—9 hension -_V tion The authors differentiate between two types of effects, immediate and enduring, based on the level of involvement when the subject was exposed to the advertising message. As far as advertisers are concerned, enduring effects produced at different levels of involvement are of critical importance. Involvement at the preattention level has resulted in a controversial body of research on the effects of subliminal advertising as well as in psychology with Cherry’s cocktail party findings (Howard, 1983, pp. 63-5). Involvement at higher levels are of special interest since the individual goes through varying levels of 34 information processing and decision making depending on the amount of attentional capacity. At the focal attention level, familiar stimuli (i.e. advertising messages) are perceived and categorized into memory traces. At the comprehension level, the stimuli result in a memory trace at the propositional level of representation. At the elaboration level, memory and attitudes are formed from specific details presented in the message. As the level of involvement increases so does the durability of the cognitive and attitudinal effects._ In their study,’ Sherrell and Shimp (1982) look at the task of measuring involvement in a laboratory setting. Their objective is to examine cognitive differences in experimental subjects who have been manipulated in to various levels of involvement. To measure the difference in cognitive activity between high and low involvement subjects, they developed three distinct measures. These measures include the personalization of the decision task, a self report of how much thought one puts into a task of how meaningful that task was; a self insight accuracy measure, a self report of how much insight one could claim for her/his cognitive operations; and the amount of time that subjects require to complete a decision task. Their results found that the two verbal methods were unable to show significant differences between the groups (group involvement manipulated using a personalization technique). Along with the poor results, were some 35 conflicting findings using the self insight accuracy measures. The elapsed decision time method, however, did detect significant difference between high and low involvement groups. They found that high involvement subjects took significantly more time which suggests that the personalizing manipulation engaged greater cognitive activity. This points to the potential usefulness of decision time measures as a means of determining level of involvement. Antil (1983) attempts to clarify the issues surrounding involvement by reviewing the literature and integrating divergent views so as to achieve some general agreement. First, he presents involvement as a directional flow between three stimuli; product, situation and communication and the individual. This means that the personal interpretation or relevance that the individual attaches to the characteristics of the product will result in involvement. The same applies to the other two stimuli, situation and communication. This is critical in that it shows the level of involvement as a result of the level of cognitive processing and not a cause of it. Second, Antil posits that there is an interrelationship between the three stimuli which combine to influence the level of involvement. For example, an experiment consisting of an advertising message (communication), brand name (product) and a laboratory setting (situation) would interact to affect the subject’s 36 interpretation hence their involvement level. The third point he makes is that involvement requires some effort on the part of the consumers. These consumers must allocate effort based on the possible benefits and costs. This, then is the foundation for deciding what level of involvement a consumer will use in a decision making situation. The final point that the author makes is that consumers have reasons for all their decisions and selections. These reasons may not be elaborate but consumers do think about their purchases. A product such as soap may be a low involvement purchase only because of the high confidence the consumer has in expected benefits from the brand purchased, not because the s/he is unconcerned about product performance. A State Orientation The second approach taken by a group of researchers views involvement as an individual level, internal state variable. This variable has motivational as well as directional properties which are evoked by a certain stimulus or stimuli. This state oriented conceptualization is founded in the social psychology literature (Sherif and Cantril 1947) and has been further elaborated by Mitchell (1978, 1980), Deighton (1982), Park and Young (1982), and Bloch and Richins (1983). These studies will be reviewed in an attempt to clarify involvement as a state oriented 3] construct. Mitchell (1978) discusses the conceptualization of involvement as a process oriented construct versus a state oriented variable. He argues against the process oriented version for several reasons. Most importantly, defining involvement as a process can be misleading in many situations because it assumes involvement is the only determinant of these processes. For example, extracting information from an advertisement, the modality, the content and the structure of the advertisement may also affect these processes. As a state variable, involvement has its roots in social psychology and has been conceptualized in terms of the relationship between an individual’s values and an issue or object. The more an issue or object becomes integrated in a person’s values, the higher the level of involvement. This is also termed ego involvement or commitment, depending on the situational context, (Lastovicka and Gardner 1978; Mitchell and Olson 1975; Newman and Dolich 1975). In his paper, Mitchell (1980) develops a conceptual model to determine the dimensions of advertising involvement. This conceptualization of involvement views it as an internal state of an individual with both intensity and directional properties. He presents 'two critical stages in the information acquisition process: attention and processing. The first stage, attention, is 38 defined as having limited capacity which causes individuals to focus their cognitive resources on a limited set of stimuli. They must decide which stimuli to attend to and how much attention to devote to each stimulus. The second stage is processing where individuals interpret information, make inferences and evaluate the information with the aid of semantic processing, elaboration schemes, scripts, just to name a few. The factors which affect attention and processing include the stimulus and the objectives of the individual while exposed to the stimulus. An application of the model follows. A group of subjects are exposed to an advertisement for a new high speed sportscar. Different individuals with varying needs will react differently to the stimulus. The content of the advertisement and the goals of the individuals will determine the direction and the amount of involvement during exposure to the stimulus. Mitchell states that the intensity of the involvement will determine how much attention is paid to the advertisement. The direction of the involvement will determine which schema is activated to aid in processing the advertisement. Finally, Mitchell suggests the existence of three levels of involvement which may result in three different types of information acquisition. One information acquisition process is caused by a high involvement condition and two are caused by low involvement conditions. 