Assessing Self-Concept Discrepancy in Consumer Behavior: the Joint Effect of Private Self-Consciousness and Self-Monitoring

Stephen J. Gould, Fairleigh Dickinson University
ABSTRACT - The measurement of self-concept in consumer behavior and marketing contexts has traditionally been plagued by problems in self-disclosure and self-awareness. This paper demonstrates how procedural aspects of self-concept, i.e. private self-consciousness and self-monitoring, affect the self-report of the contents of self-concept and related lifestyle items. As hypothesized, individuals, high in private self-consciousness and low in self-monitoring, report larger discrepancies between actual and ideal self-concepts than do others. Similar results are found when demographics are included as covariates. Finally, implications are drawn for consumer research.
[ to cite ]:
Stephen J. Gould (1993) ,"Assessing Self-Concept Discrepancy in Consumer Behavior: the Joint Effect of Private Self-Consciousness and Self-Monitoring", in NA - Advances in Consumer Research Volume 20, eds. Leigh McAlister and Michael L. Rothschild, Provo, UT : Association for Consumer Research, Pages: 419-424.

Advances in Consumer Research Volume 20, 1993      Pages 419-424

ASSESSING SELF-CONCEPT DISCREPANCY IN CONSUMER BEHAVIOR: THE JOINT EFFECT OF PRIVATE SELF-CONSCIOUSNESS AND SELF-MONITORING

Stephen J. Gould, Fairleigh Dickinson University

ABSTRACT -

The measurement of self-concept in consumer behavior and marketing contexts has traditionally been plagued by problems in self-disclosure and self-awareness. This paper demonstrates how procedural aspects of self-concept, i.e. private self-consciousness and self-monitoring, affect the self-report of the contents of self-concept and related lifestyle items. As hypothesized, individuals, high in private self-consciousness and low in self-monitoring, report larger discrepancies between actual and ideal self-concepts than do others. Similar results are found when demographics are included as covariates. Finally, implications are drawn for consumer research.

The measurement and assessment of self-concept is a central issue in consumer research, especially when considered in terms of the self-congruity between products and oneself and also in terms of the various components of one's multiple self, such as the ideal and actual self-concept (Sirgy 1982; Sirgy et al. 1991). How one sees oneself will be mirrored in what one buys (Malhotra 1988; Sirgy et al. 1991) and how one makes self-report responses in consumer and market research surveys (Sirgy 1982). Individual differences are likely to play a role in self-concept research as evidenced by Malhotra's (1988) study in which he found that cognitively complex individuals produced a better match between preference for houses and congruence with their self-concepts than did what he called cognitively simple individuals. Moreover, individual differences with respect to one's self-reported self-concept likely exist because of self-disclosure issues, which cause some consumers to be more troubled about revealing intimate generalizations concerning the self than others (Sirgy 1982). Thus, according to self- awareness (also sometimes referred to as self-focus) theory, which is concerned with the degree of focus on the self and which Sirgy mentioned as being in implicated in self-disclosure issues, some individuals may also feel uncomfortable confronting such topics concerning the self internally and may wish to avoid them or seek distractions (Duval and Wicklund 1972; Fenigstein, Scheier, and Buss 1975). Other individuals may not have clear ideas concerning their self or various aspects of it and/or may have difficulty in accessing them (Carver and Scheier 1981a).

In order to further investigate the relationship between self-reporting and self-concept, this paper will apply a theoretical approach which makes use of self-awareness theory and which has previously been applied in attitude-behavior consistency and expectancy research. This approach involves the self-relevant traits of private self-consciousness and self-monitoring, both of which are thought to affect attitude-behavior consistency (Miller and Grush 1986, 88). The rationale for using both these traits is that they allow us to account for the joint influence of both one's self-awareness of oneself and one's attunement to social norms on the self-reporting of one's self-concept.

