Measuring Consumers' Information Acquisition and Decision Behavior With the Computer-Based Information-Display-Matrix

Acknowledgement: The authors would like to thank the two anonymous reviewers as well as the editors for their help and valuable comments.

Consumer research has a wide range of methods at its disposal. In view of the fast pace of change in the conditions under which consumption takes place, these consumer research methods must be continuously reevaluated. The process-tracing technique called information-display-matrix (IDM) has been applied predominantly to investigate information acquisition behavior as part of research into the basic premises of the discipline in the 1970s and 1980s. However, the number of studies on this method, and the number using it, has remained small albeit steady since this period. In the recent past, the IDM has been utilized in a series of studies published after 2000 (e.g., Andersson, 2004; Huneke, Cole, & Levin, 2004; Newell, Weston, & Shanks, 2003; Weenig & Maarleveld, 2002), while it has been developed further by some (e.g., Anderson, 2001; Jasper & Shapiro, 2002; Johnson & Willemsen, 2004). One barrier to the more widespread use of the method, especially for applied research, might be that it has been criticized sharply with respect to the external validity of results, by reason that the information acquisition context is regarded as too artificial (Kroeber-Riel & Weinberg, 2003, p. 286).

There are good reasons to believe that IDM has the potential to increase in standing because the research environment has changed. While the method is becoming more alike to consumer choice situations, the consumer choice surrounding becomes more alike to the method as well. On the one hand, the technological developments have made an improved and more realistic research design form possible which allows a better simulation of the everyday consumer choice context, with further progress to come (Mühlbacher & Kirchler, 2003, p. 148). On the other hand, consumers today are far more familiar with the technology, especially computer hard- and software, involved in IDM research designs. It is particularly noticeable in this regard that the widespread use of the Internet has altered consumers’ information acquisition behavior (Ratchford, Talukdar, & Lee, 2007, p. 111). Interestingly, the information presentation in the IDM is comparable to the manner in which information is presented on the Internet (Peterson & Merino, 2003, p. 106). As online shopping is becoming frequent (Anderson, 2001, p. 234), precisely those businesses which have an active and popular homepage (Bauer, Grether, & Sattler, 2002, p. 265) might well show renewed interest in IDM with a view to practical research.

The aim of the paper at hand is to familiarize the reader with the method, the course of its development, as well as its advantages and disadvantages in comparison to other methods of process-tracing. In particular, it aims to show where the current disadvantages of the method might be alleviated, and where approaches for the future development of the method might be found against a background of technological progress and changing consumer behavior. In order to demonstrate the possibilities offered by IDM for analysis, this paper will close with results from a study about information acquisition behavior of different groups of fair-trade coffee consumers. The data for this study were obtained by using a newly developed, computer-based IDM-application. The major contribution of the paper is that it gives a uniquely comprehensive description of the method – up to this point frequently mentioned, but insufficiently explained in most textbooks and articles – and that it summarizes the starting points for its further development and use in a changing research environment.

A helpful way to explain the central principle behind IDM is to refer to the tables and matrices used in magazines produced by consumer organizations. Just as with this type of publication, each column is headed by a product (also called product alternative, stimulus, or option in IDM terminology), either named or imaged, and each row is led by a characteristic of the product (also known as attribute, property, or dimension in IDM). An example of an IDM is shown in Figure 1. The fields (or cells) in the matrix itself house the salient characteristics of the product in question. The fields are blank, thus, the information is hidden. Participants in an IDM research study must choose a product on a computer or laptop and select the information relevant to their decision from the table. This means that the test persons must actively access the information in the matrix: Opening the information fields in the table is done by keyboard, mouse, or touch screen. All activity on the computer is saved, allowing researchers to see which and how many data fields were accessed, and in which order they were called up. Furthermore, the length of time required by the subject and the choice made by the subject are also recorded. This information can then be used to arrive at conclusions about the cognitive processes underlying search, judgment, and choice.

This description makes clear that the method in question is an observational method which aims to document information acquisition behavior. It can thus be classified as a process-tracing method. The most important process-tracing methods in consumer research, beside IDM itself, include: direct observation, eye-tracking, and verbal or think-aloud protocols (Jacoby, Jaccard, Kuß, Troutman, & Mazursky, 1987, p. 156f; Kroeber-Riel & Weinberg, 2003, p. 282; Silberer, 2005, p. 264; Trommsdorf, 2004, p. 272; Zielke, 2001, p. 100ff).

The opposite of process-tracing techniques is, in terms of method, structural modeling. Using this method, the process itself is not observed: Instead, variations in the starting point for the experiment are compared to final decision outcomes; from this, hypotheses about information acquisition are then elaborated (Covey & Lovie, 1998, p. 57; Ford, Schmitt, Schechtman, Hults, & Doherty, 1989, p. 75; Harte & Koele, 1995, p. 49). Developed as a response to deficiencies in structural models (Ford et al., 1989, p. 76; Harte & Koele, 1995, p. 49; Jacoby et al., 1994, p. 291), process-tracing methods were also designed to remedy problems with direct surveys into the process of information acquisition (Felser, 2007, p. 463ff; Jacoby, Chestnut, & Fisher, 1978, p. 532; Kroeber-Riel & Weinberg, 2003, p. 282).

