Decision Making, Consumer Choice, and the Internet
Jeffrey M. Stibel, stibel@brown.edu
Brown University
In order for consumers to be able to make decisions over the internet (whether purchase related or otherwise), they need to be able to easily retrieve, digest, manipulate, and utilize information relevant to the decision making process. The internet, like the mind, has the potential to store vast amounts of information; the problem with both, of course, lies within the process of retrieving that information. The advantage of the human mind is that it is capable of searching its memory by processing information in parallel, using a variety of different activation cues (e.g., synonyms, antonyms and syntax). Currently, computers (and their search tools) can only search the internet in a top-down fashion, using a singular semantic hierarchy. Furthermore, interfaces and visualizations of information on the web are further limited by the constraints of HTML, Java, web browsers, and two-dimensional computer screens.
Numerous attempts have been made to bypass these problems. To deal with information overload, interfaces have utilized hierarchical category tables in multi-dimensional space (Kumar, Plaisant, & Shneiderman, 1997). Web pages, and information within these pages, have been reorganized to allow anticipated information more prominence. User interface designers have also expanded and refined their use of color-coding, consistency, instructions, and other simple heuristics that greatly benefit the user interface (Shneiderman, 1998). Overall, these changes have led to a more intuitive interface that has allowed users to obtain information in a more efficient manner.
Principles of psychology and engineering have led to many of the aforementioned enhancements. Knowledge of attention, action, color, perception, and navigation has directly influenced many of the simple manipulations that greatly enhance the usability of web sites (Norman, 1993). Nevertheless, most of the changes mentioned address and resolve specific issues and do not affect the process by which an entire site is developed, nor do they change the way users think about a site.
More robust manipulations of the user interface can also be used to provide a better representation of a system's architecture. In fact, by taking advantage of a few key principles from the decision making literature, web pages can be made far more usable (Klein, Kaempf, Wolf, Thorsden, & Miller, 1997).
Paramount among these principles is the manipulation of an interface to allow for a more natural cognitive fit with the user. This approach has been characterized as a mental models approach to user interface design: find a paradigm that is a metaphor for a given task and is intuitive to users and apply it to the user interface (Shneiderman, 1998). Essentially, you want to generate a frame, or a natural analogy, that fits both the information in your web site and the navigational needs of the user. Classic examples of frames that work are: sports analogies to describe businesses; telephony as a metaphor for the mind; and automobiles as a metaphor for the human body. Computer programmers and software designers have also taken advantage of such frames: e.g., a simple visualization of a desktop with windows represents the complex interface of an operating system; a trash bin represents file deletion. User interfaces on the web can also take advantage of these existing frames, as well as new ones.
When done correctly, these efforts can result in a more powerful, yet simpler, interface. Nevertheless, broad-reaching changes to a user interface can prove disastrous. In trying to provide a more intuitive user-interface, one must be careful not to exploit incorrect mental models. In a number of cognitive domains, framing effects and in particular, mapping indirect visual models onto complex problems, have been shown to advance biases as well as create considerable confusion with regards to navigation. Providing an interface that is unintuitive will, at best, lead to confusion. Providing an incorrect mental model, however, can lead to disastrous consequences--including incorrect decisions, navigation mistakes, and worse yet, complete confidence in such choices.
A second area of relevance involves ways in which to present complex information. In order to do so, it is essential to know the information provided and understand how that information will be used. Information can be displayed in one of many fashions and how it is displayed will impact how it is interpreted and utilized. There are essentially two ways of displaying large amounts of information on the web: in terms of a hierarchy or as a simple list. Hierarchies have the advantage of nesting information in a natural manner, but are often more complex and time consuming than lists. Lists are quick and easy to traverse, but do not show the underlying structure of the information. Depending on the information presented, either form may be sufficient. However, using the wrong mode of presentation can lead to confusion and a lack of understanding.