Private self-consciousness, which is the trait equivalent of state self-awareness or self-focus, concerns the degree to which an individual tends to focus on his own inner thoughts or feelings (Fenigstein, Scheier, and Buss 1975). It is thought to affect attitude-behavior consistency in reflecting the degree that a person is able to access and know his/her own attitudes - a person higher in private self-consciousness being more likely to exhibit such consistency than others (Carver and Scheier 1981b; Miller and Grush 1986). In addition, when experimental inductions of self-awareness, such as facing a mirror, are applied to subjects, they have been found to provide more veridical self-reports than non-self-aware controls (Carver and Scheier 1981b). Similarly, Scheier (1980) reported that individuals, high in private self-consciousness, tended to display stronger relationships between privately held beliefs and publicly stated attitudes than did others.

Another possible factor influencing one's self-reported self-concept is self-monitoring which assesses the degree to which an individual tends to guide his self-presentation in response to social cues (Snyder 1974). Low self-monitors have been found to exhibit greater attitude-behavior consistency than have high self-monitors (Miller and Grush 1986). According to Miller and Grush (1986), low rather than high self-monitors will exhibit greater attitude-behavior consistency because they act more in accord with their own attitudes rather than with social norms. But high self-monitors would tend to show higher norm-behavior consistency.

Using both private self-consciousness and self-monitoring, Miller and Grush (1986) conducted a study of attitude-behavior consistency within the theory of reasoned action context (Ajzen and Fishbein 1980). They operationalized their joint use by creating median splits of the measures of each construct and combined them to form four cells: HH, HL, LH, and LL. As hypothesized, individuals high in private self-consciousness and low in self-monitoring (HL) exhibited more attitude-behavior consistency than did the other three groups in relation to an academic study task. However, the other three exhibited greater norm-behavior consistency than did the HLs.

SELF-CONCEPT, PRIVATE SELF-CONSCIOUSNESS AND SELF-MONITORING

Both private self-consciousness and self-monitoring have been conceived and studied with respect to self-report and self-disclosure. They both may viewed as procedural elements or schema of the self-concept which trigger or determine the contents of self-concept (e.g. "I am a person who possesses good self-control") that an individual will perceive, recall, and self-report (Ingram et al. 1988; Nasby and Kihlstrom 1986). Private self-consciousness reflects the degree to which one possesses the tendency to focus on one's own inner self. In related self-awaresness research, high self-focus individuals have been found to report larger perceived discrepancies between the actual and ideal self-concept (Duval and Wicklund 1972). These discrepancies, themselves, are thought to be indicants of more self-knowledge and to lead to more accurate self-report and also to greater self-report/self-concept consistency. Larger discrepancies also produce discomfort which is predicted by both the self-focus (awareness) and self-concept discrepancy theories (Wicklund and Hormuth 1981; Higgins, Klein, and Strauman 1985). Self-monitoring in contrast to private self-consciousness reflects response to social cues and norms. Individuals who are high self-monitors will monitor the self-presentation of others in order to find cues for their own self-presentation (Snyder, 1974). According to Snyder, such individuals are more likely to give socially desirable responses than low self-monitors because they possess the skills to find out what others expect of them.

Consistent with the research of Miller and Grush (1986) and also with the concept of a network of procedural and content elements of self-concept, it is likely that individuals will display differences in their assessment of their self-concepts with the high private self-conscious and low self-monitoring individual reporting the largest discrepancy. These individuals will tend more often to be able to access discrepancies and also to find larger ones than will others, while at the same time, not being as responsive or subject to external social cues, cues which often will dictate the suppression of discrepancy reporting. [A reviewer made the interesting observation that discrepancies may either be larger for high private self-conscious and low self-monitoring individuals than others or that they may be just less likely to be suppressed in self-report. This is a question not covered by this study, but I suspect that aspects of both may be at play, depending on the individuals (i.e. some individuals actually perceive larger discrepancies within themselves than others, but other individuals, given equally large discrepancies when compared to others, are less responsive to social norms, which are negative concerning the self-report of discrepancies, and thus, report larger discrepancies).] On the other hand, those low in private self-consciousness but high in self-monitoring will be just the opposite. Their responses will tend mainly to be guided by their monitoring of external cues. The other two groups of individuals, high private and high self-monitoring individuals, and low private and low self-monitoring individuals will fall somewhere in between. The former when faced with a self-report situation will confront self-relevant information which tends to conflict (i.e. internally focused self-consciousness cues versus externally-oriented self-monitoring cues). In their responses, they will tend to suppress extreme responses so as to conform with what they perceive as expected social norms. The latter group also tend to not give extreme responses. Being low in self-consciousness, they do not see much discrepancy on the one hand, but as low self-monitors, they also have less reason to change, repress, or hide what they do perceive on the other. Another way of describing this process is that individuals in this group tend to elaborate self-relevant information less than others, at least with respect to the internal and social interaction cues represented by self-consciousness and self-monitoring. Therefore, it is hypothesized that:

