U.S. Department of Justice Office of Justice Prorams Bureau of Justice Statistics Profile of Inmates in the United States and in England and Wales, 1991 By James P. Lynch, Ph.D. Steven K. Smith, Ph.D. Helen A. Graziadei Tanutda Pittayathikhun October 1994, NCJ-145863 (The full text of this report with tables and figures is available through-- * the BJS Clearinghouse, 1-800-732-3277 * the BJS Internet Web page at http://www.ojp.usdoj.gov/bjs/ * the BJS gopher: gopher://www.ojp.usdoj.gov:70/11/bjs/ * the National Criminal Justice Reference Service Electronic Bulletin Board (set at 8-N-1, call 301-738-8895, select BJS). U.S. Department of Justice Bureau of Justice Statistics Jan M. Chaiken, Ph.D. Director Acknowledgments This report was prepared by James P. Lynch, Associate Professor of the American University and BJS Fellow, 1993-94; Steven K. Smith, Ph.D., BJS Statistician; Helen A. Graziadei, BJs Program Analyst; and Tanutda Pittayathikhun, BJs Statistical Assistant. Carol DeFrances provided statistical review. Tom Hester edited and produced the report, assisted by Priscilla Middleton, Marianne W. Zawitz, Jayne Robinson, and Yvonne Boston. Marilyn Marbrook supervised publication. Mike Hough and his staff at the Research and Planning Unit of the Home Office reviewed a draft and gave significant assistance. Contents Overview Sources of information Criminal justice in the United States and in England and Wales Types of correctional systems Offenses Criminal history Sentence length Comparisons between the incarcerated and general adult populations of the United States and of England and Wales Young offenders Race Marital status Employment Education Children of inmates Family of origin Methodology Appendix I: Standard error tests Appendix II: The number of inmates in the United States and in England and Wales, 1980-91 Selected sources * * * * * Profile of Inmates in the United States and in England and Wales, 1991, is structured around published information from England and Wales--primarily the reports to the Home Office based on a 1991 survey of inmates and reports prepared by statistical departments of the Home Office. The Bureau of Justice Statistics (BJS) gratefully acknowledges the generous help provided by the staff of Home Office departments and the staff of the School of Social Policy of the University of Manchester. * * * * * Overview Adult correctional facilities both in the United States and in England and Wales held about a third of the inmates for a violent offense--murder, rape, robbery, or assault. A larger percentage of the U.S. inmates, however, than of the inmates in England and Wales were in custody for drug offenses. Compared to their respective general adult populations, inmates in both countries were more likely to be male, minority group members, young, single, and less educated. (For definitions concerning discussions of issues and comparisons, see Methodology, page 17.) For the first time, U.S. data have been collated from all levels of the adult correctional system by using the 1991 national surveys of inmates in State and Federal prisons and the national 1989 survey of inmates in local jails. Taken together, these surveys represented about 1.2 million prison and jail inmates in the United States in 1991. The data for England and Wales come from the National Prison Survey 1991 (NPS) sponsored by the Home Office Research and Planning Unit. The results represented about 45,900 sentenced and unsentenced inmates in institutions throughout England and Wales, belonging to a single prison system. Because many issues included in the England and Wales survey match those in the U.S. surveys, detailed comparisons of the inmate populations are possible. The characteristics of inmates, their current offenses, sentence length, criminal history, and family background are examined. In cross-national criminal justice research, precise categories are not always available for all types of information. In some cases, as noted, reported data were reclassified to allow for greater comparability between the two criminal justice systems. Major findings include the following: * On average, U.S. inmates received longer sentences for comparable offenses than inmates in England and Wales. * Among the inmate population in each country, about 8 in 10 were sentenced prisoners; the rest included those awaiting trial and those convicted but not yet sentenced. * More than half the convicted inmates in each country had a record of adult custody before their current confinement. * Inmates in the United States were more than twice as likely as inmates in England and Wales to have spent their childhood living with only one parent. * Among convicted inmates in each country, about a third said that a family member, most frequently a sibling, had served time. * In both countries a majority of female inmates who had been living with their dependent children prior to incarceration reported that a grand-parent or family member other than their current or former spouse (or partner in England and Wales) was caring for their children. By contrast, male inmates were more likely to report that the children were being cared for by their current or former spouse/partner. * More than 90% of the inmate populations in each country were male. * Inmates in England and Wales were younger on average than those in the United States. * Compared to inmates in England and Wales (70%), the U.S. inmates (55%) were less likely to be single and more likely to be divorced. * On average, inmate populations in each country were less educated than the general adult population. * The inmate population in each country had a larger proportion of minorities than its respective general adult population. The U.S. inmate population was 45% black; the U.S. adult population, 11% black. In England and Wales, blacks comprised 11% of the inmates and 2% of all adults. * A greater proportion of inmates in the United States (66%) than in England and Wales (51%) were employed before entering custody. Information in this report comes from surveys of the incarcerated populations The 1991 Survey of Inmates in State Correctional Facilities, sponsored by the Bureau of Justice Statistics, is based on personal interviews with a nationally representative sample of 13,986 inmates in 275 prisons nationwide. The 1991 Survey of Inmates in Federal Correctional Facilities contains data from personal interviews with a nationally representative sample of 6,572 sentenced inmates in 53 Federal prisons. Data from the BJS 1989 Survey of Inmates in Local Jails, weighted to 1991 inmate totals, were combined with the State and Federal prison inmate surveys to represent the entire adult U.S. inmate population. Taken together, these surveys represent the approximately 1.2 million prison and jail inmates held in the Nation's Federal and