Five examples of nontraditional psychosocial treatments used for children/adolescents are reviewed: eye movement desensitization and reprocessing, electroencephalographic (EEG) biofeedback, deep pressure/touch therapies, stress-challenge treatments, and confrontational scare treatments. The generic recommendations from the September 1992 National Institutes of Health Conference on Unconventional Medical Treatments are summarized. Additional screening principles specific for psychosocial treatments are proposed and applied to the five treatments. The screens do not validate treatment efficacy or evaluate the quality of any previous research, but only facilitate decisions as to whether treatments deserve controlled investigation. Scientific evaluation of the nontraditional treatments reviewed could in general benefit from blinds (at least for assessment); control conditions matched for intensity, frequency, and duration (double blind where feasible); dose-response studies; testing of generalization and endurance supplements or boosters for quick, cheap treatments with time- or domain-limited effects; and comparing cost-effectiveness with established treatments. Two unscientific pitfalls must be avoided: embracing new treatments uncritically and rejecting them without fair examination. These pitfalls must be skirted without dissipating scarce research resources.