H1: High privately self-conscious and low self-monitoring consumers (HL) will display greater self-concept discrepancy than others while low privately self-conscious and high self-monitoring consumers (LH) will display the least self-concept discrepancy. The remaining two groups (HH and LL) will fall in between the first two.

METHOD

Sample and Procedure

The sample consisted of 337 adults in the Northeastern U.S. who returned questionnaires out of a potential quota of 340. The questionnaire was administered on a drop-off and pick-up basis by business school students for classroom credit. Each student was required to fulfill a quota of four adults: a male and female from 25 to 39 years of age and a male and female forty or older. Since some students did not fulfill their quota exactly, the final sample consisted of 51.3% females and 50.4% were 40 or older. Along with the completed questionnaires, the students handed in a separate list of respondents and their telephone numbers for validation. Approximately 20% of the respondents were telephoned for this purpose and asked whether they had answered the questionnaire and what they remembered from it. They all remembered something about the survey.

Measures

The measures assessed in this study and described in the following sections, include the two trait measures of private self-consciousness and self-monitoring, self-concept discrepancy measures, and demographic measures.

Private Self-Consciousness. This trait concerns the degree to which individuals focus on their own thoughts and feelings. It was assessed as part of the Self-Consciousness Scale and is represented by ten items (Fenigstein, Scheier and Buss 1975). A sample item is "I reflect about myself a lot."

Self-Monitoring. Self-monitoring is a trait, reflecting the degree to which a person monitors his or her self-presentation in accord with social cues. It was assessed using Snyder's (1974) scale in its refined and condensed eighteen item form (Snyder and Gangestad 1986). A sample item is "I would probably make a good actor."

Self-Concept Discrepancy Measures. The self-concept discrepancy measures consisted of thirty items which assessed how close or how far a person felt s/he was from his or her ideal state with respect to a variety of self-characteristics. The items were selected on an exploratory basis so as not only to reflect the inner psychological self-perceptions usually assessed in self-concept research (e.g. self-confidence) but also to reflect a broader range of self-relevant psychographic data, reflective of consumer lifestyles, from perceived physical well-being (e.g. health) to perceived socio-economic status (e.g. wealth) - see Table 1. This broader sample of self-concept items will provide more information in demonstrating how the proposed theoretical framework operates across various aspects of self-concept than would be the case were they excluded. At the same time, given this exploratory perspective, they do not represent so much a formal scale as much as a first attempt to construct one, and to serve the purpose of testing the hypothesized joint effect of private self-consciousness and self-monitoring over as wide a range of self-relevant material as possible. The items were measured on a 10 point scale with a "1" indicating that an individual felt his/her actual states to be "Very Close" to ideal (a small discrepancy) and a "10" indicating that s/he felt his/her actual states to be "Very Far" from ideal (a large discrepancy).

Demographics. Sex (0=female, 1=male), age, education (0 = "grade school or less" to 7 = "graduate degree") and income (0 = "less than $15,000" to 7 = "over 74,999") were also assessed.

RESULTS

Factor Analysis of the Self-Concept Measures

The thirty self-concept discrepancy items were reduced to eight factors, using varimax-rotated principal components analysis, and only including items which had loadings of .5 or better, as shown in Table 1. The items "strong character" and "masculinity if you are a man or femininity if you are a woman" and "intellectual stimulation" did not load on any factor and were dropped from further analysis. The resulting factors include: (1) "Physical Shape," (2) "Self-Efficacy," (3) "Age and Love," (4) "Wealth and Work," (5) "Excitement," (6) "Good Quality," (7) "Private Self," and (8) "Other." The items for each factor were added together using equal weights of 1 for each item which had a factor loading of .5 or greater and the resulting scores were used in further analysis. The data was handled in this way to reduce it to manageable proportions and the resulting factors are not meant to represent any more than exploratory scales as noted above. Nonetheless, all factors had coefficient alphas exceeding .5, the minimal level for exploratory reliability (Nunnally 1967), except for factor 8 which was dropped from further analysis C it also lacked face validity in combining height and amount of stress as a factor (see Table 1). The first three factors had alphas of .81, the fourth factor had an alpha of .80, the fifth, .60, the sixth, .56, and the seventh, .51.