State prisons and local jails in 1991. This total does not include the approximately 11,800 unsentenced Federal prisoners. In 1991 the Home Office Research and Planning Unit sponsored a prison inmate survey of 3,970 inmates representative of the approximately 45,900 inmates held in institutions throughout England and Wales. * * * * * Criminal justice in the United States and in England and Wales United England States and Wales -------------------------------------------- Population, 1991 General 252 million 50 million Adult (age 18 or older) 187.5 million 38.6 million Crimes recorded by the police, 1992 14.4 million 5.4 million Violent crimes 1.9 million 284,000 Homicide 23,760 689 Rape 109,062 4,142 Robbery 672,478 53,000 Burglary 2.98 million 1.35 million Adult prisoners, 1991 1.2 million 45,900 Incarceration rate (number of inmates per 100,000 adult residents) 640:100,000 119:100,000 Law enforcement, 1992 Police officers 604,000 128,000 Civilian 237,000 52,000 Number of sworn officers per 100,000 residents in 1991 240:100,000 256:100,000 Total annual expenditures of the justice system, 1990/91/92 $74.2 billion $15.5 billion (œ8.8 billion) Cost of justice system per resident in 1991 $299 $310 Note: The maps are not drawn to scale. Most numbers are rounded. For a discussion of categories and sources, see Methodology. For a presentation of trends in the number of persons incarcerated, see Appendix II (two graphics not presented in this ASCII file--see hard copy or the Acrobat Portable Document Format file on the BJS Internet Web page at http://www.ojp.usdoj.gov/bjs/). Other notes on table above: * Police in England and Wales recorded about 5.4 million crimes, the majority of which were property crimes. Crime is a notifiable offense, excluding those incidents causing less than œ20 in damages. See Information on the Criminal Justice System in England and Wales, Digest 2, Home Office Research and Statistics Department, 1993. * England and Wales reports the number of crimes recorded by the police as official police statistics. The FBI's Uniform Crime Reports compiles the number of crimes reported to the police. For U.S. crimes recorded by the police, see Crime in the United States 1992, Uniform Crime Reports, FBI, 1993. * Violent crime and sexual offenses were about 5% of the recorded offenses. Minor wounding comprised 65% of the violent offenses in England and Wales. * The number of U.S. law enforcement officers includes only full-time, general purpose personnel and does not include Federal employees. See Census of State and Local Law Enforcement Agencies, 1992, BJS Bulletin, NCJ-142972, July 1993. * For the number of police in England and Wales see The Criminal Justice System in England and Wales, Home Office Research and Statistics Department, 1993. * In 1991 the annual average exchange rate was $1.77 per pound sterling (U.S. Bureau of the Census, Statistical Abstract of the United States, 1992, table 1406). * The total expenditures for the United States include both the criminal and civil systems. The U.S. expenditures are for fiscal 1990. See Justice Expenditure and Employment, 1990, BJS Bulletin, NCJ-135777, September 1992. * The expenditures for England and Wales are presented for 1991/92. See Digest 2. * Both the United States and England and Wales conduct victimization surveys, but the results are not directly comparable. In 1991 England and Wales reported 15 million criminal victimizations or 300 per 1,000 residents of all ages. * * * * * The United States and England and Wales administer corrections in distinctive types of systems The Prison Service in England and Wales, a centrally administered system, supervises sentenced prisoners, those awaiting trial, inmates who were convicted but unsentenced, and those being held for transfer.(The 1991 average population consisted of 35,400 sentenced prisoners, 10,200 untried or unsentenced prisoners, and 3,200 noncriminal prisoners held for civil offenses including the nonpayment of fines. See Prison Statistics England and Wales 1991 [London: Government Statistical Service, Home Office], 1993, p. 36, and Information on the Criminal Justice System in England and Wales, Digest 2 [London: Research and Statistics Department, Home Office), 1993, p.87.]) The average inmate population in England and Wales decreased from 49,900 in 1988 to 45,900 in 1991. In 1991 the Prison Service in England and Wales oversaw 35 local prisons, 52 closed training prisons, 19 open training prisons, 17 remand centers, 22 Young Offender Institutions, and 12 female institutions. (Digest 2, p.87.) Remand prisoners in England and Wales "include both untried and convicted unsentenced prisoners committed to custody on criminal charges by any court so empowered. A person first enters the remand population when remanded in custody on or after his first appearance in court on a charge or summons." (Prison Statistics England and Wales 1990 [London: Government Statistical Service, Home Office], 1992, p. 5.) Responsibility for corrections in the United States is spread among levels of government and across jurisdictions or agencies within each level. State and local governments manage almost all the Nation's criminal justice operations, including corrections. Local jails hold nearly all of the prisoners awaiting trial as well as a large proportion of sentenced prisoners. The majority of sentenced prisoners, however, are in State prisons. The Federal Government has custody of persons awaiting trial or sentencing in Federal courts and less than 10% of the Nation's sentenced prison inmates. In 1991 the U.S. inmate population was about 1.22 million, including about 64,000 in Federal prisons, 728,000 in State prisons, and 426,000 held in local jails (sentenced or awaiting trial or sentencing). The 1991 figure represents an increase of more than 140,000 inmates over the 1989 combined jail and prison population of approximately 1.08 million. In 1991 the United States had about 1,200 State prison facilities, 80 Federal prison facilities, and 3,400 local jails. Percent of inmates ------------------ United England Detention status States and Wales ---------------------------------------- Total l00% l00% Convicted 84 82 Sentenced 81 79 Awaiting sentence 3 3 Unconvicted 15 17 Other* 2 1 *In the United States "Other" includes those who had not been charged. In England and Wales "Other" includes those held for civil offenses such as nonpayment of fines. Sources: Tricia Dodd and Paul Hunter, The National Prison Survey 1991 (London: Social Survey Division of the Office of Population Censuses and Surveys), 1993, and compiled U.S. surveys of State and Federal prison inmates and local jail inmates. ---------------------------------------- In both the United States and England and Wales about four-fifths of the inmates were serving a sentence Sentenced prisoners were 81% of the 1991 U.S. inmate population. Those convicted but awaiting sentence made up 3% and those unconvicted 15%. The unconvicted category includes those not arraigned, arraigned and awaiting trial, or on trial. England and Wales had a similar pattern in 1991: 79% were sentenced inmates, 3% convicted awaiting sentence, and 17% unconvicted awaiting trial. For certain offense categories, the distributions of inmates in the two countries were similar In each country about 1 in 3 inmates were in custody for a violent offense such as murder, rape, robbery, or assault (table 1). Persons charged with or convicted of murder were 10% of the U.S. incarcerated population and 8% of the population incarcerated in England and Wales. Those in custody for rape were 2% of the inmates in the United States and 5% in England and Wales; for assault, 6% in the United States and 8% in England and Wales; and for robbery, 12% in the United States and 13% in England and Wales.