Faculty Working Paper 93-0126 Effects of Knowledge Types on Objective and Perceived Choice Performance The I.Ibrar/ of t'lo MAY 3 1993 University of Illinois of D'h^nji-rhafnoalon Eloise Coupey Sunder Narayanan Department ofBusiness Administration Department ofBusiness Adnrinistration University ofIllinois University ofIllinois Bureau of Economic and Business Research College of Commerce and Business Administration University of Illinois at Urbana-Champaign BEBR FACULTY WORKING PAPER NO. 93-0126 College of Commerce and Business Administration University of Illinois at Urbana-Champaign April 1993 Effects of Knowledge Types on Objective and Perceived Choice Perfornnance Eloise Coupey Sunder Narayanan Department of Business Administration Digitized by the Internet Archive in 2011 with funding from University of Illinois Urbana-Champaign http://www.archive.org/details/effectsofknowled93126coup EFFECTS OF KNOWLEDGE TYPES ON OBJECTIVE AND PERCEIVED CHOICE PERFORMANCE Abstract Two components ofknowledge used in making choices are examined: alternative- specific knowledge and choice rule knowledge. The effects ofthese knowledge types, both singly andjointly, upon objective choice quality and perceived choice quality are examined in two studies. The results ofStudy One indicate that the knowledge types are differentially beneficial, and that subjects tend tobe more overconfidentabout theperceived quality oftheirchoices when they have alternative-specific knowledge than when they have choice rule knowledge. The hypothesis that this difference is due to subjects' greater awareness ofalternative-specific knowledge, rather than rule knowledge, is examined and supported in Study Two. EFFECTS OF KNOWLEDGE TYPES ON OBJECTIVE AND PERCEFV^ED CHOICE PERFORMANCE Eloise Coupey and Sunder Narayanan University ofIllinois at Urbana-Champaign Knowledge is a frequent and desirable component ofdecision making for choice. Knowledge can be used to guide the acquisition and organization ofrelevant information and aid in the selection ofan appropriate rule for evaluating options for choice. However, despite much research on decision making for choice, questions still remain. Forexample, can the knowledge that influences decision making forchoice be classified into different types? Ifso, are some types more beneficial than others? Are decision makers aware of different types ofknowledge? Do they recognize that one type may be more useful for making a choice ofgood quality than another type? The research described in this paper addresses these questions about knowledge, building upon pastresearch in decision making and problem solving. In subsequent sections ofthis paper we present a theory-based rationale for examining two typesof knowledge that are relevant for making achoice: alternative-specific knowledge, i.e., knowledge about the alternatives, and rule knowledge, i.e., knowledge ofa choice rule for integrating andevaluating information about alternatives. Empirical assessmentsofthese two knowledge types and their effects on objective and perceived quality ofchoice are obtained with two studies. The results ofthese studies indicate that the knowledge types are differentially beneficial, and that subjects tend to be more overconfident when they have alternative-specific knowledge than when they have rule knowledge. BACKGROUND Anderson (1983) described knowledge as consisting ofdeclarative knowledge and procedural knowledge. Declarative knowledge consists offacts about the items in a 1 problem space, such as the attributes ofthe alternatives in a choice problem, and the values ofthese attributes for different alternatives. Procedural knowledge reflects the skill component: knowing how to use the declarative infonnation. The rules typically ascribed to decision makers forarriving at choices (e.g., elimination-by-aspects (Tversky 1972)) are examples ofprocedural knowledge. They direct the use ofdeclarative information. Newell and Simon (1972) describe a similar dichotomy with their distinction between representation and method for problem solving. The internal representation reflects the current situation for the decision maker. It is contained in aproblem space, along with all otherpossible states ofknowledge. The knowledge states are nodes, connectedby processes. Processes are carried out by the activation ofappropriate operators. Ifthe internal representation enables the decision makerto selectand apply a method, such as a choice rule in a choice problem, which, when executed, will achieve a desired end, the decision maker carries out the method. Ifno method is appropriate the decision makerchanges the internal representation, adjusting it until a method can be retrieved and applied. A third, similar distinction is made by Chi, Glaser, and Rees (1981) in their study ofthe developmentofexpertise. They describe expertise as consisting oftwo components: abody ofusable information and procedural skill. They suggest that what distinguishes experts from novices in any domain is that experts have the ability to convert knowledgeof fundamental principles (as in physics problems) into procedures for problem solving. That is, a largeamount ofalternative-specific information win notautomatically make a decision makeran expert decision maker. A lotofinformation may be useless unless the decision maker knows how to evaluate it The knowledge components described above are necessary forevaluating alternatives to make achoice. It must be noted, however, that the choice process exists beyond the stage ofevaluating alternatives. Forexample, considertheprocess model depicted in Figure 1. The stage ofNeed Recognition may necessitate knowledge about