In short, all methods have their advantages and disadvantages.

Our research leads us to the following conclusions.

There is no one-size-fits-all solution. If all schools were the same, in demographically similar neighborhoods, with similar crime rates in the surrounding community, with similar-quality teachers and similarly committed staffs, and similar budgetary constraints, then we could feel safe advocating a common policy for all schools. But schools are self-evidently not like that. The ideal violence-prevention policy will likely be different for each school.

For most anti-violence interventions, the evidence is either sparse or mixed. Many programs have been imperfectly monitored, so little evidence exists, but even those programs that have been monitored work in some cases and not in other cases.

Even programs that "don't work" in some overall sense may work at individual schools:

Since programs work in some places and not in others, the only reasonable agenda for fighting school violence is to encourage individual schools to experiment and to find what "works" in their particular circumstances.

Many traditional anti-violence remedies, mostly those related to discipline and punishment, have been limited at public schools, either legislatively or judicially (through constitutional interpretation). These methods have been limited at public schools because, as the schools are government-run, the government must provide safeguards against the abuse of its power. This involves notice and hearing requirements and other procedural roadblocks to punishment--all necessary, given the public nature of the service, but all making it difficult for schools to effectively choose a disciplinarian approach. This may be one reason why private schools have less violence than public schools--they are, in essence, able to require certain behavioral norms, and establish certain disciplinary procedures, through contract as a condition of attendance--and it may be one reason to encourage private schools as educational providers.

Part 1

Introduction

School violence is a serious problem, especially in public schools. Improving the quality of American education is difficult without also addressing school violence, since regardless of how good the teachers or curriculum are, violence makes it difficult for students to learn.

School violence wears many faces. It includes gang activity, locker thefts, bullying and intimidation, gun use, assault--just about anything that produces a victim. Violence is perpetrated against students, teachers, and staff, and ranges from intentional vendettas to accidental killings of bystanders. Often, discussions of school violence are lumped together with discussions of school discipline generally, as both involve questions of how to maintain order in a school.

We divide school violence-prevention methods into three classes--measures related to school management (that is, related to discipline and punishment), measures related to environmental modification (for instance, video cameras, security guards, and uniforms), and educational and curriculum-based measures (for instance, conflict-resolution and gang-prevention programs). All methods have their advantages and disadvantages.

Our research leads us to the following conclusions.

There is no one-size-fits-all solution. As William Modzeleski of the U.S. Department of Education put it, "There is no one program, no silver bullet, so that you can get one program up and say, `Here it is--if you put this program in your school, you are going to resolve violence.'" If all schools were the same, in demographically similar neighborhoods, with similar crime rates in the surrounding community, with similar-quality teachers and similarly committed staffs, and similar budgetary constraints, then we could feel safe advocating a common policy for all schools. But schools are self-evidently not like that. The ideal violence-prevention policy will likely be different for each school.

For most anti-violence interventions, evidence of effectiveness is either sparse or mixed. Many programs have been imperfectly monitored or evaluated, so few data on results exist. Those programs that have been monitored work in some cases and not in other cases.

Yet programs that "don't work" in some overall sense may work at individual schools. Every case study of an anti-violence program that works at some school should be an individual cause for rejoicing, even if we wouldn't want to mandate that same program everywhere. Since programs work in some places and not in others, the only reasonable agenda for fighting school violence is to encourage individual schools to experiment and to find what "works" in their particular circumstances.

As in any field, out of the many hot, new solutions, some are real, and some are unsubstantiated fads. Moreover, since school violence research is sparse and mixed--and since there are so many variables that it is even difficult to recognize success or failure--the most reliable way of distinguishing between the real and the faddish is to subject individual schools, in their experimentation, to the discipline of competition. Schools choose their anti-violence programs; parents choose their children's schools.

Many traditional anti-violence remedies, mostly those related to discipline and punishment, have been limited at public schools, either legislatively or judicially (through constitutional interpretation). This is not because these methods should not be used at schools at all--if parents choose their children's school, they should be able to delegate authority to schools to use discipline measures, up to and including corporal punishment. But these methods have been limited at public schools because the government must provide safeguards against the abuse of its power in circumstances where education is compulsory and attendance at specific schools is mandatory. These safeguards involve notice and hearing requirements and other procedural roadblocks to punishment--all necessary, given the mandatory and monopoly nature of the service, but all making it difficult for schools to effectively choose a disciplinarian approach. These constraints on public schools may be one reason why private schools have less violence than public schools, and it may be one reason to encourage private schools as educational providers.