Payne and Jacoby were the first to use information-display-matrices in their work in the USA at the end of the 1970s and beginning of the 1980s (Ford et al., 1989, p. 76; Jacoby et al., 1987, p. 150). The outdated English term information-display-board (IDB) still implies that the information was, at first, attached to a board. This information was rendered accessible by having cards with writing on both sides – the individual fields in the matrix – which could be turned over, or cards which could be removed from envelopes. The focus of this early research was gathering data about the influence of various factors on decision-making processes and the information acquisition strategies employed during this process. In particular, the varying complexity of the task was looked into (Ford et al., 1989, p. 83). The results showed, among other things, that relatively little information was used and that the research subjects limited themselves to key information. This led to further discussion of and research into the phenomenon of information overload for consumers (Hwang & Lin, 1999; Kroeber-Riel & Weinberg, 2003, p. 286; Meyer, 1994, p. 306).

In the early 1980s, researchers started to use computers to display the information and then record information acquisition behavior in process-tracing exercises. Operationally, researchers made use of a variety of solutions for a computer-based IDM: A major difference in these solutions was between matrices shown on the screen (classic IDM) and applications offering pull-down menus for subjects to choose the information (menu-based IDM), as described by Jacoby et al. (1987, p. 151). The first computer-based and classic IDM-application was an MS-DOS program known as MouseLab (Jasper & Shapiro, 2002, p. 365; see Johnson, 1996 as well) written by the research group gathered around Johnson and Payne. This program was subsequently much used in research and further developed (“MouseTrace”, Jasper & Shapiro, 2002; “MouseLabWeb”, Johnson & Willemsen, 2004). Other developments in computer-based IDM displayed the menu-based approach (“P1190”, Andersson, 2001; “ComputerShop” used as in Huneke et al., 2004). The opportunities for research with the IDM-method presented by the Internet have been explored in three more recent studies (Mühlbacher & Kirchler, 2003; Roßmanith, 2001; Scholderer, Hagemann, Sorensen, & Czienskowski, 2007).

Researchers applying the IDM made use of several variations of the method in order to adapt the research design to the questions in the study at hand. In the early studies, the size of the matrix was varied by changing the number of product alternatives and characteristics listed (Kroeber-Riel & Weinberg, 2003, p. 285). Another possibility for altering the design was the number of fields that could be kept open simultaneously. While Payne, for example, left all fields open once that they had been accessed, Jacoby et al. (1987, p. 150) limited the options to one field at a time. Schopphoven (1996, p. 178), on the other hand, allowed one fifth of the fields at a time to be open. In other studies, the information offer was restricted, either by setting a time-limit on the viewability of the matrix (Weenig & Maarleveld, 2002, p. 700) or by subtracting sums of money from the participants’ final remuneration based on the amount of information they accessed (Newell et al., 2003, p. 85). A phased decision was permitted by Jasper & Shapiro (2002, p. 365), where subjects were allowed to first select a consideration set before making their final decisions. Finally, the way in which information is presented has varied in IDM studies in that symbols were used instead of the usual text (Verplanken & Weenig, 1993).

Criticism of IDM is primarily directed at questioning the external validity of the method. A central criticism in this regard is that the presentation of separate pieces of information, out of context and only available in a chronological order, does not correspond to consumer reality, where much information is absorbed simultaneously. Often, too much information is offered (Kroeber-Riel & Weinberg, 2003, p. 283; Marten, 1992, p. 131) and the information already accessed must be mentally retained by the subjects (Schopphoven, 1996, p. 177). Another fundamental criticism is that information offered in IDM is preselected and the number of separate pieces of information available is known from the start; this means that the decision-making problem is already structured for the subjects (Biggs, Rosman, & Sergenian, 1993, p. 190; Brucks, 1988, p. 118; Büttner, Rauch, & Silberer, 2005, p. 1; Jacoby et al., 1987, p. 154; Williamson, Ranyard, & Cuthbert, 2000, p. 204). Furthermore, all information presented is equally accessible and often presented only in text-form (Jacoby et al., 1987, p. 154), causing an overly rational approach to decision-making behavior in IDM-situations (Kroeber-Riel & Weinberg, 2003, p. 283). The information depiction found in the matrix itself also influences the information acquisition strategy that researchers are seeking to study, inasmuch as information is usually processed in the way it is offered (Bettman & Kakkar, 1977, p. 233; Kleinmuntz & Schkade, 1993; Lehmann & Moore, 1980, p. 450; Trommsdorf, 2004, p. 312).

