Self-Monitoring and Reactions to Image Appeals and Claims About Product Quality

ABSTRACT - Low versus high self-monitors have been shown recently to differ markedly in their reactions to image versus quality-based advertising appeals. In a series of three studies, Synder and DeBono (1985) reported results which showed low (high) self-monitors as favoring quality (image) ads, willing to pay more for products promoted via quality (image) messages, and more willing to try products marketed by quality (image) claims. The present paper describes the results of a series of several studies conducted at three separate times designed to replicate and extend that- research. The results suggest that it is premature to conclude that self-monitoring is a strong moderator of reactions to different advertising strategies.

Citation:

William O. Bearden, F. Kelly Shuptrine, and Jesse E. Teel (1989) ,"Self-Monitoring and Reactions to Image Appeals and Claims About Product Quality", in NA - Advances in Consumer Research Volume 16, eds. Thomas K. Srull, Provo, UT : Association for Consumer Research, Pages: 703-710.

Advances in Consumer Research Volume 16, 1989      Pages 703-710

SELF-MONITORING AND REACTIONS TO IMAGE APPEALS AND CLAIMS ABOUT PRODUCT QUALITY

William O. Bearden, University of South Carolina

F. Kelly Shuptrine, University of South Carolina

Jesse E. Teel, University of South Carolina

ABSTRACT -

Low versus high self-monitors have been shown recently to differ markedly in their reactions to image versus quality-based advertising appeals. In a series of three studies, Synder and DeBono (1985) reported results which showed low (high) self-monitors as favoring quality (image) ads, willing to pay more for products promoted via quality (image) messages, and more willing to try products marketed by quality (image) claims. The present paper describes the results of a series of several studies conducted at three separate times designed to replicate and extend that- research. The results suggest that it is premature to conclude that self-monitoring is a strong moderator of reactions to different advertising strategies.

The relative influence of dispositional versus situational determinants of individual behavior has been of interest to consumer researchers for some time (e.g., Belk 1974; Kassarjian 1971). This has resulted in efforts to identify enduring personality traits and/or to investigate the strength and consistency of attitude-behavior relationships. Recognizing that individuals vary in their sensitivity to social cues and the ability to adapt their behavior to the requirements of the situation, Snyder and his colleagues developed the concept of self-monitoring along with a scale to measure the construct (e.g., Snyder 1974; Snyder and Swann 1976). Self-monitoring is offered as a means of identifying those individuals for whom situational or dispositional variables have primary influence (Becherer and Richard 1978).

High self-monitoring individuals are particularly sensitive to the expression and self-presentation of others in social situations and use these cues as guidelines for managing their own behavior (Snyder and Monson 1975, p. 637). Consequently, high self-monitors exhibit situation-to-situation variability in behavior. In contrast, low self-monitors guide their behavior from personal dispositions, demonstrate more consistent behavior across situations, and display substantial correspondence between attitudes (internal predispositions) and behavior (Ajzen, Timko, and White 1982; Snyder and Swann 1976; Zanna, Olson, and Fazio 1980).

Recent research by Snyder and DeBono (1985) indicated that low and high self-monitors vary in their reactions to different types of advertisements. The strength and consistency of the relationship between self-monitoring, a personality construct, and various measures of response to advertisements make these results of particular interest to consumer research. The studies reported in this article were primarily designed to replicate Snyder and DeBono (1985) and to provide additional evidence of the relationship between one aspect of personality, self-monitoring, and consumer behavior.

SELF-MONITORING AND CONSUMER BEHAVIOR

The self-monitoring concept has been applied to a number of consumer behavior and marketing research issues. Becherer and Richard (1978) found that the relationship between individual personality characteristics and private brand proneness was strongest (as predicted) for low self-monitors. In a field study of sales representatives, Caldwell and O'Reilly (1982) reported self-monitoring-to be positively correlated with both job performance and tenure. McCann and Hancock (1983) report findings that suggest high self-monitors are likely to vary their messages in communication with other individuals in a manner that is consistent with their perceptions of the listener's position or attitude. Nantel and Strahle (1986) indicate that low self-monitors demonstrate greater consistency between stated intentions and behavior. Brinberg and Plimpton (1986) demonstrated that high self-monitors are more susceptible to reference group influence when considering luxury products.