This is nowhere more evident than in economic game theory, where researchers are just learning the consequences of presenting information in the wrong mode (Schotter, Weigelt, & Wilson, 1994). Game theorists have two methods for presenting information: normal form and extended form. In normal form, information is presented in list-like tables:
| Dell | |||
| IBM | Windows |
DOS |
|
| Windows |
Win, Win |
Win, Lose |
|
| DOS |
Lose, Win |
Lose, Lose |
|
Extended form takes the same information but incorporates a hierarchical structure, using a tree diagram:
What economists have recently found (and this was quite surprising to them) is that when the same information is presented in these two different modes, people tend to have a different perspective into the problem, and consequently perform differently (Schotter, Weigelt, & Wilson, 1994). Specifically, when the game consists of a series of hierarchical steps, people behave far more normatively when the information is presented in the extended form. In simple games where the information needed is uni-dimensional, however, the normal form seems to benefit people's understanding of the information. Not surprisingly, failure to present necessary information in a complementary mode often leads people toward sub-optimal performance.
Research into thinking and decision making has yielded similar results. In the heuristics and biases literature, for example, a harsh debate is underway focussing on the advantages of framing a problem in terms of frequencies (multiple instances), as opposed to traditional probabilities (single events) (see, for instance, Kahneman & Tversky, 1996; and Gigerenzer, 1996). Numerous studies, in fact, have shown that when people are asked questions in terms of frequencies, biases and errors quickly disappear--this is true of conjunction fallacies, base-rate neglect, and overconfidence (Casscells, Schoenberger, & Graboys, 1978; Gigerenzer, 1996; Cosmides & Tooby, 1996; Kahneman & Tversky, 1996).
Theorists have argued from an evolutionary perspective that this result is due, in part, to people's need to see things in terms of large numbers and multiple trials (Cosmides & Tooby, 1996; Gigerenzer, 1996). Current research, however, has suggested that what may be underlying this effect is the way in which the information is presented and incorporated. One explanation is that framing problems in terms of frequencies lends itself to a "nested-sets" formulation, whereby information can be viewed as a hierarchy. Probability frames, however, present information in a much more list-like manner that does not allow people to see the underlying hierarchies of information. This distinction may be underlying the effects of frequency in such problems. The implications, of course, are that the ways in which information is presented, whether in terms of hierarchies or lists, has an immense effect on how information is incorporated into the mind, and ultimately, on people's performance on tasks.
Web site designers and usability engineers can learn from these debates. Designing a user interface requires an explicit knowledge of not only the user needs, but also the information necessary to meet those needs. By understanding the different modes of presenting information, the ways in which the information will be used, and the process by which people think about that information, we will be able to provide a more intuitive and usable user interface.
References
- Casscells, W., Schoenberger, A., & Graboys, T.B. (1978). Interpretation by physicians of clinical laboratory results. New England Journal of Medicine, 299, 999-1001.
- Cosmides, L., & Tooby, J. (1996). Are humans good intuitive statisticians after all?
Rethinking some conclusions from the literature on judgment under uncertainty. Cognition, 58, 1-73.
- Gigerenzer, G. (1996). On narrow norms and vague heuristics: A reply to Kahneman and
Tversky. Psychological Review, 103, 3, 592-596.
- Kahneman, D., & Tversky, A. (1996). On the reality of cognitive illusions. Psychological Review, 103, 3, 582-591.
- Klein, G., Kaempf, G., Wolf, S., Thorsden, M., & Miller, T. (1997). Applying decision
requirements to user-centered design. International Journal of Human-Computer Studies, 46, 1-15.
- Kumar, H., Plaisant, C., & Schneiderman, B. (1997). Browsing hierarchical data with
multi-level dynamic queries and pruning. International Journal of Human-Computer Studies, 46, 103-124.
- Norman, D. (1993). Things That Make us Smart. Addison-Wesley, Reading, MA.
- Schotter, A., Weigelt, K., & Wilson, C. (1994). A laboratory investigation of multi-
person rationality and presentation effects. Economic Behavior.
- Shneiderman, B. (1998). Designing the User Interface. Addison-Wesley, Reading, MA.
© Internet Technical Group
Last update: March 6, 1999
URL: http://www.sandia.gov/itg/newsletter/mar99/decision_making.html
hosted by Sandia National Labs
Disclaimer: Neither Sandia Corporation, the United States Government, nor any agency thereof, nor any of their employees makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately-owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by Sandia Corporation, the United States Government, or any agency thereof. The views and opinions expressed herein do not necessarily state or reflect those of Sandia Corporation, the United States Government or any agency thereof.