Fishbein's Subjective Norm: Theoretical Considerations and Empirical Evidence

Myron Glassman, University of South Carolina
Nancy Fitzhenry, South Carolina Commission of Alcohol and Drug Abuse
ABSTRACT - A number of questions about Fishbein's behavioral intention model were raised and in some cases empirical findings pertaining to the questions were discussed. The psychological significance of the regression coefficients, the conceptualization of the subjective norm and the specificity of the motivation to comply were dealt with.
[ to cite ]:
Myron Glassman and Nancy Fitzhenry (1976) ,"Fishbein's Subjective Norm: Theoretical Considerations and Empirical Evidence", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 477-480.

Advances in Consumer Research Volume 3, 1976      Pages 477-480

FISHBEIN'S SUBJECTIVE NORM: THEORETICAL CONSIDERATIONS AND EMPIRICAL EVIDENCE

Myron Glassman, University of South Carolina

Nancy Fitzhenry, South Carolina Commission of Alcohol and Drug Abuse

ABSTRACT -

A number of questions about Fishbein's behavioral intention model were raised and in some cases empirical findings pertaining to the questions were discussed. The psychological significance of the regression coefficients, the conceptualization of the subjective norm and the specificity of the motivation to comply were dealt with.

HISTORY OF THE MODEL

In recent years, Fishbein's behavioral intention model has stimulated a great deal of research. (For a review see Fishbein and Ajzen, 1975). While most of the research has been promising, there remain certain methodological and empirical questions about the normative component of the model. As originally formulated (Fishbein, 1967), the model contained two normative components, the personal normative component and the social normative component. Mathematically the model was expressed as:

B = BI = (AB)w0 + (NBP x MCp)w1 + (NBsMCs)w2   (1)

Where B = overt behavior; BI = behavioral intention; AB = the attitude toward performing the behavior; NBp = the respondent s personal feelings as to whether or not he should perform the behavior; MCp = his personal motivation to comply; NBs = the respondent's perception of the social norms pertaining to the performance of the behavior; MCs = his motivation to comply with the social norm; and w0, w1 and w2 = empirically determined weights. Since a high correlation was found between the respondent's personal normative belief (NBp) and behavioral intention (BI), the personal normative component was not included in subsequent research, and the social normative component was reformulated to include the respondent's perceptions of the expectations of significant others regarding his performance of the behavior (NBi) and his general motivation to comply with each significant other (MCg). Mathematically the model was represented as:

B = BI = (AB)w1 + (SNBiMCg)w2    (2)

Most recently, the normative component has been viewed in terms of a subjective norm (SN), the respondent's perception of what "others who are important to him" feel that he should do regarding the performance of the behavior. It should be noted that this formulation of the model means that SNBiMCg is no longer viewed as a primary determinant of intention. It has been replaced in the predictor equation by SN. Instead, it is now viewed as the determinant of SN. The implications of the change in the role of SNBiMCg will be discussed shortly.

THE RELATIVE IMPORTANCE OF AB AND SN

One finding of the studies dealing with Fishbein's behavioral intention model is that there has been a tendency for behavioral intention to be primarily determined by attitudinal factors (Fishbein, 1975, p. 311). There are two explanations for this finding. The first is that since the relative weights of the two components are expected to vary as a function of the topic and population under investigation, the slight tendency for attitudinal considerations to play a dominant role in determining intention is due to the fact that for the particular topics and populations studies, attitudinal considerations in fact were dominant. The second explanation is that the impact of normative beliefs in determining intention is felt in the attitudinal component. That is, one of the determinants of an individual's favorable attitude toward buying a particular product might be that he believes that Referent X will be pleased.

Another question pertaining to the relative importance of the normative and attitudinal component concerns the psychological meaningfulness of the weights. While the weights have been shown to vary across topics and populations, one might question whether these weights are meaningful or are merely "statistical artifacts." The results of studies pertaining to this question are inconclusive. Two of the studies (Ajzen and Fishbein, 1970 and Ajzen, 1971) used a Prisoner's Dilemma game. In both studies, some subjects were told to treat the other player as a partner (cooperative condition) while others were told to try to do better than the other person (competitive condition). It might be expected that in the cooperative condition normative factors would be dominant while in the competitive condition attitudinal factors would be dominant. As seen in Table 1, the results were encouraging.