The strongest results using self-monitoring in a consumer research context were described recently by Snyder and DeBono (1985). In a series of three studies, high (low) self-monitors were shown to consistently react more favorably to image (quality) appeals. Image messages were constructed to convey information about the role of the advertised products in affecting one's looks or social position; quality messages stressed product performance and/or intrinsic product attributes. A number of implications were provided for the development of advertising and marketing strategies. These conclusions were based solely upon the findings of the study with only limited consideration given to prior research regarding the effects of advertising and other persuasive communications .

The original development of the Snyder (1974) scale attempted to address five domains of self-monitoring: (1) concern for the social appropriateness of one's self-presentation; (2) attention to social comparison information as cues to appropriate self-expression (3) the ability to control and modify one's self-presentation and expressive behavior; (4) the use of this ability in particular situations; and (5) the extent to which the respondent's expressive behavior and self-presentation is consistent or variable across situations (Snyder 1974, p. 529). Lennox and Wolfe (1984) later developed a revised version of Snyder's self-monitoring scale. The need for this re-examination of the self-monitoring construct and revision of the Snyder (1974) scale resulted from conflicting findings in tests of the cross-situational hypothesis and failure of factor analytic studies to reproduce the five components of self-monitoring (e.g., Briggs, Cheek, and Buss 1980). The principal conceptual criticism has been that a number of the construct's original components are related to social anxiety which is inconsistent with social give-and-take -- a critical aspect of self-monitoring. The revised scale of Lennox and Wolfe (1984) represents an attempt to more narrowly define self-monitoring by considering only two factors: (1) the ability to modify self-presentation and (2) sensitivity to the expressive behavior of others.

OVERVIEW

The present research involved a series of systematic replications of the Snyder and DeBono (1985) study in that alternative methods and operationalizations of variables were employed. In the original research of Snyder and DeBono (1985), low and high self-monitors were exposed to either quality or image appeals for three branded products (i.e., Canadian Club whiskey, Barclay cigarettes, and Irish Mocha Mint coffee). In three separate studies, their results revealed that low self-monitors reacted more favorably to quality-based ads. In contrast, the high self-monitors reacted more favorably to the image-based appeals.

In the present study and, as in most uses of the Self-Monitoring Scale (Snyder 1974), respondents were initially split into low and high self-monitors using a median split of respondent Self-Monitoring (SM) Scale scores (i.e., the summed scores formed by adding each subjects' responses to the 25 true-false items comprising the scale.) [In view of the possibility that measurement error in the self-monitoring measures may have resulted in enough misclassification of cases near the median to prevent self-monitoring from moderating as expected, the analyses were repeated using only subjects from the top and bottom thirds of the scale distributions (cf. Wolfe, Lennox, and Hudiburg 1983, p. 1073). The resulting pattern of means and the statistical significance of the findings were not affected in any of these extreme split comparisons.] The results of this traditional partitioning are described along with other median sample splits formed by subject responses to: the 13-item Revised Self-Monitoring Scale (Lennox and Wolfe 1984); the seven-item Ability-to-Modify-Self-Presentation subscale within the Revised SM scale; and both the Snyder SM scale and the Lennox and Wolfe Revised SM scale. In this latter analysis, only those subjects falling above (below) the median for both scales were included as high (low) self-monitors.

The results that are presented in the subsequent sections of this paper are based upon a series of studies conducted at three separate intervals or phases. These systematic replications (Phases I-III) involved jeans and shoes as product stimuli -- the two products most frequently mentioned in a pretest elicitation of 30 student subjects. For all studies, junior and senior business administration students were used as subjects. As in the research of Snyder and DeBono (1985), subjects were told that their help was needed in evaluating the relative merits of advertisements currently being studied by researchers at the university. In the first phase, the three studies of Snyder and DeBono (1985) were replicated with different products. In these studies, subjects first responded to the self-monitoring measures which were followed by exposure to the advertisements and assessment of the dependent measure. In the second phase, one of the first three studies was again conducted. However, in this instance, self-monitoring scores were assessed two days prior to exposure to the advertisements and assessment of the dependent variable. The third phase involved a fifth study in which actual brand names were employed. [Two fictitious brand names were used in the first two phases to represent neutral stimuli in an attempt to avoid positive or negative affective feelings associated with the brand names confounding the effects of message claims. In a pretest involving 50 subjects, each brand name was evaluated on three 7-placed bipolar adjective semantic differential scales (e.g., pleasant-unpleasant, like very much-dislike very much). Coefficient alpha estimates for these scales were .91 and .94 for the shoe and jean analyses, respectively. Two-tailed test comparisons of the mean responses to these scales with the corresponding midpoint did not reveal any significant differences.] [Two direct replications (in an unreported fourth phase) using the same advertisement, product, brand names, and operational measures, as the Snyder and DeBono research, were also conducted. The results of these final two studies also failed to support the predictions of Snyder and DeBono (1985).]

