Mounting social science research confirms what most Americans already know: The breakdown of the family contributes significantly to many of society's ills, including poverty, crime, drug addiction, school drop-out rates, and poor health. From this same research, Americans also are learning that when fathers are absent from their families, the rate of juvenile delinquency rises dramatically.
Americans are less likely to know that the federal and state juvenile corrections systems generally act as if this significant relationship between family breakdown and juvenile crime did not exist--as if, in other words, the absence of fathers and married parents in a juvenile delinquent's life had no effect on the incidence of juvenile crime.
Startling data available from Wisconsin--the one state that has identified some of the family background of its delinquents--indicate that the probability of incarceration for juveniles in families headed by never-married single mothers might be at least as much as 22 times higher than for juveniles in the two-parent family. Yet despite such evidence, the juvenile justice system and the U.S. Department of Justice's Office of Juvenile Justice and Delinquency Prevention (OJJDP) seem intent on ignoring the best prevention strategy in society today: a family environment in which the child's mother and father remain married.
Because tracking family background never enters the juvenile justice policy debate or underlying research, promoting strong families is never part of juvenile delinquency prevention programs. Now, however, with the OJJDP constructing a new Census of Juveniles in Residential Base, Congress has an opportunity to act quickly in the appropriations process to demand that accurate and detailed data be gathered.
In the last decade of the 20th century, Wisconsin was the only state to have conducted a study on the relation between family structure and juvenile incarceration.1 Its results should serve as an eye-opener to the juvenile justice system.
When the results of the 1994 Wisconsin study, based on data collected in 1993, are combined with census data for Wisconsin from the Current Population Survey for 1993, it is possible to go beyond the Wisconsin study to see the dramatically different probabilities of incarceration for each type of family, or household, structure. The rate of incarceration of juveniles is lowest among two-parent families, five times higher among families with married-but-separated parents, 12 times higher among divorced-single-parent families, and 22 times higher among families with always-single parents.2 (See Chart 1.)3
Dramatic though these numbers are for Wisconsin, they most likely understate the powerful effects that life in an intact family with married parents has in preventing juvenile delinquency. Though the incarceration rate for juveniles in always-single-parent families is 22 times higher than that of juveniles in two-parent families, based on the Wisconsin data, the difference in incarceration rates between juveniles from "intact married" families and "always-single-parent" families may be even greater--perhaps as much as 44 times higher. This is because the Wisconsin study includes "intact married families" in the category of "two-parent families" along with "stepfamilies" and "cohabiting two-biological-parents."
Previous research has disclosed enormous differences among these types of families in rates of child abuse, a significant precursor to delinquency and crime.4 Therefore, it is likely that an equally large difference exists in the rates of juvenile crime for these three types of two-parent families.
Research on child abuse,5 for example, shows that the rate of abuse in British stepfamilies is at least six times higher than the rate in intact-always-married families,6 while it is 20 times higher in cohabiting-biological-parent families than intact married families (see Chart 2). Chart 3 shows the rates of fatal child abuse in these different family structures. This finding indicates at least the possibility that the probable rates of incarceration shown in Chart 1 are not an exaggeration. The probable difference in the rates of incarceration of juveniles from the non-intact family structures is likely to be even greater.
One would think that everyone would have an interest in collecting complete data on the family background of juvenile delinquents. If the strong relationship between household type and juvenile incarceration shown in the Wisconsin data holds true in other states as well, a fundamental rethinking of public policy is appropriate.
Yet especially since 1993, when Congress required that the states address the issue of "Disproportionate Minority Confinement" for juveniles,7 bureaucrats at the OJJDP and in most states other than Wisconsin have shown no interest in doing so. Their desire to find out whether there is a strong causal role between juvenile delinquency and the absence of marriage and fathers seems to be virtually nonexistent.
For instance, in research reports like the OJJDP's "Disproportionate Minority Confinement: 1997 Update," the Department of Justice regularly details how the OJJDP ensures state compliance with the 1993 congressional mandate.8 But none of its reports investigate the detailed family background of juveniles who commit crime.
The 1993 mandate was Congress's answer to complaints that minorities, especially African-Americans, were being locked up at disproportionate rates, compared with the population as a whole. The response of the Department of Justice, and subsequently of the states, has been to assume that race and racism, not family structure, are the principal driving forces behind these disproportionate rates of incarceration.
Consider the example of Maryland in 1994. The Maryland Juvenile Crime Commission procured a study of juvenile crime, which concluded that racism resulted in the incarceration of a disproportionate percentage of African-American teenage juvenile delinquents from the Baltimore area. African-Americans, however, were serving as Baltimore's mayor and prosecutor, and another was a prominent official in Maryland's Department of Juvenile Justice.
