Advances in Consumer Research Volume 2, 1975 Pages 249-264
THE PROCESS OF ATTITUDE ACQUISITION: THE VALUE OF A DEVELOPMENTAL, APPROACH TO CONSUMER ATTITUDE RESEARCH
Jerry C. Olson, Pennsylvania State University
Andrew A. Mitchell, Pennsylvania State University
[Preparation of this paper was partially supported by a grant to the first author from the Center for Research, College of Business Administration, Pennsylvania State University.]
[Both authors are Assistant Professors of Marketing, The Pennsylvania State University.]
Given the current controversies in attitude and communications research, it is suggested that substantial effort be focused on the process of attitude acquisition which is the basis for the phenomena of attitude structure and change. A model of the attitude acquisition process based on behavioral learning theory mechanisms is presented. The authors suggest that any attitude model (e.g., expectancy-value models) should be described in terms of the developmental processes involved in acquiring the model components. An example is presented of the developmental aspects underlying the Fishbein attitude model. The implications of this developmental perspective for attitude and communication research are discussed.
Over the last 40 or 50 years in psychology and at least over the last decade in marketing, immense efforts have been devoted to the study of attitudes and attitude-related phenomena such as attitude change. In particular, consumer researchers, beginning perhaps with Hansen (1969), have become enamored with attitudes, especially with attitude models of the expectancy-value variety.
Consumer behaviorists have used the concept of attitude for a variety of applied purposes, ranging from predicting consumer choice behavior (e.g., Bass & Talarzyk, 1972; Ginter, 1974), to serving as a basis for market segmentation (e.g., Mitchell, 1974; Myers & Alpert, 1968), to use as a criterion for advertising effectiveness (e.g., Lavidge & Steiner, 1961; Greene & Stock, 1966; Winter, 1973), in addition to its ubiquitous inclusion as an explanatory construct in theoretical models of consumer behavior (cf. Engel, Kollat, and Blackwell, 1973; Hansen, 1972; Howard & Sheth, 1969; Nicosia, 1966; Sheth & Park, 1973).
Given the widespread use of the attitude construct in consumer behavior research, it is surprising to find few (if any) explicit discussions of the antecedent conditions and dynamic processes underlying attitude formation. On the contrary, much attitude research is cross-sectional and treats attitude as given, a variable to be measured and related to other concepts such as behavioral intention, purchase, or information search. Those studies which have examined the attitude construct in an experimental and/or longitudinal fashion generally have not explicitly discussed the factors and processes which cause attitudes, but rather seem to suggest that attitudes rather "magically" occur upon, for example, exposure to an advertising communication or after some purchase/use experience.
The above-stated comments are not intended to suggest that previous research is without value. Clearly, some operational and conceptual advances have been made in attitude research (e.g., see Day, 1973; Wilkie & Pessemier, 1973). However, we do intend to suggest that ignoring or side-stepping the basic conceptual issues underlying attitude acquisition and the related topic of attitude structure generally has been dysfunctional and has led to several problems which are clearly evident in the consumer research literature, particularly in the areas of communications/advertising research and attitude models.
For example, research concerning the effect of communication on consumer attitudes and behavior seems to have been strongly influenced by the methodology and descriptive problem orientation of Hovland, Janis, and Kelley (1953). The typical study in this "Yale approach" can be described by the theme sentence, "Who says what to whom with what effect?". In most studies the dependent variable, "with what effect", was a self-reported measure of behavior or attitude change. The independent variable usually involved either the communication source ("who"), the structure or content of the communication ("what"), or certain characteristics of the communication recipient ("to whom"). Partially because of this atheoretical influence, few studies of communication/ advertising effectiveness or attitude change have attempted to examine the process of creating or changing attitudes and the factors involved therein. (See Wright  for a notable exception, although he does not specifically discuss attitude formation.)
In our opinion, this research approach has hindered the development of a real understanding of communication and advertising effects. For instance, much advertising today seems to communicate in a nonverbal manner. For example, many so-called image advertisements merely present a brand against a pictorial background--e.g., Marlboro or Salem cigarette ads--without any direct persuasive copy. Yet, communication and advertising researchers, to the best of our knowledge, have not evolved a theory that will explain such phenomena, and in fact have studiously ignored this topic.
Since attitude has long been a primary criterion measure of advertising effectiveness, we believe that a thorough conceptual examination of the processes and factors involved in attitude acquisition and-change may provide insights regarding the effects of "image advertisements" as well as the typically studied, more directly persuasive communications. Certainly, an understanding of the processes involved in communication effectiveness would be useful to marketing management and certain regulatory agencies.
The dysfunctional effects of ignoring such attitude acquisition phenomena are perhaps even more obvious in the burgeoning research literature on expectancy-value attitude models. This research contains numerous examples of nonconsistent use of terminology, imprecise conceptualizations, and miscellaneous controversies and confusions (e.g., witness the exchanges between Cohen, Fishbein, and Ahtola, 1972; Bass, 1972; Sheth, 1972, and Talarzyk, 1972). Often consumer researchers seem to treat the components of expectancy-value attitude models as given, frequently modifying them for operational ease, without considering the conceptual and theoretical logic upon which the models are based. A recent review (Wilkie & Pessemier, 1973) indicates the extent to which such a non-theoretical orientation results in a field becoming "bogged down" in predictive operational issues (e,g., should one include an "importance" score?), while ignoring the critically important theoretical basis for the model. Consequently, seldom if ever in the marketing literature have the basic theoretical concepts underlying attitude acquisition and change (and therefore the models themselves) been put to a rigorous test. Thus, the present massive body of attitude model data is built upon a weak, unvalidated conceptual base.
