A nominal scale with 15 descriptive categories ordered in three general classes was designed. Being more complete than previous scales it captures systematically formal and content aspects of bizarreness.
The data consisted of 39 REM-dream reports and 12 presleep waking fantasies (triggered by a stimulus scene) collected from 6 subjects in 12 experimental nights. The reports were rated by 4 independent judges according to the scale and for comparison according to a global (0-5) rating in bizarreness. Interrater-reliability pairwise for each category was: 61-100%. Corrected bias for low frequencies: Cohen's kappa was 0.06-0.56. Category frequencies correlated 0.82-0.86 validly with the global judgement (0-5).