TABLE 1

ATTITUDINAL AND NORMATIVE WEIGHTS IN PRISONER'S DILEMMA GAMES

Also of relevance is a study conducted by Glassman (1971). In this field study conducted among working class housewives, 127 respondents were asked questions pertaining to their beliefs, attitudes and intentions to buy eight convenience goods which had been selected because of their particular location on the Cohen and Barban (1970) product typology. Four of the products (two brands each of coffee and laundry detergent) were classified as ego-conspicuous products, i.e., they gave the consumer a sense of personal satisfaction (ego involved) and they were in some way visible to others (conspicuousness), while four of the products (two brands each of gasoline and potato chips) were classified physiological-inconspicuous, i.e., the products did nothing more for the consumer than the jobs for which they were designed (physiological) and the products and/or the results of their use were not visible to others (inconspicuousness). In addition, respondents completed the C.A.D. personality inventory (Cohen, 1967). This instrument classifies individuals according to their interpersonal orientation. Specifically, an individual can be viewed as moving toward people (compliant), against people (aggressive) or away from people (detached). Since perceived social conspicuousness has been called the most general attribute bearing on a product's susceptibility to reference group influence (Bourne, 1963), it was hypothesized that the intention to purchase those products that were classified as being ego-conspicuous would be primarily determined by normative factors while the intention to purchase those products that were classified as being physiological-inconspicuous would be primarily determined by attitudinal factors.

TABLE 2

REGRESSION COEFFICIENTS FOR EIGHT CONVENIENCE GOODS

It can be seen in Table 2 that irrespective of product classification, attitudinal considerations were the primary determinant of intentions. In fact, for three of the four ego-conspicuous products, the regression coefficient of the normative component was insignificant. However, it was encouraging to note that normative considerations played a significant role in deter- mining intention to purchase gasoline (physiological- inconspicuous). Specifically, subsequent analyses showed that the women felt that their husbands had very definite ideas as to which gasoline should be purchased and there was a strong motivation to comply with this referent. It was also felt that the extent to which a respondent was compliant, aggressive or detached, would influence the amount of importance that she placed on each of the components of the model. Specifically, it was hypothesized that for those women who classified themselves as being compliant, intention would be primarily a function of normative considerations while for those women who were aggressive or detached, attitudinal considerations would be primary. Since 70.6% of the women classified themselves as being primary compliant it was felt that a compliant versus aggressive and detached analysis would be inappropriate. Rather, the sample was divided into three groups on the basis of their compliance score. The results did not lend support for the hypothesis.

THE SUBJECTIVE NORM

It will be recalled that the behavioral intention model specifies SN, the generalized normative belief, as the normative predictor of intention. That is, SNBiMCg is no longer placed directly in the regression equation as a predictor of BI, but rather is viewed as the determinant of SN. The results of a study conducted by Fitzhenry and Glassman (1974) support the notion that SNBiMCg is the determinant of SN. One hundred and thirty females who were enrolled in an introductory psychology course at the University of Illinois during Fall 1974 answered questions pertaining to the various components of the behavioral intention model for (1) the following topics: taking birth control pills, donating blood, engaging in premarital sex, using amphetamines, buying a car, drawing up a will and using marijuana and (2) the following referents that were chosen based on responses to an elicitation question: parents and friends. Of relevance to the discussion is the form of the following questions:

SN:    Others who are important to me think I should +3:___:___:___:___:___:-3: I should not perform Behavior X

MCg:    In general I want to do 7:___:___:___:___:___:___:1: do not want to do what Referent A thinks I should do

NB:     Referent A would think that my performing Behavior X would be Good +3:___:___:___:___:___:-3: Bad

While the measure of NB is not i6 the traditional form, it is consistent with Fishbein's (1975, p. 315) conceptualization of the variable. As seem in Table 3, the results are consistent with Fishbein's theorizing. In a similar study by Glassman and Birchmore (1974), sixty females who were introductory psychology students at the University of Illinois during the Fall of 1972 were asked to fill out a questionnaire which dealt with (1) various components of the intention model for several contraceptive behaviors: use of the birth control pill, use of a diaphragm, use of the rhythm method, use of an I.U.D., asking their boyfriend/husband to obtain a vasectomy, obtaining contraceptive foam and asking their boyfriend/husband to obtain a condom and (2) the following elicited referents: physician, mother, father, friend, planned parenthood, priest, husband/ boyfriend and women's magazines. In addition, motivation to comply was measured at three levels: general (MCg), moderate (MCm) and specific (MCs).