39 He postulates that the first low involvement condition is a cause for information acquisition much like the Petty and Cacioppo (1983) issue and product involvement manipulation. The "lowest" level of involvement is similar to Krugman’s (1965) definition of information acquisition under low involvement conditions. Two interesting findings in his paper include the following. First, Mitchell’s work follows the findings of other studies (Petty and Cacioppo 1979,1983; Mitchell, Russo, and Gardner 1980) of the effect of involvement on information processing where a reduced relationship is found between semantic information and attitudes. Second, he stresses the point that an examination of factors affecting the information acquisition process can occur at different levels. Based on his empirical work on the effects of involvement on the communication processes Mitchell (1982,1986) and Mitchell and Olson (1977,1981) make a strong case for the conceptualization of involvement as a state oriented construct. In their paper on the different types and levels of involvement, Park and Young (1982) (herein, P&Y) examine involvement’s varying effects in an effort to determine brand attitude formation. They develop guidelines to use in research when trying to identify the causes of involvement in order to understand the multiple effects of the construct on information processing. The personal interpretations or motives that an 40 individual attaches to the characteristics of a product, message or situation will result in a different level of involvement. These motives or interpretations vary in intensity and direction. Therefore, P&Y investigate involvement at the cognitive and affective level of a learning hierarchy. Their results include an emphatic suggestion that the traditional view of attitude formation needs to be reconsidered as a theoretical framework for the involvement construct. This is based on the fact that P&Y are able to disprove the traditional hierarchy of communication effects as the theoretical foundation from which to measure advertising effectiveness when using involvement as the mediating variable. The next paper to be reviewed is Deighton’s (1982) study where he identifies a set of heuristics which consumers are hypothesized to use when deriving inferences from advertising messages. These heuristics are presented in the form of a model for the low involvement consumer. The author presents the model as passing through three phases. First, the individual recognizes the problem, then a solution is proposed by the message, and finally, the solution or proposition is tested for validity. The underlying assumption in Deighton’s paper is a state oriented conceptualization of involvement. He hypothesizes that a set of heuristics will guide the uninvolved or "low" involved consumer’s information 41 processing under the influence of advertising. In other words, since the consumer is uninvolved, the depth of her/his information processing is at a minimal level. In a reevaluation of the product involvement construct, Bloch and Richins (1983) (herein, B&R) develop a model which they use to differentiate product importance and product involvement. They define product importance as importance pggggiygg by consumers rather than some objective level of importance strictly inherent within the product itself. This is in contrast to product involvement which is defined by B&R as the general level of interest in the object or the centrality of the object to the person’s ego structure. The two concepts differ in that product importance is a cognitive state of awareness whereas product involvement is a motivational, internal state variable determined partly by product importance. The authors’ conceptualization of involvement is in terms of an internal state variable with two dimensions of direction and intensity. These studies which conceptualize involvement as an internal, motivational state variable position the level of involvement as preceding the depth and direction of information processing. This is in contrast to the process oriented conceptualization which positions the level of involvement as a result of the information processing and decision making employed by the individual. The next section explores various typologies of involvement to 42 determine the framework within which advertising effectiveness can be evaluated. Central vs. Peripheral Routes of Persuasion This argument is based on the Elaboration Likelihood Model (ELM) which assumes that different methods utilized to evoke attitude change may work best depending on whether the probability of issue or product relevant thoughts occurring is high or low. It predicts that when elaboration likelihood is high, the central route to persuasion is most effective. This is in contrast to the occurrence when elaboration likelihood is low, resulting in the peripheral route as the optimal course. The important implication underlying. the ELM for advertising messages involves the effective use of different types of appeals (i.e.: serious, humorous, sexy) for different audiences. Under this model, involvement is dichotomized into two types: issue involvement (also known as ego involvement, product involvement, personal involvement and commitment) and response (or task) involvement. The first type of involvement concerns the extent to which the attitude issue or the product itself has some direct personal relevance or consequence, and people are concerned with forming a reasoned opinion (Petty and Cacioppo 1979). The second type of involvement finds that the attitude response is important and people are more concerned with expressing an attitude that will result in immediate rewards rather than 43 forming a reasoned opinion (Zimbardo 1960). In this study, issue (or product) involvement is the variable of interest since response involvement is defined as a situational state variable. High issue involvement is predicted as the result of information processing via the central route whereas low issue involvement is predicted via the peripheral route. Previous research in psychology and consumer behavior emphasizes the central route over the peripheral route in decision making. Fishbein and Ajzen (1975), among others, state that persuasiveness of a message can be increased if careful attention is paid to the message content rather than the manipulation of credibility, attractiveness, expertise of the source. The results of Petty and Cacioppo (1979,1983) show that under certain conditions (low involvement) such nonmessage cues as expertise or attractiveness of a source can have maximal impact on persuasiveness. This is contingent on the fact that extensive information processing of message content is not present. Under this situation, persuasion is determined more by a well-practiced script such as ”Experts are to be believed”. This is in contrast to the central route where persuasion occurs as a result of a subject’s attention and processing of message content. Overall, Petty and Cacioppo (1979,1983) found that low involvement resulted in the celebrity status of the endorsers being the critical determinant of attitude toward 44 the product. Under high involvement conditions, the product endorsers had no effect on attitudes but the potency and relevancy of the information about the product contained in the advertisement became a powerful determinant of product evaluations. This tells us the importance of nonmessage cues under conditions of low product and issue relevant involvement. The level of personal relevance contained in the product or issue then becomes the critical factor of the manner in which persuasion will occur - the central route or the peripheral route. The critical feature of the central route to persuasion is that attitude change occurs as a result of indepth thought regarding the information that a person feels is central to the true merits of a certain attitudinal position or product. In contrast, the peripheral route views attitude change as a result of nonmessage cues or due to simple heuristics which obviate the need to process issue or product relevant information. The results of the Petty and Cacioppo (1979) study are more in line with the ELM than with the Social Judgment Theory (Sherif, Sherif, & Nebergall 1965) or Krugman’s Sequence model of involvement (1965,1967). When involvement is defined as a state variable, the Social Judgment Model states that it is more difficult to change attitudes under high involvement conditions. This is not seen in Petty and Cacioppo (1979). In this study, they _45 found a lack of interaction effects between the endorser and the quality of the argument. In other words, Social Judgment Theory is unable to explain the lack of a relationship between nonmessage cues and the strength of the arguments made in the advertisements under high and low involvement conditions. In Krugman’s sequence formulation, he states that under high involvement conditions attitude change precedes behavioral change but that under low involvement the opposite occurs. In Petty and Cacioppo (1979) this held for high involvement conditions but not for low involvement conditions. Under low involvement, effects were found on the attitude measures but not on the measures of behavioral intentions. There should have been, theoretically, stronger behavioral measures under low involvement but this did not occur. Instead, the ELM seems to be more consistent with the findings of Petty and Cacioppo (1979, 1983). This model predicts a greater interaction between attitude and behavior under high involvement (central route) than under low involvement (peripheral route). It doesn’t state a directional or sequential relationship as the other two models attempt to do. The underlying reason for the success of the ELM to explain this relationship rests with the concept of the level of personal relevance evoked by a message. The next section explores the various typologies of involvement to determine the framework within which advertising effectiveness can be evaluated. 