Suggestions For Manipulating and Measuring Involvement in Advertising Message Content

ABSTRACT - Given the recent interest in the theoretical predictions of many involvement-driven frameworks applied in advertising (e.g., the ELM, the Aad model), successfully manipulating and measuring involvement in advertising content is of great importance. Our paper seeks to aid researchers developing their own manipulations and measures of advertising involvement by providing an operational example of manipulated involvement in advertising content. Direct manipulation checks of the manipulated ad involvement condition are provided that successfully meet the requirements of a unidimensional, reliable, and valid measure of advertising content involvement. Implications for those attempting to measure involvement in advertising research are provided.

Citation:

J. Craig Andrews and Srinivas Durvasula (1991) ,"Suggestions For Manipulating and Measuring Involvement in Advertising Message Content", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 194-201.

Advances in Consumer Research Volume 18, 1991      Pages 194-201

SUGGESTIONS FOR MANIPULATING AND MEASURING INVOLVEMENT IN ADVERTISING MESSAGE CONTENT

J. Craig Andrews, Marquette University

Srinivas Durvasula, Marquette University

ABSTRACT -

Given the recent interest in the theoretical predictions of many involvement-driven frameworks applied in advertising (e.g., the ELM, the Aad model), successfully manipulating and measuring involvement in advertising content is of great importance. Our paper seeks to aid researchers developing their own manipulations and measures of advertising involvement by providing an operational example of manipulated involvement in advertising content. Direct manipulation checks of the manipulated ad involvement condition are provided that successfully meet the requirements of a unidimensional, reliable, and valid measure of advertising content involvement. Implications for those attempting to measure involvement in advertising research are provided.

INTRODUCTION

Adequately manipulating and measuring a person's involvement in advertising message content is becoming an increasingly important issue in experimental research today. This is because of the recent advancement and testing of many involvement-driven models in advertising that have an important bearing on knowledge development in the advertising and consumer research fields (e.g., the Elaboration Likelihood Model (Petty and Cacioppo 1981a; 1986), the Attitude-Toward-the-Ad Model (Lutz 1985; Mitchell and Olson 1981; Shimp 1981) and the Integrated Information Response Model (Smith and Swinyard 1982; 1983)).

To best examine the moderating effects of involvement in advertising content, the manipulation of involvement levels is recommended (cf., Petty, Cacioppo, and Schumann 1983). A sound manipulation of involvement will help to enhance internal validity (i.e., the ability to draw cause and effect inferences) and rule out confounding extraneous variable explanations (Carlsmith, Ellsworth, and Aronson 1976; Cook and Campbell 1979). However, manipulations of involvement in advertising message content vary greatly, including instructions for the memorization of ad content, expectations of purchase decisions, implications of purchases influenced by brand differences, expectations of local product availability, and distraction to reduce involvement (cf., Gardner 1985; Gardner, Mitchell, and Russo 1978; Laczniak, Muehling, and Grossbart 1989; Leigh and Menon 1987; Park and Young 1986; Petty, Cacioppo, and Schumann 1983). To be successful, however, researchers must manipulate differing levels of involvement in ad content, while holding all other factors constant (Andrews 1988; Apsler and Sears 1968).

Perhaps more troubling is the need for psychometrically-sound measures (i.e., manipulation checks) of manipulated advertising involvement. The development of psychometrically-sound measures begins with an explication of the construct for which the measures are to be derived (Cook and Campbell 1979). Some would agree that involvement can be defined as an internal, individual state of arousal with intensity and direction properties (Mitchell 1979; 1981). Others define involvement as personal relevance (Greenwald and Leavitt 1984; Petty and Cacioppo 1986; Zaichkowsky 1985; 1986). Still others have linked the two definitions arguing that a motivational state of arousal or activation can emanate from the personal relevance of the stimulus in question (Andrews 1988; Cohen 1983).

From this point, however, a multitude of options face the researcher attempting to measure manipulations of involvement in advertising content. A first choice for some may be the use of Zaichkowsky's (1985) 20-item personal involvement inventory (PII) used to assess one's personal involvement in a product category. However, questions have arisen regarding the PII's unidimensionality (McQuarrie and Munson 1987; Mittal 1989); that it is somewhat cumbersome to use with other important advertising effect variables (e.g., ad cognitive responses, Aad, brand attitude); and that it remains validated only for product categories (as opposed to advertising message content).

