Henry and Short (1954) proposed a theory of suicide and homicide based on the frustration-aggression hypothesis. Their central assumption was that suicide and homicide were alternative aggressive responses to frustration, defined as the condition when a goal response suffers interference. Henry and Short focused on the impact of economic frustrations generated by fluctuations in the business cycle on suicide and homicide. At the level of individual psychology, Henry and Short suggested that external restraints increased the legitimization of other-directed aggression. They also suggested that child-rearing practices affected the legitimization of other-directed aggression and the choice between suicide and homicide. Since 1954, most investigations of the relationship between suicide and homicide have been based implicitly or explicitly on Henry's and Short's 1954 conceptualization. Methodological and theoretical fine tuning has occurred, but the stream analogy, which links homicide and suicide through the aggression-frustration hypothesis, has remained intact. Some researchers, however, notably Richard C. Quinney (1965), rejected this hypothesis, arguing that suicide and homicide were causally separate phenomena, based on the finding that suicide varied directly and homicide inversely with economic development. Douglas (1964)and Hendin (1967) criticized the methodology that based suicide-homicide theory on broad cross-national statistics. They argued that individual cases of suicide and homicide must be analyzed in the context of particular psychological and sociological factors that influenced the victims. Cumulative findings from such case studies, they argued, should be the empirical basis for developing theory.