46 Involvement Typologies There are basically five concepts which have all been studied under the heading of involvement in the marketing literature: ego involvement, commitment, purchase or product importance, communication involvement, and issue involvement (Muncy and Hunt, 1983). Once these concepts have been identified, the type of involvement which is relevant to determining the effectiveness of script and nonscript print advertising messages will be explored by reviewing several empirical findings. The first involvement concept to be explored is the ego involvement construct. According to Ostrom and Brock (1968) ego involvement can be defined as the degree to which an object or idea is centrally related to an individual’s value system. An example of how a product can become ego involving is seen in an individual who purchases only European cars because it is associated with the yuppie subculture and value system. The concept of ego involvement has its roots in the social psychology literature. Sherif and Sherif (1967) define ego involvement as the operationalization of an individual’s latitude of acceptance, latitude of rejection and latitude of noncommitment to stands on issues. They operationalize the construct as the range of the latitude of rejection where low involvement is characterized as a narrow latitude of rejection whereas a wide latitude of rejection is characteristic of a highly involved consumer. 47 Sherif shows that his model holds for both issues and objects (Houston and Rothschild 1977), by measuring each salient attribute, since most choice behavior occurs for objects possessing multiple attributes as choice criteria. This model predicts that individuals highly involved with the object or issue of an advertising message will be more resistant to attitude change than individuals who are not as involved. The second involvement concept found in the literature is commitment. .The distinction between commitment and ego involvement is.a gray area both in the social psychology as well as the consumer behavior literature. This is evidenced by the fact that these terms are used interchangeably. Lastovicka and Gardner (1979) used two underlying dimensions to measure ego involvement: commitment and importance. Freedman (1964) defined ego involvement as commitment to a specific issue or position. Although, ego involvement and commitment are related, ego involvement can exist without commitment and vice versa. For example, a person may become committed to a European car as a result of purchasing a BMW. On the other hand, a person’s commitment to European cars may be related to that person’s 'self—worth. In this situation, ego involvement will precede commitment. Commitment has had numerous implications to both consumer behavior thought and marketing practice and continues to be plagued with conceptual and methodological 48 issues. The third concept to be discussed is communication involvement which was first presented by Krugman (1965). There are two important characteristics which distinguish communication involvement from ego involvement. First, communication involvement occurs at a specific time (i.e. the advertising message) making it situationally specific and transitory. This is in contrast to ego involvement which involves an object or issue which is enduring. The second important characteristic of communication involvement is that the connections are made with any aspect of the person’s life, not just those which are related to the person’s central value system. In other words, a person can be involved in a message simply because of its utilitarian value. This type of involvement is particularly relevant in research on consumer information processing and it supports the idea that the effectiveness of advertising campaigns on consumer behavior can be understood in the communication involvement framework. The fourth type of involvement is purchase or product importance which was developed by Howard and Sheth (1969). It has been defined as product class specific not distinguishing between brands and variously labeled as task importance, product involvement, and seriousness of consequences. Purchase importance has also been confused with ego involvement. It must be remembered that purchase importance is a result of an individual’s ego involvement 49 with a purchase. Based on the above, a strong need emerges to isolate the exact effect of this construct on a buyer’s behavior as well as\to distinguish it from other forms of involvement. The fifth way of viewing involvement is issue involvement introduced by Petty and Cacioppo (1979,1983). They define the construct as the intrinsic importance or personal relevance a person attaches to an issue or a product. They predict that increases in involvement will enhance the importance of the message contents in producing attitude change. The underlying assumption suggests that low involvement persuasion situations may be controlled through an automatic processing mechanism whereas high involvement situations are controlled by cognitive processing on the part of the individual. This view is supported by Chaiken (1978) who found that subjects were more affected by arguments used in a message under high involvement conditions than under low involvement. Correspondingly, he also found that low involvement conditions result in subjects being more highly affected by nonmessage cues such as source attractiveness or expertise. In this study, it is issue involvement which is used as one of the independent measures of interest. This is due primarily to the fact that a subject, when exposed to an advertising message, is hypothesized to be concerned with forming a veridical opinion of the issue or product. 50 The issue of concern here is cigarette smoking and the product of concern is cigarettes. An important point to be made is that the key determinant of whether involvement will facilitate or hinder attitude change is the extent to which the information provided in the advertisement appears to contradict the subject’s initial position. In other words, does the scripted version of the advertisement portray a message of social desirability with the issue of smoking that will elicit counterarguments from subjects. In to relations to this issue, Petty and Cacioppo (1979,1981,1983) predict that if an advertisement elicits primarily counterarguments, then increased involvement should result in decreased agreement and more negative attitudes towards the message. In contrast, if the message content elicits primarily favorable thoughts then involvement will tend to enhance these positive thoughts and result in increased agreement. The hypothesized relationship presented in the next section were derived previously from the theoretical arguments advanced in this chapter. CHAPTER THREE METHODOLOGY In this chapter, the research hypotheses are presented, followed by a discussion of the product, medium and sample selected for the experiment along with the results of the pretest. Then, the research methodology is discussed including the procedures used to measure the dependent and independent variables. Research Definition and Hypotheses The research problem involves the investigation of the effectiveness of script versus nonscript print advertisements and high versus low product involvement. This is a 2 x 2 factorial design which attempts to determine the effectiveness of a message based on its presentation format and the general level of involvement the subject holds with the product. The objective is to determine the most influential format to employ in print advertisements to better attract the attention and increase memorability. This is hypothesized to depend on the degree of involvement that each subject may have with the product/brand advertised. Based on the literature reviewed previously and the research problem stated above, the following research hypotheses are proposed to: 5,1 52 H1: A scripted advertisement will result in higher product class recall for both low and high involvement subjects than a nonscripted advertisement. H2: A scripted advertisement will result in higher aided brand recall for both low and high involvement subjects than a nonscripted advertisement. H3: A scripted advertisement will result in higher unaided brand recall for both low and high involvement subjects than a nonscripted advertisement. H3A: High involvement subjects will have higher unaided brand recall for both scripted and nonscripted advertisements than low involvement subjects. H4: A scripted advertisement will be rated more favorably by both high and low involvement subjects than a nonscripted advertisement. H4A: A high involvement subject will have more favorable attitudes for both scripted and nonscripted advertisements than a low involvement subject. H5: A brand portrayed by a scripted advertisement will be rated more favorably by both low and high involvement subjects than a brand portrayed by a nonscripted advertisement. H5A: A brand portrayed by either a scripted or nonscripted advertisement will be more favorably rated by a high involvement subject than a low involvement subject. The relevant variables to be measured are defined and operationalized next. This explanation is critical to understanding the statistical analysis in the final section as well as to ensure clarity and to avoid problems with interpretations of the findings. Dependent Measures To determine the effectiveness of advertisements two dependent variables are used: 1) recall and 2) attitudes These two variables are defined as follows. 