An Investigation Into the Causal Links Between Attribution Schema and Decision-Making

ABSTRACT - An Interpersonal Replication technique (Bem, 1967) was used to test the hypothesized causal links between perceived attribution schemata and confidence in market information, belief and affect strength, and purchase intention. The results strongly support the validity of a causal link and reveal the differential impact of schema complexity upon consumer information processing.


Richard Mizerski and Stephen Green (1978) ,"An Investigation Into the Causal Links Between Attribution Schema and Decision-Making", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 126-130.

Advances in Consumer Research Volume 5, 1978      Pages 126-130


Richard Mizerski, University of Cincinnati

Stephen Green, University of Cincinnati


An Interpersonal Replication technique (Bem, 1967) was used to test the hypothesized causal links between perceived attribution schemata and confidence in market information, belief and affect strength, and purchase intention. The results strongly support the validity of a causal link and reveal the differential impact of schema complexity upon consumer information processing.

Marketing has applied the concepts of Attribution Theory in a variety of areas such as children's reactions to television advertising (Robertson and Rossiter, 1974), and advertiser credibility (Settle and Golden, 1974). While the applications may differ, they all probe the general area of how consumers process market information in order to make consumption decisions.


The behavioral nature of this process has been suggested by Fishbein and Ajzen (1975), and tested within the marketing environment by Mizerski (1975). These studies suggest that through the process of attributing the causes for events (e.g., product experience, word-of-mouth information, and advertising), beliefs about the stimulus product are formed which then might prompt the development of affect (Mizerski, 1975).

More specifically, Attribution Theory suggests that individuals more readily accept, and are more strongly influenced by, information about an entity if they attribute the report to the entity being described. Applying this to the marketplace, a consumer would perceive that information about a product was accurate and useful the more he believed the content of the information was "caused" by the product being described. For example, suppose the information that a particular brand and model of automobile has good gas economy is perceived to be the result of ("is caused by") the information source actually getting good gas economy with that automobile and telling the receiver about it. If the receiver felt this was the case, he would attribute the information to properties of the stimulus object - a stimulus attribution.

In many cases, however, the consumer may have some doubt about whether the information was "caused" by actual product performance. The receiver may suspect that other causes such as the source's conflict of interest (e.g., the source is selling his own automobile), lack of expertise or specific product knowledge, or some other individual bias (e.g., the source always prefers to speak favorably about the brand) prompted the information. In these latter situations, the receiver makes circumstance (properties unique to the situation or circumstance) and/or person (personal or dispositional properties of the person) attributions, and generally finds the information much less suitable for making inferences about the product.

Measures of Attribution

Research applying these concepts to marketing (see Settle, 1973, and Burnkrant, 1974, for a review) has had a number of limitations. First, causal attributions in marketing research are seldom measured directly. Instead, the strength of the stimulus attribution is often inferred using measures of confidence or credibility. However, research by Newtson (1973) and Mizerski (1974) seriously question the validity of this technique.

A second limitation of this research is that it typically elicits the degree of stimulus attribution only, which is not necessarily the lone factor of importance. Consumer perception of what is suitable/useful information is largely determined "... by the configuration of factors that are plausible causes of that information" (Kelly, 1973, p.108). In other words, the type and mix of causes (causal schema) the consumer perceives can also tell us a great deal about how he or she will interpret and process product information.

The Concept of Causal Complexity

In an attempt to overcome these problems, Mizerski (1975, 1977) developed a scale that elicits an individual's causal schema in terms of (1) multiple types of perceived causes, and (2) the relative importance of individual causes in the schema. These values are then transformed into a measure of entropy or an "H" statistic. This transformation provides a measure of the complexity of the subject's causal domain.

As an example, Figure 1 shows five hypothetical individuals' causal allocations among five possible causes for information about a product.



The measure of causal complexity would designate subjects A, B and C the least complex (causally simple), subject E the most complex, and locations such as D in between.

Theoretical Rationale for the Differential Influence of Causal Schema

The literature (e.g. Mizerski, 1975, 1977) suggests that the five causal schemata shown in Figure One should prompt significantly different cognitive responses in the areas of confidence in the schema, confidence in the information, belief, affect strength and behavioral intention.