A second choice for many researchers is to use ad hoc measures to determine if they have successfully manipulated involvement in ad content. A wide array of measures have been used, including recall of ad content, recall of involvement instructions, and message cognitive responses (see Laczniak, Muehling, and Grossbart 1989 for a review). However, as indicated by Cohen (1983), cognitive responses and recall can be viewed as consequences of involvement, as opposed to more direct measures of whether or not a person is involved in the ad's message content.

One alternative choice is a six-item measure of advertising content involvement proposed by Andrews (1985; 1988). As suggested by Andrews (1985), the state of involvement in an advertising message can be checked via multiple items assessing the degree to which subjects attend to, concentrate on, think about, focus on, spend effort in looking at, and carefully read the advertising message. These proposed items are consistent with the definition of involvement as an internal, individual state of arousal with intensity and direction properties and would be less cumbersome to apply than the 20-item PII in an experiment. However, the unidimensionality (via confirmatory factor analysis; Gerbing and Anderson 1988), reliability, and validity (Fornell and Larcker 1981) of this six-item, advertising message involvement scale remain to be explored.

Therefore, the purpose of our paper is to provide an operational example of manipulated involvement in advertising message content in order to offer guidance to researchers attempting to manipulate and/or measure advertising message involvement. Included in this example are: direct manipulation checks (i.e., Andrews' six proposed measures) of the advertising content involvement condition; assessments of the measure's unidimensionality, reliability, and validity; its impact on predicted consequences; and how it differs from other related constructs of advertising content involvement. Implications for those attempting to manipulate and/or measure involvement in experimental advertising research are also provided.

INVOLVEMENT AND YOU: 1000 GREAT IDEAS

As indicated in Cohen's (1983) provocative article on the conceptualization of involvement, there may very well be "1000 great ideas" on the concept of involvement (cf., Andrews 1988; Antil 1984; Batra and Ray 1983; Cohen 1983; Gardner, Mitchell, and Russo 1978; 1985; Greenwald and Leavitt 1984; Houston and Rothschild 1978; Krugman 1966-1967; Lastovicka and Gardner 1979; Mitchell 1979; 1981; Park and Young 1986; Petty and Cacioppo 1986; Wright 1973; Zaichkowsky 1986; see Andrews, Durvasula and Akhter, 1990 for a comparison of involvement conceptualizations). Various "types" or targets of involvement have been outlined, such as task involvement (i.e., the importance of adopting a position - or performing a task - that will maximize immediate rewards; Sherif and Hovland 1961; Zimbardo 1960),- personal involvement (Apsler and Sears 1968; Petty and Cacioppo 1981b; 1986), product involvement (Bloch 1981; Day 1970; Zaichkowsky 1985), and advertising message (content) involvement (Andrews 1988; Laczniak, Muehling, and Grossbart 1989; Wright 1973).

Typology labels, however, may give the wrong impression since it is the individual who is involved, not products, tasks, or advertising content. Some involvement theorists have gone beyond involvement types and focused on the underlying properties of those in the state of involvement. For example, one view that is emerging is that involvement represents an individual, internal state of arousal with intensity and direction properties (Mitchell 1979; 1981). [See Greenwald and Leavitt (1984) for the conceptualization of involvement as a process.] In this sense, involvement is more than simply "arousal," because it not only is characterized by a variety of intensity levels, but is directed toward a particular stimulus object or situation. As proposed by Mitchell (1979), intensity refers to the level (or degree) of arousal, interest, or drive in a stimulus object or situation. For example, in the case of involvement in advertising message content, one can measure the degree of focus, concentration, attention, thought, effort, etc. expended (on scrutinizing the content of a target advertisement). Direction, on the other hand, concerns the actual evoking stimulus object and/or situation. With advertising involvement, involvement direction can refer to different levels of specificity; that is, toward the ad in general, its executional features, or its message content (Baker and Lutz 1987). For the purposes of our paper, we will focus on the manipulation and measurement of involvement in advertising message content.

As suggested by Andrews (1988, p. 220), "...indicators tapping the involvement state (vs. antecedents or consequences of this state) should be used as a measure (i.e., manipulation check) of involvement." For example, numerous antecedents to advertising message involvement exist, such as the personal relevance of a product, prior product experience, risk, decision factors (e.g., time and magnitude), and personality traits (e.g., need for cognition) (Andrews 1988; Petty and Cacioppo 1986; Zaichkowsky 19&6). Similarly, numerous consequences of involvement also exist,- including one's degree of search behavior, information processing (e.g., cognitive response activity, message recall), and persuasion (Cohen 1983). These distinctions are important because there may be a temptation to simply infer from antecedents or consequences that one's advertising involvement manipulation has "taken" without assessing more direct measures of the involvement state. Certainly, however, involvement state measures should also be related to their antecedents and consequences in the experiment.