TABLE 1

FACTOR ANALYSIS OF THE SELF-CONCEPT DISCREPANCE MEASURES

Multivariate Analysis of Variance Results

A four-group variable, PSCSM, was created by dividing Private Self-Consciousness (Md = 1.5) and Self-Monitoring (Md = 8.5) at their respective medians and merging the two as described by Miller and Grush (1986, 88). The two scores were also independent and uncorrelated (r = .04). The self-monitoring scale had a coefficient alpha of .69. For private self-consciousness, the short version reported by Gould (1986) and designated by him as "private reflective self-consciousness" was more reliable (coefficient alpha = .71) than the full version of the scale (coefficient alpha = .60), so it was used in all subsequent analyses. The following describes the four groups: (1) individuals low in self-consciousness and high in self-monitoring (LH), (2) those low on both variables (LL), (3) those high on both (HH), and (4) those high on private self-consciousness and@a¡O``Aa ¡O`The overall multivariate (MANOVA) effect of PSCSM for the seven factors was significant, F(21, 893) = 2.07, p<.0033. There were significant univariate effects for all but the last factor, p<.05, as shown in Table 2. In addition, planned comparisons revealed that the HLs exceeded LHs in the size of their discrepancy for all but the last factor, p<.05. Furthermore, for all factors, the other two groups fell in between though not necessarily at levels of statistical significance. For the psychological self-concept measures of Self-Efficacy and Excitement, the planned comparisons revealed that the HLs exceeded all three other groups. This result strongly supports the idea that individuals, high in private self-consciousness and low in self-monitoring, are more likely to perceive discrepancies between their actual and ideal selves than others, particularly those low in private self-consciousness and high in self-monitoring. Thusoverall, the hypothesis proposed for this study is supported. [An alternative MANOVA was also run with separate significant main effects for private self-consciousness, p<.0279, and self-monitoring, p<.0017, while the interaction was not significant, p<.6613. This result indicates that each construct had its own separate additive effect, thus supporting the Miller and Grush method which is based on their cumulative, joint effects.]

TABLE 2

MEANS, STANDARD DEVIATIONS AND UNIVARIATE F TESTS FOR THE SELF-CONCEPT DISCREPANCY VARIABLES

When sex, age, income, and education were included as covariates in a MANOCOVA analysis, the results changed slightly. In this case, the overall MANOVA effect of PSCSM was also significant, F(21,791) = 1.67, p < .0301. There were also significant univariate effects for all the lifestyle factors, p < .05. The significant covariates included sex on Physical Shape, p < .0005; sex and education on Self-Efficacy, p < .0022 and .0070, respectively; sex, age and income on Age and Love, p < .0409, .0001 and .0008, respectively; education and income on Wealth and Work, p < .0143 and .0042, respectively; sex on Excitement, p < .0015; income on Good Quality, p < .0001; and education on Private Self, p < .0301. Finally, zero order Pearson correlations, matching the independent variables with each other and with the dependent self-concept variables, reveal that some demographic variables may be markers for private self-consciousness and self-monitoring, as well as being related to and explaining some of the variance in the self-concept variables, themselves. Age especially is correlated (negatively) with both private self-consciousness and self-monitoring. As an example of a demographic-self-concept discrepancy relationship, a higher income is associated with a smaller discrepancy in both Wealth and Work and Good Quality.