5.3 (1) Recall This variable measures the level of memorability that an advertisement has when a consumer recalls the product, brand and message. There are two types of recall measures used to evaluate the effectiveness of an advertisement. The first is unaided recall and the second is aided recall with unaided being a much stronger measure of effectiveness. The assumption about the strength of unaided recall is based on the belief that a longer lasting impression has been made when a subject freely recalls the advertising. To measure the level of recall, the amount of knowledge among consumers that can be directly related to the message must be determined. Recall is used to determine the extent to.which advertising messages have been retained by consumers. Both aided and unaided recall measures were taken in this study. First, subjects were asked unaided recall questions concerning the product categories, brands, other advertisements, and topics of the articles in the stimulus booklet. Then, subjects were asked to identify the brand of liquor, the brand of cigarettes and the branch of the military service advertised in the booklet (aided recall). For the purposes of this experiment, both recall and recognition measures were needed to validate the effect of the advertisement format in terms of product and brand memorability. (2) Attitude This variable is a measure of the liking that a 54 subject has toward a product and a brand. This variable assumes that the consumer is aware and has knowledge of the product either due to the message or actual experience with the product. Attitude tests are used to measure consumers’ dispositions toward a brand/product or the degree of acceptance of various claims made in the advertisement when used over time. Seven point semantic differential scales, previously developed and validated, were used to determine the subject’s attitude toward the product/brand. Some examples include "Please rate the advertisement on the following attributes:" and "Please rate the brand on the following attributes:". The relevant attitudinal questions were combined into a composite measure of attitude toward the advertisement, the product, and the brand. Independent Measures. The two independent variables of the study were advertisement format and level of product involvement. These two variables were defined and operationalized as follows. (1) Format of Advertisement The format of the advertisement was manipulated between a script version and a nonscript version. The script version is comprised of an advertisement with two or more visual scenes involving a sequence of events, behaviors or contexts which are familiar and frequently encountered by s_s the target audience. Since individuals are hypothesized to possess scripts for routine behaviors and situations, a restaurant script was used in the scripted version. Furthermore, restaurant scripts are found to be one of the most commonly held scripts by individuals (Abelson 1977). The nonscripted version was an advertisement with one or more unrelated scenes, involving some type of brand comparison or a visual description of the product’s attributes. In the nonscript version of the advertisement, the message was the same, as well as the copy with the only manipulation being the lack of a meaningful script. This was accomplished by removing one frame from the advertisement. These two versions were designed to determine if the script format is truly more effective in affecting attitudes and memorability of the message. These variables were measured in terms of the product, the brand'and the advertising message. The script format was evaluated to identify the relationship between memory for key points in the advertisement and the corresponding effect the advertisement may have on attitudes toward the product, the brand and the message. (2) Issue Involvement As explained in the previous section, involvement has been notoriously overdefined by some, underdefined by others and misdefined by yet others. The definition used here is the one presented by Petty and Cacioppo $6 (1979,1981,1983) due to the strong empirical justification as well as its relevance to the problem on hand. The conceptualization of involvement is based on the belief that a subject, when exposed to an advertising message, is hypothesized to be concerned with forming a veridical opinion of the issue or product. The issue of concern here is cigarette smoking and the product of concern is cigarettes. The level of involvement will determine if attitude development and cognitive responses occur due to exposure to a specific message (i.e. an advertisement) within a specific situation (i.e. flipping through a magazine). Since involvement must be conceptualized and operationalized as a continuous variable, scales were used to measure the respondents’ level of involvement due to the product. Zaichkowsky (1985) scale of involvement, which is proven to produce high reliability levels is used as a measure of the involvement construct. In the following subsection, other elements of the experimental design are explained and justified. The Medium Some relevant characteristics concerning the print medium are now discussed in order to better understand its role in this experiment. The print medium used here was the magazine. Probably the most important characteristic of print is 57 the selectivity in the choice of content that it allows readers to exercise. With magazines, the reader can read it where and when s/he wishes, take up where s/he left off and spend as much or little time as s/he wishes. This results in the reader being able to discriminate among the items in which she is interested in more readily than with the broadcast media. Another important characteristic includes the role of this medium in information processing by consumers. Research on comprehension indicates that print provides messages that are ” more easily learned and remembered ...” (Tan, 1985, p.170). The general conclusion here is that we learn more by reading than by hearing or by hearing and seeing. (Tan, 1985) Print can also be a more effective medium for more complicated messages. This is especially true when the purpose of the messages is to affect the consumer’s attitude toward the product. This is evidenced in Chaiken and Eagley’s (1976) study where subjects were also found to have a higher comprehension level for more complicated messages. This is a significant point since scripted advertisements are a more complicated format in which to present a message. This could mean that this format for messages has the potential to be more effective in print. This is further evidenced in Chaiken and Eagley’s findings that subjects exposed to easy messages on videotape 58 perceived the message as more persuasive hence, more effective (Tan, 1985, 169-77). This tells us that communication modality interacts with message complexity in determining the extent of attitude change. Since advertisements which use scripts tend towards a more difficult message then print, it may be posited as the ideal medium to use a script format. Some other facts to consider about this medium include the following. First, print has a longer life than other media. The reader has the time and opportunity to discriminate among messages. Secondly, print is used as a source of information of a serious nature rather than purely for entertainment. This implies that scripted versions have the opportunity to make consumers aware as well as increase their knowledge in regards to the advertised product/issue. Finally, magazines, which are designed to appeal to particular interest groups (i.e. housewives, tennis afficionados, etc.), are able to attract readers to their advertising because the messages are generally related to the focus of the magazine. The use of excellent photography, copy and layout and quality reproduction enable the reader to do one of two things. The reader glances at the message and disregards it if s/he is not interested or reads it thoroughly if it looks interesting. Since the reader has the ability to process the information in the message at her/his leisure, the script format can 59 provide her/him with informative as well as persuasive information about the product. The Product The product that was selected for this experiment had to have certain characteristics. First, it had to be a product that could be advertised using a script as well as a nonscript format in print. The only example found thus far in print is the new Benson and Hedges cigarette advertisement which was