Three-Dimensional Information Acquisition: an Application to Contraceptive Decision Making

Wayne D. Hoyer, The University of Texas at Austin
Jacob Jacoby, New York University
ABSTRACT - The present investigation introduces a new methodology whereby all aspects of the information environment (i e., sources x options x dimensions) are presented simultaneously in the form of a 3-dimensional matrix. This methodology was utilized in the context of a contraceptive decision making task. Key results were that: (1) medical authority was the most preferred source for information and (2) health risks, effect on later conception, and effectiveness were the most frequently accessed types of information. Further results and implications are discussed.
[ to cite ]:
Wayne D. Hoyer and Jacob Jacoby (1983) ,"Three-Dimensional Information Acquisition: an Application to Contraceptive Decision Making", in NA - Advances in Consumer Research Volume 10, eds. Richard P. Bagozzi and Alice M. Tybout, Ann Abor, MI : Association for Consumer Research, Pages: 618-623.

Advances in Consumer Research Volume 10, 1983      Pages 618-623


Wayne D. Hoyer, The University of Texas at Austin

Jacob Jacoby, New York University


The present investigation introduces a new methodology whereby all aspects of the information environment (i e., sources x options x dimensions) are presented simultaneously in the form of a 3-dimensional matrix. This methodology was utilized in the context of a contraceptive decision making task. Key results were that: (1) medical authority was the most preferred source for information and (2) health risks, effect on later conception, and effectiveness were the most frequently accessed types of information. Further results and implications are discussed.


Much of the research on consumer behavior in recent years has focused on the processes by which consumers process information in order to make a choice decision. In making this decision, one of the major steps is the acquisition of information from the external environment, and numerous studies have attempted to capture this process using either a behavioral process approach (e.g. Jacoby 1975 1977; Moore & Lehmann 1980) verbal protocols (e.a. Bettman 1971; Payne & Ragsdale 1978) or eye movement analysis (e g. Russo & Rosen 1975; Russo 1978). In most of these studies, only package information is provided to the subjects. A problem with this approach is that consumers can acquire information about choice options from sources other than a package, and as a result, ar incomplete picture of the search process may have been provided. The purpose of the present study is to introduce a methodology which permits presentation of a more thorough and complete information environment in the context of a specific decision context, the selection of a - contraceptive method.

The Information Environment

In any decision context, the information environment is defined in terms of a set of externally available cues which the individual may use in arriving at his or her decision. In general this environment can be defined in terms of the following four factors (Jacoby 1975):

1. Options - a set of alternatives (actual or hypothetical) from which an individual may make a selection (e.,.. different birth control methods)

2. Properties - a category or type of information about a set of options (e.a. chances of pregnancy, health risks)

3. Sources -- different points of origin from which the information may be obtained (e.g. doctor or friend)

4. Values - the information resulting from a specific option X property crossing. (e.g. Method A X chances of pregnancy = 1 in 1000)

Most of the previous information acquisition studies employing, process methodologies examined only three of the four factors of the task environment. Typically only the option, attribute, and values factors are considered and no attempt is made to examine the sources from which this information is accessed. While Berning & Jacoby (1974) examined the effects of different sources on new product purchases, only a source X option x value matrix was used.

The exclusion of source factors from process methodologies appears to be problematic in light of evidence on information sources from studies employing retrospective questioning (e g. Katona & Mueller 1955; Houston 1979). That is, it can generally stated that consumers rarely rely on one source for information exclusively (Engel & Blackwell 1981). Rather, different sources tend to be consulted in order to gather complementary information. Further, different sources appear to be more influential at different stages of-the choice process. Mass media seem to play a greater role at the awareness and interest phases, while word-of-mouth sources are more critical at the actual decision stages (Berning & Jacoby 1974; Houston 1979). Finally, it is possible that consumers will consult different sources for different types of information. For example, in making a contraceptive usage decision, a consumer might consult a physician for health-related information while information- related to sexual pleasure might be obtained from more personal sources (e.g. friends, counseling center).

In summary, it is clear that in order to more fully and accurately describe the process of pre-purchase information acquisition, source factors must be incorporated into the assessment procedure. The present study differs from the previous process investigations in that all four aspects of the task environment will be accommodated simultaneously. Specifically, the present investigation will provide subjects with a matrix from which they say acquire items of information regarding different types of properties (e.g , health risks, chances of pregnancy) concerning different options (e.g. birth control methods) from different sources (e.g. doctors, friends, the family planning literature).