(The assault figure for England and Wales is "wounding" and does not include the 3% in the reported category "actual bodily harm, violent disorder, and death by reckless driving," considered as "Other violent" in table 1.) In each country 3% or less were charged with the property crimes of motor vehicle theft, arson, fraud, and stolen property. A notable difference between the countries existed for drug offenses and burglary. About 24% of U.S. inmates were in custody for drug offenses, as opposed to 8% of the inmates in England and Wales. Seventeen percent of inmates in custody in England and Wales were charged with burglary, compared to 11% in the United States. In each country a larger percentage of younger inmates than older inmates were confined for property offenses. The offense distributions in each country varied by sex and race: * In the United States female inmates (35%) were more likely than male inmates (23%) to be held for drug offenses and less likely to be incarcerated for robbery and burglary (table 2). Among both males and females, 1 in 10 were held for murder. Blacks were more likely than whites to be held for robbery (16% to 8%) and drug offenses (27% to 21%). * In England and Wales, a larger percentage of black inmates, compared to white inmates, was held for drug offenses (21% to 6%) and robbery (20% to 13%). White inmates (19%) were more likely than black inmates (6%) to be convicted of or charged with burglary. * The England and Wales statistics for black inmates distinguish among national origin groups by offense. For example, among black inmates held for drug offenses were 12% of the Caribbean Islanders, 39% of the Africans, and 17% of the other black inmates. For robbery and aggravated burglary, the percentages were 26% of the Caribbean Islanders, 11% of the Africans, and 15% of the other blacks; for burglary, 6%, 2%, and 14%. * Among sentenced prisoners in England and Wales, men were more likely than women to be serving time for a robbery or burglary. By contrast, women (28%) were 3 times as likely as men (9%) to be in prison for a drug offense. These data come from annual prison statistics that use categories slightly different from those of the inmate survey: Percent of sentenced inmates England and --------------- Wales offense Men Women --------------------------------- All offenses 100% 100% Violence against the person* 24 20 Rape 5 -- Other sexual offenses 5 2 Burglary 18 4 Robbery 14 3 Theft and handling 10 19 Fraud and forgery 3 5 Drug offenses 9 28 Other offenses 12 19 Number of inmates 28,964 967 Note: The table excludes 5,002 men and 181 women for unknown offense. Percentages do not add to 100% because of rounding. --Less than 0.5%. *Includes murder, homicide, attempted homicide, assaults, cruelty to children, and other violent offenses. Source: Prison Statistics England and Wales 1991. ------------------------------------- A majority of convicted inmates in both populations had served a prior sentence to adult custody Among convicted inmates, about 55% of the U.S. inmates and 57% of those in England and Wales had previously served time as an adult (table 3). In England and Wales younger inmates were more likely to have been in adult custody than younger U.S. inmates. About 1 in 3 convicted inmates in both countries had family members who had also served a sentence to incarceration The percentages of convicted inmates with an immediate relative who had served a jail or prison sentence varied according to sex, race, and nationality (table 4). Whereas in England and Wales about 33% of the convicted inmates of either sex had a close relative with an incarceration record, in the United States 45% of the convicted women and 36% of the convicted men reported such family members. Among the convicted prisoners in England and Wales, the white inmates (38%) were more likely than the black inmates (24%) to report a relative with a record of incarceration. In the United States the racial association was reversed: Convicted black inmates (41%) were more likely than the white inmates (33%) to report such relatives. The two countries differed somewhat according to which relatives had a record of incarceration (table 5). Among convicted inmates in England and Wales, 20% reported the relative to be a brother or sister, versus 30% of their U.S. counterparts. About the same percentage of inmates in each country had a parent who had been incarcerated. U.S. inmates, on average, received longer sentences for comparable offenses than inmates in England and Wales In 1991, 34% of sentenced U.S. inmates, but 4% of those in England and Wales, had a prison sentence to a term of over 10 years but less than life (table 6). Six percent of U.S. inmates had a sentence to life or to death, compared to 8% in England and Wales. These data represent imposed sentences and not time actually served. Both the United States and England and Wales release some inmates in less time than that imposed by a judge. Some research indicates that sentence reductions operate differently in the two systems. On average, inmates in England and Wales serve slightly less than half of their imposed sentence, and those in the United States serve slightly more than a third. For specific offenses inmates serve longer on average in the United States than in England and Wales.