CMStatistics 2020: Start Registration
View Submission - CMStatistics
Title: Design of experiments on networks for decision making Authors:  Vasiliki Koutra - University of Southampton (United Kingdom) [presenting]
Steven Gilmour - KCL (United Kingdom)
Ben Parker - University of Southampton (United Kingdom)
Abstract: Design of experiments provides a formal framework for the collection of data to aid decision making, ranging from what drug treatment is most effective through the choice of wheat variety to maximise yield to the selection of an internet advertisement to optimise revenue. When such experiments are performed on connected units, i.e. linked through a network, the resulting design, analysis and decision making is more complex; e.g. is the observed response from a given unit due to the direct effect of the treatment applied to that unit, or the result of a network, or viral, effect arising from treatments applied to connected units. We propose a methodology for constructing efficient designs which controls for variation among the experimental units from two sources: blocks and network interference, so that the direct treatment effects can be precisely estimated. We provide evidence that our approach can lead to efficiency gains over conventional designs such as randomised designs that ignore the network structure. We illustrate its usefulness for experiments on networks.