Designing a Cost-Effective Field Trial for an Intervention in a Messy System With Multiple, Interacting Variables
The "randomized field trial" is a standard methodology (see Implementing Randomized Field Trials in Education - The Result of Workshop, NAP, 2004) used to discover causative agents in social science and medicine. However it is simply too expensive to imagine using a large scale random field trial to unravel the effects of multiple simultaneous interventions, none of which are common practice. We hypothesize that it is possible to obtain scientifically meaningful cause and effect data with relatively small numbers of participants by designing a particular variant of a field trial. We suggest that the field trial a) have an outcome variable that is quantitative and objective, b) reproduce the outcome of the full treatment, and c) have an experimental outcome variable that far exceeds the historic norm. We hypothesize that under the above conditions it is possible to build knowledge and a strong case for the effectiveness of an intervention while not necessarily teasing apart all the interacting factors.