Confidence in the causal schema. The literature dealing with the measure of entropy (of which causal complexity is derived) suggests that consumers would be most confident in a causally simple allocation since individuals are less certain/confident in an array with a "broad distribution'' than a "sharply peaked" one (Jaynes, 1957). If one considers a consumer's decisions on the number and allocation of cause to be a causal array (see Mizerski, 1975, for a more thorough discussion), an individual should have more confidence in causally simple schemata (receivers A, B, and C) than with the complex schema represented by E. Also, since we are suggesting schema confidence only at this point, there should be no differences between the three simple schemata on this dimension.

H1: All causally simple schemata will evoke significantly stronger confidence in the schema than a causally complex schema.

Confidence in market information received. A consumer should believe that information about a product was more accurate, and thus have more confidence in that information, the more he assumes that the content of that information was caused by the product being described. Therefore, the less causality attributed to the product evaluated (the stimulus cause), the less confidence he or she will have in the rating information.

H2: A causally simple stimulus attribution schema will evoke significantly stronger confidence in the product information than causally simple non-stimulus attribution schema (e.g., person or circumstance attributions, receivers B and C)

H3: A causally simple stimulus attribution schema will evoke significantly stronger confidence in the product information than a causally complex schema (e. g., receiver E in Figure One).

Strength of belief, affect and intention to purchase. In an approach consistent with other researchers, Fishbein (1965, p. 107) has stated that "... a belief about an object may be defined as the probability or improbability that a particular relationship exists between the object of belief and some other object, concept, value, or goal." This definition of a belief appears to be one result of an attribution (Ajzen and Fishbein, 1975; Mizerski, 1975). Kelley (1973, p. 107) notes that attribution is a part of the process" which man 'knows' his world, (and) has a sense that his beliefs and judgements are veridical."

It would seem to follow that if an individual made a strong attribution about information concerning product characteristics to a stimulus cause, that would manifest itself in a strong belief that a relationship existed between the product and those characteristics. In Fishbein terminology, the stimulus attribution attaches a higher probability that the information was related to the product. Therefore, the stronger the stimulus attribution, the more extreme the belief. In situations of complex attributions, such as receiver E in Figure One, Kelley (1973, p. 113) suggests that "...the role of a given cause in producing a given effect is discounted if other plausible causes are also present." Therefore, a large number of person or circumstance attributions should discount (in the consumer's mind) the possibility of a stimulus cause.

A substantial amount of research on expectancy - value attitude models suggests that an attitude or affect toward an object is a function of (1) the strength of the individual's beliefs about the object, and (2) the evaluative aspect of those beliefs. While expectancy - value models differ somewhat, each uses some measure of instrumentality or belief strength as a basis for predicting affect. If causally simple subjects form more extreme beliefs, the differential in belief strength should manifest itself in a stronger affect toward the product.

Just as affect has been suggested to be a function of the salient beliefs about a product, a consumer's intention to purchase has been linked to his or her cognitive and affective responses (Lutz, 1977). Although the intention to purchase can be rather distant in both time and number of intervening variables from belief and affect formation, it may also directly respond to differences in causal schemata. Therefore, the following hypotheses will be tested.

H4: A causally simple stimulus attribution schema will evoke a significantly stronger belief about, affect toward, and intention to purchase the product than causally simple nonstimulus attribution schemata.

H5: A causally simple stimulus attribution will evoke significantly stronger belief about, affect toward, and intention to purchase the product than a causally complex schema.


While the results of previous research (Mizerski, 1975, 1977) appeared to establish a correlational relationship between causal complexity and other processes such as belief formation, there was no evidence that differences in the subjects' attributional schemata caused the differences in confidence, belief strength, etc. In fact, it became apparent that the present paradigm was incapable of testing that question since all the phenomena were cognitive, unobservable behaviors and all were elicited by the same stimulus (i.e., product information). Therefore, even though the present authors strongly believe that the causal schemata preceded and "caused" the beliefs, there was little available evidence to defend that position. How could this question of cause and effect be answered?

A review of the Dissonance Literature, which had undergone a similar problem of investigation that probable causes underlying observed behaviors in experiments, revealed a unique approach developed by Bem (1965). His technique was referred to as an "interpersonal replication paradigm," and it appeared to offer a valid approach for investigating the causal links under study.

Bem's analysis of dissonance, and the present analysis of causal schemata rest upon the single experimental generalization, "that an individual's belief and attitude statements and the beliefs and attitudes that an outside observer would attribute to him are often functionally similar in that both sets of statements are partial 'inferences' from the same evidence..."(Bem, 1967, p.186). Bem gave his subjects information about the treatment a single subject received in a dissonance experiment, and then asked how that subject would have responded. He then argues that when his subjects replicate the dissonance finding, the same process that was operating for his observer was also operating in the dissonance subject. With slight modification, a similar interpersonal replication paradigm was developed for the purpose of testing the preceding hypothesis and to determine whether it seems reasonable to argue that the subjects' causal schemata in the original Mizerski research (1975, 1977) caused the differences in belief and affect.


