In his monograph "Die Akte Astrologie" ["The Astrology Dossier"], Sachs (1999) claims that statistically significant relationships exist between signs of the Zodiac and behavior. The author presents these relationships in the form of local associations in contingency tables that cross the signs of the Zodiac and numerous behavior categories from statistical registers of Switzerland. As one example, we discuss and re-analyze his data on astrology and crime that, according to Sachs, demonstrate a clear relationship. Although Sachs’ methods are largely transparent, his conclusions are not valid. The main reasons for this lack of validity are that the analyses capitalize on chance and fail to take into account the mutual dependency of statistical tests performed on the same data. Our re-analyses of Sachs’ data on criminal convictions and the signs of the Zodiac using (1) Configural Frequency Analysis and (2) two-way cluster analysis suggest that the conclusions drawn in the monograph are untenable. Statistical and substantive aspects of our results are discussed.