(David Farrington and Patrick Langan, "Changes in Punishment and Crime in England and America in the 1980s," Justice Quarterly (9:1, March 1992), pp. 6-46, and James P. Lynch, "A Cross-National Comparison of Length of Custodial Sentences for Serious Crimes," Justice Quarterly (10:4, December 1993), pp. 639-660.) This finding by other researchers suggests that differences in early release practices do not erase the differences in time served. Data on sentence length do not account for criminal history. Violent offenses: Among inmates sentenced for a violent offense, about 33% of those in England and Wales, but 10% of U.S. prisoners, were sentenced to 3 years or less (table 7). Among persons convicted of a violent offense, 71% of U.S. inmates, but 37% of British inmates, were sentenced to more than 10 years or to life. Also among violent inmates, 28% were serving a life or death sentence in the United States, compared to 35% in England and Wales with a life sentence. Robbery: In England and Wales, 24% of the inmates serving a robbery sentence had a term of 3 years or less, and 11%, more than 10 years. In the United States, 6% of such inmates were sentenced to 3 years or less, and 56% to more than 10 years. Burglary: Eighty-one percent of inmates in England and Wales incarcerated for burglary had a sentence of 3 years or less, compared to 16% of U.S. inmates. Drug offenses: Sentence lengths for drug convictions in the two countries were similar in the shorter sentence categories. Sentence length 17 months or less: 15% of U.S. inmates convicted of a drug offense and 10% of such inmates in England and Wales 18 months to 4 years: 21% of U.S. inmates and 28% in England and Wales 4 to 10 years: 36% of drug offenders in U.S. prisons and 56% in England and Wales More than 10 years: 27% of U.S. inmates sentenced for a drug offense and 6% in England and Wales. Inmates in England and Wales and in the United States differ from their general population In both countries inmates were more likely to be younger, minority members, single, and less educated than their counterparts in the general adult population. U.S. inmates were more likely to be divorced, minority members, and older, compared to inmates in England and Wales (table 8). Men formed the overwhelming majority of inmates in both systems In the United States 93% of the inmates, and in England and Wales, 97%, were men--about twice the percentage of men in the general adult population (table 9). In England and Wales male inmates were younger on average than females--41% of males were age 24 or younger, compared to 32% of females (table 10). Inmates on the whole in England and Wales were younger than those in the United States Forty percent of inmates in England and Wales were under age 25, compared to 26% in the United States. The percentage of inmates under 21 in England and Wales (18%) was twice that in the United States (9%). Sixteen percent of each inmate population were age 40 or older. The two systems differed in treatment of inmates under 21.(For a discussion of criminal justice proceedings and custodial penalties for young persons in England and Wales, see Barclay, 1993.) (See "Young Offenders" below.) * * * * * Young offenders The British Prison Service treats persons age 21 or older as adults, but inmates from age 17 to 21 serve sentences of youth custody in Young Offender Institutions, which were in the scope of the 1991 National Prison Survey. Because States vary considerably in their response to young offenders, the resulting situation in the United States is not entirely comparable with the system in England and Wales. In some U.S. jurisdictions, some persons who are between ages 16 and 18 and in custody will be in adult prisons, and others will be in juvenile facilities. The former are represented in the Survey of Inmates in State Correctional Facilities or the Survey of Inmates in Local Jails. Persons in juvenile facilities were not in the surveys analyzed here. In addition, some States have the youthful offender status that allows persons between ages 18 and 24 to be kept in separate facilities. These facilities, too, were excluded from the surveys of adult inmates. This variability in ages covered by the U.S. inmate surveys must be considered in international comparisons. Relative to the survey in England and Wales, the U.S. surveys underestimate the number of persons age 17 to 20 in secure custody. However, the estimated level of that undercoverage is small and will not greatly affect most comparisons of the two adult inmate populations. The confined offender population from age 17 to less than 21 in the United States is split among prisons and jails, institutions for young offenders, and juvenile institutions. The inmate surveys include those in jails and prisons. The Children in Custody statistical reporting program and the 1987 Survey of Youth in Custody describe persons in juvenile institutions. Estimates of this under-21 population are as follows: Prisons and jails 97,000 Institutions for young offenders 7,778 Juvenile institutions 13,000 The inmate surveys include about 81% of the confined young offender population that is, presumably, accurately represented in the England and Wales survey. * * * * * Compared to the respective adult populations in both countries, the inmate populations held a larger percentage of persons under age 25. In England and Wales those age 24 or under in the prison population (40%) were about 3 times their age group's representation in the general adult population (13%); in the United States the representation of younger persons in the incarcerated population was almost twice (26% to 14%) that in the general adult population. When inmates under age 21 are excluded from the comparisons of prison populations, the populations are more similar, although the England and Wales inmates overall are still somewhat younger: Percent of inmates who were age 21 or older ------------------------ Age of United England inmate States and Wales -------------------------------------- 21-24 19% 27% 25-29 26 27 30-39 37 27 40-49 13 13 50-59 3 5 60 or older 1 1 Number of inmates 1,037,000 37,600 ------------------------------------ Racial or ethnic minorities were overrepresented in the incarcerated populations in both countries In the United States, 51% of the inmates were white (including 14% white Hispanics and 37% white non-Hispanics), and 45% were black (including 2% black Hispanics) (table 11). In 1991 all minorities (blacks, Asians, Pacific Islanders, Native Americans, and Hispanics of any race) comprised 62% of the inmates,but 22% of the general adult popula- tion. (U.S. Bureau of the Census, Current Population Reports, 1980 to 1991, (Series P25-1095), table 1.) Non-Hispanic blacks were 43% of the inmates and 11% of all adults. In England and Wales the inmate population was predominantly white (82%), with blacks representing 11% and Asians 4%. Blacks and Asians were about 5% of all adults in the general population and 15% of all inmates. Specifically, blacks represented about 2% of the adult general population and 11% of the inmate population. Women and men in prison in the United States had similar racial distributions. In England and Wales males (11%) were less likely than females (21%) to be black. A larger percentage of U.S. inmates, compared to their British counterparts, had married at some time The proportion