Criticism goes further in stating that some aspects crucial to information processing cannot be measured by IDM: The method cannot give a sure answer to the question of whether the information accessed is actually cognitively processed or not (Mühlbacher & Kirchler, 2003, p. 149; Williamson et al., 2000, p. 204). In addition to this, it is not possible to document whether and what kind of effect internal information on the part of the subjects has on the decision process (Lehmann & Moore, 1980, p. 450; Schopphoven, 1996, p. 160). Finally, participants in IDM studies are aware of the fact that these data are being collected; this can lead to reactivity (Jacoby et al., 1987, p. 154).

Some of these criticisms, especially those directed at external validity, are true of other process-tracing methods such as eye-tracking and think-aloud protocols as well. Direct observation is the least likely of process-tracing methods to exert an influence on the behavior that is being examined; its one major disadvantage, however, and the reason for its limited use is the comparatively meagre yield of information for the purpose of the research (Kroeber-Riel & Weinberg, 2003, p. 283). We shall thus not discuss this method further.

In comparing the main advantages and disadvantages of the three process-tracing methods, it must be noted that the methods concentrate on different points of the information acquisition process: While eye-tracking traces the accession of all information – including information viewed in a focused, although not necessarily attentive manner – IDM only documents information that has been purposefully accessed. Think-aloud protocols, in turn, can only deal with information which has been internalized by the subjects and can thus be verbalized.

As opposed to IDM, eye-tracking is able to offer subjects a very realistic, holistic depiction of the information in question (e.g., on a single poster or actual packaging), as well as to present a situation in which information acquisition is done visually and with little extra effort. The disadvantages of this method are the high costs, the cumbersome technology involved (Büttner et al., 2005, p. 1), and the fact that the subjects might take badly to the equipment they have to wear on their heads.

Think-aloud protocols are, in comparison both to eye-tracking and IDM, far less complicated and expensive, requiring nothing more than an audio recording device. This also makes the research more mobile and flexible, that is, the information can be depicted in a real-life situation, such as at the actual point of sale. Further, only this method allows researchers to track the use of the information after it has been acquired and includes analysis of the use of internal information. A disadvantage of this method is that the evaluation is time-consuming and difficult (Biggs et al., 1993, p. 189; Kroeber-Riel & Weinberg, 2003, p. 284). In addition, many suspect that the process of verbalization has a heavily distorting influence on information behavior (Silberer, 2005, p. 264f).

Due to enhanced technological possibilities, increased use of information technology, and changes in consumption patterns in the last decade, the information search in the IDM has moved closer to the actual information acquisition and purchasing situations it seeks to simulate; this is especially true in terms of Internet shopping. The central criticism of IDM – that is, concerns about its external validity – is thus becoming increasingly weak. Furthermore, improvements can also be achieved through combining IDM research with other methods. Possibilities for further development are described in the following:

1. Location: Using laptops, IDM research can take place at the point of sale, which is the actual information search and decision context, and in some cases, directly before the subject wants to purchase the product in question. The research can also take place over the Internet, and indeed this has already occasionally been done. The great advantages of this way of carrying out the research are the low human resources costs and the opportunity of reaching a large sample from one central location. Rather less advantageous is the fact that analysis of timed results is, due to the varying speeds and capabilities of different computers, only of limited value; neither can the conditions of the research (interruptions, collaboration with other people) be strictly controlled. In addition, it should be remembered that people with Internet access are not demographically and socially representative of the population as a whole (Peterson & Merino, 2003, p. 107).
2. Attractiveness and relevance: If possible, the product – and indeed, ideally, the alternative selected – can be handed out or sent to the participant after the simulated purchase in order to increase both the attractiveness of participation and the relevance of the choice. This purpose can additionally be served by linking the price of the product chosen to the remuneration offered to the participants. Efforts made in the development of stated choice-tests to move close to real-life situations are of relevance here (Alfnes, Guttormsen, Steine, & Kolstad, 2006; Jacoby et al., 1987, S. 155; Lusk & Schroeder, 2004).
3. Individualized structure: The process through which the research subjects are asked to go can be adapted to the product concerned by offering staged decisions where preliminary selections can be made (Jasper & Shapiro, 2002, p. 365) or can be chosen freely by the participants. Another option is to replace decision stages with progressive elimination of product alternatives in the selection. The advantage of this is that it allows researchers to note which precise piece of information leads to a particular product being ruled out. Another method suggested by Jacoby et al. (1987, p. 155) is an individualized IDM; this takes account of the criticism of IDM based on overly stark predefined informational structures: Before the IDM-survey is carried out, each individual participant is asked which of the product attributes are important to him or her, so that only these criteria are then offered in the matrix. This does, however, have the disadvantage of making it difficult to compare some of the indicators of analysis when the matrices themselves are of varying size.
4. Layout and navigation: A wide variety of software applications allow IDM to approach real-life in terms of optics and user-friendliness. The layout can be fitted to product test reports or to online shops; it is also possible to reproduce the shelving order of the products in the shops on the screen. The criticism that in an IDM, the whole picture of the product is not given can be tackled by offering slightly blurred graphical reproductions of the whole product-as-is on screen. Participants are then offered the option of clicking on certain pieces of information (brand, price, product test results, ingredients lists, etc.) and reading the content in a bigger field (Scholderer et al., 2007, p. 10). In principle, any pictorial information offered in the real-life situation can also be in this form in the IDM (e.g., brand name and quality labels). By employing such techniques and easy, logical navigation, it is entirely legitimate to assume that the framework of the research situation slips into the background and does not exercise a particularly disruptive effect.
5. Combining methods: The criticism that IDM is incapable of recording important aspects of research (e.g., the actual processing of information, the use of internal information) can be tackled by using a combination of methods. Using, for example, retrospective, perhaps even video-recorded (Silberer, 2005) verbal protocols, or questioning subjects after the research about the information they have acquired and the state of their knowledge regarding information not depicted in IDM is therefore an advisable addition in this context. Since most IDM-measures are scaled in ratios or intervals, it is quite practicable to combine information from IDM with variables from other methods, especially questionnaires, when calculating structural models (Harte & Koele, 1995, p. 50; Jasper & Shapiro, 2002, p. 371). Finally, it should be remembered that IDM research culminates in a product being chosen and this decision being documented. For this reason, an additional analysis of the decision in the manner of a choice experiment is also possible (see e.g., Louviere, Hensher, & Swait, 2000; Tutz, 2000). As a matter of fact, most studies in the wide field of choice experiments lack information about the extent of the information search phase preceding choice: This could be alleviated by integrating a process-tracing method, such as the IDM, into the research design.

