detail, with a reported response rate by surveyed inmates of about 94 percent, but is susceptible to the failings of the subjects' self-reporting, including problems of memory and deceit. The latter survey of correctional agencies reached high levels of coverage on some matters, but encountered bureaucratic problems on others, as illustrated by the California Department of Corrections, the largest prison system in the United States, which reported national origins of "immigrants" based on data collected on citizens as well as noncitizens. We are as careful as possible in this chapter to use the term immigrant only to refer to noncitizens and to use the term noncitizen instead of immigrant when the original sources do so; but there may be some cases in which the original sources we rely on have included foreign-born citizens among immigrants without indicating they are doing so.

By 1980 the percentage of the U.S. population formed by noncitizens was estimated to be between 4 and 5 percent (see Isbister, 1996:37). The Survey of State Prisons reported that more than 4 percent, or 31,300 state prison inmates, were not U.S. citizens. The Department of Corrections survey indicated that more than 7 percent, or 71,294 state and federal prison inmates, were not U.S. citizens. The latter figure is larger in part because it includes federal prisoners and also because the survey was conducted two years later during a growth spurt in prison populations. However, there is still a disparity between these two sources that underscores the difficulties of assembling information on immigration and crime in the United States. This point is further underscored by a recent Bureau of Justice Statistics report (Scalia, 1996) that indicates that there were less than 19,000 noncitizens in federal prisons in 1994, more than 7,000 fewer noncitizens than indicated in the Corrections Compendium report from one year earlier referenced above (Wunder, 1995). Scalia's report indicates in its first figure (1996:1) that about 14,000 noncitizens were serving a sentence of imprisonment in a federal prison in 1991, whereas the last table in this report (1996:10) indicates that 9,916 noncitizens were inmates in federal prisons in 1991.

We are left with a range of estimates that between 4 and 7 percent of prison inmates in the United States are noncitizens, compared with estimates that legal immigrants constitute between 4 and 5 percent of the U.S. population, with perhaps as much as 1 percent more of the U.S. population being illegal immigrants (Passel and Woodrow, 1984). Of course, this tells us nothing about whether specific immigrant groups are over- or underrepresented in prisons. The most systematic and comprehensive published data on the national origins of noncitizens in prisons are found in the 1991 Survey of State Prisons. Nearly half of the immigrants in state prisons (47%) came from Mexico, whereas nearly another fourth (26%) were from Latin and Caribbean countries. Together, these figures indicate that a large majority of immigrant inmates in U.S. state prisons, perhaps between 70 and 80 percent, are of Hispanic origin. This distribution of the national origins of immigrants in state prisons is quite similar to that found among immigrants convicted of an offense in the U.S. district courts in 1994 (see