In this paper we present a method for participant classification, based on trait and state, in the experimental evaluation of ESP (extrasensory perception). We conducted three Ganzfeld experiments with a sample of 237 participants. In experiment I (N = 60) twenty participants ranked the target stimulus in the first position, achieving a non-significant rate of correct guesses of 33.3% (z = 1.48, p = .07, one tailed). In experiment II (N = 90) only 26.6% of the participants’ guesses were correct (z = .35, p = .36, one tailed). Weighting trials in the second experiment on the basis of the most successful predictors of the participants’ performance in the first experiment increased the rate of correct guesses from 26.6% to 36.4%. Results in a confirmatory experiment (N = 87) were not significant (32.7%, z = 1.32, p = .09, one tailed). However, overall results in this study were consistent with the effects sizes data reported in previous meta-analyses.