Usually, the indicators used in IDM-analysis are divided into measures of depth, content, and sequence (see Table 1; e.g., Ford et al., 1989, p. 81; Jacoby et al., 1987, p. 151f; Jasper & Shapiro, 2002, p. 370). These indicators are in absolute or relative numbers and are either given for individual participants or describe the average of the sample. The depth measures are useful in answering the question of how extensively the subjects attempted to acquire information, or indeed to what extent the information offered was capitalized upon. It allows researchers to deduce the importance that subjects attach to a good final decision as opposed to the opportunity costs implied in extensive information search behavior: “the decision maker will trade off effort and accuracy, seeking a decision of acceptable quality for an acceptable level of effort” (Huneke et al., 2004, p. 67f). If an extensive information search is undertaken, this might indicate a higher level of involvement on the part of the subjects (see Mittal, 1989, for a discussion of this). Furthermore, the depth of the information search is used to deduce the type of heuristic applied in making the decision: When a large amount of information is used, it is usually posited that subjects are operating according to a compensatory decision-making strategy (Ford et al., 1989, p. 81). Measuring the depth of information search by means of, for example, process-tracing methods has played an important role in the paradigmatic change away from the “homo oeconomicus” and to a differently involved consumer, because it could be shown that consumers did not necessarily access as much information as needed for a rational decision.

The content measures are used to quantify the relative weight given to the various types of information on offer, as expressed by the differences in information acquisition behavior registered between different types of information. They are useful in answering the question of how important different pieces of information are in deciding in favor of certain products, and allow researchers to pinpoint crucial key information. The underlying assumption is that information which is accessed earlier and more frequently has a comparatively higher importance (Schopphoven, 1996, p. 161; Trommsdorf, 2004, p. 311). The measures of content are useful in the broad area of analysis of consumer preferences for different brands, price-levels, labels, quality seals, etc.

Analyzing measures of sequence means examining the structure of the information acquisition process involved. Such analyses can show, for example, to what extent decisions are taken by comparing product alternatives (also called by-brand) and to what extent decisions are taken by comparing attributes of the products themselves (i.e., by-attribute); they can also show how frequently and indeed at which points in the experiment the subjects switched from one strategy to the other. Knowing the type of information strategy applied enables the researcher to infer which decision heuristic is used by the consumer when buying the product in question (Schopphoven, 1996, p. 162). If, for example, there is less variability in the number of attributes accessed per product, then it is usually assumed that a compensatory search strategy is being used (Ford et al., 1989, p. 81), and vice versa. Knowledge about the processing strategy can offer clues about the appropriateness of information type, content, and positioning. When information acquisition proceeds along noncompensatory lines, for example, the information has to be placed so as to attract attention in order for the product alternative concerned to come into question at all.

The latter two measures shown in Table 1 require a more in-depth explanation. In transition analysis, as first presented by Payne (1976, p. 376), the change from field-access n to field-access n + 1 is divided into four different possibilities: transition type 1 is the repeated accessing of the same field; transition type 2, where the subject stays with the same product alternative but changes the attribute, is interdimensional; transition type 3 is intradimensional, and is where the subject accesses the same attribute but for a different product alternative; and transition type 4 is a complete shift, that is, a change both of attribute and of product alternative (Jacoby et al., 1987, p. 152). After the transitions have been tallied, the formula (transition 2 − transition 3)/(transition 2 + transition 3) can be used to calculate an index which shows whether the transitions undertaken were primarily interdimensional (< 1 and > 0) or intradimensional (> −1 and < 0). If information is acquired interdimensionally, that is, by-brand, the strategy applied by the subject might be additive or linear, for instance. On the other hand, intradimensional information acquisition, that is, by-attribute, offers grounds to suspect a lexicographical approach or a so-called elimination-by-aspect strategy (Ford et al., 1989, p. 81; Jacoby et al., 1987, p. 152; Payne, 1976, p. 376; Trommsdorf, 2004, p. 313). Plotting or sequence graphs are purely a graphical depiction of what happened in terms of the information acquisition process (Schopphoven, 1996).