To summarize thus far, we contend that consumer researchers have ignored a critical phase in their research on attitudes and now need to "back-track" to that point and conduct several important studies. Specifically, we believe that consumer behaviorists should devote substantial research attention to the process by which attitudes are formed or acquired, a process which underlies the important phenomena of attitude structure and attitude change. Thus, when attitude models are used in our research, we should identify and validate the processes by which the components of that model are acquired by the consumer. Even if empirical validations are not conducted, we should at least force ourselves to logically describe in conceptual terms the acquisition processes involved in our attitude models.
Such theoretical and empirical attention to the attitude acquisition process could have a variety of benefits. For example, by forcing researchers to be more precise in their conceptual thinking, confusing and conflicting terminology would be clarified. Thus, our hypothetical concepts would be developed to fit the underlying theory rather than the researcher's problem area or purpose.
Another benefit derived from concentrating upon the attitude acquisition process would be to aid and encourage the formulation of an explicit, clear attitude construct and of a formal conceptualization of attitude structure. Such theoretical developments would in turn encourage researchers to validate these notions, the success of which would provide investigators with a firmer conceptual base.
Moreover, an explanation of attitude formation would be useful in developing an explanation of the attitude change process. With an explicit conceptualization of how attitudes change, communications research would be on firmer theoretical ground in the past, and perhaps findings of a generalizable nature could at last be developed. Also, with a theoretical base from which to work, advertising researchers may be more willing to study difficult phenomena such as image advertising or nonverbal communication.
Finally, an explication of the process of attitude acquisition would provide a firmer and less ambiguous relationship between information processing phenomena and attitude structure. Indeed, it seems clear that attitude development occurs through information processing. What researchers need now is a description, in theoretical terms, of the Process by which one incorporates information and acquires an attitude.
In order to stimulate research and hopefully begin to achieve some of the above-stated benefits, the following section presents a conceptual model of the attitude acquisition process based upon behavioristic learning theory. This model was chosen for its parsimony, the precision of its concepts, the amount of past research it has generated (especially in the psychology literature), and simply because it is the only model of which we are aware. Although we have made several minor conceptual modifications to this model of attitude acquisition, essentially the following discussion is a summary of the work of others (as referenced).
THE PROCESS OF ATTITUDE DEVELOPMENT, FORMATION, OR ACQUISITION
Although seemingly ignored by most marketing researchers and social psychologists, a substantial body of conceptual and empirical work in psychology has addressed the issue of attitude [In this paper, attitude is considered to be a unidimensional evaluative construct, equivalent to affect (liking) towards, or an evaluative judgment (good-bad) of, a stimulus object. This view is consistent with most of the behavioristically oriented researchers who are concerned with attitude formation, for example, Doob (1947), Eisman (1596), Staats & Staats (1958), and Fishbein (1967). However, it should be noted that the tripartate conceptualization of attitude, with affective, cognitive, and conative components, has numerous apostles and certain unique advantages (cf. Triandis, 1971) as well as disadvantages, particularly operational problems (Fishbein, 1967).] acquisition in terms more specific than the typical lip-service that "attitudes are learned through experience." The usual theoretical perspective of this research is drawn from behavioristic learning theory [Van De Geer and Jaspars (1966) have termed this theoretical perspective the neo-behavioristic approach to cognition.] and, consequently, relies heavily upon the basic learning mechanisms of classical and instrumental or operant conditioning (as well as the more complex mechanisms of generalization) for explanations of the attitudinal acquisition process.
Doob (1947), who was perhaps the first to apply learning theory principles to the attitude formation process, considered attitude to be ". . .(1) an implicit response, (2) which is both (a) anticipatory and (b) mediating in reference to patterns of overt responses, (3) which is evoked (a) by a variety of stimulus patterns (b) as a result of previous learning or of gradients of generalization and discrimination, (4) which is itself cue- and drive-producing, (5) and which is considered socially significant in the individual's society." Generally, later theorists maintained the essence of Doob's conceptualization while frequently emphasizing, after Allport (1935), that the implicit (i.e., internal) attitudinal response is of an evaluative nature (cf. Rhine, 1958; Staats & Staats, 1958). Our interest, of course, is with the process by which a subject acquires that implicit evaluative response called attitude.
