Predictive Indices of “Success” for Foster Home Placements

Abstract

The sine qua non of any instrument constructed for the purpose of prediction is its “predictive validity”—i.e. it must be demonstrated that the instrument in fact does the job which it was meant to do. For social and/or behavioural prediction the construction of predictive instruments is based on sample(s) drawn from a defined population. “Predictive validity” of the instrument is demonstrated when it effectively predicts the required events or characteristics when used on new samples drawn from the same population.