Research relied primarily on published national and State data. Fraction measures reflect the proportion of a base population confined at a given point in time; chance measures reflect the probability of entering confinement during a given time period, given membership in a specified base population. Deeper examination found that the numerator of the traditional 'rate of imprisonment' leaves much to be desired; it ignores differences among jurisdictions and changes over time in the laws and procedures under which offenders are allocated to confinement facilities. Chance measures represent the cumulation of decisions made about cases at the various stages of the criminal justice system. Two other measurement issues, time served and prevalence measures, also receive attention. The recommendations deal with slight modifications in the ways data are collected and some changes in the ways data are presented in published reports. Over 100 references, footnotes, 6 figures, and 22 tables are included.