What Do Consumers Know About Familiar Products?
ABSTRACT - Twelve consumer subjects recalled their knowledge about familiar products. Their knowledge statements were classified according to a new, theory-based taxonomy of product knowledge. The analysis revealed: (1) a majority of brand-attribute values, the lowest knowledge category; (2) a predominance of brand-based higher level knowledge and attribute-based lower level knowledge; and (3) an organizational structure mainly linked by brands, even for attribute-based knowledge. The possible sources for different types of knowledge are discussed, including advertising and past purchase decisions.
Citation:
J. Edward Russo and Eric J. Johnson (1980) ,"What Do Consumers Know About Familiar Products?", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 417-423.
[This research was partially supported by Grant DAR 76-81806 from the National Science Foundation to the University of Chicago.] Twelve consumer subjects recalled their knowledge about familiar products. Their knowledge statements were classified according to a new, theory-based taxonomy of product knowledge. The analysis revealed: (1) a majority of brand-attribute values, the lowest knowledge category; (2) a predominance of brand-based higher level knowledge and attribute-based lower level knowledge; and (3) an organizational structure mainly linked by brands, even for attribute-based knowledge. The possible sources for different types of knowledge are discussed, including advertising and past purchase decisions. INTRODUCTION When shoppers are deciding what to buy, they rely on two sources of product information. From previous experience they remember some knowledge of the available products, and at the point of purchase they obtain additional information. Recent research on consumer information processing has focussed on external information rather than on remembered knowledge. In contrast, the present study investigates the latter, internal source of information. We ask: What knowledge about familiar products is stored in memory and available to typical shoppers? When shoppers are asked to recall, i.e., to speak aloud, everything they know about a product category, a variety of statements emerges. Consider the following knowledge statements, all from a shopper speaking about butter and margarine. "And Squeeze Parkay is even more of a waste (than Soft Chiffon), because you get to the place where you want to play with it." "Land O'Lakes butter sounds good to me." "I know there's a Fleischmann's diet margarine which I buy." "I have found in stores that it [Land O'Lakes butter] is usually higher in price [than margarines]." "Squeeze Parkay is good." "Mazola, I think, is supposed to be low in cholesterol." "I tend to think that [Land O'Lakes butter] is richer in taste than margarine." How do we organize this variety of knowledge? Which statements are similar and should be grouped together? How many different types of knowledge statements should be allowed, and what is the relationship, if any, among these different types? Several recent investigations of product knowledge have confronted these problems. Olson and Muderrisoglu (1979) elicited from subjects their remembered product knowledge. Payne and Ragsdale (1978) and Rip (1979) recorded "think aloud" protocols during a purchase decision. Wright and Barbour (1975) analyzed the product information explicitly stated in advertisements. In all four cases, a classification scheme was devised and the observed frequencies were reported. Although these schemes overlap, their differences exceed their similarities. This heterogeneity is caused partly by the authors' different sources, such as advertisements versus decision protocols, but also by the absence of a more general theory that guides the classification of knowledge statements. With such a theory the analysis of remembered product knowledge ceases being ad hoc and can be embedded in a larger conceptual framework. This, in turn, may reveal theoretical linkages or interesting data analyses that might otherwise remain unnoticed. THEORY Our theory of product knowledge is adapted from the information processing view of problem solving and understanding (Newell and Simon, 1972; Kintsch, 1974). Any problem solver moves from some initial state of knowledge to a final "goal" state that contains the problem's solution. Initial knowledge can be quite basic, especially for a novel problem. Progress toward the solution requires inferences that derive higher level knowledge from the initial information. The process of generating still higher level inferences continues until finally the solution itself is reached. In a decision making task, the initial information is the set of alternatives and their descriptions, while the goal knowledge state is the conclusion that one of the alternatives is preferred over all the others. The inference process is the heart of problem solving; from it we derive the needed taxonomy of product knowledge. This taxonomy uses inferential level as the central classifying concept. Inferential Level Level of inference is a continuum