This paper concludes with a discussion of what some private schools are doing, including the results of our interviews with principals of several Catholic schools. We further suggest that compulsory education laws may be contributing to violence in public schools.

Our general conclusion is to encourage innovation and experimentation in schools through decentralization and deregulation. Incentives matter, so effectively addressing school violence must include some level of parental choice, and an emphasis on private, voluntary, contractual methods rather than compulsory ones.

Part 2

Background

A. The Extent of the Problem

In 1940, public school teachers ranked the top seven disciplinary problems at public schools. They were talking out of turn, chewing gum, making noise, running in the hall, cutting in line, dress code violations, and littering. By 1990, the top seven disciplinary problems had changed somewhat. They were now drug abuse, alcohol abuse, pregnancy, suicide, rape, robbery, and assault.

As one elementary student eloquently and succinctly put it, "My perfect school would have everything except violence things." "Violence things," or, as most researchers prefer to call it, "school violence," is a broad term, which includes, but is not limited to, assault (with or without weapons), threats of force, bomb threats, sexual assault, bullying or intimidation, arson, extortion, theft, hazing, and gang activity.

The total number of crimes committed per year in or near the 85,000 U.S. public schools has been estimated at around 3 million. Many students feel unsafe in schools. A high school student explains, "I dislike having to attend a school where there is so much violence. Our school has a big gang problem. At times I don't feel I'm safe, which is my constitutional right!" The statement is oblique and not quite accurate, but it's the thought that counts. Student drawings of the perfect school often include police helicopters and security personnel. "People who fight would be locked up," one student suggests. Many teachers feel the same way. As a middle school teacher put it, "You're on constant management and police patrol. If you let up your guard for a second, you don't know what's going to happen in the room. I try to maintain high standards in my room and I will not allow anything to go on that will infringe on a child's safety, but I go home drained because you can never rest or relax. You step outside your room for the four-minute passing, you're on more patrol than you are within your four walls."

But horror stories and personal testimonials aside, the one constant in school violence literature is that it is hard to pin down the extent of the problem. Different surveys often define victimization slightly differently, refer to a different timeframe ("Have you ever been victimized?", "Have you been victimized at least once within the past year?", "within the past month?"), or interview different populations.

According to the National Crime Victimization Survey Report, conducted in 1989 and printed in 1991, about 9 percent of all students were victimized at school at least once during a six-month period (see Table 1). For all main groups, the rate of violent victimization hovers around 2 percent, while the rate of property crime hovers around 7-8 percent. These numbers seem to hold regardless of gender or race. Hispanic students were less likely to be victims of property crimes. Victimization rates are similar in junior high and high schools, though they seem to peak among 13- and 14-year-olds (eighth and ninth graders). Overall crime rates are higher among students who have moved frequently, and seem to weakly increase with increasing income (mainly because of increased property crimes). Victimization rates also seem to be largely independent of whether the student lives in a central city, suburb, or rural area.

On the other hand, according to the National Household Education Survey, 12 percent of students have been victimized in or around school, and the MetLife survey places this number at 23 percent of students (30 percent of male students, 16 percent of female students) and 11 percent of teachers.

We should neither minimize nor exaggerate school violence. Violence is not unique to schools, nor did it begin in the postwar era, despite the movie The Blackboard Jungle, which suggested that juvenile delinquency and disruption of classes was a new phenomenon. Misbehavior, violence, and disruption have existed in schools for centuries, and school officials have rarely been happy with student behavior. Youth misbehavior is discussed in clay tablets from Sumer written in 2000 B.C. Schoolchildren in 17th-century France were often armed; they dueled, brawled, mutinied, and beat teachers. Schoolmasters feared for their lives, and others were afraid to walk past schools for fear of being attacked. Student mutinies, strikes, and violence were also frequent in English public schools between 1775 and 1836; schoolmasters occasionally sought assistance by the military. In 1797, some boys at Rugby, who had been ordered to pay for damages they had done to a tradesman, responded by blowing up the door of the headmaster's office, setting fire to his books and to school desks, and withdrawing to an island in a nearby lake. British constables finally took the island through force.

