Prison Admissions in Wisconsin 1990-1999
This page contains links to the Excel spreadsheets which we have used to analyze Wisconsin prison admissions, as well as supporting analysis notes. We need to emphasize that this is work in progress, and these are not "publication quality" spreadsheets or graphics, but we are putting them out as quickly as they are generated so they can inform public discussion and debate.
The six counties analyzed are the only counties in Wisconsin with more than 1000 black people who are not prisoners. 76% of Wisconsin's black population lives in Milwaukee County. "Wisconsin balance" is the rest of Wisconsin beyond the six itemized counties, "Big five" is the sum of Dane, Waukesha, Kenosha, Racine, Rock counties.
Admission Status. Black & white trends in admission status (new sentence only, probation/parole violation only, new sentence plus probation or parole violation). Wisconsin Totals: Status Trends Excel Spreadsheet
November 1, 2001
Analysis of Wisconsin Prison Admissions 1990-1999
Department of Corrections data file of persons who were in Wisconsin state prison system any time between January 1, 1990 and April 30, 2000. A "case" is a spell in prison; if a person is released on parole, s/he becomes a new "case" if s/he reenters prison. Cases are excluded from all analysis if they are missing information on county of sentencing or racial classification.
1) Offense is our classification of the "current governing offense," which is usually the most serious offense. Some inmates have multiple offenses. Violent offenses are homicides and assaults (including sexual, aggravated, and other). Robbery and Burglary are grouped together as serious property offenses which are understood to have a potential for violence as well. Theft includes non-violent larceny and theft, which we pull out separately because this single offense group accounts for many prison admissions. Drug includes all drug offenses. Other is all other offenses, which are a broad range of non-violent and generally less serious crimes; this category also includes what we call "derived" offenses, which include escape, bail jumping, failing to answer a warrant, etc. There is also a small number of cases for which the offense is missing in the record.
2) County of sentencing pertains to the current governing offense; people are generally sentenced in the county in which the crime was committed. Cases are excluded if the county of sentencing is missing or if the offender has been sentenced in more than one county. Nothing in the record permits us to identify the county of residence at the time of sentencing, and a person can obviously be sentenced in a county in which s/he does not reside.
3) Admission status. We have created three groups in this analysis: 1) new sentence only, 2) probation or parole violation only, 3) probation or parole violation coupled with a new sentence. This last group is very heterogeneous, as the new sentence may be more or less serious than the previous offense and may or may not be the current governing offense.
4) Geographic units. In addition to the state of Wisconsin as a whole, we analyze the six counties (Milwaukee, Dane, Waukesha, Kenosha, Racine, Rock) which have more than 1000 black residents who are not prisoners. We also analyze the "Wisconsin balance," which is the rest of Wisconsin outside these six counties, and the "big five" which are the five counties other than Milwaukee with significant black populations.
5) Population of each racial group for appropriate geographic unit is taken from the US Census population estimates for the appropriate year. These have not been adjusted for the 2000 census, as adjusted estimates are not yet available.
6) Racial/ethnic groups. All Hispanics are classified as Hispanic and are excluded from all other racial groups. We are distinguishing among whites, blacks, Asians, American Indians and Hispanics. For some counties, Asians and American Indians are grouped together as "other" because there are too few of them for separate analysis.
7) Calculation of rates. In accord with standard practice, rates are expressed as prison admissions per 100,000 population of the appropriate geographic/racial group. For example, the number of black persons admitted to prison in 1998 who were sentenced in Milwaukee County is divided by the Census Bureau's black population estimate for Milwaukee County in 1998, and then that quotient is multiplied by 100,000.
Sources of Error
Imprisonment rates calculated on small populations (e.g. minority populations outside Milwaukee) can change significantly depending on the accuracy of the population estimates, and can be substantially influenced by non-resident offenders (those sentenced in a county who did not reside there at the time of sentencing). Census population estimates do take account of migration, so, in principle, recent arrivals to a county who are residing there would be counted in the population estimate, but it is well understood by demographers that it is hard to get fully accurate estimates of migration. Small differences in rates between racial groups, counties, or years should not be over-interpreted, especially for minorities. Large differences in rates are likely to remain large even when these sources of error are considered.
Questions or Comments? Email Oliver -at- ssc -dot- wisc -dot- edu. Last updated December 25, 2004 © University of Wisconsin.