Schulte-Mecklenbeck, Michael

Risky Information Search in Databases

Abstract

The usage of databases as a research tool and decision theory as a research topic is demonstrated in a web based experiment. In two conditions a quasi naturalistic decision task was presented in a between subjects design. One group had to search a given data base for information (unstructured version). The second group was confronted with a prestructured information environment using hyperlinks as categories (structured version). Differences were found in the number of information items used (more items used in the structured than in the unstructured version), the importance of certain categories and the usage of probability information in the decision process (less usage in the unstructured version). Differences were found between the internet and a local sample. The present experiment introduces the combination of process tracing with current web-based techniques (search engines (MySQL Database, PHP queries) and hyperlink structured information). These results support the idea of active information search (Huber, Wider & Huber, 1997) that decision makers are more concerned about information on new alternatives and worst case plans than on probability information in risky tasks. A second important result is that web-based research tools offer a reliable possibility of exploring decision behaviour.