This study aimed to evaluate all known and hypothesized predictors for nightmare frequency measures in a population with frequent nightmares. A total of 666 Internet-recruited participants completed questionnaires on nightmares, sleep, and psychopathology, 146 of whom further completed a subsequent 7-day diary. In contrast to previous research, comparison of questionnaire- and diary-measured nightmare frequency revealed a significantly higher log-transformed nightmare frequency on the questionnaire: t(127) = 4.43; P < 0.001. No differences were found regarding the number of nights with nightmares, t(127) = 0.61; P= 0.54. Regression analyses showed that subjective sleep quality was the only variable significantly associated with nightmare frequency variables in the whole sample (R2adj between 10.5–11.5%; P < 0.01). These results support the notion that nightmares are independent from other mental complaints in a population of nightmare sufferers and should therefore be viewed from a sleep medicine perspective: as a sleep disorder that can and should receive specific attention and treatment.