extending from the basic, initial knowledge to the ultimate goal state. Lowest level knowledge requires no inference at all; it is usually available as information in the external environment. The highest level of knowledge is the problem's solution. It can only be inferred from information at the lower levels. Suppose that a purchase decision must be based on a matrix of brand-attribute values. This matrix is the basic information that comprises the initial state of knowledge of a naive consumer. Examples of brand-attribute values include "Tuna fish is average in taste," "Chicken is very good on price," and, in general "[Brand X] is [average] on [attribute C]." As people progress toward the goal of finding the best brand, they infer intermediate levels of knowledge. Such knowledge might be, "The cheapest one is [Brand X]," or "I definitely prefer [Brand X] to [Brand Y]." Finally, they reach the conclusion, "[Brand X] is best. That's the one I'll buy." This simple three level taxonomy (lowest, intermediate, and highest inferential levels) was used by Johnson and Russo (1978a) to study what people remembered from a purchase decision. They found that the different levels of product knowledge were differentially remembered: The higher the level of inference, the better the knowledge was remembered. We now propose a similar, but more elaborate classification scheme. Again, the lowest level of knowledge consists of brand-attribute values. When the inferential levels in the present taxonomy are numbered this becomes Level 5. Just above this are single pair comparisons and single rankings (Level 4). These propositions describe the relative status of two or more brands on a single attribute ("Fiats are cheaper than Cadillacs'') or two or more attributes of a single brand ("You buy a Cadillac for prestige, not gas mileage"). Level 3 contains whole evaluations, such as "I have always liked Fiats a lot", in which an entire brand (or attribute) is evaluated. The second level from the top contains whole rankings and whole pair comparisons. These propositions compare two or more brands with all their attributes taken into consideration ("I like Fiats better than Cadillacs") or two or more attributes over all brands ("In my business I buy a car for prestige, not gas mileage"). The final level is knowledge of the preferred brand. These levels of inference are displayed in Table 1. Also shown is an example of each, taken, whenever possible, from the statements collected for this study. In addition to the five major classifications described above, one minor distinction is made. Ranking knowledge is considered a higher level of inference than pair comparisons because more knowledge is necessary to infer a ranking. Specifically, a minimum of two pair comparisons is required before a ranking proposition can be inferred, and usually the number of necessary pair comparisons is much larger. Although both are classified at Level 2, rankings are listed above pair comparisons. As we shall see shortly, the five major classifications make important distinctions among ordinary product knowledge. The minor one does not. SAMPLE STATEMENTS CLASSIFIED BY INFERENTIAL LEVEL Is Knowledge Organized by Brand or by Attribute? Though inferential level provides a useful, theory-based taxonomy, it is not a current topic of consumer research. In the literature on product knowledge, however, one question has emerged repeatedly and without a clear-cut answer: Is product knowledge organized by brand or by attribute? As it turns out, this is really two questions, one about the content of knowledge, the other about its structure. Whether product knowledge, as content, is organized by brand or by attribute depends on the inferential origin of that knowledge. Each knowledge proposition can be inferred by comparing brands or by comparing attributes. The knowledge that "A Fiat is cheaper than a Cadillac," reflects a brand comparison, while the proposition, "I buy a car for prestige, not price," results from an attribute comparison. Thus, knowledge, in the sense of its content, can be identified by the type of comparison from which it was inferred, either a brand comparison or an attribute comparison. We call this the inferential basis of the knowledge. The taxonomy defined by inferential level can now be further differentiated by inferential basis. The knowledge propositions at Levels 2, 3 and 4 can be distinguished as brand-based or attribute-based. Table 1 contains examples of each inferential basis for these three levels. Level 5 knowledge, a brand-attribute value, is neutral with respect to inferential basis; and the preferred brand (Level 1) can only be, by definition, brand-based. Whereas the content of knowledge refers to a set of distinct propositions or facts, the structure of knowledge refers to the associative links connecting these propositions (Anderson, 1976). This network-like structure can be organized or linked by brand or by attribute. If the two statements "A Fiat is cheaper than a Cadillac" and "Fiats are sexy Italian cars" are linked, then this knowledge is brand-organized. Similarly, if "A Fiat is cheaper than a Cadillac" were linked to "But I buy what I like; price has never meant much to me," then the structure of this knowledge would be attribute-based. We call this distinction the structural basis of remembered product knowledge. Evidence on Inferential Basis There has been considerable speculation over the relative amounts of brand-based and attribute-based knowledge. The main question is which one is the predominant (inferential) basis of product knowledge. By and large, we live in a brand-based world. Advertising, with the minor exception of comparison advertising, is organized about individual brands. Our use experience also tends to be with single brands. It is the rare shopper who simultaneously purchases, uses and compares two brands of peanut butter or laundry detergent. There is some use of multiple brands, e.g., breakfast cereals; but this is the exception, not the rule. The point-of-purchase environment itself is largely brand-organized. The available information, usually just the printing on the packages and the price tag on the shelves, is separated by brand. Many people have argued that this makes it difficult to make comparisons across brands. At least one study reorganized a single attribute, price, across all brands and found major changes in purchase behavior (Russo, 1977). In contrast to a brand-organized world, results from decision studies suggest that attribute-based information may predominate. Johnson and Russo (1978a) found that the intermediate level of knowledge produced during a purchase decision was more attribute- than brand-based. To the extent that product knowledge stored in memory was derived from past purchase decisions, attribute-based knowledge can be expected. This may be especially likely for familiar products that are frequently repurchased. In general, the relative frequency of brand-based versus attribute-based knowledge remains an open question. Evidence on Structural Basis The evidence, though incomplete, would seem to favor a brand-based structure. As noted above, the external environment offers brand-based sources of product information. Johnson and Russo (1978b) have demonstrated that the organization of the input structure (the attribute- or brand-based external environment) determines the organization of the resulting memory structure. Another way of stating this result is that there does not seem to be a single preferred structure of knowledge into which all information will be transformed. Thus, given a brand-structured purchase environment, one might expect a brand structure in memory. Comparison to Other Taxonomies of Product Knowledge We propose a taxonomy of product knowledge with the following characteristics: 1. The major classifying concept is level of inference, with five levels distinguished. 2. A second classifying concept is the inferential basis of knowledge, brand-based or attribute-based. This distinction applies to knowledge at inferential Levels 2, 3 and 4. 3. The structural basis of knowledge is distinguished from its inferential basis. The structure of product knowledge can be linked by brand or by attribute. This scheme for classifying product knowledge differs considerably from others that have been described in the literature. Several researchers have recorded verbal protocols and classified individual statements into types of product (and other) knowledge. Payne and Ragsdale (1978) and Rip (1979) recorded protocols during a purchase decision task. Payne and Ragsdale's scheme makes fewer distinctions than ours, especially at the lower and middle inferential levels. Rip's taxonomy is more similar to ours in the number of different knowledge types and even in the specific categories. It does not, however, recognize as strict an inferential hierarchy or systematically distinguish brand-based from attribute-based knowledge. Both taxonomies are oriented toward purchase decisions rather than remembered product knowledge. Olson and Muderrisoglu (1979) use a recall or "free elicitation" technique to expose stored knowledge. Their goal, however, is primarily methodological, an analysis of the recalled statements for frequency, rate, and reliability over subsequent elicitations. Wright and Barbour (1975) investigate the content of advertisements. They examine the statements in advertisements, i.e., the stimulus, rather than the response of what people believe or remember as a result of exposure to an advertisement. Their classification scheme is similar enough to ours to enable comparison of the results they report with our own. In general, the nature of product knowledge and its various sources are receiving considerable attention. We believe that a factor essential to the success of such investigations is a theory of types of knowledge that can provide a systematic taxonomy. METHOD Experimental Rationale To identify the product knowledge that consumers retain in memory, we chose a free recall technique. Participating subjects are asked to speak aloud while they remember everything they know about a specified topic, in this case, a familiar product