'Behavioral contagion' refers to an increased tendency for a behavior to be performed when socially related persons have already performed it. So understood, behavioral contagion may be involved in criminality, conduct disorder, drug abuse, suicide, and teenage pregnancy. In a recent paper we showed that contagion, when present, generates highly distinctive result patterns in survey data. These patterns allow one to infer that contagion is present even though it has not been directly observed. The present paper extends this approach to case-control designs, where, it will be shown, contagion has equally distinctive but different consequences. The analysis is illustrated with a Canadian study of conduct disorder and further evidence is presented that this condition is contagious. Limitations of the approach and alternative interpretations of results are discussed.