Before examining the specific learning processes by which one acquires a particular attitude toward a stimulus object, it should be noted that the conceptual perspective of the present paper is grounded in the theoretical positions of Osgood (e.g., Osgood, Suci, & Tannenbaum, 1957) and Fishbein (cf. 1967), among others, in that all stimulus objects are considered to have associated with them two implicit, mediating responses (see Figure 1). That is, a stimulus object elicits two internal responses from a subject. One of these mediating responses represents the identification or categorization of the stimulus object and is often termed a concept-labeling response, while the other internal response is the attitudinal, affective or evaluative response. This attitudinal response may be positive or negative in varying degrees of intensity or, in the case of novel, unfamiliar stimuli, the evaluative response may be neutral. The critical aspect of this theoretical notion is that the elicitation of the evaluative attitude response (either positive, negative, or neutral) is "automatic" upon exposure to and identification/categorization of the stimulus object (see Fishbein, 1967, for a more detailed presentation and justification of this position).
A STIMULUS OBJECT ELICITS TWO IMPLICIT RESPONSES, AN EVALUATIVE RESPONSE WITH ASSOCIATED STIMULUS PROPERTIES (re --s) AND A LABELING RESPONSE WITH CUE PROPERTIES (rso --s), BOTH OF WHICH BECOME CONDITIONED TO, AND TEND TO ELICIT, ONE ANOTHER.
To summarize, note in Figure 1 that a stimulus object elicits two internal, mediating responses, one a concept-labeling response and the other an evaluative, attitude response. Furthermore, over time, through classical conditioning, the labeling and evaluative responses become conditioned to, and are elicited by, one another. Note, however, that this theoretical position merely states that persons have attitudes (in the form of evaluative responses) to all discriminable stimuli, but does not account for the acquisition of a specific attitude -- i.e., one with a particular degree of positive or negative direction. Before discussing that acquisition process, let us first examine the processes by which one may acquire the concept-labeling response to a stimulus object.
Concept Formation Through Direct Conditioning
Before an attitudinal evaluative response to a stimulus object can be acquired, one must first be able to discriminate, identify, and categorize that stimulus. These latter abilities are developed as one comes to recognize that a set of specific stimuli "belong together" or form a discriminable stimulus pattern, and thus are given a specific meaning, [Note that this process of assigning meaning to stimulus patterns is essentially equivalent to the formation of chunks in information processing (cf. Newell & Simon, 1972; Olson, 1974, p. 34).] typically in the form of an internal labeling response. Such learning has been termed concept formation (cf. Rhine, 1958). For instance, in our culture a stimulus object with four legs, a back, and a seat is usually recognized as a chair while a four-legged, backless object with a seat is commonly labeled a stool. The process of learning such concepts (i.e., attributing meaning to stimuli) takes place over a period of time and is a function of both classical conditioning and reinforcement contingencies. For instance, parents may continually point out and verbally identify objects for their children and, moreover, may provide positive reinforcement when the child demonstrates concept learning by making the correct overt, verbal labeling response. Thus, as depicted in Figure 2, first-order concept formation may be conceptualized as the acquisition of an implicit response to a stimulus or set of stimuli (cf. Osgood, 1952). The response is typically considered to be an internal, labeling response with associated cue (stimulus) properties which may in turn mediate (or guide) other internal or overt responses.
FIRST-ORDER CONCEPT FORMATION OCCURS WHEN A SPECIFIC PATTERN OF STIMULI (S1, S2, AND S3) COME TO ELICIT, THROUGH CLASSICAL AND/OR INSTRUMENTAL CONDITIONING, AN IMPLICIT LABELING RESPONSE (rc) WHICH POSSESSES CERTAIN CUE PROPERTIES (s), AND WHICH MAY MEDIATE OTHER RESPONSES.
Concept Formation Through Mediated Generalization
It is perhaps more realistic conceptually, especially for multi-attribute brands and other complex marketing stimuli, to discuss a somewhat more complicated concept formation process based upon the principle of mediated or secondary generalization (e.g., Cofer & Foley, 1942; Eisman, 1955). This more involved type of concept acquisition is essentially analogous to direct conditioning. However, instead of external, physical stimuli coming to elicit an internal, labeling response, in this case the concept labeling response is a generalization of previously learned, internal, concept responses. Since the generalized, higher-order labeling response is a function of other internal, mediating responses, the process is termed secondary or mediated generalization.
To illustrate secondary generalization of a concept labeling response, consider in Figure 3 a relatively complex stimulus object such as a product brand "X" which possesses three discriminable attributes A, B, and C (e.g., price of $4.98, red color, large package size). Each of these three attributes is, in fact, an array of stimuli which, through previous conditioning, has come to elicit its own concept labeling response and, of course, its own implicit evaluative response (see Figure 2). If the three attributes A, B, and C are frequently experienced together, i.e., contiguously as they would be if they were stable characteristics of a brand, the three labeling responses and their respective stimuli (i.e., ra --s, rb --s, and rc --s) may come to be associated with and thus elicit a secondary, generalized, internal labeling response (rx --s) which represents the meaning of the overall stimulus object, brand X. Usually this internal labeling response will be the brand name. Of course, concept formation through mediated generalization is facilitated by both classical conditioning (frequent exposure to brand X) and reinforcement contingencies (rewards or punishments for overt, verbal labeling responses or other overt behaviors).
FORMATION OF A SECOND-ORDER CONCEPT RESPONSE (rx --s) TO A STIMULUS OBJECT BY SECONDARY, MEDIATED GENERALIZATION OF THE SET OF FIRST-ORDER LABELING RESPONSES (ra --s, ETC.) ASSOCIATED WITH THE ATTRIBUTES OF THE STIMULUS OBJECT (X).
