Factors Lnfluencing Consumer Strategies in Information Processing

Klaus P. Kaas, [Professor of Marketing, Johann Wolfgang Goethe-Universitat. Mertonstr 17. 6000 Frankfurt/M,W.Germ.]
ABSTRACT - By means of data from an information monitoring experiment an investigation was made of the influence of three factors: the number of alternatives, number of attributes and degree of habituation, on consumer information processing. The two task-related factors have significant effects on the intensity and the structure of information processing. The degree of habituation has only a minor effect on the dependent variables. The results are discussed in light of theoretical considerations and previous empirical findings. The main focus is on the use of single or multistage information processing strategies. In conclusion the implications of this work for future research are outlined.
[ to cite ]:
Klaus P. Kaas (1984) ,"Factors Lnfluencing Consumer Strategies in Information Processing", in NA - Advances in Consumer Research Volume 11, eds. Thomas C. Kinnear, Provo, UT : Association for Consumer Research, Pages: 585-590.

Advances in Consumer Research Volume 11, 1984      Pages 585-590


Klaus P. Kaas [Professor of Marketing, Johann Wolfgang Goethe-Universitat. Mertonstr 17. 6000 Frankfurt/M,W.Germ.]

[The author gratefully acknowledges the financial support of the "Deutsche Forschungsgemeinschaft", Bonn-Bad Godesberg, W. Germany.]


By means of data from an information monitoring experiment an investigation was made of the influence of three factors: the number of alternatives, number of attributes and degree of habituation, on consumer information processing. The two task-related factors have significant effects on the intensity and the structure of information processing. The degree of habituation has only a minor effect on the dependent variables. The results are discussed in light of theoretical considerations and previous empirical findings. The main focus is on the use of single or multistage information processing strategies. In conclusion the implications of this work for future research are outlined.


Over the last decade research in the field of consumer information processing has made substantial progress. Data collection techniques have been developed, such as information boards and verbal protocols, and data analysis techniques such as transition indices and decision nets (see, for an overview, Bettman 1979). By means of these as well as other techniques, a wealth of knowledge has been gained on consumer information processing.

Unfortunately, this knowledge is not always clear-cut or consistent. If one thing is certain, it is that in multiattribute choice settings consumers make use of a whole range of information processing strategies. There are, for example, "(1) Single Stage versus Multistage Process Strategies, (2) Processing by Alternative versus Processing by Attribute, (3) Compensatory versus Noncompensatory Processing, and (4) Strategies using Attribute Weights versus those not using Attribute Weights" (Raju and Reilly 1980, p. 190). These are criteria that can be used to characterize numerous separate decision rules, such as the linear-additive rule, the conjunctive and disjunctive rule, the lexicographical rule, the additive difference rule and elimination by aspects to name but a few of them (see Bettman 1979).

In view of this multiplicity of real information processing strategies one problem of research must be lent greater importance: what are the factors that determine the choice of type of strategy? This question is of significance for anyone who wishes to explain, forecast, or influence consumer behavior, i.e. for marketing researchers and marketeers, for those interested in consumer policy, for psychologists and others. Factors that would come into consideration for an analysis can be grouped in different ways (Moore and Lehmann 1980). The simplest classification would be according to task-related and person-related factors. The influence of the first means that consumer information processing must be regarded as adaptive. The existence of the last set of factors signifies that consumers differ in this activity, as they also do in many other respects.

Task-related factors, such as the format and the amount of information presented seem to exert an important influence on consumer information processing (Payne 197,, Staelin and Payne 1976, van Raaij 1977, Bettman and Zins 1979, Lussier and Olshavsky 1979, Malhotra 1982). However, the influence of person-related factors such as perceived risk or cognitive complexity has turned out to be not as clear cut (Capon and Burke 1980, Moore and Lehmann 1980, Malhotra 1983). This is especially true for product similarity or product experience (Park and Shaninger 1976, Jacoby et al. 1978; Bettman and Park 1980; Moore and Lehmann 1980; Tan and Dolish 1981: Johnson and Russo 1981 )


The present study will endeavor to examine the influence of the following factors: number of alternatives (brands), number of attributes and the habit forming stage on consumer strategies in information processing by means of an information display experiment. In comparison with research thus far, it is hoped that progress will be made in various respects

Firstly, on the side of the dependent variables certain operationalizations were applied that should allow for a more precise identification of information processing strategies.