The use of a three-dimensional matrix allows the examination of all issues addressed by two dimensional operationalizations as well as issues unique to a three-dimensional manipulation. Thus, this methodology provides an advantage over previous approaches because a more complete description of the search process is provided. In other words, in addition to identifying the types of information subjects acquire, data regarding source preferences (e. g. where subjects would like to get information from) is provided.

In order to operationalize the three-dimensional matrix. a computer presentation technique was developed. Briefly, all the choice information is stored in the computer and can be accessed by typing a special-code number corresponding to the source, option and property desired. This information then appears on a television monitor which is viewed by subjects. The advantages of this technique are that: (1) information can be presented to subjects more quickly and efficiently (with little effort), (2) the visual bias inherent in the presentation of a physical matrix is eliminated, (J) since no physical matrix is present, there is greater flexibility in constructing information environments (e.g. all four aspects of the information environment can be accommodated simultaneously and (4) due to the fact that the search process is automatically recorded by the computer. data collection is more efficient and reliable.

Research on Fertility Related Behaviors

Due to the rapidly rising birth rate in other areas of the world and to the increasing number of unwanted teenage pregnancies in our country, scientists have become increasingly interested in fertility related issues. However, most of the research has been demographic in nature, and researchers have only just begun to realize the importance of psychological factors in explaining this type of behavior.

Unfortunately, early research which has attempted to relate psychological variables to fertility behavior was generally disappointing. Two major fertility studies-the Indianapolis study (Whelpton & Kiser 1946-1958) and the Princeton Study (Westoff, Potter & Sagi 1967)--were both unable to find meaningful relationships between psychological variables and indices related to fertility. In both studies, most of the meaningful findings were related to socioeconomic status.

It is important to note that these early studies were plagued with theoretical and methodological problems (i.e. haphazard assessment of variables due to a lack of theory guiding the research effort). More recent efforts, however, have attempted to utilize psychological theory to examine fertility related behaviors. For example, one series of studies (Jaccard & Davidson 1972; Davidson & Jaccard 1975) employed a structural model developed by Fishbein (1972) to study fertility-related intentions. The results were promising in that an average of 60 percent of the variance in behavioral intentions was explained by the model.

Notably absent from the contraceptive literature are studies which attempt to examine the process of contraceptive decision making and it is in this area that the theories and methods developed for the study of consumer decision making may be quite useful. For example, Wright and Weitz (1977) examined the effects of women's time horizons on the types of evaluation strategies used. In short, it was found that women making imminent yet leisurely judgments used more complex evaluation strategies than women who made hasty decisions or distant decisions.

In light of these promising results, the present study will examine the process of contraceptive decision making using a behavioral process approach (e.g. Jacoby 1977) In brief, this methodological places subjects in a decision situation and closely monitors the process of information acquisition leading to a choice

Due to the absence of previous research in the area of contraceptive decision making, few hypotheses can be generated. However, many descriptive issues remain to be addressed. Important questions include: (1) what types of information are most frequently accessed regarding the available contraceptive options? (2) What types of sources are most frequently consulted? and (3) Are there specific patterns of individual search behavior such that different sources are consulted for different kinds of information?

Also of interest is the extent to which particular sources are consulted for specific types of information. That is, certain properties (e.g. health risks, side effects) may be classified as more technical than others. It is postulated that individuals would consider a physician, with his technical training, as the most dependable source for this type of information. Likewise, other information is more subjective or based on personal opinion (e.g. interrupts the sex act, effect on sexual pleasure). It is suggested that more intimate sources (e.g. a knowledgeable friend or sex partner) would be more frequently used for this type if information.

Additional issues of interest include the relationships between search behavior and: (1) past experience and attitudes toward sex, and (2) past experience and attitudes toward contraception. Specifically, the questions to be examined are: (1) Do individuals with positive attitudes toward sex and contraception engage in more extensive information search than those with negative attitudes? and (2) Do individuals with greater past experience toward sex and contraceptives acquire more information than those with less experience?