Acquisition of the Evaluative, Attitudinal Response
Now let us turn our attention to the process by which one acquires a specific attitude response to a stimulus object (i.e., an evaluative response with a particular degree of positive or negative intensity). The attitude acquisition process may occur in essentially two ways: (a) through direct conditioning of the evaluative response by classical or reinforcement conditioning mechanisms, or (b) through more indirect conditioning of the evaluation response by mediated generalization mechanisms, in addition to classical and instrumental conditioning influences.
Attitude Acquisition Through Direct Conditioning. A particular evaluative response to a stimulus object may be acquired directly through classical conditioning mechanisms (cf. Staats & Staats, 1958). For example, a positive attitude toward brand "X" may be established by repeatedly and contiguously presenting a subject with brand "X" and another stimulus which elicits the desired evaluative response either innately (e.g., food) or through previous conditioning (e.g., colorful sunsets). Conversely, a negative evaluative response could be conditioned to brand "X" by repeatedly pairing it with a stimulus which elicits the desired negative response (e.g., electric shock [Bandura & Rosenthal, 1966]).
Alternatively, attitude responses may also be acquired through reinforcement mechanisms. For example, by rewarding overt, verbal evaluations of a positive nature (e.g., Eisman, 1955), a positive, internal, evaluative response may be conditioned.
Attitude Acquisition Through Mediated Generalization. For multi-attribute stimuli of major interest to consumer researchers, the usual process by which a particular evaluative response is acquired may be considered to be "higher-order" conditioning based upon the principle of mediated generalization, in addition to classical and operant conditioning mechanisms. To illustrate that process, consider the acquisition of a specific attitude (of a particular direction and intensity) towards a Product brand, X (see Figure 4).
Suppose that brand X has three attributes or characteristics (A, B, and C) associated with it, each of which elicits its own unique, previously learned labeling and evaluative responses (e.g., r --s and r --s). The reader may recall from the discussion above, that a second-order, concept labeling response (rx --s) to brand x is acquired through the process of secondary, mediated generalization (and perhaps also classical and reinforcement conditioning). By a similar process of mediated generalization, an overall evaluative response may be acquired towards brand X. That is, the evaluative attitudinal response to brand X (rex --s) is generalized from, or derived from, the various mediating evaluative responses associated with the attributes comprising brand X (see Figure 4). Thus, through the mechanisms of mediated generalization, and perhaps classical and reinforcement conditioning, as well, a consumer acquires a particular evaluative response. an attitude towards brand X.
BRAND X HAS THREE CHARACTERISTICS OR ATTRIBUTES (A, B, AND C), EACH WITH PREVIOUSLY-ACQUIRED LABELING AND EVALUATIVE RESPONSES. THE OVERALL EVALUATIVE RESPONSE TO BRAND X (rex--s) IS A FUNCTION, THROUGH GENERALIZATION, OF THE COMBINATION OF THE SEPARATE EVALUATIVE RESPONSES TO THE ATTRIBUTES ASSOCIATED WITH BRAND X. MOREOVER, THROUGH EXPOSURE TO AND EXPERIENCES WITH BRAND X, (I.E., CLASSICAL AND INSTRUMENTAL CONDITIONING, RESPECTIVELY), BRAND X MAY COME TO ELICIT THE GENERALIZED OVERALL EVALUATIVE AND LABELING RESPONSES DIRECTLY.
The notion that the overall, evaluative, attitudinal response to a stimulus object is a generalization from specific evaluative responses to the attributes (or concepts) which are associated with that stimulus object is a basic theoretical position of many attitude researchers (e.g., Osgood & Tannenbaum, 1955; Peak, 1955; Rosenberg, 1956; Fishbein, 1963), although some do not explicitly recognize this basic assumption (e.g., Thurstone, 1928). Not only does this notion make good logical sense, but it has also received indirect empirical validation (e.g., Fishbein and Hunter, 1964), although the specific form of the mediated generalization (i.e., the combination of evaluative responses) is not agreed upon (see Anderson & Fishbein, 1965; for a "test" of two alternative combination forms).
However, irrespective of the generalization/combination procedure favored (additive, averaging, differentially weighted, etc.), we believe that it is extremely important, conceptually and operationally, to remind ourselves of the basic theoretical premise that the formation of a specific attitude toward a complex stimulus object is a function of the combined evaluative responses to the attributes of that stimulus object.
Evidence for Validity of the Attitude Acquisition Process
There is considerable evidence in the psychological literature which supports the attitude acquisition process described above. Staats and Staats demonstrated in several studies that attitudes may be acquired through higher-ordered classical conditioning, that is, by pairing the attitude object with a stimulus previously conditioned to elicit a particular evaluative response (Staats & Staats, 1957, 1958, 1959; Staats, Staats, & Biggs, 1958; see also Staats, 1967). Furthermore, Staats, Staats, & Crawford (1962) successfully demonstrated first-order classical conditioning of a negative attitude by pairing words (the CS) with shocks or loud noises (the UCS), as has Maltzman (1968). Evidence also exists to support an instrumental conditioning explanation of attitude acquisition (e.g., Eisman, 1955; Scott, 1957, 1958), although Scott's results have been questioned (Dahlke, 1963; Kiesler, 1965).