Secondly, unlike in most studies, a convenience sample was not used. The subjects were 360 young mothers with a first child, the product investigated was baby diapers. In this manner the external validity is increased.

Thirdly, great emphasis was placed on the influence of product familiarity as a variable. This variable was operationalized more stringently than in previous studies. Furthermore, product familiarity was varied at two levels, in addition to the number of alternatives and attributes, according to a 3x3x2-between-subject-factorial design. This enables an analysis to be made not only of the variable's main effect on the dependent variables but also of the interaction effect with task-related factors



The data was collected in the summer of 1981 by means of a version of the information board, presented open in interviews conducted in the homes of the subjects. The board was laid out according to the pattern used by the German consumer product testing agency. The subjects were asked to choose one of the brands of diapers whereby they could look at as many information units (cells) as they wished. Each cell was numbered and the subjects were asked to tell the interviewer the sequence of the numbers of the cells viewed (for details. see Kaas and Hofacker 1983)

Independent Variables

The number of alternatives (brands' was predetermined with the values 4, 8 and 16, the number of attributes as 5, 9 and 16. In this context the larger information boards contained the smaller ones as subsets. 40 mothers with a first child were assigned to each of the resulting 9 cells, i.e. each subject had to work through only one information board. 20 mothers had babies up to five and half weeks old. the babies of the remaining 20 mothers were between 22 and 33 weeks old. The cohorts of mothers formed random samples of the births registered in the City of Frankfurt, and were each randomly assigned to treatment cells.

This operationalization of the habit forming stage has three advantages. Firstly, it is based on an objective, rather than a subjectively reported experience of buying diapers. It guarantees, secondly, that other product-related variables are controlled. Thirdly, it is embedded in a wider theoretical context. According to Howard's ( 1977 ) theory the "young" mothers are in the phase of "Extensive Problem Solving Behavior" (EPS), whereas the "older" ones are in the phase of "Routinized Response Behavior" (RR3). This implies that they differ from one another in a whole series of variables (e.g. information seeking behavior, perceived risk). This was checked empirically and proved to be correct (details cannot be given here, cf. Dieterich 1983. Kaas 1982)

Dependent Variables

Numerous variables were calculated from the raw-data in order to characterize the information processing strategies. In this paper the following will be analyzed:

1. The intensity, measured by the amount of information units acquired and the duration of the process.

2. Five transition indices, which embrace the occurrence of

- processing by alternatives (brands) (J1)

- processing by attributes (J2)

- paired comparison processing (J3)

- structured shifts (J4), and

- unstructured (random processing (J5).

These indices are not based, as is the case in most previous studies (Payne 1976, van Raaij 1979, Weinberg and Schulte-Frankenfeld 1983) on paired transitions, but rather on triple transitions. The triple-transition-analysis naturally involves more effort, but possesses, however, important advantages (for details, see Hofacker, in this volume). Paired comparisons, such as occur in the additive difference rule, can be identified, and structurally-determined changes in alternative or attribute can be differentiated from the unstructured transitions.

3. Single stage and multistage information processing strategies. These were identified by means of an improved technique for establishing changes in strategy (for details, see Hofacker, in this volume). The sequence of a subject's transitions was investigated by means of a "moving window" for structural breaks. The window spanned a certain number of transitions. The indices mentioned above were calculated for each position of the window. The distances between two adjacent window positions could then be quantified by a Minkowski-Supremum-Metric and then interpreted as indicators of strategy changes.