TABLE 3

THE RELATIONSHIP BETWEEN SN AND SNBiMCg

Of relevance to this discussion is the wording of the following questions:

SN:    Others who are important to me think I should +3:___:___:___:___:___:-3: should not perform Behavior X

NB:     Referent X thinks that  I should +3:___:___:___:___:___:-3: I should not perform Behavior X

MCg:    In general I want to do 7:___:___:___:___:___:___:1: do not want to do what Referent X thinks I should do

MCm:    With respect to sexual behavior I want to do 7:___:___:___:___:___:___:1: do not want to do what Referent X thinks I should do

MCs:    With respect to matters of birth control I want to do 7:___:___:___:___:___:___:1: do not want to do what Referent X thinks I should do

TABLE 4

CORRELATIONS BETWEEN SN AND SNBiMC

As seen in Table 4, the results generally support the notion that SN is determined by the sum of the normative beliefs weighted by the general motivation to comply.

While Fishbein normally stresses the importance of specificity, he opts for a general motivation to comply. His argument (Fishbein, 1975, p. 306) is that if MC were measured at a specific level it would merely reflect the normative component's regression coefficient. However, French and Raven (1959), the same source that Fishbein uses as a foundation for the motivation to comply suggest that an individual's influence is likely to be topic specific. In fact, given that SNBiMC is no longer placed in the regression equation that predicts intention, the argument relating a specific motivation to comply (MCs) to the normative regression coefficient seems irrelevant.

In its new role, SNBiMC can be conceptualized as follows. The respondent, in arriving at his perception of the subjective norm, can be viewed as either consciously or unconsciously taking into account the desires of relevant referents and weighting each referent's desires by the relative importance of the referent's opinions. In this light, the determinants of SN can be viewed in terms of the respondent's perceptions of the desires of "important others" and a set of weights. Within this framework, measuring MC at a general level may be inappropriate in that at best MCg can be viewed as a measure of general importance, i.e., if a person generally tends to comply with an individual it can be assumed that the individual is generally important to that person. However, it is very likely that with respect to a specific behavior, this person's opinions may not be important. Similarly, while a person's opinions about a specific behavior may be important, in general they may not be.

While a specific motivation to comply may be more appropriate given the above framework, there is the possibility that questions using an MCs format will be responded to in terms of the respondent's intention and his perception of the desires of others. If the respondent intends to perform the behavior and views the referent as thinking that he should perform the behavior, he will be motivated to comply with the referent regarding the performance of the behavior. Likewise, if the respondent intends to perform the behavior and he feels that the referent does not want him to perform it, he will not be motivated to comply with the referent regarding the performance of the behavior.

It would seem that based on the previous discussion both MCg and MCs are inappropriate measures of the motivation to comply. Returning to Table 4, it appears that from an empirical point of view it makes little difference as to whether the motivation to comply is measured at a general, moderate or specific level.

Another issue pertaining to the motivation to comply is whether or not it should be used as a weight for SN. Again, Fishbein has argued that the motivation to comply with "important others," irrespective of the level of specificity, would merely reflect the regression coefficient of the normative component. Some evidence relating to this question comes from the Glassman and Birchmore (1974) study. Table 5 reports the average multiple regression coefficients for the seven intentions studied.

TABLE 5

AVERAGE MULTIPLE REGRESSION COEFFICIENTS

It can be seem that the addition of MC irrespective of the level of specificity does nothing to improve the predictability of the model. As an additional note, all multiple regression coefficients were significant except those dealing with the use of the diaphragm. These non-significant findings were due, in part, to the lack of variability in the intention to perform the behavior.