class. These protocols are tape recorded for later analysis. Often this recall from memory is prompted by external cues or an appropriate cover story. It is important, however, to facilitate complete recall and not to bias recall toward particular types of knowledge (Ericsson and Simon, 1978). This technique is quite familiar in experimental psychology and is seeing increased utilization in consumer research (e.g., Olson, 1978). Procedure For the primary experimental task, we asked subjects to imagine that they had an English speaking friend from a foreign country who needed to know about the products available in the United States. The subjects' task was to tell their friend everything she or he would need to know to make informed purchases in a specified product class. The friend was characterized only as a newcomer who spoke English and knew nothing about American brands (but did have a basic knowledge of common products like automobiles, headache remedies, etc.). The few subjects who pressed the experimenter for details were told that in other respects, the imaginary friend "is someone like you." To insure more complete protocols, subjects began the experiment by practicing "thinking aloud" on a different task, their thoughts in response to advertisements. During the free recall protocols, whenever there was a prolonged silence, the experimenter used one of several standardized, neutral prompts (Payne and Ragsdale, 1978) such as "Is there anything else you know about these kinds of products?" To aid subjects' recall, we provided a list of five representative brands and a second list of five relevant attributes. At the same time these lists aided subjects' recall, they may also have biased the recall protocols by causing an overrepresentation of the listed brands and attributes. After generating recall protocols for two product classes, subjects completed a short demographic questionnaire. Subjects The participating subjects were a convenience sample of 12 staff members, mainly secretaries, of the Graduate School of Business of the University of Chicago. They were exclusively female; and all were primary shoppers for their families, making a median of 1 or 2 major shopping trips per week. Their mean expenditure was between 40 and 50 dollars per week. The mean family income of the group was $14,500 (median = $10,500), which is slightly below the average for the Chicago area. The sample had a mean of 13 years of formal schooling and averaged between 30 and 40 years of age. They were paid for the one hour session at the same salary rate of their current position. The generalizability of this sample to the entire U.S. population is problematical. Although its primary characteristics (age, education and income) conformed well to national averages, they were all employees of the University of Chicago. At worst, this selection homogeneity may introduce an unknown but important bias. At best, however, the sample does not overrepresent high education and high income shoppers; and their familiarity with the experimenters permitted a more relaxed and frank protocol of product knowledge. Products The product categories were chosen to be highly familiar, so that typical consumers would be able to generate an informative recall protocol. Product categories were selected on the basis of high levels of familiarity for representative consumers interviewed in a local supermarket. The 20 consumers interviewed looked at sets of different brands and attributes for a number of product classes, which they then rated on overall familiarity. Since the four product categories that scored highest in familiarity also covered a wide range, they were selected. These were: automobiles, butter and margarine, dinner entrees (meat, fish, and fowl) and headache remedies. Each subject recalled her product knowledge for 2 of the 4 categories. This limit was imposed to reduce the chance of fatigue. The assignment of products to subjects was randomized within a counterbalanced design. Each of the 4 product categories was recalled by 6 subjects, yielding a total of 24 recall protocols. Protocol Coding The free recall protocols were divided into short segments each representing a complete thought (Newell and Simon, 1972). These were then coded by 2 judges, both blind to the experimental hypotheses, according to a scheme that reflected our a priori ideas of the possible types of knowledge (see Theory above). A total of 42 knowledge categories were coded, including distinctions that turned out to be uninteresting, and are not included in our taxonomy. The coders agreed on the classification of 83% of the statements. Disagreements were resolved by discussion. Altogether, 80% of the statements could be classified into one of the 12 categories listed in Table 1. Almost all of the remaining statements were of one type, an isolated mention of a single brand or of a single attribute with no indication of the knowledge associated with it. We believe that there were two causes of such statements. First, subjects failed to think aloud, i.e., to state the knowledge they were recalling about the brand or