MEASURING MANIPULATED INVOLVEMENT IN ADVERTISING MESSAGE CONTENT: SOME SUGGESTIONS

In experimentation, researchers face the difficult challenge of providing a close correspondence between their constructs (and proposed manipulations) at a conceptual level and their measures (e.g., manipulation checks) at the operational level. While correspondence rules between the conceptual and operational levels can be assessed through the content validity of the measures, it is clear that more rigor is needed in subjecting our measures of manipulated advertising involvement to accepted procedures of unidimensionality, reliability, and validity assessment (Churchill 1979; Fornell and Larcker 1981; Gerbing and Anderson 1988; Perdue and Summers 1986; Peter 1981). For example, internal consistency and exploratory factor analysis are argued nor to be explicit tests of the predicted dimensionality of a construct (Gerbing and Anderson 1988, p. 189, 190). Rather, in order to confirm whether the number of predicted dimensions can be verified empirically, a confirmatory factor analysis is strongly suggested in scale construction (Churchill 1979; Gerbing and Anderson 1988). Also, in order to enhance the researcher's confidence in the causal explanations of the experiment, the potential confounding of involvement manipulations should be avoided (Perdue and Summers 1986). Confounding occurs when a manipulation of one theoretical construct (e.g., involvement) is found to represent more than one construct (e.g., involvement and opportunity to respond due to distraction instructions). Therefore, confounding checks (Perdue and Summers 1986) are needed to determine if related constructs (e.g., opportunity to respond) have been inadvertently affected by the involvement manipulation. Also, manipulations of other variables in the experiment (e.g., ELM-related variables, such as source credibility and argument strength) should not have an influence on involvement manipulation checks. With this in mind, we now present an overview of our experiment manipulating involvement in advertising message content. Theoretical predictions and results for measures of the manipulated involvement condition are then discussed.

AN OPERATIONAL EXAMPLE OF MANIPULATING INVOLVEMENT IN ADVERTISING MESSAGE CONTENT

Experiment Overview

Consistent with involvement antecedents suggested by Petty and Cacioppo (1986), subjects' involvement in the content of a target, mock advertisement was manipulated by varying the personal relevance of a low alcohol beer brand. The targeted mock ad appeared in a booklet of nine mock and real ads presented to 186 undergraduate students of legal drinking age. The overall purpose of the experiment was to examine the persuasive effects of advertising content involvement, source characteristics, and argument strength treatments. These variables represent important elements in ELM-related research (see Petty, Cacioppo, and Schumann 1983 for similar manipulations). A series of four pretests were first conducted varying involvement procedures suggested by Petty, Cacioppo, and Schumann (1983), number of booklet ads (0-9 ads), and viewing time constraints (1-7 minutes). The last pretest was successful and a final manipulation was developed in which both high and low involvement subjects were given five minutes to view the nine ads in the booklet. High involvement subjects expected the target advertised brand to be test marketed in their city, expected a gift choice in the advertised brand's product category, and the possibility of an interview to determine if they carefully read the target ad's claims. Low involvement subjects, on the other hand, only expected the target advertised brand to be test marketed in a distant region and a product gift choice in an unrelated product class.

Theoretical Predictions

Consistent with our conceptualization of involvement, the theoretical predictions regarding the advertising content involvement manipulation and manipulation checks are as follows. First, an immediate check on the personal relevance instructions should indicate, for example, that high involvement subjects did indeed expect the low alcohol beer brand to be test marketed in their ow city to a greater extent than low involvement subjects (see Petty, Cacioppo, and Schumann 19 However, these types of checks do not indicate whether or not high involvement subjects were indeed more involved in the content of the advertisement than low involvement subjects. To accomplish this, the summation of six, 9-point items (coefficient alpha = .95) was used as a check to measure the intensity of manipulated involvement in the message content of the target advertisement. Subjects indicated their level of agreement/ disagreement to whether- they were paying attention to, concentrating on, thinking about, focusing o spending effort looking at, and carefully reading content of the target advertisement (Andrews 198 These items were originally developed and purified following Churchill's (1979) first four steps in sc development. For example, regarding content validity, an attempt was made to sample items that tapped the involvement state definition of an individual's internal state of arousal in advertising, content. Coefficient alphas during scale purification ranged from .89 to .95. The focus of the present study is on the evaluation of the resulting six-item measure on the basis of Churchill's: step #5 (collection of data on the proposed measure), an assessment of dimensionality, step #6 (reliability: assessment) and step #7 (validity assessment).