substantially altered to avoid any complications of brand and advertising familiarity. Secondly, it was important that the product could invoke low involvement as well as high involvement among the subjects in relations to the product, (cigarettes) and the issue, (cigarette smoking). This would allow the scripts presented in the message to activate processing of the information whether the individual consciously or unconsciously deemed it necessary. Since activation of a script is believed to occur automatically through the situational context, the script version performs the role of activation. An example is the script version of the Benson and Hedges cigarettes which presents two scenes for the RESTAURANT script. The two scenes in the script version are designed to evoke a familiar experience, whether direct or vicarious, which provides information on how to select a behavior (which cigarettes to buy/whether to smoke or not to smoke). In 60 other words, it can be posited that once a subject views the script version, s/he associates smoking cigarettes with a fun, social encounter within a socially desirable context (meeting a good looking person). The two scenes identify the RESTAURANT script in the context of creating a feeling of social acceptance for smoking. The experiences evoked by this script are assumed to be highly social as well as familiar to the subjects. In the nonscript version, the advertisement has the modest objective of attempting to remind those users of the product and this specific brand. The Sample The sample selected for this experiment included undergraduate students at two midwestern universities. These students were chosen because they are believed to be the primary target market for the advertisements used in the experiment. The student were recruited from Marketing, Advertising and Economics classes. Participation in the experiment was completely voluntary and no compensation was provided. A total of 76 male and female undergraduate students participated in the experiment. Half of the subjects (38) received the booklet with the scripted advertisement and the other half received the nonscript version. The booklets were randomly distributed in each class to ensure random assignment to each format cell (script and nonscript). 61 The Experimental Materials Two booklets were prepared for the experiment. The first booklet contained the advertising stimulus (scripted or nonscripted advertisements) embedded in a set of articles and advertisements. The second booklet contained the questionnaire with the dependent measures and the independent variables. The first booklet was presented as a newspaper insert entitled CAMPUS NEWS. It was prepared as a booklet to be periodically inserted in the school newspaper. Each issue was proposed to provide pertinent information on a specific topic of interest to the students. The topic of this booklet was job hunting for students preparing to look for either career or summer employment. Each booklet contained either a scripted or a nonscripted stimulus advertisement with everything else being the same. (See Appendices B and C for copies of the booklets.) The advertisements in the booklet were all directed to the target market, students. The first advertisement was a liquor advertisement, the second was the advertising stimulus, High>Life, and the third was a Coast Guard enlistment advertisement. The booklet was designed to appeal to the subjects as well as to be cohesive in its presentation of information. Given the importance of the format (script versus nonscript), manipulation in this study, a pretest was conducted to assess the degree to which these 62' advertisements were able to provide this manipulation. During the pretest, the subjects were presented with one of two versions of the High>Life cigarette advertisement. Version 1 contained the scripted advertisement while the second version contained the nonscripted advertisement. Using the existence of script versus nonscript as the dependent variable, a sample of 26 students were used to complete the pretest questionnaire. The results indicated a successful manipulation of the format (script/nonscript) variable. The basic assumptions behind this pretest are twofold. First, it is presumed that people rely heavily on scripts that are well-established in memory. Second, these script- like structures in memory tend to be fairly general and some variation exists in the way in which specific information is interpreted and used in decisionmaking. This pilot study resulted in a more detailed description by the subjects who were given the nonscript version. The variation that arose with the nonscript version included such statements as "all are attractive and nicely dressed” and "the second woman has her back to us". Among the statements made, there were several which were common to both versions. These statements were also most often provided as a description for the advertisements. This includes statements such as "one girl is smoking and one girl isn’t", "two women go to lunch", and ”both women notice the good looking man at the next 63 table." In order to determine the existence of a script, it was necessary to analyze subjects’ responses to the pretest questionnaire. An independent judge was used to determine the number of irrelevant actions/events in both the nonscript and script advertisements. Once the number was determined for both, a t-test of difference of means was performed. It was found that there was a significant difference (t = 2.34, alpha = .05) between the number of irrelevant actions in the script and the nonscript advertisements. In the final analysis of both versions, it was found that a story line or script existed for Version 1 but not for Version 2. The consistent pattern of steps can be seen in Table l where the corresponding number of times each step was mentioned is also shown. This consistent script did not exist for Version 2. Instead the subjects provided much detail about the one frame with more irrelevant actions and variation as seen in Table 2. 64 TABLE 1 Script Format (1) A restaurant scene with two 9 women having lunch together and a man sitting at the adjoining table. (2) One woman is smoking while the other 12 is not. The man at the other table smokes. (3) The two women are discussing 10 the man at the other table. (4) The brunette and the man end up talking 10 together and getting to know each other. (5) The second woman is not smoking 6 and is left out of the conversation. TABLE 2 Nonscript Format (1) Three people in a restaurant 11 with two smokers and one nonsmoker. (2) Brunette smoker is checking out 8 the man. (3) Both women are looking at the man. 11 (4) All are attractive and nicely dressed. l (5) Man is at table alone and reading. 3 (6) Man is oblivious to women. 4 (7) Women go to lunch at a nice restaurant 11 (8) Second woman has her back to us. 2 65 This pretest was performed to verify statistically the existence of a script in the High>Life cigarette advertisement. It was imperative that the script version of the advertisement be perceived by the subjects as a script therefore this pretest was performed solely for that purpose. The Procedures The subjects were in a classroom environment and were given the first booklet. They were told that the Advertising Department was interested in their opinions of the booklet as a whole as well as on the topics of the articles, the- advertisements, the overall appearance and other pertinent information. Hence, they were required to familiarize themselves with the booklet first to be able to express their views and opinions about it. Since the subjects paced themselves through the booklet, they could spend as much time as they wished evaluating each article and each advertisement. This avoided the time pressure created by external pacing and allowed the subjects to use their own personal motivations to guide their information processing behavior. Once a subject had finished examining a booklet, s/he closed it, raised her/his hand and was provided with the second booklet which contained the questionnaire. The first page of the questionnaire explained the purpose of the study again and provided instructions on how to proceed 66 through the booklet. The booklet contained a series of questions intended to measure the dependent variables of the study as well as the manipulation checks and other relevant scales. The subjects were told to go through the questionnaire at their own pace and to raise their hand when they were finished. They were then debriefed and thanked for their participation in the study. The questionnaires were then marked to identify those respondents who evaluated the script format from those who evaluated the nonscript format. This was used later as one of the checks to determine if the format had been successfully manipulated. This will be further discussed in Chapter Four. Independent Variables (1) Advertisement Format The advertisement