Moreover, there was substantial African-American representation in the police force as well as the judiciary,9 making the allegation of racism seem even more unlikely. And the director of the study, Lakshmi Iyengar, when questioned by the chairman of the Juvenile Crime Commission, Lieutenant Governor Kathleen Townsend, had to acknowledge that her research had not controlled for family structure factors, such as single parenthood.10
Leaving marriage and family structure data out of the discussion in establishing juvenile crime policies is completely at odds with social science literature. Criminology studies have long asserted a strong link between juvenile crime and family breakdown, as Francis Fukuyama noted recently in The Great Disruption11 and as this author has reviewed in "The Real Root Causes of Violent Crime: The Breakdown of Marriage, Family, and Community."12 Yet none of the standard federal crime surveys looks at this fundamental connection, even though they scrutinize almost every other aspect of crime.13
Although there is a large disparity between the crime rates of whites and those of African-Americans,14 any perception that this is based solely on racism quickly evaporates when family structure is considered in the data. When researchers control for marriage, the rates of crime for whites and blacks in each family group are very similar.
Douglas Smith and G. Roger Jarjoura, for instance, found in a major 1988 study that "the percentage of single-parent households with children between the ages of 12 and 20 is significantly associated with rates of violent crime and burglary." They went on to dismiss as false the popular assumption that there is an association between race and crime. Instead, they found that the absence of marriage, combined with the failure to form and maintain intact families, explains the incidence of high crime in both white and black neighborhoods.15
Some social scientists have said that once socioeconomic variables (poverty or family income) are considered in the analysis, the relationship between family type and delinquency disappears, but this is not the case. According to University of Illinois sociologist Robert J. Sampson, writing in an earlier study on the differential effects of poverty and family disruption on crime in the prestigious American Journal of Sociology,
Overall the analysis shows that rates of black violent offending, especially by juveniles, are strongly influenced by variations in family structure. Independent of the major candidates supplied by prior criminological theory (e.g. income, region, size, density, age and race composition) black family disruption has the largest effects on black juvenile robbery and homicide.... The effects of family structure are strong and cannot be easily dismissed by reference to other structural and cultural features of urban environments.... The effect of family disruption on black violence is not due to the effect of black violence on family structure.16
"In fact," Sampson concludes, "the predictors of white robbery are in large part identical in sign and magnitude to those for blacks."17
The Office of Juvenile Justice and Delinquency Prevention's failure to track family data in its survey on incarcerated juveniles, Juveniles Taken Into Custody, does not appear to be corrected in the more robust Census of Juveniles in Residential Base that the OJJDP is now constructing. While data will be gathered on family and education, they will be misleading because such information will not be specific enough on family structure.
This same oversight was found to be a problem with the National Incidence Surveys of Child Abuse and Neglect administered by the U.S. Department of Health and Human Services (HHS) in 1980, 1986, and 1993. These HHS studies were considered less than successful because, among other things, the surveys on child abuse commingled data on three two-parent family structures (always-intact married family, stepfamily, and cohabiting biological parents) by collapsing the three into one category, the "two-parent family." Mixing these data hid more truth than it revealed.
For instance, comparing data from the United States with data from Great Britain yields a similar progression in the rates of abuse among different family structures (see Chart 4), but the British data are even more detailed than the data for the United States. Consequently, researchers and analysts have data from Britain on the three family structures with the "two-parent family" broken out. They are able to see the dramatically different rates of abuse among the three different family types within the mingled "two-parent" category.
Researchers cannot do the same analysis with the U.S. data because no attempt was made to break down the data in the "two-parent" and "single parent" categories to reflect everyday family situations. There is some likelihood that the true U.S. picture of child abuse by family type is akin to that of Britain (see Chart 2 and Chart 3), but the case cannot be argued persuasively in the absence of authoritative U.S. data.
If the OJJDP decides, in its new Census of Juveniles in Residential Base, to collapse the data on different family types into broad categories such as "two-parent families," the result will be yet another misleading picture of juvenile crime that will cause policymakers to overlook what may be the most critical factor in juvenile delinquency--family structure. In addition, the failure to collect accurate data could well have the devastating effect of perpetuating divisive race-based theories.
The solution is simple. Congress has the authority to set the guidelines for OJJDP data collection. During the appropriations process, the juvenile justice bill conference report (for S. 254 and H.R. 1501, the Violent and Repeat Juvenile Offender Accountability and Rehabilitation Act of 1999), or other vehicles, Congress should direct the Office of Juvenile Justice and Delinquency Prevention to gather specific data on the family background of the juveniles in the Census of Juveniles in Residential Base.