Although no previous research in the fertility area has examined these questions, Jacoby, Chestnut, & Fisher (1978) provide evidence regarding the relationship between information acquisition and past experience. Briefly, it was found that subjects with greater relevant experience acquired more information than individuals with less experience.

Alternatively, however, Bettman & Park (1980) found an inverted-U relationship between acquisition and search. Thus, based on these contradictory findings, two competing hypotheses will be tested.


Subjects. The sample was composed of 61 undergraduate females enrolled in Introductory Psychology at a large Midwestern university whose participation served as partial fulfillment of course requirements.

Apparatus. An IBM 5100 portable computer was used to present the stimulus information (i.e. the task environment) and to collect the subjects' search responses. A Sony portable television monitor was connected to the computer in order to display the information to subjects.

The Information Environment

The decision environment consisted of a three-dimensional (source X option X property) matrix. In this particular investigation, the matrix contained 3 options, 5 sources, and 19 properties for a total of 180 unique values.

Options. The options were three hypothetical birth control methods. Hypothetical methods were used because a major purpose of the study was to examine the content of search; that is, what kinds of information and what kinds of sources individuals access when making a decision regarding unfamiliar methods of birth control. Previous-studies (e.g. Jacoby & Chestnut 1977; Jacoby et al 1978) have shown that not only do subjects acquire less information when brand name is present, but also the percentage of immediate choices (i.e. no external search) is greater. Thus, in order to more fully examine the search process, it was felt that hypothetical methods were the most desirable choice options. In addition, it should be mentioned that, in the search task, subjects were told that these were new methods of birth control.

Sources. Five different types of information sources were included. Due to the absence of previous studies regarding source effects for birth control information, the five sources were selected on the basis of a pretest of source importance. These sources were: medical authorities, a knowledgeable friend, published literature, one's sex partner, and counseling centers. In addition, in order to add to the realism of the source manipulation, a list of physical costs (i.e. the effort which must be expended to reach the source) was provided as part of the description of the sources. Subjects were asked to carefully consider these costs and behave as they normally would when making such a decision.

Properties. Two criteria were applied in selecting the properties to be used in the study. First, it was considered desirable to include the most important attributes as identified in a pretest of attribute importance. In addition, an attempt was made to include an adequate number of "subjective" (i.e. based on personal experience) as well as "technical" (i.e. requires personal training) information properties in order to test the hypotheses regarding the accessing of specific sources for specific types of information. The list of properties can be viewed in Table 2.

Values. The actual values of information were developed by the experimenters in the following manner. First, the values for each of the options were balanced for positive and negative information. That is, in order to generate rich external search data, it was desirable that no one birth control method be obviously better than the others. Second, it was felt that there should be little disagreement across sources regarding the values of information they convey due to the confusion this phenomenon might cause. As a result, each of the sources were made to provide information which did not contradict the information provided by any of the other sources, yet at the same time was not identical to the information given by the other sources.


A 35 item questionnaire was administered in order to assess past experience and attitudes toward sex and contraception. This questionnaire contained a battery of questions developed by the experimenter to measure a wide variety of topics relevant to sexual and contraceptive attitudes and experience (e.g previous sexual experience, previous use of contraceptives, discussions or sex with family members as a child sexual and contraceptive opinions. etc.).


Upon entering the experimental room, subjects were seated and read a short introduction to the experiment. Each subject was then given the questionnaire which measured attitudes and past experience toward sex and contraception. Following the completion of this form, subjects participated in the information acquisition task in which they were asked to choose one of three new birth control methods which were not yet on the market Subjects were told that they could select as many or as few or the information values as they wished and could re-access any one of these values as often as they wished.

All of the information (i.e. the option- source-property values) was stored in the IBM computer With the aid of an attached television monitor, subjects were able to see any particular information by calling out a code number corresponding to cell in which the desired information was stored. That is, subjects were presented with three Lists: a list of sources, a list of options, and a list or properties, and associated with each source, option, and property was a code number. Included on these lists were short descriptions of each of the items (sources and properties only, since the options were hypothetical and designated by the numbers 1, 2, and 3). In order to acquire an item of information, the subject read the option number, source number, and property number (e g. "I'll have 3-1-12"). The experimenter then typed this number into the computer and the information appeared on the television screen. It should be noted that while the description of this procedure suggests a tedious process, in actuality the procedure was executed with considerable ease and speed. Further, the search process was recorded by the computer as it occurred and the relevant search statistics were computed easily upon completion of the session. Finally, the session was concluded with a post-task questionnaire designed to assess subjective reactions to the task (i e anxiety, difficulty of task).