of divorced inmates in the United States was nearly double that in England and Wales (17% to 9%) (table 12). Inmates who had never married were a lower percentage of the incarcerated U.S. population (55%) than of England and Wales inmates (70%). (In the England andWales survey, "Single" includes 20% of inmates who said they were living with a boyfriend or girlfriend.) Among inmates age 40 or older, the rate of unmarried persons in the British population was nearly double that in the United States (33% to 18%). However, a slightly larger percentage of the entire U.S. inmate population was currently married (19%), compared to the inmate population in England and Wales (15%). Younger inmates in both countries were more likely than older inmates to be single. From a third to a half of inmates, depending on the country, were unemployed around the time of their arrest About 2 in 3 of the U.S. inmates reported having had a job or a business during the month before their arrest (table 13). In England and Wales about half the inmates said they did not have any paid work just before they came to prison. Across all age categories, a higher percentage of U.S. inmates than of those in England and Wales reported having work before entering prison. A higher unemployment rate for inmates in England and Wales might reflect higher overall unemployment rates. (For example, for 1991, see Barclay, 1993, p. 7, reporting a British unemployment rate of 8.3%, and U.S. Bureau of the Census, Statistical Abstract of the United States 1992, p. 381, reporting a U.S. unemployment rate of 6.6%.) Inmates who reported no paid work in England and Wales included those out of the labor force such as retirees. As noted in the original Home Office analysis of the survey, "Six percent of male prisoners and 12% of female prisoners had never been in paid employment." (Digest 2, p. 63.) In England and Wales black (59%) and Asian (61%) inmates were more likely to be working than white inmates (50%). In the United States 62% of black inmates said they had been working, compared to 70% of white inmates and 74% of Asian inmates. * * * * About the same proportion of inmates in both systems reported having prison jobs In England and Wales 59% and in the United States 57% reported having work assignments while in custody (table not shown). The distributions of work assignments by age were similar: While in custody, older inmates were more likely to be employed than younger ones. About equal percentages of U.S. male (58%) and female (55%) inmates worked, compared to 59% of the males and 72% of the females in England and Wales. Inmates of all races had about the same likelihood of prison work in the United States, but in England and Wales a greater proportion of white inmates (61%) than blacks (50%) or Asians (52%) were employed. * * * * The educational level of inmates in each country was generally lower than that of the respective general adult population Education levels are not directly comparable between the two countries, but the inmates' background can be related to their society's general educational attainment. In the United States, 47% of the inmates and 22% of the adult population said they did not graduate from high school, although they may have earned a General Equivalency Degree (GED) (table 14). Sixteen percent of inmates said they had some college, compared to 43% of the general adult population. In England and Wales, slightly more than 4 in 10 inmates and adult residents said they had left school before age 16. Nine percent of inmates had remained in school past age 16, compared to 21% of the adults in general. In both countries about a third of the inmates were living with dependent children before incarceration In both England and Wales and the United States, a higher proportion of female than male inmates had at least one dependent child living with them before incarceration. Percent of inmates living with child --------------------- Inmate United England States and Wales --------------------------------- Male 31% 32% Female 52 47 Note: The table was based on the total inmate populations. For England and Wales, information was limited to inmates who were living with their child(ren). Information on the total number of inmates with children was not available. Sources: The National Prison Survey 1991 and compiled U.S. surveys of State and Federal prison inmates and local jail inmates. --------------------------------- In England and Wales the male inmates who had been living with their dependent offspring usually (91%) reported that their child(ren) was/(were) being cared for by a spouse/partner (64%) or ex-spouse/ex-partner (27%) (table 15). Among female inmates with at least one dependent child who had previously lived with them, the offspring was much more likely to be in the care of other relatives (52%) than a current or former spouse/partner (23%). U.S. female inmates were also more likely (66%) than males (9%) to report their dependent children's being cared for by a relative other than their current or former spouse. Growing up in a one-parent family was more than twice as likely for U.S. as for British inmates Over two-thirds of inmates in England and Wales said they had spent most of their childhood living with both their parents (includes 5% who said parent and stepparent) (table 16). An additional 19% said they lived most of their childhood with one parent, and 4% identified other relatives such as grandparents. (United States: "When you were growing up, who did you live with most of the time?" England and Wales: "How did you spend most of the time as a child?" using codes from "As a child, up to the age of 16, who did you live with?" and "Up to age 16 did you spend any time in any kind of institution such as children's home, borstal, young offenders' unit (before being remanded for this offense)?") The British inmates younger than 25 were generally more likely to have lived with only one parent during childhood than were older inmates. Eight percent reported having spent most of their childhood in an institution such as a children's home, borstal (youth reformatory), or young offenders' unit. In England and Wales a larger percentage of white inmates (69%) than blacks (59%) said they had spent most of their childhood living with both parents. A large majority (86%) of Asians had lived with both parents. Among U.S. inmates, 45% said they lived with both parents most of the time while growing up. Forty-one percent spent most of their time with one parent; 1% in an institution. About half of black inmates said they grew up in a one-parent household, compared to a third of white inmates. Methodology The data on U.S. inmates were obtained from the Survey of Inmates in State Correctional Facilities 1991 (SISCF), the Survey of Inmates in Federal Correctional Facilities 1991 (SIFCF), and the Survey of Inmates in Local Jails 1989 (SILJ). The 1991 SISCF was conducted for the Bureau of Justice Statistics by the U.S. Bureau of the Census. Through personal interviews during June, July, and August 1991, data were collected on individual characteristics of prison inmates, current offenses and sentences, criminal histories, family background, characteristics of victims, gun possession and use, prior drug and alcohol use and treatment, educational programs, and other services provided while in prison, as well as other personal characteristics. Similar surveys of State prison inmates were conducted for BJS in 1974, 1979, and 1986. The 1989 SILJ was conducted for the Bureau of Justice Statistics by the U.S. Bureau of the Census. Through personal interviews conducted during July, August, and September 1989, data were collected on individual characteristics of jail inmates, current offenses and sentences, criminal histories, characteristics of victims, jail activities and programs, prior drug and alcohol use and treatment, and health care services provided while in jail. Similar surveys of jail inmates were conducted for BJS in 1972, 1978, and 1983. The data from the 1989 survey were weighted to match the 1991 population of jail inmates. The 1991 SIFCF includes information on a representative sample of sentenced inmates in Federal prisons in March 1991. Interviewers asked questions concerning the inmates' age, race, sex, education, criminal history, current offense, prior offenses, prior alcohol and drug use, and health care. This was the first survey of its kind conducted in the U.S. Bureau of Prisons (BOP) institutions. The 1991 National Prison Survey (NPS) in England and Wales was carried out by the Social Survey Division of the Office of Population Censuses and Surveys on behalf of the Home Office Research and Planning Unit. Sample design The Home Office's NPS was administered to a randomly selected sample of adult prisoners in Prison Service facilities in January and February of 1991. The sample of all prisoners age 17 or older in all of the 125 England and Wales Prison Service establishments included about 4,432 eligible respondents. These were about 10% of the male prisoners and 20% of the female prisoners, because women were intentionally oversampled. The response rate for the survey was 90%. Of inmates not responding, 50% refused to complete the interview, 10% were persons whom the interviewer was advised to avoid, and 40% had left the prison system before being interviewed. The interview lasted about 40 minutes. The data obtained were weighted to reflect the distributions in the population of about 45,900 inmates. (For a more complete discussion of the sample design and the interview methodology, see The National Prison Survey 1991, pp. 1-5. Appendix II. ) The SISCF used a two-stage, stratified random sample of inmates in State correctional facilities. Two sampling frames were employed--one each for women and men. Within each frame correctional facilities were stratified by size and then chosen within each strata. In the second stage inmates were systematically selected within each of the institutions. This sample design yielded a sample of about 15,000 inmates in 275 institutions. Interviews were conducted with 13,986 inmates. The nonresponse was 6.3% of State inmates. The sample of the Federal inmate population was selected from a universe of 81 male facilities with a total of 48,166 sentenced inmates and 14 female facilities with 4,015 sentenced inmates. The institutions were stratified by size and by whether they held men or women. The largest institutions were selected with certainty. The smaller institutions were grouped in size strata, and one institution was randomly selected from each stratum. Within the sample institutions inmates were selected systematically, starting with a random seed. Of the 5,520 male inmates selected 4,991 were interviewed. Of the 1,701 females selected, 1,581 participated in the survey. The 1989 Survey of Inmates of Local Jails was based on a two-stage, stratified random sample of inmates in 3,312 jails containing 301,470 males and 27,053 females. In the first stage of selection, six strata were formed based on the number of inmates in the jail separately for males and females. All of the facilities in the largest size strata for men and women inmates were selected for the sample. Different sampling ratios were used to select institutions in the other four strata. In the second stage of sampling, systemic samples of male and female inmates were chosen. These procedures resulted in a sample of 6,146 eligible inmates of whom 5,675 were interviewed. Of those not inteviewed, 471 refused, were in court, sick, or released before they could be interviewed. The resulting data were weighted to reflect the population of inmates in local jails on June 30, 1989. The weighting procedure included an adjustment for nonresponse in all three surveys. The weighted estimates from the SISCF 1991, the SIFCF 1991, and the SILJ 1989 were combined to provide an estimate for the entire inmate population of the United States. State prison inmates comprised 59.5% of the total 1991 custody population, local jail inmates, 36.5%; and Federal prisoners, 4.5%. The 1989 SILJ estimates were adjusted to reflect change in the jail inmate population between 1989 and 1991. The distributions obtained from the SILJ 1989 were multiplied by the population of jail inmates reported in the Census of Jails 1991. Comparability of survey methods In general, the two countries used similar survey methods. Both surveys employed scientific sampling methods and in-person interviews in a private setting. Both relied on experienced interviewers under the auspices of the government--the Office of Population Censuses and Surveys in England and the U.S. Census Bureau. The interview content differed considerably between the two. The U.S. survey included much more information on the criminal activity of inmates, while the British survey emphasized conditions of confinement. The greatest potential source of noncomparability in these surveys lies in the question wording and ordering. The text of the questions used in the respective surveys is provided where necessary. Accuracy of the estimates The accuracy of the estimates presented in this report depends on two types of error: sampling and non-sampling. Sampling error is variation that may occur by chance because a sample rather than a complete enumeration of the population was used. Nonsampling error can be attributed to many factors, such as selective nonresponse, differences in question wording or the interpretation of questions by respondents, the behavior of the interviewer, recall difficulties, and data processing errors. It is difficult to determine the full extent of nonsampling errors in