PHASE I

Three hundred twenty-three undergraduates participated in this phase of the research. Subjects were randomly assigned to one of three studies. Sample sizes for studies 1-3 were 107, 108, and 108, respectively. On the basis of a median split, participants were categorized as either high or low self-monitors. The methods used to partition the subjects into high and low categories will be referred to hereafter as: (1) original, Snyder's (1974) original Self-Monitoring Scale; (2) revised, Lennox and Wolfe's (1984) Revised SM Scale; (3) the self-presentation factor, the ability-to-modify-self-presentation factor within the revised scale; and (4) combined, the overlapping groups formed by median splits using both the original and revised scales. The average internal consistency reliability estimates (either coefficient alpha or KR-20) were .57, .81, and .80 for the original, revised, and self-presentation factor scales, respectively.

Three covariate scaled measures -- self-esteem (Rosenberg 1965), gender and measures of product involvement (Zaichlowsky 1985) -- were also assessed. These measures were used as covariates in a series of parallel analyses to those conducted to replicate the findings of Snyder and DeBono (1985). Self-esteem and gender have been argued to be related to persuasibility. Likewise, product involvement has been shown to affect information processing and reactions to marketing appeals (e.g., Petty, Cacioppo, and Schumann 1983). Examination of the covariate mean scores reveals high levels of self-esteem across studies and greater product involvement for jeans in comparison to the running shoes. The average reliability estimates across the three studies were .85, .95, and .97 for the measures of self-esteem, jean involvement, and shoe involvement.

As part of a criticism of the self-monitoring theory, the revised SM scale of Lennox and Wolfe (1984) was developed to more closely reflect Snyder's (1974) original theoretical structure. The new measure was designed to assess only sensitivity to the expressive behavior of others and the ability to modify self-presentation. Consequently, it was anticipated that the original and revised measures would be only moderately correlated and that median splits formed by use of the scores from the two variables would result in somewhat different groups of subjects. An analysis of the overlap in subjects between the different median splits reveals just those anticipated results. For example, the average correlations across the three studies between the original SM scale and the revised scale was .30. The low/high sample sizes for studies 1-3 formed by selecting only those subjects below/above the median on both the original and revised SM scales were 35/50, 27/30, and 21/30.

Experimental Advertisements and Manipulation Check Results

The four advertisements to be evaluated by the subjects were professionally constructed. The two jean and two shoe ads differed only in the single verbal claim associated with each advertisement. For both products, the layout depicted wear appropriate for both males and females. The "quality" and "image' statements were developed from a series of pretest elicitations. The image claims stressed social approval and expressiveness; the quality claims centered on intrinsic product attributes (e.g., durability, comfort). For example, for the quality version, the following statement was used: "Navello (jeans), a comfortable fitting, five-pocket jean with extra stitching for durability." As a part of each study, manipulation checks were included after the experimental treatments and dependent variable measures (cf. Perdue and Summers 1986) to verify subject perceptions of the advertisements. In these tests, the definitions of image and quality claims used by Snyder and DeBono were provided. Subjects were required to rate each of the ads on 9-place bipolar scales labeled an image/quality appeal. Both the direction of the mean scores and the t-test estimates reflecting the significance of the differences (E2 < .01) between scores suggest that the ad claims were interpreted as intended. For example, the average t-value was 10.07 for the jean comparisons. The corresponding mean scores for the image claim (1) and the quality claim (9) were 2.08 and 6.67, respectively. Similar results were found for the shoe analysis. That is, the average, t-value was 9.95. The corresponding average mean scores were 2.79 and 7.31 for the image and quality treatments, respectively.

Study One

Method. Subjects were presented with two advertisements (i.e., a quality and an image claim) for a single product with the order of presentation randomly varied across subjects. (In the Snyder and DeBono study, each subject received a pair of advertisements for all three products.) The sample sizes for the jean and shoe exposures were 50 and 54, respectively. After each set, subjects responded to a 12-item survey similar to that employed by Snyder and DeBono that required evaluative comparisons between the two advertisements in each set. (The exact 12-item measure of Snyder and DeBono is used later in Phase IV.) Example items include: "Which ad appeals to you most?" "Which ad generates more interest?" An index of favorability was constructed by assigning 0 each time the quality version was preferred and 1 each time the image version was favored. The resulting KR-20 internal consistency estimates for the favorability measures were .83 and .89 for the jean and shoe responses.