4. Six strategy clusters. All the strategies found (n= 573) - global strategies and the individual phases of multistage strategies - were submitted to a cluster analysis on the basis of the index values pertaining. In this manner six "natural" strategy types were found (see Hofacker, in this volume):

- processing by attributes (lexicographical rule, elimination by aspects) (C1)

- processing by alternatives (long sequences, e.g. linear-additive rule) (C2)

- processing by alternatives (short sequences, e.g. conjunctive, disjunctive rule) (C3)

- paired comparison processing (additive difference rule . attribute dominance rule) (C )

- mixed processing (C5)

- random processing (C6)

The cluster types allow a more comprehensive characterization to be made of the phases as separate indices, because these types represent typical, theoretically meaningful "mixtures' of all five indices.

Sequence of the Results

In the following section the findings on the influence of both task-related factors will be presented first, after which that of the habit forming stage will be described. With respect to the dependent variables, the following sequence will be adhered to:

- Intensity (amount and duration),

- over all-indices, i.e. calculated for the transition sequences as a whole,

- number of strategy phases,

- single stage strategies, characterized by the transition indices and by the clusters, and

- multistage strategies characterized in the same manner, but separately for the first and second phases.


Task-Related Factors

Intensity: The influence of the complexity of the task on amount and duration of information acquisition is clearly in line with previous results (Payne 1976, Staelin and Payne 1976, Capon and Burke 1980). Both an increase in the number of alternatives as well as attributes leads to a significant augmentation of the amount of information acquired and to a lengthening of the decision time, whereas in comparison the respective relative measures significantly decrease (see Tables 1 and 2).



Over All-Indices: Firstly, the five indices were computed for the whole transition sequences which were not phased. The analysis of variance (ANOVA) demonstrated that with an ascending number of alternatives a significant decrease in processing by alternatives (J1) (a = 0.009) and of paired comparison (J3) (a = 0.009) occurred, whereas the processing by attributes (J?) became more frequent (a = 0.000). This confirms the findings of previous studies (Svenson 1974; Payne 1976; Payne and Braunstein 1976). This pattern can be interpreted as follows. The consumers master an increasing task complexity by changing from rather more selective to more eliminative processing. This point will be returned to below, in the phase analysis.

Theoretical arguments for a similar effect in the case of an increasing number of attributes are not as convincing. Nevertheless, Staelin and Payne (1976) found a decrease in the variance of attributes per alternative, i.e. of compensatory procedures. In our data neither the main effect of the number of attributes nor the interaction effect with the number of alternatives are of significance for any transition index.

Number of Phases: In the random sample as a whole of the 356 subjects who could be evaluated, 58% showed a single stage strategy, 29% a two stage and 13% a strategy with more stages (cf. Hofacker, in this volume, Table 2). A whole row of variables were correlated with the number of the phases in a manner to be theoretically expected, so that the construct validity of the phase structure is very high (cf. Hofacker 1983 for particulars). For example, in those cases where two stage strategies occurred in the first phase processing by attribute and in the second phase processing by alternative were seen to dominate. In the second phase above all compensatory rules and the paired comparison were found to exist. How does this picture change under the influence of the numbers of both alternatives and attributes? The analysis of variance indicated as expected that the mean number of phases grew significantly with both factors (cf. Table 3).


The effect of the alternatives is clearest here as with respect to the other results mentioned. This is also shown in the frequency analysis. An increasing r.umber of alternatives is accompanied by a continuous decrease in the number of single stage strategies, while two stage, three stage and multistage strategies - each individually - increase just as uniformly. This is only true for the attributes if all the strategies with two or more phases are collected together in one group (see Table 4).


The results presented accord with those of previous work by Lussier and Olshavsky ( 1979 ) and Malhotra (1982). The latter discovered, however, that an increasing number of attributes was connected with a tendency to single stage information processing. His findings, nevertheless, contradict his own hypothesis, one based on sound argumentation. He himself explains it as having been "due to a built-in feature of the algorithm" that was used.