any data collection. Sampling error, as measured by an estimated standard error, varies by the size of the estimate and the size of the base population. Estimates of standard errors have been calculated for the surveys of inmates in the United States and in England and Wales. These standard errors may be used to construct confidence intervals around percentages from each survey presented in this report. For example, the 95-percent confidence interval around the percentage of inmates who have been incarcerated previously as an adult is approximately 54% plus or minus 1.96 times .7% (or 52.9% to 55.1%). These standard errors may also be used to test the statistical significance of the differences between two sample statistics by pooling the standard error of the two sample estimates. For example, the standard error of the difference in the percentage of inmates who have been incarcerated previously in the United States, compared to England and Wales, is 0.16% (or the square root of the weighted sum of the standard errors in each nation). The confidence interval around the difference is 1.96 times 0.16% (or 0.32%). Since the observed difference of 3.0% ( 57.0% in England minus 54.0% in the United States) is greater than 0.32%, this difference would be considered statistically significant. Three types of comparisons are made in this report, and each requires a test for statistical significance. The first type involves differences between subgroups within a nation. For example, differences between the proportion of male as opposed to female inmates who have been incarcerated previously in the United States must be tested for statistical significance. If these differences do not exceed 1.96 times the standard error of the difference, then the difference may be because of sampling error. The second type of comparison involves differences between the inmate populations in the United States and in England and Wales. For example, the difference between the percentage of inmates having been previously incarcerated in England and Wales and in the United States must be tested for statistical significance. Finally, comparisons were made between the percentage of the general adult population in each nation with a certain characteristic. For example, the proportion of black residents in England and Wales would be compared to the proportion of black U.S. residents in the general adult population, and tests of statistical significance would be made. Each of these tests of statistical significance requires the pooling of standard errors as described in the preceding paragraph. The specific procedures used in this pooling are described in Appendix I. Unless otherwise noted, all comparisons discussed in this report were significant at the 95% confidence level. Comparability issues Young offenders. The similarity of the component of the confined young offender population included in the survey and that excluded can be derived by comparing the Survey of Youth in Custody with the combined inmate surveys. Distribution from ----------------- Inmate Attribute surveys SYIC ------------------------------- Race White 44.5% 54.0% Black 52.0 40.4 Other 3.4 5.3 Sex Male 93.0% 93.0% Female 7.0 7.0 Charge Violence 32.0% 35.5% Drugs 21.1 11.3 The largest differences identified here are about 10 to 12 percentage points. Assuming that the SYIC data accurately describe that component of the youth population not in the inmate surveys, then including that component would change the race and charge distributions by 2 percentage points. Educational level. The education systems differ in the two countries. The U.S. system traditionally measures educational achievement by whether a person completed high school, attended at least some college, and did some postgraduate studies. In contrast, the British system traditionally measures educational achievement on the basis of "age when left school." The categories usually include before age 16, age 16, age 17, age 18, and age 19 or over. Children are legally required to attend school between the ages of 5 and 16 ("Education in Britain," Central Office of Information reference pamphlet, 15/RP/84, p. 142). Race and ethnicity. The racial and ethnic minorities in England and Wales generally consist of blacks and Asians. Blacks are primarily from the Caribbean islands, Guyana, and Africa. Asians consist primarily of persons from the Indian subcontinent. Among the black inmate population, 57% are considered black Caribbean, 29% black African, and 14% other. The Asian inmate population that contributes 4% of the total inmate population consists of 75% from the Indian subcontinent (India, Pakistan, and Bangladesh) and 25% other (Chinese). Appendix I Standard error calculation and tests of significance This report employed six different samples--four surveys of the inmate populations and one survey of the general population in each nation. As noted above, the SIFCF, SISCF, and SILJ were used to characterize the incarcerated population in the United States. The NPS was used to describe the inmate population in England and Wales. The general population in the United States was characterized by the Current Population Survey (CPS) and the general population in England and Wales by the General Household Survey. Standard errors were calculated for statistics derived from each of these samples. In the case of the United States, the standard errors from the three inmate surveys were combined to form a standard error for the entire incarcerated population. Tests of statistical significance for differences in proportions within and across samples were also made. These signifi- cance tests require that the standard errors for the individual samples be pooled. The specific type of pooling involved depends on the comparison made. Comparisons across subgroups within a particular sample, for example, require the pooling of stand- ard errors for the subgroups. Comparisons of proportions across nations means that the standard errors for the samples in each country must be combined to form a standard error for the difference in the proportions. The specific procedures used to calculate standard errors and to pool them in computation tests of statistical significance are described below. Computing standard errors for point estimates The standard errors for the percentages in the U.S. incarcerated population were computed by calculating the standard error for the SISCF, SIFCF, and the SILJ and taking a weighted average of the three estimates. Standard errors of proportions in the three inmate surveys are calculated with a general variance estimation formula. For the SISCF that formula is-- S(p) = Sqrt(bp(1-p)/x) x = total number in base of % p = % a = design effect parameter 1 b = design effect parameter 2 Table A1 presents values for the design effect parameters for the entire