The results of this test of the validity of a Swedish translation of the MCAA show that all items tended to tap into the same general criminal attitude construct. Together with the high item-total correlations to the subscales, this suggests that the items are homogeneous in terms of content, which provides conditions for high reliability. This was confirmed by the homogeneity analysis as well as by the retest reliability check for the small group of offenders who took the test twice. The tests showed that the MCAA effectively measured a general criminal attitude and at least three relatively independent subfactors that distinguished between the attitudes of criminal offenders and a nonoffender group. The test was also sufficiently sensitive to measure therapeutic change. Future research should address whether the test can predict future criminal behavior. The two samples used in the study consisted of 184 criminal offenders and 556 participants who completed the questionnaire through an Internet Web site open to the general public. The offender group consisted primarily of men (n=169). The Internet sample, which consisted of people who chose to visit the site and complete the questionnaire, consisted of 243 men and 313 women. 2 tables, 1 figure, and 17 references