Original Study

In Mizerski's (1975) original experiment, subjects were provided with evaluative rating information on one of three salient attributes for either a fictitious automobile or motion picture. The information was presented in the form of personal ratings that were supposedly made by another individual who was randomly chosen to test and evaluate the product. Following the information treatment, the subjects were asked to allocate the probable cause for the bogus "rater's" opinion (the information treatment) on a causal complexity scale shown in Figure 2. They were then asked to indicate their confidence in the allocation, beliefs about the product, and affective response to the product.


Interpersonal Replication I

Following the basic approach of Bem (1967), an experiment was developed in which subjects (University of Cincinnati undergraduate Business students) would be provided with the same product information as in Mizerski's original study. They were then told to assume that they made one of four possible causal allocations: (1) causally simple to the product - a stimulus attribution, (2) causally simple to a person attribution - "the influence of other people's opinions," (3) causally simple to a circumstance attribution - "the personality of the product evaluator ...," or (4) a causally complex attribution which provided approximately equal allocation to each possible cause (refer to Figure 2). Ninety four percent was allocated to each simple attribution, with one or two percent allocated to the remaining causes. In the complex treatment, between 18 and 22 percent were allocated to each of five causes. [Originally, the simple causal allocation treatments had the total 100% allocated to the relevant simple cause, with an equal 20% allocated to each of the five potential causes in the causally complex treatment. However, pretests showed that the subjects inferred that the causal allocator " ...did not take enough time" with the task using those percentage allocations.]

After being provided with the rater's information that the automobile tested rated "superior" on gas mileage and the treatment causal allocation, the subjects were asked to respond to several questions. The first question asked "How much confidence would you have in your assignment of percentages for the product rater's opinion about the automobile's gas mileage" (confidence in the schema). The second question asked "How confident do you think you would be that the product rater's opinion of the automobile's gas mileage was accurate; i.e., that the automobile actually had superior gas mileage?" This measured the subjects' confidence in the rater's opinion. The subjects' responses were gauged on a nine point scale that ranged from "No confidence" (#1) to "Complete confidence'' (#9).

They were then asked a question about the strength of their belief that the automobile had superior gas mileage (unlikely--#1, to likely --#9), and the strength of their affect toward the stimulus product (extremely low appeal --#1, to extremely high appeal--#9). Finally, they were asked," ... If you felt that an automobile's gas mileage was very important, what do you think is the probability that you would actually consider purchasing this automobile if you were in the market when it is introduced?" Their response was scored on a scale that ranged from "very low probability" (0%) to "very high probability" (100%). All hypotheses were tested with a one-way analysis of variance, with each replication tested separately.

Interpersonal Replication II

After the initial interpersonal replication experiment was completed, it was felt that the experiment should be expanded so that results from treating all the potential simple causal schemata could be examined. Therefore, a second experiment treated the following allocations: (1) causally simple to the circumstance - "An effort to pleas or antagonize the interviewer;" (2) causally simple to the person - "A general bias for or against automobiles;'' (3) the same causally simple to the product; and (4) causally complex schemata provided the subjects in the first interpersonal replication study.

The first interpersonal replication was run on two undergraduate marketing classes; the second experiment used two undergraduate management classes. Both samples closely approximated the backgrounds of the individuals in Mizerski's original study. The treatments were randomly assigned to the subjects, with each subject exposed to one of four treatments in each experiment. Because of the similarity in methodology, the results of the two studies will be discussed together in the text.


Confidence in the Schema

In an effort to simplify presentation of the data, the results of both experiments are provided in Table I. Standard deviations of the mean scores are shown in parentheses, with cell sizes reported to the left.

The first hypothesis proposed that there would be no significant differences between the simple causal schemata, that they would all prompt stronger confidence in the schema than the complex schema. The mean scores from both studies (the first column of Table I) show only partial support for the hypothesis. The simple attribution to the product was the only simple schema that prompted significantly stronger confidence than the causally complex allocation (t=2.34, df = 46, p<.01; and t=3.43, df = 58, p<.001 for study I and II, respectively).