To begin, because the six-item advertising involvement check was developed to measure the intensity of the involvement state, it is expected that the check be unidimensional in nature. In o to verify the check's undimensionality, data was collected and subjected to a confirmatory factor analysis using LISREL VI (Joreskog and Sorbom 1983). The specific assessment of undimensional can then be examined through fit indices (e.g., X goodness-of-fit index; adjusted goodness-of-fit index; and root mean square residual), normalized residuals, and significance of the t-values of each manipulation check item from the confirmatory factor analysis results (Gerbing and Anderson 19, Next, the measurement properties and predicted structure of the advertising involvement check item can also be, examined through the results of the confirmatory factor analysis. In particular, the reliability of the measures can be evaluated via it, and construct reliability estimates, with a desired minimum standard of .50 (Fornell and Larcker 19 p. 45).

It was also expected that the check of the advertising involvement state be construct valid (Peter 1981). That is, the six-item measure should demonstrate content validity (as previously discussed), convergent validity, discriminant validity, and predictive validity within its nomological network of relationships (see Peter 1981; Zaltman, Pinson, and Angelmar 1973, p.44 a for validity definitions). The proportion of variance extracted (i.e., shared variance, Fornell and Larcker, 1981, p. 46) from the multiple items provides an estimate of the convergent validity of the advertising involvement construct. Again, the minimum cutoff is to explain at least 50 percent of the variance. The convergence between the advertising involvement manipulation and the advertising involvement check can be assessed by the degree to which high involvement subjects do, in fact, indicate higher levels of involvement on the check than low involvement subjects (Perdue and Summers 1986). Discriminant validity can be estimated by showing that the shared variance estimates within two related constructs (e.g., advertising involvement and argument strength) are greater than the squared correlation between the two constructs (Fornell and Larcker 1981, p. 46). In addition, the predictive validity of the advertising involvement manipulation can be established if important theoretical predictions regarding the influence of the manipulation on criterion variables (e.g., message-oriented thoughts, message argument -recall) are confirmed. For example, the ELM (Petty and Cacioppo 1986, pp. 36, 37) predicts that greater message-oriented thoughts and message argument recall should occur under high versus low involvement.

Consistent with our previous conceptualization of involvement, the directionality (i.e., toward the content, as opposed to the sources) of the advertising involvement manipulation can be assessed through a nine-point measure of relative concentration, anchored by "I concentrated most on the claims in the ad" versus "I concentrated most on the people in the ad" (Wright 1973). It is expected that the relative concentration on claims (vs. people) in the ad will be greater for high versus low involvement subjects in the experiment. Finally, the argument strength and source manipulations should not influence (i.e., confound) the involvement manipulation check. An example of a confound check (Perdue and Summers 1986) is also provided in a second advertising involvement experiment to determine if a related construct (i.e., opportunity to process) had been inadvertently influenced by the involvement manipulation.

Results

First, a check on the personal relevance instructions was made by asking subjects, on a nine-point measure, the likelihood that the low alcohol beer described in the advertisement will soon be available in their region. As expected, high involvement subjects (M = 7.71, SD = 1.73) expressed a stronger likelihood of the product's availability than low involvement subjects (M = 5.85; SD = 2.34; F(1,184 I = 34.67, p < .001). However, as previously indicated, in order to assess whether subjects were then more involved in the content of the ad, the six-item measure of advertising involvement intensity was examined.

To help examine the unidimensionality of the manipulation check of advertising involvement intensity, overall model fit statistics (e.g., x2, GFI, AGFI, and RMSR) are estimated from the confirmatory factor analysis and provided in Table 1. However, due to the sensitivity of x2 to sample size, it is generally agreed that x2 should only be used as a guide rather than an absolute assessment of fit (Bagozzi and Yi 1988; Bearden, Sharma, and Teel 1982; Fornell and Larcker 1981; Hayduk 1987; Shimp and Kavas 1984). The goodness-of-fit and adjusted goodness-of-fit indices, as well as the root mean square residual, indicate a relatively good fit of the predicted one-factor model for the advertising involvement check (see Table 1). [For comparison purposes, a principal components analysis of the six-item, advertising content involvement measure indicated that one factor accounted for 78.9 percent of the total variance. All factor loadings were .82 or greater on this single factor.] Furthermore, all normalized residuals are less than .61 indicating an appropriate specification of the model. (Normalized residuals greater than 2 indicate a need for respecification; cf., Gerbing and Anderson 1988). Finally, each indicator t-value exceeded 12.30 (p < .001). Therefore, the confirmatory factor analysis results provided support for the predicted unidimensionality of the manipulation check of advertising involvement intensity in the experiment.

