Demographic-specific estimates of crime rates are used as indicators of criminality for different population subgroups, and multivariate analyses relate subgroup criminality rates to various structural factors. Basic input data are from the Uniform Crime Reports and the National Crime Survey. Age and race alone are major factors in crime rate variations for burglary, robbery, and murder. Period and boom, however, are also significant in variations in age-by-race crime rates for the three crime types. The interaction of race with age or with period further accounts for crime rate variations. The race effect is in the expected direction, with higher crime rates for nonwhites than for whites. Crime rates tend to increase through juvenile and early adult years and then decrease through older adult years. Models used to evaluate the data are noted, and mathematical formulas are given. 56 references, 6 tables, and 33 figures