A study of hospitalization anxiety among 408 medical and surgical patients in four voluntary, nonprofit, general hospitals is reported. The hospitals included a major teaching center, a large city hospital and two small community hospitals. A new Hospitalization Anxiety Scale was tested in a pilot project and revised for use in this study. Patient anxiety levels were highest in the university hospital and lowest in the two small community hospitals. The large city hospital fell between these two groups. With the exception of malignancy, the diagnoses correlated poorly with the anxiety levels. Cancer patients tended to be more anxious than the other patients. In discussing possible causes of the observed differences in arrested anxiety, four aspects of the hospitalization process must be considered: the patient's antecedent environment, hospital characteristics, the nature of the illness and the person who is the patient. Some applications of this study to the training of health personnel are emphasized.