Regarding the measurement properties of this six-item measure of advertising content involvement, results of the confirmatory factor analysis in Table 1 indicate that all item-and construct reliabilities, as well as the shared variance estimate, exceed the minimum standard of 50 for the scale. This manipulation check also revealed that high involvement subjects were significantly more involved in the target advertisement (M = 45.13; SD - 8.35) than low involvement subjects (M = 35.67; SD = 12.28; F(1,184) = 33.99; p < .001), thereby providing evidence for the convergence between the advertising involvement manipulation and its associated check of the manipulation. The discriminant validity of the advertising involvement check was evaluated through its comparison with checks of the two other manipulations in the experiment: source characteristics (the sum of ten, 9-point items; a = .96) and argument strength (the sum of four, 9-point items; a = .91). The shared variance of the advertising involvement (.749) and source characteristics (.661) constructs were both greater than the squared correlation between the two constructs (.062). In turn, the shared variance estimates of the advertising involvement (.749) and argument strength (.695) constructs were also greater than the squared correlation between the two respective constructs (.180), thereby demonstrating the discriminant validity of the advertising involvement check versus the checks of the other two related manipulations in the experiment. In addition, the source and argument strength manipulations did not have a confounding effect (i.e., a significant influence) on the advertising involvement check, as desired. In support of the predictive validity of the advertising involvement manipulation and consistent with central route predictions of the ELM, high involvement subjects recalled a significantly greater number of message arguments from the advertisement (M = 2.01; SD = 1.19) than did low involvement subjects (M = 1.32; SD = 1.02; F(1,184) = 17.89; p < .001). In addition, high involvement subjects generated significantly greater message-oriented thoughts (M = 1.70; SD = 1.18) than low involvement subjects (M = 1.05; SD = 1.04; F(1, 84) = 15.41; p < .001), as predicted.

TABLE 1

MEASUREMENT MODEL AND FIT STATISTICS FOR THE ADVERTISING CONTENT INVOLVEMENT SCALE

To examine the predicted directionality (i.e., toward the content) of the involvement in the target ad, subjects' relative concentration in the ad was measured (Wright 1973). As expected, the relative concentration on "claims in the advertisement" (as opposed to "people in the advertisement") was significantly greater for high (M = 6.54; SD = 2.46) versus low involvement subjects (M = 5.50; SD = 2.52; F (1,184) = 7.95, p < .005). In conjunction with previous results indicating greater message-oriented thoughts (i.e., support and counter-arguments regarding the message content of the ad) under high (vs. low) involvement, this finding provides support for the intended directionality of the involvement manipulation.

Confound Checks

In addition to providing valid manipulation checks that successfully measure the degree to which subjects are involved in the goal-related object, it is important to also include confounding checks as well. A confounding check is a special type of manipulation check that tests for a divergence of measures and manipulations of related, but distinct "things" (Perdue and Summers 1986). Related, but distinct "things" that may be inadvertently manipulated or found operating in an advertising involvement study include one's opportunity and/or ability to process information (Andrews 1988; Batra and Ray 1986). For example, Wright (1974) found important interactions between ad content involvement and media conditions, serving to limit one's opportunity to process the advertised message.

A second advertising experiment can be used as an example of the use of this type of confound check for involvement. This experiment manipulated both the involvement in ad content, as well as the distinctiveness of the ad (via color). An opportunity to process confound check was included ('The study coordinator gave me enough time/ opportunity to look at the advertisement" on a 7-point, Likert-type scale) to determine whether or not the manipulations had inadvertently limited the subjects' processing of the target ad. The results indicated that neither manipulation had impacted the opportunity to process confound check, as hoped. However, if they were to have an effect, it is suggested that one look at the relative size of the confound effect (e.g., via omega-squared for ANOVA) to determine if it is greater in magnitude than the predicted effects for the study's dependent variables (cf., Perdue and Summers 1986).