The IDM-application described in this paper is based on the Microsoft Access database program. It allows between one and ten product alternatives or product attributes to be listed. If participants have been questioned beforehand, it is possible to display only those attributes desired by the subjects. The content of the field opened appears each time as an enlarged pop-up. The pop-up solution represents an improvement in comparison to previous computer-based IDM-applications in that it offers more space for longer field contents and increases legibility. The information fields, as well as the line- and column-headings, can be filled with either text or graphics; if the researcher wishes to allow that, up to five fields can be opened at a time. These then appear in reduced size and in the order of the accession in a memory-column at the edge of the matrix. The time for the field-opening phase, as well as the total number that can be opened, can be limited. Further to this, participants can be given the opportunity of ruling out product alternatives or attributes by eliminating them from further information searching (the column or line that has been eliminated is kept in the matrix, but looks faded). Position of product alternatives or attributes in the matrix is randomized. The subjects can take a test-run before the actual experiment begins in order to familiarize themselves with the system. Help-texts are included to help the participants navigate their way through the system, and users are asked to confirm their decision, allowing unintentional decisions to be undone. The optical presentation of the matrix is orientated toward the structure of reports produced by consumer organizations, as well as that of online shops. After the IDM experiment has run its course, a sheet can be automatically called up which contains a questionnaire. This questionnaire can be filled in by the participants themselves (although this was not done in the study for the purpose of comparing methods: Questions were asked by the interviewer during a face-to-face-interview). For the analysis, simplified evaluation is enabled through preprogramming of crucial measures, which can be selected by the researcher from a list. The results of this stage of the experiment are automatically exported to the Excel program.

The data set in this study comes from 150 consumers buying fair-trade coffee who participated in the laptop-based survey at the point of sale. Of the 150 consumers, 90 people were asked outside supermarkets (or food retail stores, FRS), and 60 outside one-world-shops (OWS). The IDM contained six products (labeled A–F); information about these products was divided into six attributes, which were: product price, production system (conventional/organic), environmental standards in the coffee production, price premium as benefit to the coffee producers (the fair component of price), origin of the product (with regard to geographical region or kind of producer organization), and child protection (regarding child-labor prevention/children’s rights) in the coffee production stage. The overall aim of the study was to analyze both the importance of the various attributes, as well as the preferences within the scope of the attributes themselves. Furthermore, the study aimed to compare the information acquisition behavior of the two samples (FRS and OWS). Based on the assumption that consumers in the specialized shops (OWS) are different with regard to several variables, for example, that they are characterized by a higher involvement toward the issue of fair-trade, several hypotheses were developed. It was hypothesized that OWS-consumers show a more extensive information search behavior, tend to conduct a more compensatory search strategy, and give less importance to price but more importance to the issue-related information, as measured by the indicators of content. Independent from the type of customers, it was further hypothesized that information acquisition behavior differs in depth, content, and sequence between participants of different gender, age, and education, assuming that men search less intensively than women (see Ibanez, Czermak, & Sutter, 2008) and that there is a relation between information acquisition behavior and age or education level, possibly because of the differing familiarity with computer technology. Lastly, it was expected that the importance given to an attribute is also reflected in a choice of a product with this attribute.

Some of the results are shown in Table 2. In terms of the depth of the information acquisition process, customers of OWS took, although not significantly, longer to search for information than the customers approached outside the supermarkets. Customers of OWS used an average of 3.0 min, while customers approached outside the supermarkets used 2.6 min on average. There were, however, clear differences in terms of the content the consumers accessed during the information acquisition. Measured in terms of which fields were accessed first, country of origin was significantly, χ2(1, N = 150) = 3.85, p = .050 more important for consumers at the OWS than for consumers at the supermarket, while for the attribute price, the contrary applies, χ2(1, N = 150) = 4.47, p = .035. Regarding the sequence of information acquisition, no significant differences between the two groups could be noted in terms of either the variability of number of attributes considered per product alternative or the preference for an interdimensional search structure (measured with the transition analysis index). The OWS-customers, however, had a higher likelihood of starting the information acquisition according to usual reading behavior: a higher ratio of persons in this group started with the field at the top of the left column (52.5% compared to 36.4% in the supermarket-customer group). Generally, the vast majority of respondents (92.5%) started accessing fields of the respective column on the left side, and most of them chose the upper-left field (42.9%). This underlines the fact that randomizing the order of rows and columns, as it is done in the computerized IDM, is an essential requirement.

