Mild and moderate insomniacs were randomly assigned to one of five group therapy conditions: relaxation without muscle-tension, stimulus control plus relaxation without muscletension, placebo, self-monitoring, and waiting list. Mild insomniacs were differentiated according to the duration of their sleep disturbance. Insomniacs were told that improvement would not be noticeable until after the final therapy session. During this counterdemand period, the two relaxation groups showed significantly greater decreases in sleep onset latency than the control conditions. In addition, it appeared that duration affected treated outcome. The effectiveness of treatment packages and self-monitoring in alleviating insomnia is briefly discussed. © 1979.