It was found that the simple attribution to the product also prompted significantly stronger confidence in the schema than the other simple attribution (t=3.20, df = 46, p<.001, and t=5.64, df = 58, p<.001). This latter finding was not expected since the literature strongly suggests that the more simple the attribution, the more confidence the attributer should have in the schema (i.e., a simple attribution is a simple attribution for this dimension).



Confidence in the Accuracy of the Rater's Opinion

The second hypothesis suggested that the subjects would have more confidence in the accuracy of the rater's opinion when using schemata reflecting simple allocations to non-product causes. The mean scores in both studies are in the appropriate direction, (See Table 1) and the results of planned comparisons (t=4.60, df=46, p<.001; and t=7.13, df=58, p<.001) support the hypothesis.

The third hypothesis, that subjects who received the simple causal allocation to the product, would have more confidence in the accuracy of the rater's opinion than those who received a causally complex schema, was also supported (t=2.73, df=46, p<.O05; and t=5.86, df=58, p<.001).

Strength of Belief, Affect, and Intention to Purchase

The fourth hypothesis proposed that subjects receiving a causally simple stimulus attribution schema will produce stronger belief, affect, and intention to purchase mean scores than subjects receiving causally simple person or circumstance schemata. Hypothesis Five went on to suggest that the same differential strength of the simple stimulus attribution schema would also be reflected over those subjects who received a causally complex schema. The easiest way to evaluate the results is to discuss one element at a time, across the two experiments, since significant differences were observed.

Both hypotheses were supported in the two experiments for the measure of belief strength. Subjects receiving the simple stimulus attribution schema formed significantly stronger beliefs than subjects under the simple circumstance and simple person treatments (t=3.57, df = 46, p<.001; and t=5.18, df = 58, p<.001), and stronger than subjects receiving the causally complex schema (t=1.82, df = 46, p<.037; and t=3.93, df=58, p<.001).

As noted earlier, the formation of beliefs has been discussed as a direct response of the attribution process. On the other hand, the development of affect and intention to purchase is a partial function of the beliefs about the stimulus object. One would expect somewhat weaker relationships with the attribution schema used.

As hypothesized, the subjects receiving the simple stimulus attribution schema formed significantly stronger affect than subjects with simple non-stimulus schemata (t=2.31, df=46, p<.013; and t=3.11, df=58, p<.O02). However, the prediction that a simple stimulus attribution schema would also prompt stronger affect than a causally complex schema was not fully supported in both experiments. While the mean scores in the two experiments were in the appropriate directions (see Table 1), they failed to reach significance (t=1.15, df=46, p<.128) in experiment #1, while there was strong support for the hypothesis (t=3.81, df=58, p<.001) in the second experiment.

Similar results were found for the subjects' intention scores. Although the mean scores comparing the simple stimulus schema to other causally simple non-stimulus schemata (hypothesis 4), and to causally complex schema (hypothesis 5) were in the appropriate direction, hypothesis four was supported marginally (t=1.43, df=46, p>.079) and hypothesis five was not supported (t=0.77, df=46, p>.44) in the first experiment. Both hypotheses, however, were supported in the second experiment, (t=1.81, df=58, p<.038, and t=1.74, df=58, p<.044).


The results strongly support the hypothesis that differences in the complexity of causal schemata can cause differences in other cognitive responses. Specifically, those individuals using a causal schema that reflected simple allocations to the product (stimulus attribution) responded with significantly higher levels of confidence in the accuracy of the information they received, stronger beliefs that the product possessed the characteristics assigned it by the information source, and a stronger affective response to the product than individuals using a complex causal schema or schemata reflecting a simple non-stimulus attribution. In addition, those subjects using a simple-product schema demonstrated a stronger intention to purchase the product than subjects using a simple-circumstance or simple-person attribution schema.

The comparison of simple-product and complex schemata on intention to purchase, however, yielded equivocal results. Although it is not clear why the subjects using the simple-product schema formed stronger intentions to purchase than subjects using a complex-schema in the second study only, one could speculate that several factors might have been operating. First, as previously noted, intention to purchase is only partially a function of beliefs and therefore not as directly related to schemata. Thus, it may be that this phenomenon is more tenuous and difficult to capture experimentally. Also, one should note that all the findings are stronger with the second study than in the first. This discrepancy may be due to the fact that Study One was conducted in marketing classes where the students were more familiar with marketing concepts and research, and therefore may have been "test wise" and wary in their responses.