Another way of approaching the data is not to look at groups which have been previously defined, but rather to compare groups which differentiate themselves during the study relating to their information acquisition behavior. In the example at hand, for each indicator listed in Table 2 a variable was produced which subdivided the sample into three groups with low, medium, and high value. These groups were then compared regarding differences in age, gender, and education level. The respective groups did not differ significantly regarding these variables. However, as participants in the group with a high value for the total number of fields accessed or a high variability in the number of attributes considered per product tended to be from the younger half of the sample, the difference in age was explored more deeply by calculating and comparing the search duration per number of fields accessed. As one might expect, participants from the older half of the sample spent relatively longer on individual data fields, Mann-Whitney U (N = 150) = 1,554, p < .001. Thus, the possible influence of age ought to be taken into account at the relevant points in designing studies with the IDM-method and interpreting the results.

The IDM results can be linked to the final choice as well as to the results of the face-to-face-interview. The assumption behind recording the percentage at which a certain attribute is called up first is that the earlier information is accessed, the more important it is in the decision-making process. The respective results described in Table 2 show that in the study supermarket-customers assign price a higher importance. As this importance might be expressed in choice as well, it can be expected that the alternatives with the lower price are chosen more frequently in the supermarket-customers group than in the OWS-customers group. Although the frequency of choosing the two lowest-priced alternatives was, at 16%, higher than the number of such decisions recorded for the OWS-customers, at 8%, the relation is not significant in the example. The data gathered in the face-to-face-interview carried out after the test itself were expected to provide a plausible context for differences between the two groups, as they can be noted in the IDM results. Fifty-two percent of the OWS-customers reported that they had visited a developing country; in comparison, only 37% of consumers approached in supermarkets had traveled to a developing country. This might help to explain why the country of origin attribute was far more important for the former group (see Table 2). However, this relation is not significant.

In the study, 80% of the participants chose an organic fair-trade coffee, although only 50% of the six stimuli were organic. Logistic regression analysis was applied with the aim of testing the usefulness of the IDM-measures for predicting choice decision outcome. The results are given in Table 3. It appears that the metric measure of content which describes the number of fields accessed per attribute organic (ACCESSORG) explains choice: Participants who had viewed a higher number of fields in the attribute organic were also more likely to choose an organic product. This result strengthens the assumptions underlying these IDM-measures that more frequently accessed information is of a higher importance. It has to be noted that the sociodemographic variables (children in the household, age, sex, household income, education level) clearly did not yield significant results.

In order to allow for a method-comparison of IDM with the face-to-face-interview that followed, the subjects were asked in the latter to place the six product attributes in order of importance. This ranking can also be calculated using the measures of content in the IDM-method (see Table 4). It is interesting to note that the order of importance given by the subjects matches that deduced from the IDM results down to one exception: price. In the face-to-face-interview, only 36% of respondents named this as one of the top three important attributes; in the IDM, however, 49% of subjects opened a larger-than-average selection of fields of the price attribute. In the face-to-face-interview, price came fifth overall in terms of importance, but was in third place according to the IDM-data. This discrepancy in the importance of price in the face-to-face-interview and the IDM fits with expectations because in the IDM information is acquired without the participation of another person. When conducting face-to-face-interviews about products carrying socially desirable attributes (here, fair-trade coffee), researchers need to be aware that respondents will tend toward giving answers they see as correspondingly socially desirable; these answers may even stand in direct contrast to their actual behavior. This would appear to suggest that IDM results are more valid in this regard than results based on face-to-face-interviews.

Summing up the results of the study, it was possible to confirm that customers from OWS differ in their information search behavior from fair-trade customers purchasing their products in supermarkets. Thus, the potential to detect differences in the depth, content, and sequence with the help of the IDM could be demonstrated. While the study indicated that the possible connection between age and familiarity with the computer technology should be kept in mind for interpretation, it could also be evidenced that the IDM appears to have the advantage of being less distorted by social desirability. However, the aim of the study was not to put all the ideas described in chapter five of a further development of the method into practice. The most important role of the study is basically being an example which shows what can be done and analyzed with the tool of the IDM in order to better understand the method, while the list of ideas for further development in the previous chapter has the role of being a “tool-box” for future applications of the method.

The criticism of IDM has been mainly directed at its lack of external validity due to the way information is presented. Yet it is argued that several measures resulting from technological progress in computer hard- and software and increased information technology uptake in the population at large, as well as a changing purchase environment can be taken to alleviate this deficiency, or indeed to turn it to advantage. From the possibilities for further development of the IDM-method discussed in this contribution, some but not all could be put into practice in the study described as an example, highlighting the potential of the method. The major contribution of the IDM-application used in the study is its easy handling and improved layout, mirroring that of Internet shops. Future studies of this kind should further explore the influence of the liberation from predefined structures, as in an “individualised” IDM, and the use of a whole but blurred picture on consumers’ information acquisition behavior. Furthermore, it is argued that the IDM is an obvious choice as part of a method combination of different, yet complementary process-tracing techniques with the aim of building up a comprehensive picture of information acquisition behavior. The IDM can also deliver data that can be used for evaluation along with structural models and experimental choice-tests.

