Studies of cross-national incarceration rates have suffered from at least three flaws in their research designs: (1) the use of 'static' rather than 'flow' designs which cause the length of sentence to be confounded with rates of imprisonment; (2) the lack of control over definitions of criminal behavior across countries; and (3) the use of the total population rather than the true population at risk in calculating incarceration rates, which confounds the rate of imprisonment with the rate of criminality. The flow design used in this study describes the sequential processing or flow of persons through the criminal justice system over a given period of time. The study was restricted to the offenses of murder, manslaughter, rape, robbery, aggravated assault, burglary, larceny, and auto theft. Results show that when cross-national variation in the incidence of crime is taken into account, the United States is no more likely to imprison offenders than Great Britain and somewhat more likely to incarcerate than Canada. The most striking feature of the comparison is that the differences among the countries in the use of imprisonment do not approach the magnitude suggested by earlier cross-national comparisons. Imprisonment rates for individual offenses studied are provided. Data sources and methods of computation are discussed in the appendix. Footnotes are given.