It must be noted that the experimental designs used in the typical attitude conditioning study have been criticized as containing inherent demand characteristics (Orne, 1962) which can provide an alternative explanation for the results (cf. Kiesler, Collins, & Miller, 1969). However, recent work has minimized the demand characteristics of the experimental procedures and has still obtained a conditioned evaluative response (e.g., Zanna, Kiesler, & Pilkonis, 1970).
Finally, the impression formation literature (see review by Tagiuri, 1968) provides additional evidence to support the attitude acquisition process presented above. In the typical experiment, subjects receive information--often in the form of adjective traits (which elicit certain positive or negative evaluative responses)--about a hypothetical person (initially neutrally evaluated) and then record their impressions (evaluation) of that person. As Chambers (1969) has noted, impression formation is very similar to higher-ordered classical conditioning and has yielded similarly successful results (cf. Anderson & Hubert, 1963). Moreover, research in impression formation has also examined repetition and order effects, variables generally ignored by the neo-behaviorist attitude researchers but factors high's relevant to advertising (e.R.. Chambers. 1971; Shapiro and Tagiuri, 1958).
In summary, the above-cited studies indicate that specific attitudes can be acquired and changed through the mechanisms of classical and operant conditioning and mediated generalization. Thus, the basic principles of the conceptual attitude acquisition model presented above have been shown to have some validity.
APPLICATIONS AND IMPLICATIONS OF A PROCESS APPROACH TO ATTITUDE DEVELOPMENT
In the absence of an alternative theoretical explanation of the process of attitude acquisition which possesses the precision and comprehensiveness of the learning approach presented above, we suggest that this behavioristic theory be used--very explicitly--in future attitude and communication research by consumer behaviorists. The remainder of this paper briefly describes some of the ways in which an explicit consideration of the attitude acquisition process can improve consumer attitude research.
Development and Validation of Attitude Models
It is, of course, desirable that consumer researchers continually evolve new models and develop modifications of existing models of attitude structure. However, we contend that scientific progress in our discipline is retarded if such "newcomers" are not subjected to certain standards. As we suggested at the beginning of this paper, a minimal "standard" should be a conceptual (and logical) description of how the components of a proposed model could have developed and been acquired by a consumer. Moreover, given the theoretical perspective on attitude acquisition presented here and elsewhere (cf. Fishbein, 1967), one might expect developers of new models to compare and contrast their attitude models with earlier work. Ideally, such inter-model comparisons would involve not only empirical (predictive) differences, but also would identify conceptual distinctions.
A comparative orientation to attitude model research will, in our opinion, limit the current proliferation of attitude models to those whose-components and developmental processes are clearly different--conceptually--from those presented here. Rigorous comparison may reveal several currently-used models to be merely the "old" model with new or modified terminology, but essentially the same concepts. Moreover, a rigorous comparison tradition may find that certain "attitude models" are not actually models of attitude, but rather are descriptions of some other construct. The whole process of positioning one's hypothetical constructs vis-a-vis previously postulated constructs is a scientifically healthy practice and could be applied profitably to areas of consumer research other than attitude formation.
An example. One currently popular and occasionally controversial attitude model which can stand up well under the scrutiny advocated above is that of Fishbein (e.g., 1961, 1963), a basic expectancy-value attitude model similar to those of other researchers (Peak, 1955; Rosenberg, 1956; Vroom, 1964). Fishbein's attitude model is, in fact, completely consistent with the acquisition process presented above, since its development seems to have been influenced by the neo-behavioristic, learning theory approach to attitude formation (e.g., Staats & Staats, 1957, 1958; Rhine, 1958). Since Fishbein has presented the developmental processes underlying his model quite clearly (especially see Fishbein, 1967) and because most consumer researchers are familiar with that model, the following only summarizes his rationale.
Fishbein (cf. 1967) considered attitude to be a unidimensional evaluation, thereby removing the cognitive dimension from the attitude construct. Instead, he conceptualized a belief construct which is separate from, but related to, the evaluative concept and which incorporates cognitive aspects of the stimulus object. To Fishbein, a belief is the subjective perception of association between the belief object and some other object, attribute, or concept, and thus beliefs are often operationalized as the rated "probability" or "improbability" that a relationship exists between the object of belief and the other concept (Fishbein & Raven, 1962).
The notion of belief fits into the attitude acquisition process in the following way. Given that the belief object is a stimulus and related objects, attributes, or concepts are represented psychologically as internal, mediating, labeling responses (see Figure 3), a belief may be considered as an S-r association. From this perspective a belief will be stronger (i.e., the subjective probability of relationship will increase) as the S-r association becomes stronger between a belief object and its related attributes; for example, as a function of the number of contiguous S-r experiences or with an increase in the magnitude or number of rewarding stimuli following a S-r association.
Clearly, the beliefs one has regarding a stimulus object may be hierarchically arranged according to strength to form a belief-family-hierarchy (see Fishbein, 1963). Thus, there are at least two ways of operationalizing strength of beliefs: (a) self-reported probability of association between the belief object and another concept, and (b) order in the belief-family-hierarchy (perhaps measured by order of free association responses). It should be emphasized that Fishbein includes only "salient" beliefs in his attitude model-i.e., those beliefs high in the hierarchy, the strongest beliefs.