Single Stage Strategies: A similar result of the transition analysis is obtained within the single stage strategies (n=206) as was gained by an undifferentiated view of the whole random sample: the consumers process by attribute (J2) significantly more frequently (a = 0.020) with an ascending number of alternatives, whereas processing by alternative (J1) distinctly recedes (t = 0.044). Simultaneously the transition sequences become longer (a = 0.001). Conversely, the number of attributes has no influence on any of the five transition indices, only the length of the strategies increases in the same way (a = 0.029). No interaction effect occurs.

A second way of characterizing the information processing strategies would be to examine the frequency of strategy Types C1 to C. A crosstabulation revealed a significant relation neither with respect to alternatives nor to the attributes. Relations within the data only became evident after the formation of starker contrasts. In the collated group with 4 and 8 alternatives Type C2(linear-additive rule) dominates over Type C2 (lexicographical rule, elimination by aspects), while in the group with 16 alternatives the opposite is the case (n = 123, x2 a = 0.039, C = 0.202).

With respect to the cluster types, the change is not as pronounced as it is in the indices. This was to be expected for a change in one single index does not necessarily cause a switch to another cluster type.

Two Stage Strategies: The main effects and interactive effects of the task-related factors on the first and second phases of the transitions are not significant except for in one case. With an increase in the number of alternatives the number of paired comparisons (J3) in the first phase recedes (a = 0.043). With regard to the clusters there is no significant result.

At first sight these results would seem to contradict the theoretical expectations. One must not forget, however, that the two stage phases in the random sample as a whole show roughly the expected structure - namely the 1st Phase: Elimination, the 2nd Phase: Examination of Details (cf. above and Hofacker, in this volume), and that consumers switch to multistage processing when task complexity increases. Taking into account, that the structure of multistage processing is almost inaffected by the task complexity, only one conclusion can be drawn: Increasing task complexity leads to more multistage processing - but there is no one specific phase sequence for small information boards and one for large information boards.

Habit Forming Stage

Intensity: EPS and RRB consumers do not differ in the intensity of information processing (see Tables 1 and 2). This finding accords with a hypothesis by Bettman and Park (1980). They suggest that consumers with a medium amount of experience (LPS, let out in our experiment) search for a lot of information. Very familiar (here: RRB) consumers and ones very unfamiliar (here: EPS) both search for little information. The former because they have no need of the information, the latter because they have too little experience. This explanation is rather speculative. There are, after all, other hypotheses. Johnson and Russo (1981) propose the "rich-get-richer" hypothesis, i.e. the more experienced the consumers are the more they can make use of additional information. Jacoby, Chestnut and Fisher (1978) provide a counter-hypothesis (similar to Moore and Lehmann 1980). They suggest that consumers need the less information the more experienced they are. Their empirical results, however, support the "rich-get-richer" hypothesis. Clearly this question needs to be researched much further

Over All-Indices: The transition indices do not indicate significant differences between the EPS and RRB sample if the underlying transition sequences were treated as a whole.

Number of Phases: The influence of the habituation degree on the number of information processing phases is shown in Table 3. In the RRB sample the middle number of phases is greater, but the difference is not significant. The frequency analysis indicates that in the RRB sample all strategies with more than two phases are more frequent. If one groups the single stage and two stage strategies together, then the relation is significant, but not very strong (see Table 4).

A further result is that the interaction effect of habit forming stage x alternatives is significant. With an increasing number of alternatives the number of phases in the RRB sample increases more greatly than in the EPS sample (a = 0.018).

Single Stage Strategies: The habit forming stage has no significant effect on the dependent variables (indices and clusters).

Two Stage Strategies: The same holds true for both phases.