CONCLUSION

The primary objective of our paper is to assist researchers in the development and validation of advertising involvement manipulation checks by providing an example of an advertising involvement manipulation and corresponding checks of the induced involvement state. Specifically, we experimentally manipulated advertising content involvement through personal relevance instructions for an advertised brand (an antecedent); measured the intensity and direction of the involvement manipulation; and examined the impact of the involvement manipulation on cognitive response and message argument recall activity (consequences). The use of an advertising involvement confound check (i.e., opportunity to process) is also demonstrated in an example from a second experiment.

Our study provides advertising involvement researchers with another option in the form of a unidimensional, reliable, and valid scale of advertising content involvement. The scale can be of assistance to those examining theoretical predictions of advertising involvement, such as those based on the ELM's central and peripheral routes to persuasion (Petty and Cacioppo 1986). For example, under high message content involvement, subjects are expected to scrutinize the content of the advertising message (Petty and Cacioppo 1981b). Therefore, it is important to first determine if these subjects are involved in the message content before any interpretation of predictions regarding their cognitive and affective responses to the ad. This understanding also holds true for advertisers and public policy officials. The design of relatively strong ad copy and the effectiveness of message arguments in alcohol and drug awareness campaigns may depend upon whether message recipients are involved in the examination of message content (cf., DePaulo, Rubin, and Miller 1987). However, if it can't be determined whether or not the target audience is actually involved in the content of the message, otherwise persuasive and believable arguments may simply turn out to be ineffective ones.

Future research should also examine the reliability, validity, and dimensionality of other advertising involvement manipulations and manipulation checks. For example, it is quite possible that physiological measures of involvement arousal (e.g., EEG patterns--Alwitt 1985; voice-pitch analysis; eye cameras, etc.) can be used in conjunction with the measures reported in our study to strengthen the insight into underlying properties of involvement. Also, manipulation check measures of advertising execution (vs. content) involvement could be developed and assessed using the techniques employed in our study. Efforts in this regard will help to enhance our understanding of the role that involvement plays in advertising research.

REFERENCES

Alwitt, Linda F. (1985), "EEG Activity Reflects the Content of Commercials," in Psychological Processes and Advertising Effects: Theory, Research, and Application, Linda P. Alwitt and Andrew A. Mitchell, eds., Hillsdale, NJ: Lawrence Erlbaum Associates, 201-217.

Andrews, J. Craig (1985), Tests of the Elaboration Likelihood Model Involving Persuasive Marketing Communication, unpublished doctoral dissertation, University of South Carolina, Columbia, SC.

Andrews, J. Craig (1988), "Motivation, Ability, and Opportunity to Process Information: Conceptual and Experimental Manipulation Issues," in Advances in Consumer Research, XV, Michael J. Houston, ed., Provo, UT: Association for Consumer Research, 219-225.

Andrews, J. Craig, Srinivas Durvasula, and Syed Akhter (1990), "A Framework for Conceptualizing and Measuring the Involvement Construct in Advertising Research," Journal of Advertising, 19(4)

Antil, John H. (1984), "Conceptualization and Operationalization of Involvement," in Advances in Consumer Research, XI, Thomas C. Kinnear, ed., Ann Arbor, MI: Association for Consumer Research, 203-209.

Apsler, Robert and David O. Sears (1968), "Warning, Personal Involvement, and Attitude Change," Journal of Personality and Social Psychology, 9, 162-166.

Bagozzi, Richard P. and Youjae Yi (1988), "On The Evaluation of Structural Equation Models," Journal of the Academy of Marketing Science, 16 (Spring), 74-94.

Baker, William E. and Richard J. Lutz (1987), "The Relevance-Accessibility Model of Advertising Effectiveness," in Nonverbal Communication in Advertising, S. Hecker and D.W. Stewart, eds., Lexington, MA: Lexington Books, 59-84.

Batra, Rajeev and Michael L. Ray (1983), "Operationalizing Involvement as Depth and Quality of Cognitive Response," in Advances in Consumer Research, X, Richard P. Bagozzi and Alice M. Tybout, eds., Ann Arbor, MI: Association for Consumer Research, 309-313.

Batra, Rajeev and Michael L. Ray (1986), "Situational Effects of Advertising Repetition: The Moderating Influence of Motivation, Ability, and Opportunity to Respond," Journal of Consumer Research, 12 (March), 432445.

Bearden, William A., Subhash Shanna, and Jesse E. Teel (1982), "Sample Size Effects on Chi Square and Other Statistics Used in Evaluating Causal Models," Journal of Marketing Research, 19 (November), 425430.