American schools, historically, have also had their share of violence, sex, drugs, and gambling. In colonial times, students mutinied at over 300 district schools every year, chasing off or locking out the teacher. One observer commented in 1837 (a year when nearly 400 schools in Massachusetts were broken up as a result of disciplinary problems), "There is as little disposition on the part of the American children to obey the uncontrollable will of their masters as on the part of their fathers to submit to the mandates of kings." It is hard to trace the evolution of school violence, since reporting procedures have never been consistent. But some analysts are not sure that student misbehavior was worse in the 1970s than it was in the 1890s.

Table 1: Students Reporting at Least One Victimization at School, by Personal and Family Characteristics

Student Characteristics Total # of Students Percent of Students Reporting Victimization at School

Total Violent Property

Sex

. Male 11,166,316 9 2 7

. Female 10,387,776 9 2 8

Race

. White 17,306,626 9 2 7

. Black 3,449,488 8 2 7

. Other 797,978 10 2* 8

Hispanic Origin

. Hispanic 2,026,968 7 3 5

. Non-Hispanic 19,452,697 9 2 8

. Not Ascertained 74,428 3* -- 3*

Age

. 12 3,220,891 9 2 7

. 13 3,318,714 10 2 8

. 14 3,264,574 11 2 9

. 15 3,214,109 9 3 7

. 16 3,275,002 9 2 7

. 17 3,273,628 8 1 7

. 18 1,755,825 5 1* 4

. 19 231,348 2* -- 2*

Number of times family moved in last 5 years

. None 18,905,538 8 2 7

. Once 845,345 9 2* 7

. Twice 610,312 13 3* 11

. 3 or More 1,141,555 15 6 9

. Not Ascertained 51,343 5* 5* --

Family Income

. < $7,500 2,041,418 8 2 6

. $7,500 - $9,999 791,086 4 1* 3

. $10,000 - $14,999 1,823,150 9 3 7

. $15,000 - $24,999 3,772,445 8 1 8

. $25,000 - $29,999 1,845,313 8 2 7

. $30,000 - $49,999 5,798,448 10 2 8

. $50,000 and over 3,498,382 11 2 9

. Not Ascertained 1,983,849 7 3 5

Place of Residence

. Central City 3,816,321 10 2 8

. Suburbs 10,089,207 9 2 7

. Non-Metropolitan Area 5,648,564 8 1 7

Source: Bastian and Taylor, School Crime, p. 1

Just as the extent of the problem is hard to define, the evidence is also mixed on whether school violence has actually been increasing or decreasing. Onetime L.A. Councilman (now L.A. County Supervisor) Zev Yaroslavsky used to say that his daughter Mina, who graduated from North Hollywood High School in 1996, saw guns on campus "all the time." Others believe, however, that either Zev or Mina was exaggerating. The percentage of twelfth graders who reported that they were victimized at school during the previous year seems to have stayed more or less constant since 1980 (see Figure 1). Moreover, the Safe School Study of the late 1970s, one of the most important studies of school violence, concluded that while school violence was "considerably more serious than it was 15 years ago," it was "about the same as it was 5 years ago."

B. Congressional Initiatives

Congress has passed a number of laws designed to deal with school violence. In one year alone, 1994, no fewer than four programs--each with money attached--were enacted, including the Safe Schools Act, the Safe and Drug-Free Schools and Communities Act, the Family and Community Endeavor Schools Act, and the Community Schools Youth Services and Supervision Grant Program.

These congressional initiatives all have a laudable goal--to reduce school violence--but they should be viewed with caution.

These initiatives result from a determination by Congress that some activities are better than others. The grants are mainly targeted to those particular specified activities. The result of these grant programs is to encourage those activities, at the expense of nonapproved alternatives. This paper, however, will question the claim of any particular program to produce across-the-board reductions in violence rates. Some programs may produce marginal benefits at best; others may be downright harmful; some programs that do not work well may produce a false sense of security and may forestall the development of other, better options. The thesis of this paper is that there is no one-size-fits-all solution, and that the correct anti-violence policy is probably different for each school. Congress is in no position to determine what this program is for each school. Nor has Congress determined whether encouraging any school to adopt these particular policies would be beneficial or harmful.

It is often said that Congress' natural inclination is to "throw money at a problem." Does throwing money at a problem actually do harm? In these cases, it might. Public schools (especially in poor areas, where their clientele, generally unable to afford private school tuition, is essentially captive) have a perverse incentive to exaggerate their violence problem to get more grant money. It is difficult to determine how often this occurs, but what is clear is that when Congress provides a generous grant program, many schools find it foolish to turn away what essentially seems like free money. "Getting a federal grant has become simple," says John Devine, "just start your own conflict-resolution program."