DISCUSSION

The results of this study support the hypothesized relationship between self-concept actual-ideal contents, on the one hand, and private self-consciousness and self-monitoring, as the procedural elements of the self-concept network, on the other. Thus, they also support the approach of Miller and Grush (1986, 88). Specifically, it was found that the differences were clearly most pronounced for HLs (people high in private self-consciousness and low in self-monitoring) and LHs (people low in private self-consciousness but high self-monitoring) in that HLs consistently reported greater self-concept discrepancies than did LHs. Also as predicted, the two other groups, HHs and LLs generally tended to fall in between. Moreover, when the demographics were included in a MANOCOVA analysis, a self-consciousness/self-monitoring effect remained over and above them. In order to assess these results, a number of limitations and implications need to be discussed.

Limitations

Threats to validity exist because of the some of the procedures employed in this study. Since the sample was not random but rather a convenience, quota sample, it is possible that biases existed in the self-selection of respondents. It is also possible that some students administering questionnaires were delinquent in some way in this procedure although great effort was made to control and verify this study. However, even with these limitations, it was thought better to conduct this exploratory study on a broader, more inclusive demographic base than merely to use students as is often done in such studies. This enhances the external validity although the sample is still limited in terms of representativeness and randomness. Nonetheless, the results remain intriguing because they represent additional support for a theory which has previously been shown to be predictive and because they extend the theory into the area of self-concept and self-report consistency. Another limitation applies to what was considered in the study. Other measures could have been included which either assess other aspects or measures of self-concept, product image and self-concept or other related attitudinal or behavioral measures. These remain for future research.

TABLE 3

PEARSON CORRELATIONS FOR THE INDEPENDENT VARIABLES IN THE STUDY WITH THE SELF-CONCEPT DISCREPANCY VARIABLES AND WITH EACH OTHER

Interpretation and Implications of the Study

Private Self-Consciousness/Self-Monitoring and Self-Concept. The interpretation of these results revolve around questions of what the self-concept and discrepancies in it mean to different consumers. One question to be asked concerns whether personality types, based on the combined private self-consciousness/self-monitoring variable, vary in the size of their perceived self-concept discrepancy because their gaps between actual and ideal are really different, or because their degrees of self-monitoring cause them to report them differentially, even though in effect, their gaps are the same. In other words, does a large gap for HLs in some way mean the same thing in its cognitive structure or behavioral consequences as a smaller does one for LHs? If so it would mean that some sort of adjustment has to be made in our analysis of the self-concept discrepancy, which requires us to consider the four separate private self-consciousness/self-monitoring groups, and to look for changes or deviations in each group's self-concept when manipulating it experimentally or assessing it in a survey. On the other hand, it seems more likely, in conformity with self-monitoring and self-consciousness theory, that LHs felt it was socially desirable to report a smaller self-concept discrepancy, as well as finding a smaller discrepancy to begin with. Likewise, HLs would act in accord with their self-consciousness and monitoring traits and find a larger discrepancy. They would also find less reason in terms of social norms to obfuscate and reduce that discrepancy in the process of survey self-report. Finally, whatever the case for the meaning of self-concept and its discrepancies to various consumers, researchers need to consider that these differences exist and relate them to other dimensions of personality and a wide-range of attitudinal measures (cf. Miller and Grush 1986).

Self-Concept Scales. The self-concept items used in this study represent an attempt to move in the direction of broadening the range of items included in self-concept measures. More work is indicated in continuing in this direction, testing other dimensions not included here, and relating them to prior self-concept research.

Product Use. Beyond the potential improvement in assessing self-concept, it is possible (although not addressed here) that the combined private self-consciousness/self-monitoring assessment can aid in predicting the use of various types of products and services, themselves. HLs may actually seek out products or services that make them more self-aware or cause them to focus on themselves. For example, Fenigstein, Scheier, and Buss (1975) had originally related higher self-consciousness to the individual's tendency to participate in such activities as transactional analysis, encounter groups, sensitivity training, and meditation. LHs, on the other hand, may be especially sensitive to image-related advertising appeals and image-related products (Snyder and Debono 1985).

Demographic Relationships. Further research seems indicated for demographics in relation to private self-consciousness, self-monitoring, and self-concept. While the results found here should be interpreted with great caution because they were not predicted a priori, they nonetheless suggest that demographic differences may exist for various aspects of one's self-concept and also for self-consciousness and self-monitoring, which serve as processual aspects of the perceiving and schematizing of it. Consumer researchers should conduct field studies grounded in the theory of each of those demographics to verify and possibly extend these results, both between and within specific categories.

