In accordance with the theory of mental tests in this study it was tried to improve the reliability of diagnostic interview data by enhancing the number of items concerning the same biographical topic (= homogeneous items). Reliability was defined as stability of the statements in repeated interviews. The reliabilities of three interview items concerning different topics were calculated. The interviews were carried out under two different conditions: Under the experimental condition (30 subjects) the item, of which the stability was calculated, was the last in a series of ten homogeneous items, under the control conditions (30 subjects) the item was the last in a series of ten heterogeneous items. Under the experimental condition the reliabilities of two items, concerning the diseases of children and concerning the school record, were significantly improved. Not improved was the reliability of the appraisal of parent's educational methods. The reliabilities of the interview data not seemed to depend on the age of the subjects.