Consistent with the conceptualization of the attitude formation process discussed in the present paper, Fishbein's model specifies that the overall evaluative, attitudinal response to a stimulus object is a function of the various evaluative responses to the concepts or attributes believed associated with the object. Moreover, the Fishbein model explicitly states that the extent to which single evaluative responses to attributes of a stimulus object affect the overall attitude toward the stimulus object is weighted by the strength of the belief regarding the relationship between the stimulus object and that attribute. Fishbein precisely specifies the form of this mediated generalization process by proposing that the overall evaluative response to a stimulus object is determined by summing the product of the strength of belief and the intensity of the evaluative response to each attribute associated with the stimulus object. Mathematically, his model is:
"o = the attitude toward object o
Bi = the strength of belief i about o, that is, the "probability" or "improbability" that o is associated with some concept xi
ai = the evaluative aspect of Bi, that is, the evaluation of xi
n = the number of beliefs about o, that is, the number of salient beliefs in the individual's belief-family-hierarchy
To our knowledge, no other attitude model has been so clearly described in terms of the processes involved in acquiring the components of the model. We reiterate our earlier suggestion: An explicit conceptualization of the attitude acquisition process would be useful, especially heuristically, for all attitude researchers.
Implications for Attitude Modeling
As one illustration of the heuristic implications for attitude modeling of a developmental perspective, note that an explicit theoretical explanation for the acquisition of attitude model components clearly reveals the origins and causal factors underlying the model. From such a perspective it is relatively easy to develop experimental manipulations of model components and thus create experimental tests of the model. For example, it is now obvious how both the belief and evaluative-aspect-of-belief components in Fishbein's model might be manipulated in an experimental test of the model's predictive power. Such a perspective is obvious only because we now have a conceptual paradigm which explains the developmental processes underlying the model.
Implications for Communications Research
When one considers an expectancy-value model such as Fishbein's from a developmental process perspective, certain interesting and theoretically important communication phenomena become more obvious. In fact, combining the developmental processes and the basic Fishbein model, for example, may form the basis for a theoretical explanation of how communications affect attitudes. For instance, exposure(s) to a communication which simply presents a stimulus object-attribute pair could create (through classical conditioning) a belief that the particular attribute is associated with the stimulus object. As this belief becomes stronger the evaluative aspect of the belief would have an increasing influence on the overall attitude toward the stimulus object. Thus, conditioning this belief to the stimulus object, through communication messages in this case, would cause an increase or decrease in the favorability of the overall attitude toward the object (depending upon the evaluative aspect of the belief). Note that a similar change in overall attitudinal favorability could be affected by increasing or decreasing the evaluative aspect of an already-conditioned belief. Research on such phenomena would not only benefit consumer behaviorists interested in attitude change phenomena, but should also stimulate communications researchers who seem to lack specific criterion measures for the effects of their independent variables.
Moreover, the attitude acquisition process presented above encourages the examination of communications such as image advertising. For example, we know that if two stimuli are presented contiguously over a period of time one will acquire an S-R association between the two, and we know that a belief is a type of verbalized S-R association. Therefore, even non-verbal, image advertising may create beliefs. Such beliefs can be measured and the effect of their acquisition on overall attitude can be empirically validated, given the concepts and operational procedures specified by Fishbein, for example (cf. Fishbein & Raven, 1964).
Need for Longitudinal Research Designs
Another consideration prompted by the developmental perspective to attitude acquisition, is the clear need for longitudinal research designs. In order to test many of the notions discussed above, most of which are entirely within subject phenomena--that is, the processes occur within an individual subject-we need multiple measures of an individual's beliefs and the evaluative aspects of those beliefs at several points in time, as a function, for example, of exposure to a series of communications. Only with a longitudinal design can changes in the determinants of attitude and subsequent attitude change be adequately inferred.
Finally, it is of distinct interest to note that the data from a set of experimental manipulations of model components, in a longitudinal within-Ss design, would provide a more direct test of attitude models than the cross-sectional studies typically conducted. Moreover, if manipulations of the components of competing models (e,g., ideal-point model vs. the Fishbein model) be included in the same study, one would have the makings of a critical test (cf. Platt, 1964) which could provide relatively unambiguous information regarding the validity and predictive value of each model. Hopefully, a program of such comparative research would provide a better understanding of attitude change processes and influences as well as a set of "better-validated" models of attitude structure. The present authors are currently conducting such a programmatic series of studies.
In this paper, we have argued for an examination of the processes underlying acquisition of a specific attitude. The model presented and advocated here is derived from the behavioristic learning theory approach to attitude formation (see Doob, 1947; Fishbein, 1967; Staats, 1967). This model makes the following basic points. (a) The hypothetical construct of attitude is most usefully considered as unidimensional evaluation or affect. (b) Any stimulus object automatically elicits two implicit responses, (i) a concept labeling response and (2) an evaluative response, both of which possess stimulus properties capable of eliciting other implicit or overt responses. (c) These implicit responses may be acquired either through the mechanisms of classical conditioning, instrumental conditioning, or mediated generalization, or some combination thereof.
