In a within-subjects design, 18 subjects listened to white noise, the intensity of which was controlled by themselves. They were instructed to choose the intensity they experienced as "clearly annoying." Three kinds of white noise were used: continuous, intermittent regular, and intermittent irregular. In the two intermittent conditions, the proportion of time with the noise on was 50%. The duration of on- and off-periods in the regular condition was 1.15 sec.; in the irregular condition, it varied between 0.25 and 1.65 sec. The subjects chose their "clearly annoying" level three times for each noise type. The results showed that the mean level chosen was 83.9 dB for continuous, 90 dB for intermittent regular, and 89.6 dB for intermittent irregular noise. Pairwise comparisons indicated a significant difference between continuous noise and each of the two intermittent noises while there was no difference between the two intermittent noises. The results are interpreted as indicating that noise-induced annoyance may be a function of the over-all amount of noise rather than the mere presence or absence of intermittency, at least when no concurrent demanding task is performed and when the required annoyance level is set by the subjects themselves. The results further showed that the intensity chosen by the subjects correlated negatively with scores from Weinstein's Noise Sensitivity Scale while the intensity chosen was unrelated to extraversion or neuroticism scores as measured by Eysenck Personality Inventory.