These results are somewhat hard to interpret in light of previous findings. Bettman and Kakkar (1977) as well as Arch, Bet.man and Kakkar (1978) found in experienced shoppers a tendency towards more processing by alternative. Jacoby et al. (1976) supported the thesis that a high usage rate would lead to processing by alternative. Park and Shaninger (1976) found that familiar consumers used a weighted rather than an unweighted linear-additive model. Tan and Dolich (1981) only found the same results as in this article - by means of a regression approach: no clear influence of familiarity on the form of information processing could be ascertained!

Raju and Reilly (1980) studied whether product familiarity caused single or rather multistage strategies. They ascertained: "There are no strong reasons to believe that usage of one stage or two stage process will itself be influenced by product familiarity (p. 193). Their own results confirm this statement as do those presented in this paper.

On the other hand, Raju and Reilly's (1980) data shows that product familiarity has an influence on the type of decision rule adopted in the second phase of the two stage processes. Such differences could not be established in the present study.


Task Complexity

Previous findings on the subject are partly confirmed, partly rendered more precise and partly supplemented by this study. The influence of task complexity on the intensity and duration of information processing belongs to the first group and need not be commented on further.

The transition indices, computed as over-all indices without accounting for phased strategies, indicate that consumers obviously master the growing task complexity by proceeding from rather more selective decision rules (e.g. linear-additive, maximax- and maximin-rules, paired comparisons) to rather more eliminating rules (e.g. lexicographical rules, elimination by aspects). This finding, which agrees with previous studies (e.g. Svenson 1974, Payne 1976, Payne and Braunstein 1376), is, however, only the tip of the iceberg. The analysis of multi stage processes shows that subjects adjust themselves to the increasing number of alternatives in highly differing ways:

1. Well over half of them retain a single stage processing strategy, which is, however, altered both quantitatively and qualitatively, just as is the case in a global view of the whole random sample. These consumers master the wealth of information by proceeding throughout in a rather more eliminating manner. The question as to the cause of this remains unanswered. Can they solve a complex problem without dividing it into sub-problems? Or do they just muddle through, being unable to figure out a simplified, that is a multistage strategy?

2. Almost half the consumers change over to multistage strategies, whereby two stage strategies form by far the largest proportion of these. These strategies contain, as is to be expected, a first phase consisting of a rather more eliminating procedure, and a second phase of rather more comparative, selective processing. Almost nothing changes in this picture if the task complexity increases. Only the paired comparisons by attribute become rarer in the first phase. This is not surprising, for this decision rule is obviously not suited to a crude, rapid elimination process.

3. A third possible cause is not given, namely that an adaptation within the multistage strategies occurs. This points to the fact that the range of empirical multistage information processing patterns cannot, after all, be very large (cf. Raju and Reilly 1980, p. 207).

Asymmetry between Alternatives and Attributes

The fact that the effect of the number of alternatives is continuously more clearly pronounced than is that of the attributes agrees with our knowledge hitherto. There are convincing explanations for this. After all, the task is not "choose an attribute!", but "choose an alternative!". An additional alternative alters the decision situation, whereas an additional attribute only changes the level of information. In the first case there is a chance of improving the best possible choice, in the second merely the opportunity of realizing the best possible choice. To this extent it is hardly surprising that consumers react with more sensitivity to a change in the number of alternatives than they do to an alteration of the number of attributes

In this context, the result of the number of phases in multistage strategies is also worthy of note. An increasing number of alternatives leads to a decrease in single stage processes and a uniform increase in all the multistage processes. On the other hand, an increase in the number of attributes causes more of a shift from single or two stage processes to ones with three or more. The first point indicates the occurrence of a successive, usually two stage narrowing of the decision problem. This is supported by the concept of the evoked set (Howard 1977) , but is also clear from the setting of the task: at the end only one sinGle alternative is to be chosen!

This consideration can in no way be transferred to a deliberation on the attributes: at the end one has to (nolens volens) choose a whole set of attributes! The pattern according to which consumers adapt to a growing number of attributes is therefore not as clear as is the case with the alternatives.