Summary Statistics on Information Search

Summary statistics were computed based on the information (values) acquired during the search task. These statistics are presented in Table 1. In general, subjects acquired an average of 24.42 values before making a decision. If repeat acquisitions are omitted, an average of 22.18 (or 12%) of the total values were acquired. However, it must be noted that a large portion of the unsearched cells are the result of not examining information from all sources. Also, subjects examined an average-of 3.00 options, 3.09 sources, and 9.22 properties.



By far, the most frequently accessed source was the medical authority (Y=13. 74), while the counseling center was second in the number of acquisitions (X=5.19). In other words, based on the average search length or 24.42 cues, 59% of the information was obtained from the medical authority source and another 19% were obtained from the counseling center. The most frequently considered properties were chances of pregnancy (X=3.37), affects ability to later conceive children (X=3.26), and health risks (X=3.21) . The means for the entire matrix are presented in Table 2.

An additional measure of interest was the sequence in which subjects acquired information. In previous studies using the behavioral process paradigm, the patterns of search were described using four basic transition vectors which essentially are nominal classifications based on a consideration of the change in option and property from one acquired information value to the next (i e. from" to n+1; see Jacoby, Chestnut, Weigl, & Fisher 1976). However, these measures of sequence are only applicable in the two-dimensional matrix case. A three-dimensional matrix requires eight such transition types (and these are diagrammed in Appendix A).



The first four transitions correspond to the previously used transitions. A Type I transition represents simply reexamining the same piece of information on the immediately following accessing opportunity. A Type II transition would indicate looking at the same source, same option, but different property (from" to n+1) and is referred to as "within source, within option search." This is commonly referred to as "brand processing" in the consumer literature. If an individual examines the same property from the same source regarding a different option, a Type III transition has occurred. This is referred to as "within source, within property" search (or more commonly referred to as attribute processing). If a subject acquires the same source but different property and different option, a Type IV transition has taken place.

Transition Types V-VIII are similar to Types I-IV with the exception that the information has been accessed from a different source. That is, a Type V transition involves examining the same option and property but from a different source. In other words, subjects engaging in this type of search would be comparing information from different sources. A Type VI transition involves a different source, the same option, and different property and is called "different source, within option search." This type of search would be most likely to occur when different sources are consulted for different types of information (but for the same brand). The final two transitions are not as theoretically meaningful but are included for the sake of completeness. A Type VII transition involves a different source, different option, and the same property ("different source, within property" search) and acquiring a different source, option, and property altogether is simply a Type VIII transition

In reference to sequence of search, subjects tended to stay within the same source for information That is, 70% of the average search was devoted to within-source search transitions. The most frequent type of search patterns were same source, within option search (32%-Type II), and same source, within property search (26%-Type III). Also, the comparison of information across sources was almost nonexistent as this type of search (Type V) accounted for only 4% Of the average search process

Finding Bearing on the Content of Search

It was suggested earlier that different sources may be consulted for different types of information ( i.e. technical vs. subjective). A test of this notion requires that an interaction between levels of sources and properties be present. In order to establish this fact, a 5 X 12 completely within subjects analysis-of-variance was run on the data. The dependent measure was the number of times each source/Property combination was accessed.

A significant sources X properties interaction was found (F=22.07, p .01) indicating that the number of values examined differs across source and property levels. A test of simple main effects was then employed to examine the various levels of each factor and to test the two hypotheses.

The simple main effects analysis for each of the properties revealed a significant simple main effect in all cases. A further comparison of means using a series of Newman-Kuels tests determined that the medical authority was accessed significantly more often for each type of technical information than and other source (p 01). In other words, subjects by far preferred the medical authority for information on health risks, effect on later conception, chances of pregnancy, necessity for continuous attention, and prevention of venereal disease.