The failure of hypothesis one, that all simple schemata will evoke stronger confidence in the schema than the complex schema, may have been due to the nature of that question more than any "real" psychological process. In pretesting the instruments, it was found that the subjects had a great deal of difficulty understanding the intent of the question concerning confidence in the schema. After many iterations on that measure, it was felt that it had been made clear. Debriefing, however, revealed that subjects had confused the confidence in the schema measure with the confidence in the accuracy of the rater's opinion measure (these two responses correlated .54, p<.001, when the treatment means were removed from the scores). Therefore, any interpretation of the confidence in the schema data appears problematic.

Nonetheless, it is apparent that differences in causal schemata complexity can cause differences in other important cognitive processes. To the extent that the process revealed in our "observers" is the same process that occurs in individuals who generate their own causal schema (as Bem, 1967, and these authors would argue), there are clear implications for marketing strategy. If one can devise advertising formats that elicit a simple-product causal schema, as opposed to simple nonproduct or complex allocations, the consumer should be more likely to be confident that the product information is accurate, as well as form stronger beliefs about the product's characteristics. Similarly, the consumer should have a stronger affective response toward that product and might (but not necessarily) develop a stronger intention to purchase the product.

Research is presently underway in an attempt to identify the varying capacities of advertising formats to elicit a simple-product schema in consumers. Other research might focus on the effects of different media presentations upon the complexity of evoked schemata or investigate the possibility of causal complexity existing as an individual difference construct.


Bem, D. J. "An Experimental Analysis of Self-Perception," Journal of Experimental Social Psychology, 1 (1965), 199-218.

Bem, D. J. "Self-Perception: An Alternative Interpretation of Cognitive Dissonance Phenomena," Psychological Review, 74 (1967), 183-200.

Burnkrant, R. E. "Attribution Theory in Marketing Research: Problems and Perspectives," in M. J. Schlinger (Ed.), Advances in Consumer Research, Vol. 2, Proceedings of the Association for Consumer Research, 1974.

Fishbein, Martin. "A Consideration of Beliefs, Attitudes, and Their Relationship," in Ivan D. Steiner and Martin Fishbein, eds., Current Studies in Social Psychology. New York: Holt, Rinehart, and Winston, Inc., 1965.

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

Jaynes, E. T. "Information Theory and Statistical Mechanics,'' Psychology Review, 106 (1957), 620-631.

Kelley, Harold H. "The Processes of Causal Attribution," American Psychologist, 28 (1973), 107-128.

Lutz, Richard J. "An Experimental Investigation of Causal Relations Among Cognitions, Affect, and Behavioral Intention," Journal of Consumer Research, 3 (1977), 197-208.

Mizerski, Richard. "A Test of the Relationship Between Trait and Causal Attribution," in J. Schlinger, ed., Advances in Consumer Behavior, Vol. 2, Proceedings of the Association for Consumer Research, 1974.

Mizerski, Richard. "An Investigation Into the Differential Effects of Causally Simple and Complex Attributions,'' in B. Anderson, ed., Advances in Consumer Behavior, Vol. 3, Proceedings of the Association for Consumer Research, 1975.

Mizerski, Richard. "Causal Complexity: A New Measure of Consumer Causal Attribution," Working Paper, University of Cincinnati, 1977.

Newtson, Darren. "Attribution and the Unit of Perception of Ongoing Behavior," .Journal of Personality and Social Psychology, 28 (1973), 28-38.

Robertson, Thomas and John Rossiter. "Children and Commercial Persuasion: An Attribution Theory Analysis," Journal of Consumer Research, 1 (1974), 13-20.

Settle, Robert and Linda Golden. "Attribution Theory and Advertiser Credibility," Journal of Marketing Research, 11 (1974), 181-185.



Richard Mizerski, University of Cincinnati
Stephen Green, University of Cincinnati


NA - Advances in Consumer Research Volume 05 | 1978

Share Proceeding

Featured papers

See More


O12. When do People Waste Time? Testing a Mechanism for Parkinson’s Law.

Holly S Howe, Duke University, USA
Tanya Chartrand, Duke University, USA

Read More


Communicating Limited Financial Resources Increases Perceived Trustworthiness and Interpersonal Connection

Grant E. Donnelly, Harvard Business School, USA
Anne Wilson, Harvard Business School, USA
Ashley V. Whillans, Harvard Business School, USA
Michael Norton, Harvard Business School, USA

Read More


The Impact of Price and Size Comparisons on Consumer Perception and Choice

Jun Yao, Macquarie University, Australia
Harmen Oppewal, Monash University, Australia
Yongfu He, Monash University, Australia

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.