You asked (1) how prison inmates counted for census purposes, (2) how the method of counting prison inmates and hospital and skilled nursing facility residents affects legislative redistricting, and (3) if the Department of Correction has a way to determine an inmate's town of residence.

SUMMARY

The U.S. Census Bureau counts people at their “usual residence.” For prisoners, this means in the prison or jail. This practice is consistent with how it counts other people who are living in group quarters. For example, college students living in dormitories are counted as living in those dormitories and long-term patients are counted as living in a health care facility.

For redistricting, this means that localities in which large correctional facilities or universities are located gain population and localities in which a prisoner or student had been living at the time of admission lose population.

The Department of Correction (DOC) collects and maintains address information at time of admission. Based on the data DOC provided, it apparently collects mailing address information. It does not verify the reported address. In the 2000 census, 34 towns were the residence of at least 50 prisoners.

The issue of whether to count prisoners at their “permanent home of record” rather than at their usual residence is becoming more controversial as the number of prisoners continues to rise. Congress directed the Census Bureau to study the issue and it reported that counting prisoners at their home of record would be expensive for both the bureau and the correctional facilities. Further, it could not be performed as accurately as counting prisoners at their usual residence.

The bureau also established a panel to examine its rules and practices relating to determining residence, including those relating to prisoners. Its most important recommendation about prisoners, that the Census Bureau conduct a research and testing program, is based on its conclusions that there is merit to the argument that prisoners should be counted at their home of record but that the ability to do so does not now exist.

COUNTING PRISON INMATES FOR CENSUS PURPOSES

The Census Bureau counts people in their usual residence, and considers an inmate's usual residence as the facility in which he is housed. This is true whether the inmate is in a prison, jail, detention center, or correctional facility with any other designation. The practice is based on the bureau's interpretation of its authorizing legislation.

The first Census Act, adopted before the 1790 census, provided:

“That every person whose usual place of abode shall be in any family on the aforesaid first Monday in August next, shall be returned as of such family…and every person occasionally absent at the time of the enumeration, as belonging to that place in which he usually resides in the United States.”

OTHER GROUPS WITH ATYPICAL LIVING SITUATIONS

The concept of “usual residence” has been followed consistently since the first census. It is customarily defined as the place where a person lives and sleeps most of the time. It can be different than a person's voting or legal residence. It is applied to all people with atypical living arrangements, not just prisoners.

The usual residence concept is difficult to apply to some living situations. Table 1 displays how a few of these groups are counted.

EFFECT ON REDISTRICTING

The effect of counting incarcerated and institutionalized people at the location of their quarters is to (1) add population to the localities hosting the facility and (2) to deduct from the locality where they lived at the time of incarceration or institutionalization.

DEPARTMENT OF CORRECTION

The Department of Correction asks prisoners to report their address at the time of incarceration. The address includes street with number, city, state, and zip code. Based on a review of DOC data, it appears that mailing addresses are collected rather than actual residence. This means that data are collected for areas that are not towns. For example, DOC reports data for Rockville (a part of Vernon), Danielson (a part of Putnam) and Willimantic (a part of Windham). DOC does not verify the accuracy of the reported address. DOC data show that the number of prisoners more than doubled from 1990 (7,998) to 2000 (16,765). For prisoners incarcerated on April 1 in 1990 and 2000, Table 1 displays their reported town of residence at time of admission. For the purpose of this table's categorization, data from parts of reported towns (i.e. Danielson) have been added to data from the balance of the town (i.e. Putnam).

TABLE 1: Where Prisoners Came From

BACKGROUND

Census Bureau Study

Congress directed the Census Bureau to study counting inmates at their “permanent home of record” rather than at their usual residence (P.L. 109-108). The bureau reported that the change would (1) increase costs to the bureau and agencies operating correctional facilities, (2) decrease accuracy, and (3) hurt efforts to count those in other types of group quarters ((“U.S. Census Bureau Report: Tabulating Prisoners at Their “Permanent Home of Record” Address,” February 21, 2006).

The bureau's report states that if it were required to tabulate prisoners at their “permanent home of record,” it would have to (1) interview each prisoner and (2) verify the existence of a residence at the address and the validity of counting the prisoner at that address. Using Bureau of Justice Statistics, it estimated that it would cost $250 million to collect the address information of the 2.6 million incarcerated adults and juveniles. Even if funded, the bureau would face difficulties in, among other things, (1) developing safe and confidential procedures for interviews, (2) determining a valid address when a provided address is unusable for census purposes, and (3) determining whether the change should affect other groups living in group quarters.

Panel on Residence Rules

Following the most recent census, the Census Bureau convened three panels to examine different aspects of census taking. It's Panel on Residence Rules was convened by the Committee on National Statistics of the National Research Council in 2004. It submitted its report in 2007 (Once, Only Once, and in the Right Place: Residence Rules in the Decennial Census). A significant portion of the report was devoted to studying populations living in group quarters, including prisoners.

In 2000, the Census Bureau counted nearly 2 million people in correctional facilities. This represents about one-quarter of the population in the criminal justice system. The remainder is under some type of community supervision. The panel noted that “prisoners are arguably tied to more than one residence location and yet defining them as resident of any location is troublesome” (p. 83). They do not live in the community from which they entered prison, but their possible return may create a need for services in those communities. The panel reported that the procedure for enumerating prisoners in prisons is long standing and reviewed procedures for doing so dating back to 1850. Recently, the number of prisoners has grown rapidly and so the issue of how to count them is becoming more significant. The issue is controversial, the panel stated, because prisoners are 90% male and include a disproportionate number of racial minorities. Secondly, is controversial because it involves “tension between urban and rural areas for equitable shares of representation and resources” (p. 86) and that “the siting of prisons in rural areas can draw representation away from urban centers” (p. 89).

The panel reviewed problems encountered while enumerating group quarters populations in the 2000 census. These included failing to reconcile rosters of group quarters residents with the bureau's Master Address File, locating group quarters such as a college dormitory at an incorrect address (including an address in a different town), obtaining high levels of enumerations from facility records, and ineffective processing of group quarters questionnaires. The panel concluded that group quarters enumeration is unacceptably bad (Finding 7.1). It also found that the concept of “group quarters” no longer makes sense (Finding 7.2). The panel recommended “a system in which all group quarters and non-household residents are approached and enumerated in the same manner as the general household population” (p. 238). To make this possible, the panel recommended that the bureau produce alternative census forms to collect a common core of information for

different types of group quarters (Recommendation 7.1, p. 241). It also recommended that the bureau establish a program to evaluate the feasibility and cost of assigning incarcerated and institutionalized individuals to another location (Recommendation 7.2, p. 243).

DD:ts