We believe that a close examination of the attitude construct from a development process perspective, such as that provided by the present model, will enrich consumer attitude research, in the following ways. (a) Researchers will be forced to be more precise, conceptually and operationally, in their use of attitude and attitude-related constructs. (b) The process approach will encourage the development of models of attitude structure, such as expectancy-value models, which are theoretically consistent with these acquisition concepts (or with some other conceptualization of attitude formation). (c) Theoretically-based attitude models encourage researchers to more directly test the basic notions (now explicitly stated) which underlie the model, thus yielding validated frameworks. (d) Attitude models which are based upon a validated developmental process may be more easily used to structure and predict a variety of attitude-related phenomena (e.g., attitude change, non-verbal communication effects). Interesting, heuristic perspectives seem to open up when one bases his research on a theoreticallY sound conceptual framework.
[Footnote 6: The processes by which attributes become associated with or are linked to a stimulus object are of enormous interest to consumer researchers and are discussed later in this paper. To give only two brief examples, however, the associations between the attributes and brand X (N.B., conceptualized as beliefs by Fishbein, 1963) may be formed by either presenting contiguous brand X-attribute combinations through advertising communications, or through the "natural" reinforcement and contiguity mechanisms inherent in product usage experiences.]
Allport, G. W. Attitudes. In C. Murchison (Ed.), Handbook of social psychology. Worchester, Mass.: Clark University Press, 1935, 798-884.
Anderson, L. R., & Fishbein, M. Prediction of attitude from the number, strength, and evaluative aspect of beliefs about the attitude object: A comparison of summation and congruity theories. Journal of PersonalitY and Social Psychology, 1965, 3. 437-443.
Anderson, N. H., & Hulbert, S. Effects of concomitant verbal recall on order effects in personality impression formation. Journal of Verbal Learning and Verbal Behavior, 1963, 2, 379-391.
Bandura, A., & Rosenthal, T. L. Vicarious classical conditioning as a function of arousal level. Journal of Personality and Social Psychology, 1966, 3, 56-67
Bass, F. M. Fishbein and brand preference: A reply. Journal of Marketing Research, 1972, 9, 461.
Bass, F. M., & Talarzyk, W. W. An attitude model for the study of brand preference. Journal of Marketing Research, 1972, 9, 63-72.
Chambers, D. K. Repetition and order effects in attitude formation. Journal of Personality and Social Psychology, 1971, 17, 219-228.
Cofer, C. N., & Foley, J. P. Mediated generalization and the interpretation of verbal behavior: I. Prolegomena. Psychological Review, 1942, 49, 513-540.
Cohen, J. B., Fishbein, M., & Ahtola, O. T. The nature and uses of expectancy-value models in consumer attitude research. Journal of Marketing Research, 1972, 9, 456-460.
Dahlke, A. E. The effects of reinforcement and punishment on attitude change. Unpublished doctoral dissertation, University of Minnesota, 1963.
Day, G. S. Theories of attitude change and structure. In S. Ward & T. Robertson (Eds.), Consumer behavior: Theoretical sources. Englewood Cliffs, N. J.: Prentice-Hall, 1973.
Doob, L. W. The behavior of attitudes. Psychological Review, 1947, 54, 135-156.
Eisman, B. S. Attitude formation: The development of a color preference response through mediated generalization. Journal of Abnormal and Social Psychology, 1955, 50, 321-326.
Engel, J. F., Kollat, D. T , & Blackwell, R. D. Consumer behavior. (2nd edition) New York: Holt, Rinehart, & Winston, 1973.
Fishbein, M. A theoretical and empirical investigation of the interrelationship between beliefs about an object and the attitude toward that object. Unpublished doctoral dissertation, University of California, Los Angeles, 1961.
Fishbein, M. An investigation of the relationships between beliefs about an object and the attitude toward that object. Human Relations, 1963, 16, 233-240.
Fishbein, M. A behavior theory approach to the relations between beliefs about an object and the attitude towards the object. In M. Fishbein (Ed.), Readings in attitude theory and measurement, New York: Wiley, 1967, 389400.
Fishbein, M., & Hunter, R. Summation versus balance in attitude organization and change. Journal of Abnormal and Social Psychology, 1964, 69, 505-510.
Fishbein, M., & Raven, B. H. The AB scales: An operational definition of beliefs and attitude. Human Relations, 1962, 15, 35-44.
Ginter, J. L. An experimental investigation of attitude change and choice of a new brand. Journal of Marketing Research, 1974, 11, 30-40.
Greene, J. D., & Stock, J. S. Brand attitudes as measures of advertising effects. Journal of Advertising Research, 1966, 6(2), 14-22.
Hansen, F. Consumer choice behavior: An experimental approach. Journal of Marketing Research, 1969, 6, 436-443.
Hovland, C. I., Janis, I. L., & Kelley, H. H. Communication and persuasion. New Haven. Conn.: Yale University Press, 1953.
Howard, J. A., & Sheth, J. N. The theory of buyer behavior. New York: Wiley, 1969.