A simple main effects analysis for each of the subjective properties (i.e. difficulty of use, impact on pleasure, impact on psyche, and causes discomfort or pain) produced simple main effects in two of the four cases (difficulty of use, and causes discomfort or pain) These simple main effects, however, were not in the expected direction. In both cases, the medical authority was more preferred than all other sources (p .01); including the two personal sources. It should be noted, however, that while medical authority was the most frequently accessed source across all properties, in the cases of the knowledgeable friend and the sex partner, the most frequently examined properties were impact on psyche and impact on pleasure.

Relationship Between Past Experience & Search

As mentioned previously, a variety of past experience and attitudinal measures were measured in order to examine their relationship to information acquisition. Unfortunately, only two variables exhibited any degree of association These were: (1) discussions of sex with the father as child (r=.47, p < .001), and (2) a self rating of sexual experience (r=.34, p < .01). Thus, subjects who discussed sex with their father and who perceived themselves as more sexually experienced, have a slight tendency to search for more information.


In employing a three-dimensional matrix, it was hoped t that rich data would be provided regarding the types of information subjects used when making a contraceptive decision. Not only was it of interest to determine just which properties subjects acquired, but also the sources from which this information was sought. In this regard, several interesting findings are worth noting. First, the strongest finding is the overwhelming desire for birth control information from the medical authority. For 10 of the 19 properties, the medical authority was accessed significantly more often than all other sources. It is thus apparent that birth control information is considered to be of a medical nature and individuals strongly prefer to receive this information from a knowledgeable medical source. Second, the results clearly indicated that the most frequently acquired types of information were of a technical nature. More particularly, subjects most often accessed information regarding health risks and the effect of the method on one's ability to later conceive children. This would imply that the major concern for college females in deciding on whether to use a particular birth control method is the extent to which the method has a detrimental effect on one's health. This finding is not surprising given the widespread knowledge of the short and long-term side effects of the birth control pill (the most widely used means or contraception).

It must be emphasized, however, that generalizability of these results are limited by several important factors. First, an obvious restriction is due to the nature of the subject population. The sample was composed entirely or college women, and it would be quite hazardous to generalize the findings to the population as a whole.

A second restriction is that the subjects' search behavior may reflect more of an ideal rather than their actual behavior. That is, the nature of the search task essentially equalizes the costs required to obtain information from each of the different sources. In a "real-world" decision context, however, the effort and resources necessary to acquire information from a medical source (e.g. make an appointment, pay a fee, suffer social embarrassment) are quite different from those expended to consult a friend or sex partner (e.g. merely talk to the individual). Thus, it may be that ideally individuals desire to obtain birth control information from a medical authority, but in reality may not be able to do so for various reasons. In particular, this limitation may account for the relatively light search from the counseling center and published literature sources. Future investigations which employ this methodology might add a personal cost dimension to the procedure in order to approximate real world constraints. For example, subjects could be provided with a sum of money at the start or the search process and be charged a different cost depending on which source the information is acquired from. Alternatively, time costs could be paired with each source such that slower responses would accompany more costly sources (given that subjects are allotted a finite time to make a decision), thereby imposing more realistic constraints on the decision task

Third, given the sensitive nature of the choice task, there is always the threat of demand bias. Even though subjects did not elicit any negative reactions on the post-task questionnaire, the e still may have been an inhibition in their search for more personal types of information (i.e. effect on pleasure and psyche) Perhaps this problem could be lessened in future studies by allowing subjects to engage in the task without the experimenter present. At any rate, future research which attempts to minimize these problems is needed before any definite conclusions regarding contraceptive decision making can be drawn.

Finally, a somewhat disappointing finding was the relative inability of factors relating to sexual and contraceptive past experience and attitudes to account for any meaningful variance in search behavior.

This failure may be due to the fact that a contraceptive decision may be very important for any female regardless of her past experience with sex and contraceptives. As a result, individuals with negative attitudes may be as likely to acquire information as those with more positive attitudes.


It was hoped that the employment of a three dimensional matrix would provide insight into subjects' search strategies when the information was available from several different sources. Unfortunately, this process was somewhat obscured due to the preponderance of information acquisition from a medical authority. It must be emphasized that this problem may be unique to the nature of this particular decision task. As a result, this methodology may prove to be quite useful in alternative decision contexts where: (a) different sources are quite valuable and able to provide the desired information (e.g. an automobile, appliance or audio equipment purchase); and (b) the topic under study is not as highly sensitive. In addition, the previously mentioned source cost manipulations would add to the effectiveness of this technique-. At any rate, given the importance of different sources of information in the search process, it appears that future research employing this methodology is warranted.




