Evaluation of unconstrained sorting data

Two statistics are offered for evaluating unconstrained sorting performance in a specific task when categories are E-defined. One statistic is based upon empirically estimated sampling distributions and can be used for determining sorting performance significantly deviant from chance for any number of S-defined categories; the second statistic can be used to evaluate consistencies between S- and E-defined categories, regardless of the number of categories used by S. The present procedure provides a statistical basis for evaluation of performance where no adequate evaluation procedures are now available.