Computer simulated shopping experiments open up new opportunities for marketing research, allowing researchers to collect purchase data in a tightly controlled yet realistic environment, at relatively low cost and with a high degree of flexibility. While a number of authors have used computer simulations to generate purchase data, research on the validity of these outcomes is scarce. The present paper aims to improve insights into the validity of computer simulation experiments, by providing a replication and extension of the work by Burke et al. (1992). Experimental purchase data are generated and compared with real life scanner data to provide additional evidence for major biases examined by Burke et al., and to assess the effect of repetitive choice tasks and incomplete product assortments on the realism of purchase quantity and choice data. The results show that, overall, the experimental data correspond quite closely to actual purchase data. Consumers exhibit realistic purchase variation, and provide meaningful data throughout the experimental periods. Yet, some biases are found in category purchase quantities -which are overestimated - and in purchase shares of generic products - which are underestimated. While promotion quantity effects appear to be slightly overestimated, more research is needed on this issue.
Campo, K, Gijsbrechts, E and Guerra, F (1999), "Computer Simulated Shopping Experiments for Analyzing Dynamic Purchasing Patterns: Validation and Guidelines", Journal of Empirical Generalisations in Marketing Science, Vol. 4, No. 2