Results. In the original Snyder and DeBono (1985) study, an ANOVA design was employed in which SM was a between subject factor and product represented a within-subject factor. In our study, the design was a simple one-factor (low versus high self-monitoring)design analyzed for each product. These results are depicted in Table 1 along with those of Studies 2 and 3 for each of the four median splits. None of the anticipated self-monitoring main effects were observed. This finding also held when gender, self-esteem, and product involvement were included as covariates. One interesting finding involving the covariates, however, was found in the shoe results. That is, product involvement and the favorability index-were negatively correlated (-.38). This suggests that, at least for one product, higher involvement is associated with preferences for the quality advertisement (e.g., those messages stressing product attributes).

Study Two

Method. This study involved a test of the hypothesis that low (high) self-monitors would be willing to pay more for products promoted via quality (image) appeals. Each subject was presented with either two image or two quality appeals; the order of presentation was counterbalanced as in the previous study. Following exposure to the advertisement, respondents were requested to indicate "How much would you be willing to pay for this?" As in the Snyder and DeBono (1985, p.590) study, two price ranges were given to avoid extreme responses. The price ranges were selected to reflect the range of prices typical of most brands in the two product categories.

TABLE 1

PHASE 1 RESULTS

Results. In these analyses, self-monitoring and type of advertisement were both treated as between-subject factors. The results of separate product 2x2 ANOVA analyses with and without the covariates did not reveal the anticipated significant self-monitoring by advertisement type interaction. While the predicted interaction failed to develop, a significant main effect for the type of advertisement was found in the shoe analyses. In both the original (F = 3.80, p < .05) and the revised (F = 4.51, p< .05) SM scale split analyses, the quality-based appeals resulted in higher estimates regarding the amount willing to pay. A similar significant main effect (F = 4.44, p < .05) was found for the self-presentation factor analysis. However, the results again do not support the conclusion that low versus high self-monitors vary in their reaction to image versus quality claims.

Study Three

Method. In the third study, subjects reacted to the same appeal/product configuration as the prior two studies. Each of 108 subjects received a quality message for one product and an image message for the other product. Again, the order of presentation was varied. The dependent variable (as in the Snyder and DeBono study) was an index of willingness to use the product. The index was constructed as the sum of standardized scores for two separate items. For both the shoe and jean results, the simple correlation between the two items was .78. The items were:

Which of the following verbal descriptions best describes your willingness to use the product advertised? (1) Definitely Not, (2) Probably Not, (3) Unsure, (4) Probably Yes, and (5) Definitely Yes.

With 0% indicating not at all and 100% indicating definitely willing, what percentage do you think best describes your willingness to try this new product?

Results. A significant 2x2 interaction was anticipated with high scores for the willingness index predicted for the low SM/quality and high SM/image combinations. The results of these analyses for both jeans and shoes across the various median split partitionings are summarized in Table 1. A significant predicted interaction was found for the original SM measure analysis for shoes (F = 6.6, p < 01). In this case, the high self-monitoring individuals did appear more willing to try the product following evaluation of the image claim while the low self-monitors reacted more favorably to the quality-based appeal. Similar findings were also found when the analyses were repeated including the covariates. The parallel interactions for jeans and for the remaining self-monitoring split interactions were not significant.

PHASE II

Given our inability to reproduce, with one exception, the consistent findings of Snyder and DeBono (1985), we reexamined our research methods. One major difference between the original study and this replication was the identification of subjects (in the studies reported here) as low and high self-monitors from questionnaire results obtained immediately prior to administration of the image and quality claims and assessment of the dependent measures. Consequently, the "price willing to pay" study was again conducted in an attempt to account for this difference in methods.

Method

Fifty-six male and female undergraduate business students were each presented with either two image or two quality ads for jeans and shoes. As in the prior experiments, the order of the product presentations was varied. Similar introductory wording and price ranges were employed. Two changes were made to the methods employed in this replication of Snyder and DeBono's Study 2. First, it was found in Phase I that the distribution of the price estimation scores was affected by the tendency of individuals to choose prices involving some multiple of five. Consequently, a unit listing of the possible prices within the "realistic" range was provided as a response scale. Second, data were collected regarding the self-monitoring measure two days prior to the evaluation of the advertisements.