attribute. Alternatively, they were sometimes doing no more than reading from these five brands or five attributes listed. The 20% of the individual utterances that did not possess a knowledge proposition had to be excluded from all analyses of the content of knowledge. However, they could be reinstated for the analysis of the structure of knowledge. RESULTS AND DISCUSSION Inferential Level What do consumers know about familiar products? We begin by considering the inferential level found in their recall. Aggregated across all 24 recall protocols, the frequencies of each knowledge level are shown in Table 2. FREQUENCY OF STATEMENTS AT EACH INFERENTIAL LEVEL The most salient feature of this distribution is the preponderance of brand-attribute values (single evaluations). Overall, this most basic level of knowledge accounted for 62% of the 464 analyzable statements. What accounts for this dominance of brand-attribute values? We believe that the source of this knowledge is probably external, i.e., received directly as a message from the external environment. The Johnson and Russo (1978a) study of purchase decisions found that brand-attribute values were the least likely knowledge to be retained in memory as a result of a purchase decision. Thus, internal decision processes probably play little or no role in the retention of this knowledge in memory. In contrast, brand-attribute values are just the type of knowledge that ought to be available externally. Unlike higher level knowledge, brand-attribute values are the least dependent on personal values; they are the most objective and easily communicated knowledge. Wright and Barbour's (1975) survey of the content of advertising supports the hypothesis that the source of brand-attribute values is external. Because their category boundaries differ from ours, brand-attribute values may occur in several of their categories. In addition, they report their results in terms of the percent of advertisements with at least one occurrence of a knowledge type, rather than in terms of the total number of knowledge statements. Nonetheless, it is apparent from their data that the majority of advertising claims are brand-attribute values, probably 70% or more. Higher level knowledge occurs less frequently, and its occurrence decreases as the level of inference increases. Thus, the results of Wright and Barbour's analysis of advertisements are compatible with our finding that product knowledge is dominated by brand-attribute values. Before concluding that the predominance of brand-attribute values probably reflects input from the external environment, we must consider one artifactual explanation for the frequency distribution of Table 2. There are many more potential brand-attribute values (Level 5) than, say, best brand statements (Level 1). With m attributes and n brands, there are m x n possible brand-attribute values, while there is only one best brand statement regardless of m and n. Possibly the greater frequency of brand-attribute values only reflects the greater opportunity for such statements. To test this alternative explanation, we calculated for each inferential level the maximum number of different statements as a function of m and n. To find appropriate values of m and n, we calculated the mean number of different attributes and of brands mentioned over the 24 protocols. These values were 7.8 attributes and 5.6 brands. These were rounded to the nearest integer and inserted into the appropriate combinatorial expressions for the maximum possible number of each of the five types of statements. The calculated maxima are reported in Table 3. The actual frequencies are also reported and compared with these maxima. RATIOS OF OBSERVED AND MAXIMUM FREQUENCIES BY LEVEL OF INFERENCE The table makes clear that the maximum base rates differ widely across inferential level. Taking them into account places the observed frequencies in a different perspective. Although brand-attribute values are still relatively frequent, only one-fourth of all possible statements at this level are recalled. In contrast, best brand statements are overrepresented. Note that the present task did not require the identification of the most preferred brand (as would a purchase decision), yet subjects voiced their first choices almost always and often repeatedly. On balance, although the comparison to maximum base rates alters the relations in Table 2, the predominance of brand-attribute values remains. The relative frequencies of the four higher levels of knowledge shown in Table 2 are best understood in conjunction with inferential basis. We turn now to that classifying concept. Attribute-Based Versus Brand-Based Knowledge Inferential basis interacts strongly with inferential level. All knowledge statements above the levels of brand-attribute values can be identified as brand-based or attribute-based. Level 1 knowledge (best brand) can only be brand-based, but Levels 2-4 may be either (see Table 1, for example). The recall protocols were coded for inferential basis, and the resulting frequencies are displayed in Table 4. FREQUENCY OF STATEMENTS