1. The scores obtained with a set of three "life situation tests" in an unselected group of 233 patients correlate closely with the average length of residence in the community. No correlation was found in the manic-depressive group of patients.2. The high degree of correlation suggests that validity may be claimed for the tests.3. The tests may be said to be reliable and to yield similar or parallel scores if repeated from three to six months after the original application.4. The test scores seem to afford a fairly valid basis for predicting the probability of continuous residence in the community after discharge from the hospital.