Simulation of Experiments for Data Collection – a replicated study

Abstract

Simulations can be used as a means for extension of data collection from empirical studies. A simulation model is developed, based on the data from experiments, and new data is generated from the simulation model. This paper replicates an initial investigation by Münch and Armbrust, with the purpose of evaluating the generality of their approach. We replicate their study using data from two inspection experiments. We conclude that the replicated study corroborates the original one. The deviation between the detection rate of the underlying experiment and the simulation models was 2% for the original study and is 4% in the replicated study. Both figures are acceptable for using the approach further. Still the model is based on some adjustment variables that are not directly possible to interpret in terms of the original experiment, and hence the model is subject to improvement.