For this reason, information display matrices are suitable instruments for use in consumer research and have the potential to be expanded in the future. There is a need for more research into validating the measures put forward in terms of realistic IDM setups. Furthermore, methodological comparisons of process-tracing techniques should follow further developments in the field. The main applications of IDM in practical consumer research would appear to lie in identifying the key information for consumers, in checking the contents of product- or service-descriptions for relevance to consumers, as well as in optimizing the design of communication content and images.

1  The various measures of content lead to slightly different results. First accession was selected for presentation because it is less influenced by a differing extent of information search.

Alfnes, F., Guttormsen, A. G., Steine, G., & Kolstad, K. (2006). Consumers’ willingness to pay for the color of salmon: A choice experiment with real economic incentives. American Journal of Agricultural Economics, 88, 1050–1061.

Anderson, P. (2001). P1198. Software for tracing decision behavior in lending to small businesses. Behavior Research Methods, Instruments, & Computers, 33, 234–242.

Andersson, P. (2004). Does experience matter in lending? A process-tracing study on experienced loan officers and novices decision behaviour. Journal of Economic Psychology, 25, 471–492.

Bauer, H. H., Grether, M., & Sattler, C. (2002). Internetspiele als Marketinginstrument für Low-Involvement Produkte. [Online games as a marketing instrument for low-involvement products]. Marketing ZFP, 24, 265–276.

Bettman, J. R., & Kakkar, P. (1977). Effects of information presentation format on consumer information acquisition strategies. Journal of Consumer Research, 3, 233–241.

Biggs, S. F., Rosman, A. J., & Sergenian, G. K. (1993). Methodological issues in judgement and decision-making research: Concurrent verbal protocol validity and simultaneous traces of process. Journal of Behavioral Decision Making, 6, 187–206.

Brucks, M. (1988). Search monitor: An approach for computer-controlled experiments involving consumer information search. Journal of Consumer Research, 15, 117–121.

Büttner, O. B., Rauch, M., & Silberer, G. (2005). Consumer cognition at the point of sale: Results from a process tracing study. Contribution to Tracking Research. No. 11, Institute of Marketing and Retailing, Göttingen, Germany: University of Göttingen.

Covey, J., & Lovie, A. (1998). Information selection and utilization in hypothesis testing: A comparison of process-tracing and structural analysis techniques. Organizational Behavior and Human Decision Processes, 75, 56–74.

Felser, G. (2007). Werbe- und Konsumentenpsychologie. [Consumer psychology and advertising]. Heidelberg, Germany: Springer.

Ford, K. J., Schmitt, N., Schechtman, S. L., Hults, B. M., & Doherty, M. L. (1989). Process tracing methods: contributions, problems, and neglected research questions. Organizational Behavior and Human Decision Processes, 43, 75–117.

Harte, J. M., & Koele, P. (1995). A comparison of different methods for the elicitation of attribute weights: Structural modeling, process tracing, and self-reports. Organizational Behavior and Human Decision Processes, 64, 49–64.

Huneke, M. E., Cole, C., & Levin, I. P. (2004). How varying levels of knowledge and motivation affect search and confidence during consideration and choice. Marketing Letters, 15, 67–79.

Hwang, M. I., & Lin, J. W. (1999). Information dimension, information overload and decision quality. Journal of Information Science, 25, 213–218.

Ibanez, M., Czermak, S., & Sutter, M. (2008). Searching for a better deal – on the influence of group decision making, time pressure and gender on search behaviour. Journal of Economic Psychology, 30, 1–10.

Jacoby, J., Chestnut, R. W., & Fisher, W. A. (1978). A behavioral process approach to information acquisition in nondurable purchasing. Journal of Marketing Research, 15, 532–544.

Jacoby, J., Jaccard, J. J., Currim, I., Kuß, A., Ansari, A., & Troutman, T. (1994). Tracing the impact of item-by-item information accessing on uncertainty reduction. Journal of Consumer Research, 21, 291–303.

Jacoby, J., Jaccard, J., Kuß, A., Troutman, T., & Mazursky, D. (1987). New Directions in behaviorial process research: Implications for social psychology. Journal of Experimental Social Psychology, 23, 146–175.

Jasper, J., & Shapiro, J. ( 2002). Mouse Trace: A better mousetrap for catching decision processes. Behavior Research Methods, Instruments, & Computers, 34, 364–374.

Johnson, E. J. (1996). The last whole MouseLab manual. Retrieved October 30, 2007, from http://www.cebiz.org/mouselab/mouselab.pdf.

Johnson, E., & Willemsen, M. C. (2004). MouselabWEB. Retrieved October 30, 2007, from http://www.mouselabweb.org/index.html.

