[Different steps in construct validation of the Basel Drug and Alcohol Questionnaire (BDA)]

In a first study we developed a test (BDA) for measuring the degree of dependence on alcohol, medicaments and drugs in alcoholics as well as drug addicts. The present study analysed this test according to the principles of construct validity. The aim was to create a test with those items that optimally fulfil some of the following criteria. This was achieved with 22 out of the original 59 items of the test. The shortened test contains those items that optimally differentiate both externally (the addicted population from the rest of the sample: depressives, neurotics and normals) and internally (the degree of dependence). The score obtained with 22 items was more efficacious in differentiating the addicts from the rest of the sample than the total score of 59 items. 86% of the 22 items are sufficiently sensitive to measure changes during hospitalisation in a psychiatric clinic. However only 33% of the 22 items can effectively discriminate between alcohol-, medicaments- und drug-dependence. Intercorrelation and factor analyses with other variables that were also measured indicate that a large part of the variance arises from unspecific "general psychopathology" and only a small part is specific for drug-dependence. The test is sufficiently reliable (rtt = .89), unifactoral, and with clear mean test-item correlation (rit = .54), to justify further investigation. We recommend this 22 items test for scientific research into dependence, in spite of the questions that still remain open. The English version of the 22 items is given in the appendix.