Results

The internal consistency reliability estimates for the original 25-item SM scale, the revised 15-item scale, and the self-presentation factor were .69, .77, and .75, respectively. The correlations between the Snyder (1974) scale and the revised measure of Lennox and Wolfe (1984) was .37. The manipulation check results again suggested that the image and quality appeals were perceived as intended. The results of separate 2x2 ANOVA analyses (presented in Table 2) did not produce the significant interaction predicted by Snyder and DeBono (1985). Further, no main effects for either product were revealed.

PHASE III

The earlier studies by Snyder and DeBono (1985) considered three products along with the use of existing brand names. It was our intent to avoid any confounding of results via associations with existing brand names. Consequently, neutral and fictitious brands were selected for use in our efforts to examine the moderating effects of self-monitoring on reactions to quality versus image claims. This aspect of our methods differed from those of the Snyder and DeBono studies. In a further attempt to replicate their findings, a fifth study was conducted in which actual brand names (e.g., Reebok, Calvin Klein) were included in the advertisements. The study is similar to Study 1 (i.e., the favorability index study) of Phase I.

TABLE 2

PHASE II AND PHASE III RESULTS

Method

Subjects were presented with both a quality and an image claim for a single product with the order or presentation again varied across subjects. The sample sizes for the jean and shoe treatments were 51 and 56 subjects, respectively. The KR-20 reliability estimates for the favorability measures were .90 and .92 for the jean and shoe responses. The paired t-test comparisons for the manipulation check measures resulted in t-values of 11.39 (p < .05) and 18.90 (p < .05). These significant values and the direction of mean scores support earlier interpretations of the message content. The reliability estimates for the SM, revised and self-presentation factor scales were .59, .77, and .78. The pairwise correlation between the original and revised scales was .43.

Results

As in Study 1, the design was a simple one-factor (low versus high self-monitors) design analyzed for each product. The results are presented in Table 2. None of the predicted self-monitoring main effects were observed for any of the four comparisons. This finding also held when the three covariates were included. Though product involvement was significantly related to the favorability measure in Study 1, none of the covariates were found significant in this study.

DISCUSSION

Self-monitoring has provided a useful paradigm for evaluating the tendency of individuals to vary in their concern for the appropriateness of social behavior, their sensitivity to important cues, and their self-regulation of behavior. The concept has generated a number of studies and has been found to moderate the strength of attitude-behavioral relationships in both social psychology and consumer behavior research. Consequently, the construct does indeed offer promise for investigating the relative roles of situational and dispositional influences in consumer behavior. Yet, in spite of the recent study by Snyder and DeBono, very strong conclusions regarding the role of self-monitoring in predicting advertising effects (i.e., that high (low) self-monitors react more favorably to image (quality) appeals) seem premature. As with any effort to support or refute a theory or hypothesis, alternative explanations for unexpected results can arise. Possibly, some undetected differences existed between the original and replication research (e.g., student subjects, the presentation of stimulus materials). However, in this research. alternative explanations tied to research methods seem unlikely. Given the product familiarity pretest data used to select the stimuli for the first five studies and the frequent use of students as the subject population for test of self-monitoring theory, questions regarding the products and subjects would not seem to cast doubt on the present findings regarding self-monitorings' failure to explain advertising effects in the tests reported in this article.

A more plausible explanation involves a theoretical interpretation of high and low self-monitors' evaluation of image and quality appeals. That is, low self-monitors, being relatively insensitive to their social surroundings, may place some value on social interpretation offered directly by ads. In contrast, high self-monitors value social approval, but can assess it from their own social environment and so do not place much value on social appeals within ads. Thus, low and high self-monitors may respond in about the same manner (as found in our studies).

Limitations and Future Research

Self-monitoring has been treated here as a characteristic of the individual and was not manipulated experimentally. This is the most frequently reported use of the self-monitoring construct (i.e., differences between nonexperimentally formed groups are focused upon) and such designs are often used in studies of personality in consumer behavior (Kassarjian 1971). However, in spite of common practice, this procedure involves a correlational approach, even though it may be conducted in the laboratory with a high degree of control. Further, nonexperimentally manipulated variables may be correlated with demand proneness via confidence, suspiciousness, or other individual traits (Sawyer 1975).

