In multiple regression analyses, the monthly time-series suicide & homicide rates for the USA from 1957-1986 were associated with unemployment & marriage rates. In contrast, annual time-series suicide & homicide rates were not as strongly associated with the social variables. (11 refs.)