ADDITIONAL CONSIDERATIONS

Given the change in the role of SNBiMCg there is a question as to whether the traditional question used to elicit referents, "Where would you go for more information or advice concerning Behavior A?" is appropriate. Within the context of the previously mentioned framework, it would seem that the major interest would be in referents whose opinions concerning the performance of the behavior are important. It would seem best if respondents were now asked, "With respect to your performance of Behavior A, whose opinions are important to you?" While there could be a considerable correspondence between the results of the two elicitation questions it is likely that differences will occur. For example, while a girl may not consider going to her mother for information or advice about using the birth control pill because of possible fear or embarrassment, it is likely that she would know how her mother felt about her use of the pill, even if this information were based solely on inferences, and it is likely that she would consider her mother's opinions as being important. While the first elicitation question would not be expected to evoke the response "mother," the second one would.

A final point has to do with the way in which the motivation to comply is phrased. Usually respondents are asked whether they "want to do" or "do not want to do" what a particular referent thinks that person should do. If scored from +3 (want to comply) to -3 (do not want to comply) a spite model is assumed. That is, if the respondent feels that Referent A thinks he should not perform the behavior (NB = -3), and he is not motivated to comply with Referent A (MC = -3), Referent A will contribute positively toward his intention to perform the behavior (-3 x -3 = 9). If scored from +6 (want to comply) to 0 (do not want to comply), the fact that the respondent does not want to comply with referent A (MC = 0) who thinks that he should perform the behavior (NB = +3) implies that the wishes of Referent A make no contribution to the respondents intention to perform the behavior (+3 x 0 = 0). Research to date indicates these different methods of scoring have little influence on the regression weights and multiple correlation coefficient.

SUMMARY

In general, the findings support the most recent conceptualization Fishbein's behavioral intention model. There is evidence that the regression weights are psychologically meaningful; SN was shown to be a function of SNBiMC irrespective of the specificity of MC; and intention was shown to be predictable from AB and the subjective norm, irrespective of whether or not SN was weighted by the motivation to comply.

REFERENCES

Icek Ajzen, "Attitudinal vs. Normative Messages: An Investigation of the Differential Effects of Persuasive Communications on Behavior," Sociometry, 34(1971), 263-280.

Icek Ajzen and Martin Fishbein, "The Prediction of Behavior from Attitudinal and Normative Variables," Journal of Experimental Social Psychology, 6(1970), 466-487.

Frank S. Bourne, "Different Kinds of Decisions and Reference-group Influence,: in P. Bliss (ed.), Marketing and the Behavioral Sciences (Boston: Allyn and Bacon, 1963).

Joel B. Cohen, "An Interpersonal Orientation to the Study of Consumer Behavior," Journal of Marketing Research, 4(1967), 270-278.

Joel B. Cohen and Arnold M. Barban, "An Interactive Consumer-Product Typological System: A Progress Report and Partial Evaluation," Working Series in Marketing Research, College of Business Administration, Penn State University, 1970, Paper No. 12.

Martin Fishbein, "Attitudes and the Prediction of Behavior,'' in Martin Fishbein, ed. Readings in Attitude Theory and Measurement. New York: John Wiley and Sons, Inc., 1967, 477-492.

Martin Fishbein and Icek Ajzen, Belief, Attitude, Intention and Behavior, An Introduction to Theory and Research, (Reading, Mass.: Addison-Wesley, 1975).

Nancy Fitzhenry and Myron Glassman, "The Prediction of Intention," Unpublished Manuscript, University of Illinois, 1974.

John French and Bertram Raven, "The Bases of Social Power," in D. Cartwright, ed., Studies in Social Power, (Ann Arbor: University of Michigan Press, 1967).

Myron Glassman, "The Effects of Personality and Demographic Factors on the Formation of Attitudes Towards Convenience Goods," Unpublished masters thesis, University of Illinois, 1971.

Myron Glassman and Melinda Birchmore, "The Relationship Between Subjective Norms and Normative Beliefs," Unpublished Manuscript, University of Illinois, 1974.

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