format could be either scripted or nonscripted. This variable was manipulated by using a Benson and Hedges advertisement found in the February through April issues of various magazines such as Life, Vogue, Glamour, People, Time, and Penthouse. For the purpose of avoiding familiarity with the advertisement and manipulating the advertisement format variable, the advertisement, the name and the appearance of the advertised product as well as the copy was altered to develop a new advertisement for a new hypothetical brand of cigarettes, High>Life. 62 The script version resulted in a package redesign along with changes in the copy wording and placement (see Appendix B). The nonscript version had the second frame removed from the advertisement and a larger version of the redesigned cigarette package was inserted in its place (Appendix C). The copy placement and size was also changed to create a new advertisement for a new brand of cigarettes, High>Life. The manipulation of subjects into script and nonscript cells was accomplished through randomly passing out the booklets. However, the questionnaire contained an item which asked subjects to describe the High>Life advertisement as if trying to explain it to someone who could not see it. This question was used as a manipulation check, the results of which will be discussed later. (2) Involvement In this experiment, involvement was not manipulated, but only measured. While the manipulation of this construct results in high internal validity, its major drawback is the lack of external validity associated with the resulting measure. (See for example Petty and Cacioppo 1983). Another method, less commonly used, is to allow subjects to assign themselves to either high or low involvements groups. This method tends to be high in external validity. but lacking in internal validity. It does not control for such variables as amount of prior 68 information, extremes in attitudes, and so on. In the present study a compromise is proposed in order to measure involvement without confounding its external or internal validity. A twenty item scale called the Personal Involvement Inventory (PII) developed by Zaichkowsky (1985) is used to measure involvement. An example of the P11 is found on page six of the questionnaire in Appendix A. This semantic differential scale was used to capture the concept of involvement with the product of interest, cigarettes. The scale’s reliability over time has been checked over two subject populations for four different products resulting in Cronbach’s alphas ranging from 0.95 to 0.97. The reliability of the P11 was checked in this experiment and was found to have a Cronbach’s alpha of .9889. Dependent Measures (1) Recall Both aided and unaided recall measures were obtained. For unaided recall, the subjects were asked to list all the product categories for which they saw advertisements and then to list all the brand names of the products advertised. Last, they were asked to list all the topics of the articles in the booklets. The final unaided recall measure, level of recall, is comprised of the first two questions and is explained next. The subjects were asked to freely recall the product category and name of the advertised brand. The variable, 69 level of recall, was a linear transformation of these two variables used to measure free recall. If the subjects were unable to recall both the brand and the product category, the variable was coded 0; if s/he recalled either one of these correctly, the variable was coded l; and if s/he answered both correctly, the variable was coded 2. This resulted in a ratio level measure of unaided recall. The second measure of recall was presented on the third page where respondents were asked to name the brand of liquor and cigarettes advertised and the branch of the military service presented in the insert. (2) Attitude Attitude toward _the advertisement and brand were measured by a set of three scales. Two different scales were used to measure advertisement attitude in an attempt to increase reliability. The scale used to measure brand attitude was adapted to fit the product and the advertising stimulus. All three attitude scales displayed high reliability estimates with Cronbach’s alphas ranging from .8744 to .9289. Next, some advertisement familiarity measures were taken to determine if there existed any confounding influence due to confusion and/or prior exposure to the Benson and Hedges advertisement. There was a surprisingly high (32.9%) number of subjects who claimed to have seen the advertisement previously but only 2 out of 76 (2.6%) had seen the brand prior to exposure to the experimental 7o_ stimulus. Another measure taken was an attempt to determine what the subjects may have learned due to their exposure to the script or nonscript advertisements. The results of these measures will be discussed in the next chapter. Coding Procedures A procedure developed by Holsti (1969) was used to compute interjudge agreement and reliability. Three naive and independent judges were used to analyze subjects’ response to the existence of the script/nonscript format. A respondent may have been exposed to a script format but if s/he didn’t perceive the advertisement to tell a story then the script characteristic failed to elicit a difference in cognitive processing. Each judge was trained in the definition and characteristics of a script and a nonscript format. They were then instructed to read each description given by the respondents and to determine how they perceived the advertising format. After evaluating each response, the judges marked it as either containing a story, not containing a story, or that they were unsure if it contained a story or not. The results will be discussed under the section on Manipulation Checks in Chapter Four. CHAPTER FOUR EMPIRICAL RESULTS Manipulation Checks A manipulation check was needed for the script versus nonscript format since subjects were randomly assigned to either one of these two categories. To assess the effectiveness of this manipulation, two measures were used. First, the three judges’ evaluation scores were computed to assess the success of the manipulation. Holsti (1969) developed a procedure to determine interjudge agreement and reliability when performing content analysis. Using this method, the average interjudge agreement was found to be 0.86 and reliability was 0.95, for this study. These results indicate that the coding procedure used by the three independent judges provided a reliable basis for additional analyses. The second check used was a comparison between the above results and the method used to identify those respondents who evaluated the scripted advertisement and those who evaluated the nonscripted advertisement. Only those respondents whose responses were found to be unanimously agreed upon as either a script or a nonscript versions were then checked by examining the notation made by the researcher on their respective questionnaires. This ensured that the format of the advertisement was accurately 21 72 manipulated. The empirical results will now be presented in relations to each hypothesis. Hypothesis One. The first hypothesis states that a scripted advertisement will result in higher product class recall for both low and high involvement subjects than a nonscripted advertisement. All 38 of the subjects exposed to the scripted format reported cigarettes and 35 of the 38 (92%) subjects exposed to the nonscripted format reported cigarettes as the correct product class. A difference of proportions test was conducted to determine if the format had a significant impact on product class recall. 