CONCLUSION

This paper has assessed the utility of Miller and Grush's (1986; 88) combining of self-monitoring and private self-consciousness as an independent variable in predicting self-concept discrepancy. Whether or not the demographic covariates were included, the results supported the basic hypothesis that HLs (LHs) will generally display greater (less) self-concept discrepancy than others, and demonstrated that the responses on self-reported self-concept measures are subject to individual differences based on the merger of these two variables. These results could be very important in interpreting the larger network of procedural and content elements of consumers' self-concepts. They are also suggestive of how psychological trait theories, often tested in laboratory experiments with students, apply to field research settings, where individual differences, based both on traits and demographics, may be found.

REFERENCES

Ajzen, Icek and Martin Fishbein (1980), Understanding Attitudes and Predicting Social Behavior, Englewood Cliffs, NJ: Prentice-Hall.

Carver, Charles S. and Michael F. Scheier (1981a), Attention and Self-regulation: A Control Theory Approach to Human Behavior. New York: Springer-Verlag.

Carver, Charles S. (1981b). "Self-consciousness and Reactance," Journal of Research in Personality, 15 (March), 16-29.

Duval, Shelley and Robert A. Wicklund (1972), A Theory of Objective Self-awareness. New York: Academic Press.

Fenigstein, Allan, Michael F. Scheier and Arnold H. Buss (1975), "Public and Private Self-consciousness: Assessment and Theory," Journal of Consulting and Clinical Psychology, 43 (August), 522-527.

Gould, Stephen (1986), "The Self-consciousness Scale: A Confirmatory Analysis," Psychological Reports, 59 (October), 809-810.

Higgins, E. Tory, Ruth Klein and Timothy Strauman (1985), "Self-concept Discrepancy Theory: A Psychological Model for Distinguishing among Different Aspects of Depression and Anxiety," Social Cognition, 3 (Spring), 51-76.

Ingram, Rick E., Debra Cruet, Brenda R. Johnson and Kathleen S. Wisnicki (1988), "Self-focused Attention, Gender, Gender Role, and Vulnerability to Negative Affect," Journal of Personality and Social Psychology, 55 (December), 967-978.

Malhotra, Naresh K. (1988), "Self Concept and Product Choice: An Integrated Perspective," Journal of Economic Psychology, 9 (March), 1-28.

Miller, Lynn E. and Joseph Grush (1986), "Individual Differences in Attitudinal Versus Normative Determination of Behavior," Journal of Experimental Social Psychology, 22 (May), 190-202.

Miller, Lynn E. (1988), "Improving Predictions in Expectancy Theory Research: Effects of Personality, Expectancies, and Norms," Academy of Management Journal, 31 (March), 107-122.

Nashby, W., and John F. Kihlstrom (1986), "Cognitive Assessment of Personality and Psychopathology," in R. E. Ingram (ed.), Information processing approaches to clinical psychology. Orlando, FL: Academic Press, 217-239.

Nunnally, Jum C. (1967), Psychometric Theory, New York: McGraw-Hill.

Scheier, Michael F. (1980), "Effects of Public and Private Self-Consciousness on the Public Expression of Personal Beliefs," Journal of Personality and Social Psychology, 44 (September), 627-644.

Sirgy, M. J. (1982), "Self-concept in Consumer Behavior: A Critical Review," Journal of Consumer Research, 9 (December), 287-300.

Sirgy, M. J., J. S. Johar, A. C. Samli and C. B. Claiborne (1991), "Self-Congruity Versus Functional Predictors of Consumer Behavior," Journal of the Academy of Marketing Science, 19 (Fall), 363-375.

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

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

Snyder, Mark and Steve Gangestad (1986), "On the Nature of Self-monitoring: Matters of Assessment, Matters of Validity," Journal of Personality and Social Psychology, 51 (July), 125-139.

Wicklund, Robert A. and Stefan Hormuth (1981), "On the Functions of the Self: A Reply to Hull and Levy," Journal of Personality and Social Psychology, 40 (June), 1029-1037.

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