Berning, C. A. and Jacoby, J. (1974), "Patterns of Information Acquisition in New Product Purchases," Journal of Consumer Research, 1, 18-22.

Bettman, J. R. (1971), "The Structure of Consumer Choice Processes," Journal of Marketing Research, 1971, 8, 465-71.

Bettman, J R. and Park, C. W. (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 (3), 234-48.

Davidson, A. R. and Jaccard, J. (1975),"Population Psychology: A New Look at An Old Problem," Journal of Personality and Social Psychology, 1971, 31(6), 1073-82.

Engel, J. F. and Blackwell, R. D. (1981), "Consumer Behavior, Fourth Edition. Hinsdale, Illinois: The Dryden Press.

Fishbein, M. (1972) "Toward An Understanding of Family Planning Behaviors," Journal of Applied Social Psychology, 9, 214-227.

Houston, M. A. (1979), "Consumer Evaluations and Product Information Sources," In J. H. Leigh and C. R. Martin (Eds.), Current Issues and Research in Advertising 1979, Ann Arbor: University of Michigan. 135-144.

Jaccard, J. J. and Davidson, A. R. (1979), "Toward An Understanding of Family Planning Behaviors: An Initial Investigation. Journal of Applied Social Psychology, 7, 228-35.

Jacoby, J. (1975), "Perspectives on a Consumer Information Processing Research Paradigm," Communication Research, 2, 203-15.

Jacoby, J. (1977), "The Emerging Behavioral Decision Process Technology in Consumer Decision Making Research," In W. D. Perrault, (Ed.), Advances in Consumer Research, Vol. 4, Atlanta: Association for Consumer Research, 263-5.

Jacoby, J. and Chestnut, R. W. (1977), "Amount, Type, and Order of Package Information Acquisition in Purchasing Decisions," Technical Report submitted to the National Science Foundation.

Jacoby, J. Chestnut, R. I., and Fisher, W. A. (1978), "Information Acquisition in Nondurable Purchase," Journal of Marketing Research, 15 (4), 532-44.

Jacoby, J., Chestnut, R. W., Leigh, R., and Fisher, W. A. (1976), "Prepurchase Information Acquisition: Description of a Process Methodology, Research Paradigm, and Pilot Investigation," In B. B. Anderson, (Ed.), Advances in Consumer Research, Vol. 3, Cincinnati: Association for Consumer Research, 306-14.

Katona, G. and Mueller, E. (1955), "A Study of Purchase Decisions," In L. H. Clark (Ed.), The Dynamics of Consumer Reactions, New York: New York University Press, 30-87.

Moore, W. L. and Lehmann, D. R. (1980), "Individual Differences in Search Behavior for a Nondurable," Journal of Consumer Research, 7 (3), 296-307.

Payne, J. W. and Ragsdale, E. K. E. (1978), "Verbal Protocols and Direct Observation of Supermarket Shopping Behavior: Some Research Findings and a Discussion of Methods," In H. K. Hunt (Ed.), Advances in Consumer Research, Vol 5, Ann Arbor: Association for Consumer Research, 571-77.

Russo, J. E. (1978), "Eve Fixations Can Rule the World: A Critical Evaluation and a Comparison between Eye Fixations and Other Information Processing Methodologies," In Y. R. Hunt (Ed.), Advances in Consumer Research, Vol. 5 Ann Arbor: Association for Consumer Research, 561-70.

Russo, J. E. and Rosen, L. D. (1975), "An Eye Fixation Analysis of Multialternative Choice," Memory and Cognition, 3, 267-76.

Westoff, C., Potter, R. S. and Sagi, C. (1967), "The Third Child," Princeton, N. J.: Princeton University Press.

Whelpton, P R. and Kiser, C V. (Eds.) (1946-58), Social and Psychological Factors Affecting Fertility (5 vols.). New York: Milbank Memorial Fund.

Wright, P. and Weitz, B. (1977), "Time Horizon Effects on Product Evaluation Strategies," Journal of Marketing Research, 14, 429-43.