THE ANALYSIS OF SPATIAL PARAMETER VARIATION WITH JACKKNIFED PARAMETERS

AS STATISTICAL MODELS ARE APPLIED WITH INCREASING FREQUENCY TO SOCIAL PROCESSES, IT BECOMES IMPERATIVE TO DEVELOP STRONG DIAGNOSTIC MEASURES OF MODEL PERFORMANCE. THE AUTHOR DEMONSTRATES THAT JACKKNIFED PARAMETERS, AS DEFINIED IN HIS STUDY, CAN BE USEFUL AND POWERFUL INDICATORS OF MODEL PERFORMANCE ACROSS SPACE. HE COMPARES JP'S TO OTHER METHODS IN MONTE CARLO EXPERIMENTS EMPLOYING REALISTIC SPATIAL DATA.