Sensitivity analysis for nonrandom dropout: A local influence approach

Diggle and Kenward (1994, Applied Statistics 43, 49-93) proposed a selectio n model for continuous longitudinal data subject to nonrandom dropout. It h as provoked a large debate about the role for such models. The original ent husiasm was followed by skepticism about the strong but untestable assumpti ons on which this type of model invariably rests. Since then, the view has emerged that these models should ideally be made part of a sensitivity anal ysis. This paper presents a formal and flexible approach to such a sensitiv ity assessment based on local influence (Cook, 1986, Journal of the Royal S tatistical Society, Series B 48, 133-169). The influence of perturbing a mi ssing-at-random dropout model in the direction of nonrandom dropout is expl ored. The method is applied to data from a randomized experiment on the inh ibition of testosterone production in rats.