sample and relevant subgroups. Table A1. Generalized variance parameters for the 1991 SISCF a b Total Total -.000175 124.8 White -.000177 126.1 Black -.000184 130.7 Hispanic -.000126 90.0 Males Total -.000196 132.0 White -.000197 132.7 Black -.000205 137.7 Hispanic -.000140 93.9 Women Total -.000513 19.9 White -.000500 19.4 Black -.000454 17.6 Hispanic -.000456 17.7 The general variance estimation formulas for the SIFCF and the SILJ are the same as that employed in the SISCF. The design effect parameters for the Federal and local surveys differ, however, from those of the State survey. The parameters for the SIFCF and the SILJ are presented in tables A2 and A3. Table A2. Generalized variance parameters for the 1991 SIFCF a b Total -.000229 12.4 Males -.000268 13.3 Females -.00568 2.4 The general variance estimation formula for the Survey of Inmates in Local Jails is the same as that used in the Survey of Inmates in State Correctional Facilities. The design effect parameters for the former survey are different from those of the SISCF. These parameters are presented in table A2. Table A3. Generalized variance parameters for the 1989 SILJ a b Total -.000198 101 Males -.000233 105 Females -.000715 29 The estimates of standard errors obtained from these general variance formulas were combined by weighting each survey estimate by the proportion of the total incarcerated population covered by the survey and adding the two estimates of the standard errors. The SISCF employed a simple random sample, and no general variance estimation formula was developed for the survey. Hence, the following formula was used to compute standard errors. S(p) = Sqrt ((p*q)/n) p = proportion q = 1 minus the proportion n = the base of the proportion The confidence intervals presented in table 10 were computed by multiplying the estimates of standard errors in each nation by 1.96. Standard errors for the Current Population Survey in the United States were computed using the same general variance estimation formula that was used in the inmate surveys. The design effect parameters to be used in this formula are presented in table A3. Table A4. Generalized variance parameters for the Current Population Survey Annual March Supplement, 1991 a b Total -.000016 2,312 White -.000016 2,312 Black -.000186 2,600 Hispanic -.000153 2,600 Significance tests Since most of the comparisons made in this report are between proportions or percentages across samples, the T statistic was used to test statistical significance. The T statistic for the difference in proportions from two independent samples is computed as follows: P2-P1/Qp2-p1 P1 = proportion from first sample P2 = proportion from second sample Qp2-p1 = standard error of the difference p2-p1 = Sqrt((Q1/N1)+(Q2/N2)) For example, the proportion of inmates with a prior custodial sentence is 54% in the United States and 57% in England and Wales. The standard error for that proportion is 0.8% in the United States and 0.75% in England and Wales. The standard error of the difference in proportions is computed by dividing the standard error from each nation by the sample size in that country. The results are summed, and the square root of that sum is the standard error of the difference in proportions (.16%). The T statistic is obtained by subtracting the proportion from one sample from the proportion from the other and dividing by the standard error of that difference. If the value of the T statistic exceeds 1.96, then there is only a 5% chance that the difference between proportions from each sample is due to sampling error. Because the T statistic in this instance is 15.6, we can be confident that the observed difference is not due to sampling error. As noted above, T tests were computed for three different types of comparisons--comparisons across groups of inmates within nations, comparisons of inmate populations across nations, and comparisons of the general populations across nations. The calculation of the T statistic differs slightly for each of these comparisons. When subgroups are compared within a particular correctional population, the proportions will differ, as will the standard errors for each subgroup and the sample size for each subgroup. Selected sources Publications issued by the Home Office, London, England: Information on the Criminal Justice System in England and Wales, Digest 2. (Research and Statistics Department) 1993. Prison Statistics England and Wales 1990. (Government Statistical Service) 1992. Prison Statistics England and Wales 1991. (Government Statistical Service) 1993. Barclay, Gordon C. The Criminal Justice System in England and Wales. (Research and Statistics Department) 1993. Dodd, Tricia, and Paul Hunter. The National Prison Survey 1991: A Report to the Home Office of a Study of Prisoners in England and Wales. (Social Survey Division of the Office of Population Censuses and Surveys) 1993. Walmsley, Roy, Liz Howard, and Sheila White. The National Prison Survey 1991 Main Findings. (Research Studies, 128) 1992. Publications issued by the Bureau of Justice Statistics, ent of Justice: Beck, Allen. Profile of Jail Inmates, 1989. NCJ-129097, April 1991. Beck, Allen, et al. Survey of State Prison Inmates, 1991. NCJ-136949, March 1993. Jankowski, Louis. Jail Inmates 1991. NCJ-134726, June 1992. Snell, Tracy, and Danielle Morton. Prisoners in 1991. NCJ-134729, May 1992. Publication issued by the Bureau of Prisons, U.S. Department of Justice: Saylor, William G., and Gerald Gaes. Prep Study Links UNICOR Work Experience with Successful Post-Release Outcomes. January 8, 1992. Other sources: Farrington, David and Patrick Langan. "Changes in Punishment and Crime in England and America in the 1980s." Justice Quarterly. Vol. 9:1, March 1992, pp. 4-46. Lynch, James P. "A Cross-National Comparison of Length of Custodial Sentences for Serious Crimes." Justice Quarterly. Vol. 10:4, December 1993, pp. 639-660. * * * * * Reviewers Mike Hough and his staff at the Research and Planning Unit of the Home Office thoroughly reviewed the draft report and provided many helpful comments. Professor Ken Pease and Amarjit Sandhu of the School of Social Policy of the University of Manchester, in Manchester, England, reviewed the report and provided general population census data for England and Wales. * * * * * For further information-- The Bureau of Justice Statistics Clearinghouse maintains a repository of criminal justice statistical reports from a variety of nations. For information on this collection, call 1-800-732-3277. For Home Office statistical publications, contact the following: * Research and Statistics Department, Home Office, Lunar House, 40 Wellesley Road, Croyton, Surrey CRO9YD, United Kingdom. Telephone 0981-760- 2850. * HMSO Publications Centre, P.O. Box 276, London SW8 5DT, United Kingdom. Telephone 071-873-9090. END OF FILE