Document Type


Publication Date



Part I of the full article briefly describes the history and current slate of research into children's suggestibility. In this part, we argue that, although psychological researchers disagree considerably over the degree to which he suggestibility of young children may lead to false allegations of sexual abuse, there is an overwhelming consensus that children are suggestible to a degree that, we believe, must be regarded as significant. In presenting this argument, we respond to the contentions of revisionist scholars, particularly those recently expressed by Professor Lyon. We show that there is good reason to believe the use of highly suggestive questions remains very common, and hat these questions present a significant possibility that children will make false allegations even on matters such as sexual abuse. Part II develops a framework, using Bayesian probability theory, for considering the findings described in Part I. We argue that there is merit to the traditional - and constitutionally compelled - view that an inaccurate criminal conviction is a far worse result than a failure to reach an accurate conviction, and that this perspective should inform the design of legal systems. With this in mind, we explain that even relatively slight probabilities of false allegations are potentially significant. Moreover, we show that the very substantial probability that a child who has been abused will fail to reveal the abuse tends, perhaps counterintuitively, to diminish the probative value of an allegation of abuse when it is actually made. In the discussion below, taken from Part III of the longer article, we turn to discussion of the legal implications of our analysis.