BY INFERENTIAL LEVEL AND INFERENTIAL BASIS The striking aspect of these data is the strong interaction between level and basis. Lower level knowledge (Level 4) is exclusively attribute based (94 to 0). Conversely, at the three highest levels, the statements are preponderantly brand-based (75 to 9). Even excluding the 33 best brand statements, brand-based knowledge predominates (42 to 9) at these highest levels of inference. As would be expected, the difference in proportions of types of statements is statistically significant (x12(1) = 105.01, p < .005, excluding the 33 best brand statements.) Returning momentarily to the original question, there is relatively little main effect of attribute versus brand. Although the observed majority of attribute-based knowledge statements, 103 to 75, or 58%, is significantly above one-half (p < .05), the major finding is the interaction between inferential level and basis, not a main effect of basis. The observed interaction agrees quite well with the pattern of knowledge generated by a purchase decision process (Johnson and Russo, 1978a). According to this process, the brand-attribute values are immediately subjected to an attribute-based inference process that generates the first higher level knowledge (Level 4 in the present scheme). This is followed by a shift to brand-based inference as the brand-oriented goal, a first choice, is approached (Levels 3 and 2). This pattern suggests that if basic knowledge (Level 5) must be obtained from the external environment, all higher level knowledge (Levels 1-4) is generated internally, possibly in the normal course of repeated purchase decisions. Organization of Product Knowledge Besides asking about the content of retrievable knowledge, we are also interested in the organization of that knowledge. Specifically, are the individual knowledge nodes linked in memory by common attributes, common brands or neither? To answer this question we have available the transitions between contiguous coded statements. In order to assume contiguity, we included in this analysis those statements in which only a brand or attribute was mentioned without further elaboration. To exclude them would have left numerous gaps in the sequence of statements, forcing us either to reduce the data base radically or to pretend that two statements that were separated by an uncoded statement were really contiguous. Neither alternative seemed as reasonable as simply including the uncoded statements. The data indicate a clear predominance of brand-based (N=275) over attribute-based (N=75) transitions. The respective percentages are 45% and 12%. This pattern holds over all four product categories, with the ratio of brand to attribute transitions ranging between 3.1 for dinner entrees to 4.9 for headache remedies. Furthermore, and contrary to expectation, the two types of transitions are not reciprocal. As shown in Figure 1, product categories with higher proportions of brand transitions also enjoy higher proportions of attribute transitions. PROPORTION OF ATTRIBUTE-BASED AND BRAND-BASED RECALL TRANSITIONS The roughly 4 to 1 predominance of brand-based transitions raises the question: How are these data to be reconciled with the clear presence of attribute-based knowledge? The first thing this apparent conflict does is reinforce the distinction between content of knowledge and its structure. To explore the content-structure distinction further, we partitioned the transitions by inferential level. A transition was assigned to a given level if it either originated or terminated at this level. The proportions of transitions for each of the five inferential levels are shown in Table 5. At the highest levels, we expect few attribute transitions because there are so few attribute-based statements. The data for Levels 4 and 5, however, are very revealing. At each of these levels the transition can be either by brand or by attribute, because each statement must include at least one brand and one attribute. For the brand-attribute values of Level 5, which are the most neutral statements with respect to structural basis, brand transitions clearly predominate. The same dominance, although reduced, occurs for Level 4 statements. Recall that these knowledge propositions were exclusively attribute-based. Thus, the majority of brand-based transitions at Level 4 underscores both the distinction between content and structure and the dominance of brands in the organization of memory. Level 4 statements, though exclusively attribute-based in content, are linked primarily by brand in memory. RELATIVE FREQUENCIES OF TRANSITIONS BY INFERENTIAL LEVEL The question remains: Why is product knowledge linked by brand? Our data cannot provide a definitive answer. Only another study that traces the development of product knowledge over time can do that. However, we believe that the following occurs. As familiarity with a product category increases, two things happen. First, the amount of knowledge increases; and second, that knowledge is connected by an increasingly dense network of links. These linkages are created by all our product