Habit Forming Stage

The influence of the degree of habituation presents no clear picture. Further research is necessary here. The main difficulty lies in the fact that hardly any clear hypotheses can be formulated. As we mentioned above, two opposing tendencies are possible. On the one hand experienced consumers (RRB) possess better prerequisites for a more efficient processing of information: better previous knowledge, a prestructured decision pattern etc. On the other hand, precisely these prerequisites mean that they do not need to process in such a manner. To put it somewhat over-finely: experts can do better, but do not need to, beginners need to do better, but cannot.

Added to this there are validity problems. One cannot exclude that in an information monitoring experiment those consumers possessing previous knowledge only acquire numerous information units in order to check whether these correspond with the information they have stored (Arch, Bettman and Kakkar 1978) . There is another danger in that the task to be solved may break up the habitual behavior of the consumers, e.g. so that they are lead to revise their product class concept. This could be explained by means of such an approach as "psychology of complication" (Sheth and Raju 1973) or "variety-seeking behavior" (Raju 1983) .

If any conclusion can be drawn from our data, it is that the first of the above-mentioned tendencies, perhaps reinforced by the task-induced artifacts, is dominant. The experienced consumers tend more than novices to use multistage processes, a feature that becomes more pronounced with the occurrence of an increasing number of brands.

Further Research

Further research would be necessary to solve the problems outlined. Above all the influence of prior knowledge on information processing should be investigated in greater detail. First steps in this direction have been taken. Marks and Olson (1981) have investigated the relation between product familiarity and the cognitive structure of prior knowledge. In an article that appeared very recently, Chi (1963) proves in a very impressive manner that these relations are much more complex than has hitherto been assumed in consumer research. Thus, experts would appear not to have better problem solving techniques than novices. They just make better use of them. Seen from this angle, the somewhat 'lean' results of this study with regard to the factor of degree of habituation are not surprising.


Arch, David C., Bettman, James R. and Kakkar, Pradeep (1978), "Subjects' Information Processing in Information Display Board Studies," in Advances in Consumer Research, Vol. 5, ed. H. Keith Hunt, Ann Arbor: Association for Consumer Research, 555-560.

Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Reading: Addison-Wesley.

Bettman, James R. and Kakkar, Pradeep (1977), "Effects of Information Presentation Format on Consumer Information Acquisition Strategy," Journal of Consumer Research, 3, 233-240

Bettman, James R. and Park, C. Whan (1980), "Effects of Prior Knowledge and Experience and Phase of the Choice Process on Consumer Decision Processes: A Protocol Analysis," Journal of Consumer Research, 7, 234-248

Bettman, James R. and Zins, Michel A. (1979), "Information Format and Choice Task Effects in Decision Making," Journal of Consumer Research, 6, 141-153.

Capon, Noel and Burke, Marian (1980), "Individual, Product Class and Task-Related Factors in Consumer Information Processing," Journal of Consumer Research, 7, 314-326.

Chi, Michelene T.H. (1983), "The Role of Knowledge on Problem Solving and Consumer Choice Behavior," in Advances in Consumer Research, Vol. 10, ed. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor: Association for Consumer Research, 569-571.

Dieterich, Michael (1983), "Habitualisierungsphasen beim Kauf von Konsumgutern," Unpublished Working Paper, Frankfurt: Johann Wolfgang Goethe-Universitat.

Hofacker, Thomas (1983), "Entscheidung als InformationsverarbeitungsprozeB - dargestellt anhand der Verarbeitung von Produktinformationen durch Konsumenten," Doctoral Dissertation, Johann Wolfgang Goethe-Universitat Frankfurt am Main (West Germany), in Press.

Howard, John A. (1977), Consumer Behavior: Application of Theory, New York: McGraw-Hill.

Jacoby, Jacob, Chestnut, Robert W., Weigl, Karl C. and Fisher, William (1976), "Pre-Purchase Information Acquisition: Description of a Process Methodology, Research Paradigm, and Pilot Investigation," in Advances in Consumer Research, Vol. 3, ed. Beverlee B. Anderson, Ann Arbor, Association for Consumer Research, 306-314.