1 The results found that (Z=l.78, p<.07) the script format produced significantly higher product class recall among both high and low involvement subjects. Hypothesis Two The second hypothesis proposes that a script advertisement will result in higher aided recall for both low and high involvement subjects than a nonscripted advertisement. They were asked to name the brand of cigarettes advertised in the insert. A difference of proportions test was performed to determine if there was a statistically significant difference in aided recall levels between subjects exposed to the scripted version and those exposed to the nonscripted version. Out of 38 subjects who viewed the scripted version, 34 accurately recalled the 73 brand name High>Life (90%). This is in contrast to 19 out of 38 (50%) subjects who could not freely recall the brand name advertised in the booklet. The calculated Z-value is 3.81 (p<.0002) which results in a statistically significant difference between script and nonscript formats when freely recalling brand names. In the first two hypotheses, tests of difference of population proportions were used due to the fact that both variables developed to measure unaided and aided recall were categorical in nature. This fact precluded the use of a more powerful test such as analysis of variance. Once the variables were converted to proportions, it was possible to conduct a differences of proportions test on each variable. The other hypotheses, though, had interval or ratio level dependent variables, allowing for the use of analysis of variance to interpret the data. Hypothesis Three. In this hypothesis, a script advertisement is postulated to result in higher unaided recall for both low and high involvement subjects than a nonscripted advertisement. The subjects were asked to name the product and the brand name of the product advertised in the booklet. The resulting measure was a ratio level variable whose operationalization is described in the previous chapter. The analysis of variance results found the advertisement format to have a significant main effects 74 (F=34.568, p<.000) to predict unaided recall (see Table 3). This implies that a scripted advertisement may be more effective in unaided recall. This is primarily due to the commonly held belief that free recall is a 'result of a longer lasting impression in memory than aided recall. The involvement variable did not contribute significantly to main effects in terms of explaining recall variation (F=.053, p<.60). The two-way interaction of scripts by involvement was also significant (F=5.ll4, p<.07) indicating that the effects of these two variables are not additive. Examination of this ordinal interaction shows that in the scripted format, recall was lower .for high involvement subjects than for low involvement subjects. The opposite was true for the nonscript case where recall was lower for low involvement subjects than for high involvement subjects. TABLE 3 Analysis of Variance Results for Hypothesis Three ‘ SUN 0’ III! SIG 5002C! OP VIRIATTON 3002338 DP 50013! P O? OIAII EFFECTS 8.947 2 4.474 17.386 0.0 SCRIPTS 8.895 1 8.895 34.568 0.0 NUINV 0.053 1 0.053 0.205 0.6 02-WIY INTERACTIONS 1.316 1 1.316 5.114 0.0 SCRIPTS NUINV 1.316 1 1.316 5.114 0.0 OZXPLAINID 10.263 3 3.421 13.295 0.0 031810033 18.526 72 0.257 OTOTAL 28.789 75 0.384 75 Hypothesis Four. The hypothesized relationship is that a scripted advertisement will be rated more favorable by both high and low involvement subjects than a nonscripted advertisement. To measure favorability, a seven item attitude scale with a Cronbach’s alpha of .9289 was used. The analysis of variance results indicated a significant format main effects in the creation of favorable attitudes toward the advertisement (F=3.277, p<.07). As predicted by the hypothesis, involvement was not a significant variable in explaining the variation in the relationship between attitude and advertisement format (F=2.061, p<.10). The two-way interaction effect was also insignificant (F=2.061, p<.155). See Table 4 for the statistical results. TABLE 4 Analysis of Variance Results for Hypothesis Four - SO! 0? HEAR SIG SOURCS OP VRRIRTTOR 5001333 DP SQURRI P OP 0'11! EPPICTS 9.197 2 4.599 2.669 0.0 SCRIPTS 5.646 1 5.646 3.277 0.0 NUINV 3.551 1 3.551 2.061 0.1 02-Hl! INTIRACTTOHS 3.551 1 3.551 2.061 0.1 SCRIPTS NUINV 3.551 1 3.551 2.061 0.1 OSXPLAINID 12.748 3 4.249 2.467 0.0 ORISIOUAL 124.045 72 1 723 OTOTRL 136.794 75 1 824 7_6 Hypothesis Five. The last hypothesis predicts that a brand portrayed in a script format will be rated more favorably for both low and high involvement respondents than a brand portrayed in a nonscript version. The result of the analysis of variance did not support this hypothesis. There was no significant main effects (F=.8l2, p<.448) with no significant difference in attitude toward the brand between the script and nonscript versions (F=.050, p<.823). An examination of the cell means (Table 5), however, are in the expected direction. They support the results obtained from the previous analysis in that the script format results in a more favorable attitude (3.84 for script/low involvement subjects versus 4.18 for nonscript/low involvement subjects). TABLE 5 Means For Each Experimental Cell On The Attitude Index Low Involvement High Involvement Scripts 3.84 3.84 (n=19) (n=19) Non- 4.18 3.35 Scripts (n=19) (n=19) Note: Attitude scores represent the average rating of the brand High>Life on a six item, seven point semantic differential scale anchored at +1 to +7 (i.e. high quality - low quality, very good - very bad, etc.). 77 In the last chapter, a discussion is presented on the empirical results in terms of managerial and public policy implications, as well as theoretical implications. Lastly, a discussion of the limitations of this experiment and the direction of future research is presented. CHAPTER FIVE DISCUSSION AND CONCLUSIONS Thus far the theoretical foundation for this study has been developed along with the experimental design and procedures used to conduct the experiment. Next, the empirical results were presented in the context of each hypothesis. In this final chapter, the theoretical, managerial and public policy implications will be discussed as well as the limitations of the study and suggestions for future research. Theoretical Implications This study was first presented as an attempt to address the current controversy in the literature concerning advertising effectiveness and consumer response to marketing communications. It presented script theory as a new framework from which to develop a new theoretical base to better understand and improve advertising effectiveness and tested the mediating role of scripts and involvement in this context. Specifically, the hypotheses proposed the script format as an effective framework to initiate and foster consumers’ advertising recall and recognition in an attempt to create changes in consumers’ affective and cognitive structures. 78 79 Scripts have been defined as routine, sequential sets of vignettes that individuals use to guide their behavior and to reduce the amount of cognitive processing needed to perform a task. As presented in Chapter Two, scripts have been found to aid in conveying a message that is readily processed and is more likely to be recalled due, in part, to the scripted format of the advertisement. This was empirically established in this study through the confirmation of Hypotheses l, 2, and 3. The hypothesized mediating effect of the involvement construct, however, was not substantiated (Hypotheses 3A, 4A, and 5A). That is the degree of product involvement a subject had did not affect his/her free recall of the advertisement and his/her brand attitude. This may be due in part to the controversial nature of the product (cigarettes) and the issue (cigarette smoking). In the first hypothesis, “it was found that the script format resulted in significantly higher unaided product class recall than the nonscript format regardless of the involvement level of the subject. The second hypothesis was also supported in its claim that script is a more effective format for aided brand recognition than a nonscript format. The third hypothesis found that the script format was significantly more effective in successfully eliciting free, unaided recall from subjects. Involvement level was not found to have any such effect. This last finding is 80 the most critical since free recall is believed to be the most effective measure of recall. The importance lies in the reinforcement of the belief in cognitive psychology and marketing that unaided recall reflects a much more detailed memory therefore, resulting in a longer lasting and more easily accessible impression (Howard, 1985). Another intrinsic concern with the advertising communications is the effect that scripted messages can have on consumers’ attitude toward the brand, as well as the advertisement. Since scripted messages provide individuals with concrete rather than abstract information, Abelson (1981) states that scripts play an important role in forming attitudes toward an object. In this study, the scripted version is the "concrete" information that the individual can use to more readily form potentially favorable attitudes. A person’s attitude toward a product is the outcome of a series of episodic scenes relating the experience, either in actual or vicarious terms, with the attitude object (Abelson, 1981). The result of the subject’s exposure to the scripted advertisement when questioned about her/his attitude toward the product, is to recall the episodic vignettes and respond accordingly. All but one of the hypotheses (H5) confirmed the hypothesized impact of scripts on recall and attitude variables. Even in the case of the unsubstantiated hypothesis (H5), cell means were found to be in the expected direction supporting the proposed impact of 8.1 scripts. The hypothesized role of product involvement, however, was not' supported except for the script x involvement interaction effects observed in hypothesis 3. The results obtained in this study imply that the script format can be an effective method to convey print advertisement messages. Consistent with the propositions of the script theory, consumers’ extent of recall of and attitudes toward the advertisement and the advertised brand may be affected by the actual or vicarious experience evoked by the scripted advertisement. In other words, as evidenced by nPetty and Cacioppo (1983), an individual exposed to such an advertisement may be influenced by the peripheral cues, in this case the story in the scripted advertisement. This finding has important managerial and public policy implications which will be discussed next. Managerial Implications Probably the critical managerial question to be addressed in the script framework is this: should the more costly script format be used extensively in print advertisements?. As noted earlier, the three important goals of an advertiser include increasing the consumer’s memory for the message, developing a favorable disposition toward the product, brand, and message, and consequently, increasing purchase likelihood. In this study, it was empirically confirmed that the script format was significantly more effective in eliciting 8?: higher recall and recognition measures as well as more favorable advertisement attitudes. This has important implications in terms of the ability of the advertiser to affect secondary demand among consumers as well as to develop and maintain brand loyalty. Consumers who are in the process of deciding what brand to purchase, (i.e. secondary demand), can be more effectively influenced by an advertisement using the script format. In other words, to create top-of—the-mind awareness for a certain brand and message, users and nonusers can be more effectively reached by the script format. These results may in effect, override the increased costs associated with the production of a scripted advertisement. Furthermore, the fact that the scripted format resulted in more favorable advertisement attitudes implies that if a consumer likes an advertisement s/he may spend more time processing it and thus be more likely to recall the brand when preparing to purchase in that product class. In sum, scripted advertisements as opposed to nonscripted ones certainly seem to promise higher advertising effectiveness in terms of product and brand recall and favorable attitudes. Managers should consider utilizing this powerful approach to advertising message presentation. 3.3 Public Policy Implications One of the major implications in this study is relevant for contributing to the regulatory guidelines needed to aid advertisers and consumers in certain areas. There are certain products such as tobacco and liquor which have created much public controversy with many consumer protection groups and politicians calling for an outright ban on the advertising of these products. On the other side of the fence are the advertisers and the advertising agencies who feel it is a violation of their First Amendment right to freedom of speech to ban advertising of these products. Then, there are other products/issues such as condoms and social diseases, which require public information but due to the lack of regulatory guidelines, are either not at all advertised or advertised ineffectively. They may be advertised but the methods used to disseminate the information/message may not be in an effective format, medium, etc. The findings of this study have the potential to help public policy makers develop some guidelines for the advertising industry to use in preparing messages for any media, in this case, the print medium. Since the scripted format is found to be effective in associating an actual or vicarious experience with heightened awareness and recall, this could provide a a set of guidelines for public service announcements (p.s.a.s). These p.s.a.s can be developed to disseminate 84 information regarding such topics as condoms and social diseases, as well as teenage smoking and drunk driving issues. On the other side of the issue, is the need to provide guidelines to the tobacco and liquor industries for the use of the script format which, according to the present findings, is a powerful tool. In other words, the question remains as to if and how the scripted advertisements can be used by advertisers to encourage and increase the consumption of socially and physically hazardous goods and services. If the effectiveness of the script format is indeed demonstrated across other media and product classes, then perhaps the public policy makers should at least be aware of this fact. In summary, the script format, if used correctly can be an effective tool for public service announcements yet at the same time can continue to contribute to the controversy surrounding certain questionable products and services. Limitations Since this research was conducted in an experimental setting, it has a limitation in external validity like any other experimental study. Even though the real purpose of the experiment was not disclosed to the subjects, they might have given more attention to the booklet and its contents than they would normally do in a natural setting. 85 While this limitation constrains the generalizability of the results, some attempts were made to improve external validity. For example, an advertisement of a real product was used by altering the actual print advertisement. In addition, the advertisement was professionally embedded in a booklet along with other advertisements and articles. Perhaps the major limitation of this study is the highly controversial nature of the product class chosen as the advertising object. It is highly conceivable that the results obtained herein would have been different had the opportunity been available to use another scripted advertisement of an actual product which had less controversy attached to it. In such a case, the hypotheses are expected to be even more highly significant. Finally, the generalizability of these findings to other media is limited since the study focused on the print medium, expressly magazines. However, given the exploratory nature of this study, it should be borne in mind that these limitations constitute the future research opportunities which will be discussed next. Future Research One interesting extension of this study would be to compare the script format across different media such as television and radio. While this was not attempted in this study due to the expense involved in creating and producing advertisements, it is an issue that needs to be addressed 86 in order to generalize the present findings. Another research avenue concerns testing the present hypotheses across various other product categories, preferably less controversial durable and nondurable products. It is expected that such a strategy will produce even higher statistically significant results. This in turn would provide a stronger evidence for the effective and efficient use of the script format by the advertising industry. A further extension of this study could be to measure its dependent variables after multiple exposures to a scripted versus a nonscripted advertisement. In this regard, the impact of repetition and other variables that are previously shown to mediate advertising effectiveness can also be incorporated into the research design to increase the internal and external validity of the findings. APPENDICES APPENDIX A QUESTIONNAIRE INSTRUCTIONS The Advertising Department is developing an insert to be included in the STATE NEWS periodically which will provide information on various topics of interest. You are being provided with a sample copy and we are interested in your opinions on the articles as well as the advertisements. Take as much time as you need to go through the articles and the advertisements. Please, do not skip any pages. When you have finished looking through the booklet, close the booklet, and raise your hand. You will then be given a questionnaire. (1) Do not spend too much time reading articles; just skim through them. You don’t have to read the articles if you choose not to. (2) You will be asked questions regarding your opinions about the entire booklet including its appearance, the advertisements, etc. (3) Don’t switch between pages of the questionnaire. Follow the order of the questions carefully. (4) Don’t re-examine the booklet after you begin answering the questionnaire unless you are told to. F Please be careful with the booklet because it will be reused. 1 87 RESPONDENT ID 1) Please list all the PRODUCT CATEGORIES (i.e.: detergents, cereals, frozen foods, etc.) that were advertised in the booklet. 2) Please list all the BRANDS that were advertised in the booklet. 3) Please list all the TOPICS of the articles that were presented in the booklet. 88 13) l4) 15) What was the name of the BRAND of liquor advertised in the booklet? What was the name of the BRAND of cigarettes advertised in the booklet? What was the BRANCH in the booklet? of the military service that was advertised Circle the number that best indicates how accurately each scale item describes or fits the advertisement for BIGE