Bloch, Peter H. (1981), "An Exploration into the Scaling of Consumers' Involvement with a Product Class," in Advances in Consumer Research, VI, Kent R. Monroe, ed., Ann Arbor, MI: Association for Consumer Research, 61-65.

Carlsmith, J. Merrill, Phoebe C. Ellsworth, and Elliot Aronson (1976), Methods of Research in Social Psychology? Reading, MA: Addison-Wesley.

Churchill, Gilbert A., Jr. (1979), "A Paradigm for Developing Better Measures of Marketing Constructs," Journal of Marketing Research, 16 (February), 64-73.

Cohen, Joel C. (1983), "Involvement and You: 1000 Great Ideas," in Advances in Consumer Research, X, Richard P. Bagozzi and Alice M. Tybout, eds., Ann Arbor, MI: Association for Consumer Research, 325-328.

Cook, Thomas D. and Donald T. Campbell, Quasi-Experimentation: Design and Analysis Issues for Field Settings, Boston: Houghton Mifflin.

Day, George (1970), Buyer Attitudes and Brand Choice Behavior, New York: Free Press.

DePaulo, Peter J., Mary Rubin, and Brenton Milner (1987), "Stages of Involvement with Alcohol and Heroin: Analysis of the Effects of Marketing on Addiction," in Advances in Consumer Research, Vol. XIV, Melanie Wallendorf and Paul Anderson, eds., Provo, UT: Association for Consumer Research, 521-525.

Fornell, Claes and D.F. Larcker (1981), "Structural Equation Models with Unobservable Variables and Measurement Error," Journal of Marketing Research, 18 (February), 39-50.

Gardner, Meryl Paula (1985), "Does Attitude Toward the Ad Affect Brand Attitude Under a Brand Evaluation Set?" Journal of Marketing Research, 22 (May), 192-198.

Gardner, Meryl Paula, Andrew A. Mitchell, and J. Edward Russo (1978), "Chronometric Analysis: An Introduction and an Application to Low Involvement Perception of Advertisements," in Advances in Consumer Research, V, H. Keith Hunt, ed., Ann Arbor, MI: Association for Consumer Research, 581-589.

Gardner, Meryl Paula, Andrew A. Mitchell, and J. Edward Russo (1985), "Low Involvement Strategies for Processing Advertisements," Journal of Advertising, 14 (2), 4-12, 56.

Gerbing, David W. and James C. Anderson (1988), "An Updated Paradigm for Scale Development Incorporating Unidimensionality and Its Assessment," Journal of Marketing Research, 25 (May), 186-192.

Greenwald, Anthony G. and Clark Leavitt (1984), "Audience Involvement in Advertising: Four Levels," Journal of Consumer Research, 11 (June), 581 -592.

Hayduk, Leslie A. (1987), Structural Equation Modeling with LISREL: Essentials and Advances, Baltimore, MD: Johns Hopkins University Press.

Houston, Michael J. and Michael L. Rothschild (1978), "Conceptual and Methodological Perspectives in Involvement," in Research Frontiers in Marketing: Dialogues and Directions, S. Jain, ed., Chicago: American Marketing Association, 184- 187.

Joreskog, Karl G. and Dag Sorbom (1983), LISREL Yl: Analysis of Linear Structural Relationships by Maximum Likelihood and Least Squares Methods, Mooresville, IN: International Educational Services.

Krugman, Herbert E. (1966-1957), "The Measurement of Advertising Involvement," Public Opinion Quarterly, 30 (Winter), 583-596.

Laczniak, Russell N., Darrel D. Muehling, and Sanford Grossbart (1989), "Manipulating Message Involvement in Advertising Research," Journal of Advertising, 18 (2), 28-38.

Lastovicka, John L. and David M. Gardner (1979), "Components of Involvement," in Attitude Research Plays for High Stakes, John C. Maloney and Bernard Silverman, eds., Chicago, IL: American Marketing Association, 53-73.

Leigh, James H. and Anil Menon (1987), "Audience Involvement Effects on the Information Processing of Umbrella Print Advertisements," Journal of Advertising, 16 (3), 3-12.

Lutz, Richard J. (1985), "Affective and Cognitive Antecedents of Attitude Toward the Ad: A Conceptual Framework," in Psychological Processes and Advertising Effects: Theory, Research and application, Linda F. Alwitt and Andrew A. Mitchell, eds., Hillsdale, NJ: Lawrence Erlbaum, 45-63.