The Federal Reserve Board's Survey of Consumer Finance, which does incorporate such data, is the best model for the OJJDP to use. In the Federal Reserve survey, seven categories of family background are specified: three two-parent types ("always intact two-parent family," "step family," and "cohabiting two-parent family"), and three single-parent types (widowed, always-single, and divorced-single). The Federal Reserve survey also looks at one other important category: "neither parent raises the child." Social science literature suggests that this category is a very significant variable in juvenile crime.
However, even the Federal Reserve does not collect data for another category that is of great importance in understanding juvenile crime: "single mother cohabiting with a boyfriend." This category yields the highest rate of serious child abuse in data from Britain: 33 times higher than the "always intact two-parent family" for serious child abuse and up to 73 times higher for fatal child abuse.18 U.S. criminologists, policymakers, and justice officials should have comparable data to see how strongly this category is connected to juvenile crime in the United States.
The OJJDP census should consider at least eight family categories to yield data that will drive the juvenile justice community in the right direction in long-term prevention efforts. Those categories are "always-intact two-parent family," "stepfamily," "cohabiting two-parent family," "widowed," "always-single parent," "divorced-single parent," "mother cohabiting with a boyfriend," and "child living with neither parent." Congress should make sure that the OJJDP makes the needed adjustments in the census it is designing to collect data for these eight family types.
To be sure, many in Congress and academia will disagree--some of them vehemently--with the suggestion that comprehensive data will reflect Wisconsin's pattern. Even though the available evidence suggests that they are incorrect, no one can know for certain unless accurate data are collected. It is in everyone's best interest to ensure that the new Census of Juveniles in Residential Base is designed with the modifications proposed above.
Good research helps produce good public policy, and much research has demonstrated the powerful effect that intact married families have on children in preventing juvenile delinquency. Congress needs to find out just how powerful that impact is by making sure that the necessary data are collected in Department of Justice research.
Congress should act now to ensure that this correction is made in the new survey now under construction. Collecting the data will be the first step toward guiding the nation in rebuilding this important institution--the family--and helping to drive down the rates of juvenile delinquency.
Patrick F. Faganis the William H. G. FitzGerald Fellow in Family and Cultural Issues at The Heritage Foundation.
4. See Patrick F. Fagan, "The Child Abuse Crisis: The Disintegration of Marriage, Family, and the American Community," Heritage Foundation Backgrounder No. 1115, June 3, 1997, for an overview of relevant studies. A 1998 paper by Cynthia C. Harper of the University of Pennsylvania's Department of Sociology and Sarah McLanahan of Princeton University's Department of Sociology on "Father Absence and Youth Incarceration," presented at the 1998 annual meeting of the American Sociological Association in San Francisco, analyzed a national data set in the Department of Labor's National Longitudinal Survey of Youth (NLSY). Harper and McLanahan found that, controlling for all other factors, the single-parent family is twice as likely to experience the incarceration of a juvenile; moreover, the stepfather family does not deliver the same protective benefits that an intact married family does. This conclusion supports the argument of this paper; but because the data did not permit Harper and McLanahan to control for family structure at the level of specificity that is advocated here, the magnitude of the difference between all the family structures is not attempted. Also, that study takes a national probability sample and works forward to the rate of incarceration. This paper takes a different approach, taking a national sample of incarcerated juveniles and working backward to find family background.
6. This is a very conservative estimate when the work of Professors Martin Daly and Margo Wilson of the Department of Psychology, McMaster University, is used as a guidepost. These Canadian social psychologists are known as the premier researchers in this field.
8. See Heidi M. Hsia and Donna Hamparian, "Disproportionate Minority Confinement: 1997 Update," Office of Juvenile Justice and Delinquency Prevention, U.S. Department of Justice. See the reference section detailing other work done on disproportionate minority confinement.
9. Robert M. McCarthy, J.D., and David B. Muhlhausen, The Dissent: How the Townsend Report Fails to Address the Roots of Juvenile Crime and What to Do About It (Baltimore, Md.: Calvert Institute, 1997), p. 14.
11. Francis Fukuyama, The Great Disruption (New York: Free Press, 1999). Fukuyama's data tables are available at http://mason.gmu.edu/~ffukuyam/data%20page01.htm.
13. See the table of contents of the annual Sourcebook of Justice Statistics from the U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics. There are hundreds of items in this table, but none bears on this issue.