Kiesler, C. A. Choice, reward and attitude change. An unpublished paper, 1965.
Kiesler, C. A., Collins, B. E., & Miller, N. Attitude change: A critical analysis of theoretical approaches. New York: Wiley, 1969.
Lavidge, R. J., & Steiner, G. A. A model for predicting measurement of advertising effectiveness. Journal of Marketing, 1961, 25 (4),59-62.
Maltzman, I. Theoretical conceptions of semantic conditioning and generalization. In T. Dixon & D. Horton (Eds.), Verbal behavior and general behavior theory. Englewood Cliffs, N. J.: Prentice-Hall, 1968.
Mitchell, A. A. The use of expectancy models for market segmentation. Working series in marketing research, College of Business Administration, Pennsylvania State University, 1974, Paper No. 25.
Myers, J. H., & Alpert, M. I. Determinant buying attitudes: Meaning and measurement. Journal of Marketing, 1968, 32 (43,13-20.
Newell, A., & Simon, H. A. Human problem solving. Englewood Cliffs, N. J.: Prentice-Hall, 1972.
Niscosia, F. M. Consumer decision processes. Englewood Cliffs, N. J.: Prentice-Hall, 1966.
Olson, J. C. Cue properties of price: Literature review and theoretical considerations. Working series in marketing research, College of Business Administration, Pennsylvania State UniversitY, 1974. PaPer No. 20.
Orne, M. T. On the social psychology of the psychological experiment: With characteristics and their implications. American Psychologist, 1962, 17, 776-783.
Osgood, C. E. The nature and measurement of meaning. Psychological Bulletin, 1952, 49, 197-237.
Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. The measurement of meaning. Urbana, Ill.: University of Illinois Press, 1957.
Osgood, C. E., & Tannenbaum, P. H. The principle of congruity in the prediction of attitude change. Psychological Review, 1955, 62, 42-55.
Peak, H. Attitude and motivation. In M. Jones (Ed.), Nebraska symposium on motivation. Lincoln: University of Nebraska Press, 1955.
Platt, J. R. Strong inference. Science, 1964, 146, 347-353.
Rhine, R. J. A concept-formation approach to attitude acquisition. Psychological Review, 1958, 65, 362-370.
Rosenberg, M. J. Cognitive structure and attitudinal affect. Journal of Abnormal and Social Psychology, 1956, 53, 367-372.
Scott, W. A. Attitude change through reward of verbal behavior. Journal of Abnormal and Social Psychology, 1957, 55, 72-75.
Scott, W. A. Attitude change by response reinforcement: Replication and extension. Sociometry, 1959, 22, 328-335.
Shapiro, D., & Tagiuri, R. Some effects of response context on trait inferences. Journal of Personality, 1958, 26, 42-50.
Sheth, J. N. Reply to comments on the nature and uses of expectancy-value models in consumer attitude research. Journal of Marketing Research, 1972, 10, 462-465.
Sheth, J. N., & Park, C. W. A theory of multidimensional brand loyalty. In S. Ward & P. Wright (Eds.), Advances in consumer research: Volume I. Urbana, Ill.: Association for Consumer Research, 1973, 449-459.
Staats, A. W. An outline of an integrated learning theory of attitude formation and function. In M. Fishbein (Ed.), Readings in attitude theory and measurement. New York: Wiley, 1967, 373-376.
Staats, A. W., & Staats, C. K. Effect of number of trials on the language conditioning of meaning. Journal of General Psychology, 1959, 61, 211-223.
Staats, A. W., Staats, C. K., & Biggs, D. A. Meaning of verbal stimuli changed by conditioning. Journal of General Psychology, 1958, 71, 429-431.
Staats, A. W., Staats, C. K., & Crawford, H. L. First-order conditioning of meaning and parallel conditioning of a GSR. Journal of General Psychology, 1962, 67, 159-167.
Staats, C. K., & Staats, A. W. Meaning established by classical conditioning. Journal of ExPerimental Psychology, 1957, 74-80.
Tagiuri, R. Person perception. In G. Lindzey and E. Aronson (Ed.), The handbook of social psychology. Vol. III. Reading, Mass.: Addison-Wesley, 1968.
Talarzyk, W. W. A reply to the response to Bass, Talarzyk, and Sheth. Journal of Marketing Research, 1972, 9, 465-467.
Thurstone, L. L. Attitudes can be measured. Journal of Sociology, 1928, 33, 529-554.
Triandis, H. C. Attitude and attitude change. New York: Wiley, 1971
Vroom, V. H. Work and motivation. New York: Wiley, 1964.
Wilkie, W. L., & Pessemier, E. A. Issues in marketing's use of multi-attribute attitude models. Journal of Marketing Research, 1973, 10, 428-441.
Winter, F. W. A laboratory experiment of individual attitude response to advertising exposure. Journal of Marketing Research, 1973, 10, 130-140.
Wright, P. L. The cognitive processes mediating acceptance of advertising. Journal of Marketing Research, 1973, 10, 53-62.
Zanna, M. P., Kiesler, C. A., & Pilkonis, P. A. Positive and negative attitudinal affect established by classical conditioning. Journal of Personality and Social Psychology, 1970, 14, 321-328.