McQuarrie, Edward F. and J. Michael Munson (1987), 'The Zaichkowsky Personal Involvement Inventory: Modification and Extension," in Advances in Consumer Research, XIV, Melanie Wallendorf and Paul Anderson, eds., Provo, UT: Association for Consumer Research, 3640.

Mitchell, Andrew A. (1979), "Involvement: A Potentially Important Mediator of Consumer Behavior," in Advances in Consumer Research, VI, William Wilkie, ed., Ann Arbor, MI: Association for Consumer Research, 191-196.

Mitchell, Andrew A. (1981), "Dimensions of Advertising Involvement," in Advances in Consumer Research, VIII-,- Kent B. Monroe, ed., Ann Arbor, MI: Association for Consumer Research, 25-30.

Mitchell, Andrew A. and Jerry C. Olson (1981), "Are Product Attribute Beliefs the Only Mediators of Advertising Effects on Brand Attitudes?" Journal of Marketing Research, 18 (August), 318-332.

Mittal, Banwari (1989), "A Theoretical Analysis of Two Recent Measures of Involvement," in Advances in Consumer Research, XVI, Thomas K. Srull, ed., Provo, UT: Association for Consumer Research, 697-702.

Park, C. Whan and S. Mark Young (1986), "Consumer Response to Television Commercials: The Impact of Involvement and Background Music on Brand Attitude Formation," Journal of Marketing Research, 23 (February), 11-24.

Perdue, Barbara C. and John O. Summers (1986), "Checking the Success of Manipulations in Marketing Experiments," Journal of Marketing Research, 23 (November), 317-326.

Peter, J. Paul (1981), "Construct Validity: A Review of Basic Issues and Marketing Practices," Journal of Marketing Research, 28 (May), 133-145.

Petty, Richard E. and John T. Cacioppo (1981a), Attitudes and Persuasion: Classic and Contemporary Approaches, Dubuque, IA: William C. Brown.

Petty, Richard E. and John T. Cacioppo (1981b), "Issue Involvement as a Moderator of the Effects on Attitude of Advertising Content and Context," in Advances in Consumer Research, VIII, Kent B. Monroe, ed., Ann Arbor, MI: Association for Consumer Research, 20-24.

Petty, Richard E. and John T. Cacioppo (1986), Communication and Persuasion: Central and Peripheral Routes to Attitude Change, New York: Springer-Verlag.

Petty, Richard E., John T. Cacioppo, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 (September), 135-146.

Sherif, Muzafir and Hadley Cantril (1947), The Psychology of Ego-Involvements, Social Attitudes and Identifications, New York: John Wiley & Sons, Inc.

Shimp, Terence A. (1981), "Attitude Toward the Ad as a Mediator of Consumer Brand Choice," Journal of Advertising, 10 (2), 9-15, 48.

Shimp, Terence A. and Alican Kavas (1984), "The Theory of Reasoned Action Applied to Coupon Usage," Journal of Consumer Research, 11 (December), 795-809.

Smith, Robert E. and William R. Swinyard (1982), "Information Response Models: An Integrated Approach," Journal of Marketing, 46 (Winter), 81 -93.

Smith, Robert E. and William R. Swinyard (1983), "Attitude-Behavior Consistency: The Impact of Product Trial Versus Advertising," Journal of Marketing Research, 20 (August), 257-267.

Wright, Peter (1973), 'The Cognitive Processes Mediating Acceptance of Advertising," Journal of Marketing Research, 10 (February), 53-62.

Wright, Peter (1974), "Analyzing Media Effects on Advertising Responses," Public Opinion Quarterly, 38, 192-205.

Wright, Peter (1980), "Message-Evoked Thoughts: Persuasion Research Using Thought Verbalizations," Journal of Consumer Research, 7 (September), 151 -175.

Zaichkowsky, Judith Lynne (1985), "Measuring the Involvement Construct," Journal of Consumer Research, 12 (December), 341-352.

Zaichkowsky, Judith Lynne (1986), "Conceptualizing Involvement, " Journal of Advertising, 15 (2), 4-14, 34.

Zaltman, Gerald, C.R.A. Pinson, and R. Angelmar (1973), Meta-theory and Consumer Research, New York: Holt, Rinehart, and Winston.

Zimbardo, Philip G. (1960), "Involvement and Communication Discrepancy as Determinants of Opinion Conformity," Journal of Abnormal and Social Psychology, 60 (1), 86-94.

----------------------------------------