Experimental control versus realism: Methodological solutions for simulator studies in complex operating environments

Abstract

The fundamental presumption behind this thesis is that classical experimental psychology provides insufficient methodological solutions for realistic simulator studies in complex operating environments. A complex operating environment is a potentially hazardous human-machine system, where human operators interact with advanced technology to achieve a commercial objective, such as producing nuclear power, pumping oil and gas, or transporting passengers between destinations.

The specific topic of the dissertation is the conflict between experimental control and the realism of simulator studies in complex operating environments. In order to draw reliable causal conclusions an experiment has to be controlled, but experimental control has a tendency to impose artificial constrains on the research setting and thereby reduce the level of realism. This trade-off is intolerable for simulator studies in complex operating environments, since a high level of realism is necessary for the transfer of experimental results from simulation to real operation. The ambition of the thesis is to demonstrate that the conventionally accepted conflict between realism and experimental control is a simplistic model, and a possible barrier to successful experimentation in applied research environments. It is considered necessary to rethink the relationship between control and realism in order to generate methodological solutions that can support practical human-machine research. Hence, it will be suggested that realism and control can be fully compatible if existing experimental methods are used strategically and developed further. In this respect, the thesis will propose a package of experimental design solutions and human performance assessment techniques that provide sufficient experimental control during realistic simulator studies in complex operating environments.