Jacoby, Jacob, Chestnut, Robert W. and Fisher, William A. (1978). "A Behavioral Process Approach in Nondurable Products, Journal of Marketing Research, 15, 532-544.

Johnson, Eric J. and Russo, 3. Edward (1981), "Product Familiarity and Learning New Information," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research. 151-155

Kaas, Klaus P. (1982), "Consumer Habit Forming, Information Acquisition, and Buying Behavior," Journal of Business Research, 19, 3-15.

Kaas, Klaus P. and Hofacker, Thomas (1983), "Informationstafeln und Denkprotokolle - Bestandsaufnahme und Entwicklungsmoglichkeiten der ProzeBverfolgungstechniken," in Innovative Marktforschung, ed. Forschungsgruppe Konsum und Verhalten Wurzburg: Physica.

Lussier, Denis A. and Olshavsky, Richard W. (1979), "Task Complexity and Contingent Processing in Brand Choice," Journal of Consumer Research, 6, 154-165.

Malhotra, Naresh K. (1982), "Multi-Stage Information Processing Behavior: An Experimental Investigation," Journal of the Academy of Marketing Science, 10, 54-71.

Malhotra, Naresh K. (1983), "On Individual Differences in Search Behavior for a Nondurable'" Journal of Consumer Research, 10. 125-131.

Marks, Larry J. and Olson, Jerry C. (1981), "Toward a Cognitive Structure Conceptualization of Product Familiarity," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research, 145-150.

Moore, William L. and Lehmann, Donald R. (1980), "Individual Differences in Search Behavior for a Nondurable," Journal of Consumer Research, 7, 296-307.

Park, C. Whan and Shaninger, Charles M. (1976), "The Identification of Consumer Judgmental Combination Rules: Statistical Prediction vs. Structured Protocols," in Advances in Consumer Research, Vol. 3, ed. Beverlee B. Anderson, Ann Arbor: Association for Consumer Research, 184-190.

Payne, John W. (1976), "Task Complexity and Contingent Processing in Decision Making: An Information Search and Protocol Analysis," Organizational Behavior and Human Performance, 16, 366-387.

Raaij, W. Fred van (1977), Consumer Choice Behavior: An Information Processing Approach, Voorschoten (N.L.).

Raju, P.S. (1983), "Measurement and Modeling of Variety-Seeking Behavior: Observations and Implications," in Advances in Consumer Research, Vol. 10, ed. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor: Association for Consumer Research. 84-87.

Sheth, Jagdish N. and Raju, Puthankurissi, S. (1973), "Sequential and Cyclical Nature of Information Processing Models in Repetitive Choice Behavior, " in Advances in Consumer Research, Vol. 1, ed. S. Ward and P. Wright, Association for Consumer Research, 348-358.

Staelin, Richard and Payne, John W. (1976), "Studies of the Information-Seeking Behavior of Consumers", in Cognition and Social Behavior, ed. J.S. Caroll and J.W. Payne. Hillsdale (N.J.). 185-202.

Svenson, Ola (1974), "A Note on Thinking Aloud Protocols obtained During the Choice of a Home," Working Paper, University of Stockholm.

Tan, Chin Tiong and Dolish, Ira J. (1981), "The Moderating Effects of Cognitive Complexity and Prior Product Familiarity on the Predictive Ability of Selected Multiattribute Choice Models for Three Consumer Products," in Advances in Consumer Research, Vol. 8, ed. Kent B. Monroe, Ann Arbor: Association for Consumer Research, 140-144.

Weinberg, Peter and Schuite-Frankenfeld, Helmut (1983), "Informations-Display-Matrizen zur Analyses der Informationsaufnahme von Konsumenten," in Innovative Marktforschung, ed. Forschungsgruppe Konsum und Verhalten, Wurzburcm Phvsica.