Kleinmuntz, D. N., & Schkade, D. A. (1993). Information displays and decision processes. Psychological Science, 4, 221–227.

Kroeber-Riel, W., & Weinberg, P. (2003). Konsumentenverhalten [Consumer behavior]. Munich, Germany: Vahlen.

Lehmann, D. R., & Moore, W. L. (1980). Validity of information display boards: An assessment using longitudinal data. Journal of Marketing Research, 14, 450–459.

Louviere, J. J., Hensher, D. A., & Swait, J. D. (2000). Stated choice methods. Analysis and application. Cambridge, United Kingdom: Cambridge University Press.

Lusk, J. L., & Schroeder, T. C. (2004). Are choice experiments incentive compatible? A test with quality differentiated beef steaks. American Journal of Agricultural Economics, 86, 467–482.

Marten, U. (1992). Die Bedeutung von Information-Display-Matrices-Verfahren für die Konsumentenforschung[Significance of information display matrix for consumer research]. Dissertation, Bremen, Germany: University of Bremen.

Meyer, J.-A. (1994). Multimedia in der Werbe- und Konsumentenforschung[Multimedia in advertising and consumer research]. In Forschungsgruppe Konsum und Verhalten (Eds.), Konsumentenforschung (pp. 305–320). Munich, Germany: Gabler.

Mittal, B. (1989). Must consumer involvement always imply more information search?. Advances in Consumer Research, 16, 167–172.

Mühlbacher, S., & Kirchler, E. (2003). Informations-Display-Matrix: Einsatz und Analysemöglichkeiten[Information display matrix: Application and possibilities for analysis]. Der Markt: Zeitschrift für Absatzwirtschaft und Marketing, 42, 147–152.

Newell, B. R., Weston, N. J., & Shanks, D. R. (2003). Empirical tests of a fast-and-frugal heuristic: Not everyone “takes the best”. Organizational Behavior and Human Decision Processes, 91, 82–96.

Payne, J. W. (1976). Task complexity and contingent processing in decision making: An information search and protocol analysis. Organizational Behavior and Human Performance, 16, 366–387.

Peterson, R. A., & Merino, M. C. (2003). Consumer information search behavior and the Internet. Psychology & Marketing, 20, 99–121.

Ratchford, B. T., Taludkar, D., & Lee, M.-S. (2007). The impact of the Internet on consumers’ use of information sources for automobiles: A re-inquiry. Journal of Consumer Research, 34, 111–119.

Roßmanith, T. (2001). Informationsverhalten und Involvement im Internet. Eine Labor- und Feldstudie zu den Determinanten der Informationsnachfrage im World Wide Web[Information acquisition behavior and involvement online. A laboratory and field experiment about the determinants of information search in the World Wide Web]. Dissertation, Karlsruhe, Germany: University of Karlsruhe.

Scholderer, J., Hagemann, K., Sorensen, T. B., & Czienskowski, U. (2007). Consumer use of information about novel foods and food technologies, NOFORISK Deliverable 34, Project Paper No. 01/07. Berlin, Germany: MAPP, Aarhus, Denmark; Max Planck Institute for Human Development.

Schopphoven, I. (1996). Messung von Entscheidungsqualität. Konzeptualisierung, Operationalisierung und Validierung eines Meßinstrumentariums für Entscheidungsqualität[Measuring of decision making quality. Conceptualisation, operationalisation, and validation of a measurement]. Frankfurt am Main, Germany: Peter Lang.

Silberer, G. (2005). Die videogestützte Rekonstruktion kognitiver Prozesse beim Ladenbesuch[The video-based reconstruction of cognitive processes at the point of sale]. Marketing ZFP, 27, 263–280.

Trommsdorf, V. (2004). Konsumentenverhalten[Consumer behavior]. Stuttgart, Germany: Kohlhammer.

Tutz, G. (2000). Die Analyse kategorialer Daten. Anwendungsorientierte Einführung in Logit-Modellierung und kategoriale Regression[Analysis of categorical data. Application-oriented introduction to logit modeling and categorical regression]. Munich, Germany: Oldenbourg.

Verplanken, B., & Weenig, M. (1993). Graphic energy labels and consumers decisions about home appliances: A process tracing approach. Journal of Economic Psychology, 14, 739–752.

Weenig, M., & Maarleveld, M. (2002). The impact of time constraint on information search strategies in complex choice tasks. Journal of Economic Psychology, 23, 689–702.

Williamson, J., Ranyard, R., & Cuthbert, L. (2000). A conversation-based process tracing method for use with naturalistic decisions. An evaluation study. British Journal of Psychology, 91, 203–222.

Zielke, S. (2001). Kundegerechte Sortimentsgliederung am Point of Sale. Ansätze zur Erhebung kognitiver Strukturen als Richtgrößen für Warenplatzierungen und Category Management[Consumer oriented assortment policy at the point of sale. Approaches for measuring cognitive structures as a guideline for placement and category management]. Marketing ZFP, 23, 100–116.