Several issues warrant additional research. First, it appears that reactions to quality and image claims can vary, although not as predicted by self-monitoring theory. Some evidence was found that product involvement and preference for quality-based advertisements are related. This rather straight-forward result may warrant additional research and certainly has implications for the design of advertising strategy across consumer segments. And, some resolution is needed regarding the conceptualization of self-monitoring, its measurement, or both. Given the modest degree of overlap between groups identified by the original and revised scales, unresolved issues remain concerning construct definition and the identification of low and high self-monitors.

REFERENCES

Ajzen, Icek, Christine Timko, and John B. White (1982), "Self-Monitoring and the Attitude Behavior Relation," Journal of Personality and Social Psychology, 42 (3), 426-435.

Anderson, Paul F. (1983), "Marketing, Scientific Progress, and Scientific Method," Journal of Marketing, 47 (Fall), 18-31.

Becherer, Richard C. and Lawrence M. Richard (1978), "Self-Monitoring as a Moderating Variable in Consumer Behavior," Journal of Consumer Research, 5 (December), 159-162.

Belk, Russell W. (1974), "An Exploratory Assessment of Situational Effects in Buyer Behavior," Journal of Marketing Research, 11 (May), 156-163.

Briggs, Stephen R., Jonathan M. Cheek, and Arnold H. Buss (1980), "An Analysis of the Self Monitoring Scale," Journal of Personality and Social Psychology, 38 (4), 679-686.

Brinberg, David and Linda Plimpton (1986), "Self Monitoring and Product Conspicuousness on Reference Group Influence," in Advances in Consumer Research, Vol. 13, ed., R.J. Lutz, Provo, Utah: Association for Consumer Research, 297-800.

Caldwell, David P. and Charles A. O'Reilly, III (1982), "Boundary Spanning and Individual Performance: The impact of Self-Monitoring," Journal of Applied Psychology, 67 (1), 124-127.

Edell, Julie A. and Richard Staelin (1983), "The Information Processing of Pictures in Print Advertisements," Journal of Consumer Research, 10 (June), 45-61.

Holbrook, Morris, B (1978), "Beyond Attitude Structure: Toward the Informational Determinants of Attitude," Journal of Marketing Research, 15 (November), 545-556.

Kassarjian, Harold H. (1971), "Personality and Consumer Behavior: A Review," Journal of Marketing Research, 8 (November), 409-418.

Lennox, Richard D. and Raymond N. Wolfe (1984), "Revision of the Self-Monitoring Scale," Journal of Personality and Social Psychology, 46 (6), 1349-1364.

McCann, C. Douglas and Rodney D. Hancock (1983), "Self-Monitoring in Communicative Interactions: Social Cognitive Consequences of Goal Directed Message Modification," Journal of Experimental Social Psychology, 19, 109-121.

Nantel, Jacques and William Strahle (1986), "The Self Monitoring Concept A Consumer Behavior Perspective," in Advances in Consumer Research, Vol. 13, ed. R.J. Lutz, Provo, Utah: Association for Consumer Research, 83-87.

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

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.

Popper, Karl (1962), Conjectures and Refutations, New York: Harper.

Rosenberg, M. (1965), Society and the Adolescent Self-Image, Princeton, NJ: Princeton University Press.

Sawyer, Alan, G. (1975), "Demand Artifacts in Laboratory Experiments in Consumer Research," Journal of Consumer Research, 1 (March), 20-30.

Snyder, Mark (1974), "Self-Monitoring of Expressive Behavior," Journal of Personality and Social Psychology, 30 (4), 526-537.

Snyder, Mark and Kenneth G. Bono (1985), "Appeals to Image and Claims About Quality: Understanding the Psychology of Advertising," Journal of Personality and Social Psychology, 49 (3), 586-597.

Snyder, Mark and Thomas C. Monson (1975), "Persons, Situations, and the Control of Social Behavior," Journal of Personality and Social Psychology, 32 (4), 637-644.

Snyder, Mark and William B. Swann, Jr. (1976), "When Actions Reflect Attitudes: The Politics of Impression Management," Journal of Personality and Social Psychology, 34 (5), 1034-1042.

Wolfe, Raymond, Richard Lennox, and Richard Hudiburg (1983), "Self-Monitoring and Sex as Moderator Variables in the Statistical Explanation of Self-Reported Marijuana and Alcohol Use," Journal of Personality and Social Psychology, 44 . (5), 1069-1074.

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

Zanna, Mark P., James M. Olso, and Russell H. Pazio (1980), "Attitude-Behavior Consistency: An Individual Difference Perspective," Journal of Personality and Social Psychology, 38 (3), 432-440.

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