MICROWORLDS FOR EXPERIMENTAL RESEARCH - HAVING YOUR (CONTROL AND COLLECTION) CAKE, AND REALISM TOO

Microworlds (MWs) are dynamic computer-generated environments that sub jects interact with in the laboratory and that simulate conditions enc ountered in the field. Precise levels of experimental control and impr oved accuracy and efficiency of data collection procedures are charact eristic of MWs. It is proposed that these benefits are achieved with c oncomitant gains in internal validity (afforded by high levels of expe rimental realism) and external validity (afforded by the replication o f the temporal-interactive nature of most field phenomena). To illustr ate these ideas, three sets of MW studies are described that investiga ted rumor and behavior in the stock market (BROKER), escalation behavi or (INVE$TMENT CHOICE$), and the application of foraging theory to int ernet shopping (CYBERSHOPPER).