experience, including advertising, word-of-mouth communication, and actual usage, as well as by the purchase decisions that are responsible for higher level knowledge. The observed data are well explained if we assume that this product experience is brand-based and that it primarily creates linkages among existing knowledge rather than generating new knowledge. Thus, serving tuna fish to one's family may be accompanied by the thoughts, or even a conversation, that tuna is high in protein, not expensive, and excellent for sandwiches. We see such processing as creating or strengthening links among existing knowledge, rather than creating new knowledge. This is not to say that no new knowledge is created in this way, but that this is less often the case than the creation or strengthening of links among existing knowledge propositions. Summary The analysis of recall protocols reveals three major characteristics of consumers' knowledge of familiar products. 1. Product knowledge can be separated by level and source. The lowest level knowledge, brand-attribute values, accounts for most knowledge and probably its source largely in the external environment. Higher level knowledge results from inferences that may occur mainly during purchase decisions. 2. The attribute- versus brand-based knowledge controversy must be indexed by inferential level. The highest level knowledge is predominantly brand-based, while that at lower levels is overwhelmingly attribute-based. 3. The structure of knowledge, in the sense of the links among knowledge propositions, is primarily brand-based. This result emphasizes the distinction between the content and structure of knowledge. Future Research This initial study of product knowledge has generated as much speculation as confirmation. Especially needed are investigations of the development of product knowledge. How do people gain knowledge over time? Does knowledge at the higher inferential levels derive primarily from purchase decisions? Is the effect of other product experience confined mainly to the learning of brand-attribute values and the creation of a brand-based network of knowledge? We believe that these questions are amenable to realistic experimental studies. REFERENCES Anderson, John R. (1976), Language, Memory and Thought, Hillsdale, N. J.: Lawrence Erlbaum. Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Reading, Mass.: Addison-Wesley. Ericsson, K. Anders and Simon, Herbert A. (1978), "Retrospective Verbal Reports as Data," C. I. P. Working Paper No. 388, Carnegie-Mellon University. Johnson, Eric J. and Russo, J. Edward (1978a), "What Is Remembered after a Purchase Decision?" Technical Report, Center for Decision Research, University of Chicago. Johnson, Eric J. and Russo, J. Edward (1978b), "The Organization of Product Information in Memory Identified by Recall Times," in Advances in Consumer Research, Vol. 5, edited by H. Keith Hunt, Ann Arbor: Association for Consumer Research. Kintsch, Walter (1974), The Representation of Meaning in Memory, Hillsdale, N. J.: Lawrence Erlbaum. Newell, Alan and Simon, Herbert A. (1972), Human Problem Solving, Englewood Cliffs, N. J.: Prentice-Hall. Olson, Jerry C. (August 1978), "Use of Free Elicitation Procedures to Probe the Content and Organization of Knowledge Structures Stored in Semantic Memory." Paper presented at the Marketing Educators' Conference, American Marketing Association, Chicago. Olson, Jerry and Muderrisoglu, Aydin (1979), "The stability of Responses Obtained by Free Elicitation: Implications for Measuring Attribute Salience and Memory Structure," in Advances in Consumer Research, Vol. 6, edited by William L. Wilkie, Ann Arbor, Mich.: Association for Consumer Research. Payne, John and Ragsdale, E. K. Easton (1978), "Verbal Protocols and Direct Observation of Supermarket Shopping Behavior: Some Findings and a Discussion of Methods," in Advances in Consumer Research, Vol. 5, edited by H. Keith Hunt, Ann Arbor: Association for Consumer Research. Rip, Peter D. (June 1979), "The Extent and Basis of Insight in Decision Making: A Study in Consumer Behavior,'' unpublished doctoral dissertation, Stanford University. Russo, J. Edward (1977), "The Value of Unit Price Information,'' Journal of Marketing Research, 14, 193-201. Wright, Peter L. and Barbour, Frederic (1975), "The Relevance of Decision Process Models in Structuring Persuasive Messages," Communication Research, 2, 246-59. ----------------------------------------
Authors
J. Edward Russo, University of Chicago
Eric J. Johnson, Carnegie-Mellon University
Volume
NA - Advances in Consumer Research Volume 07 | 1980
Share Proceeding
Featured papers
See MoreFeatured
A1. Trusting and Acting on Chance Online
Shivaun Anderberg, University of Sydney, Australia
Ellen Garbarino, University of Sydney, Australia
Featured
Explaining the Attraction Effect: An Ambiguity-Attention-Applicability Framework
Sharlene He, Concordia University, Canada
Brian Sternthal, Northwestern University, USA
Featured
When Consumers Choose for Others, Their Preferences Diverge from Their Own Salient Goals
Olya Bullard, University of Winnipeg