If one's view is that there is a direct relationship between the amount of money spent and the results in terms of school violence reductions, this is all to the good. However, if the relationship is more complicated, and depends more on the actual nature of the school's problems, the attitudes of the administration, support from the community, and other factors, the amount of money is not necessarily beneficial. If schools set up programs for no other reason than for extra funding, the programs may end up being downright harmful. Many hastily instituted programs use untrained staff and give the administration a false sense of security. Some schools do best with an inexpensive program, as the experience of some public schools and many private and religious schools suggests. (One of the authors of this paper went to a private, secular school, where tuition was approximately equal to California per-pupil public school expenditure, with no security guards, no metal detectors, and not a word about violence prevention in any class or in any part of the curriculum.) Some schools that would be best served, for example, by adopting a hard-line disciplinarian approach may be tempted to forego such an approach, in favor of a more expensive, and less effective, violence-prevention curriculum.

C. A Reflection of Society?

While we accept that there are many causes of violence, and that general crime-prevention policies have their place in society (and that successful crime-prevention policies will probably also reduce school violence), we concentrate on what schools can do about the problem. We do not expect schools to reduce violence to zero, nor do we expect schools to solve all our problems, but this will not stop us from exploring the effectiveness of different school policies.

Still, it is interesting to consider broader societal causes of school violence. Just about every possible cause has been suggested by someone--society-wide violence (juvenile and otherwise) can spill into the school; poverty and discrimination can contribute to school violence to the extent that it contributes to violence in general, just as can illegitimacy and the breakdown of families. Domestic violence and child abuse have been blamed for fostering behavior problems, frustration, and retaliation. The drug culture and its violent distribution network have been blamed for encouraging students to arm themselves. Immigration, especially from countries where formal education is less valued, can contribute to the problem. So can population mobility, which creates an atmosphere of anonymity. Some have blamed violent cultural imagery, from TV shows to sympathetic news coverage of militaristic foreign policy, for numbing children to the effects of violence; materialism and advertising, for creating a culture where children are manipulated and feel exploited; and competitiveness and high parent expectations, for making children lose the identity and uniqueness of childhood before their time.

These possible explanations (presented in no particular order) run the gamut from the plausible to the ridiculous. Since school violence rates are generally lower than violence rates in society at large (see Table 2), it is not clear that the same causal factors are at work in both cases. But we will let readers decide for themselves which are which. They are outside of the scope of this paper, and we doubt that some of them significantly explain school violence. School violence is complicated and determined by many factors. This does not mean that schools should do nothing, nor does it mean that schools should do everything. Schools cannot mandate love, make poor people rich, break up gangs, or change the composition of TV programming.

Table 2: Violent Crime Rates (per 100,000 persons) in Selected Areas, 1991-1993

Area Overall Crime Rate School Crime Rate

Homicide

. Dallas 48.60 0.71

. Houston 36.50 0.48

. Chicago 29.30 0.24

. Los Angeles 29.30 0.12

. Florida 8.70 0.91

. Dade County, Fla. 17.38 1.02

. Hillsborough County, Fla. 8.35 0.00

Aggravated Assault

. Dallas 1308.00 16.00

. Houston 657.00 38.00

. Chicago 1502.00 325.00

. Los Angeles 1324.00 47.00

. Florida 350.80 21.39

. Dade County, Fla. 907.90 115.30

. Hillsborough County, Fla. 470.90 5.51

Robbery

. Florida 350.80 21.39

. Dade County, Fla. 907.90 115.30

. Hillsborough County, Fla. 470.90 5.51

Rape

. Florida 101.10 7.82

. Dade County, Fla. 85.02 7.82

. Hillsborough County, Fla. 100.60 0.00

Source: Adapted from Irwin A. Hyman, School Discipline and School Violence (Boston: Allyn and Bacon, 1997), Tables 10.2 and 10.3, pp. 312-313. Texas, Chicago, and Los Angeles overall numbers are from 1991. Texas school numbers are from 1993. Chicago and Los Angeles school numbers are from 1992. All Florida numbers are from 1993.

Increasing violence rates may or may not indicate a failing school-violence policy; even a successful policy might lead to increased violence, if it is implemented in a community where other factors would otherwise make violence rates increase even faster. Add this to the already sparse set of valid evaluations of school-violence programs, and the conclusion emerges that we should be extremely careful before deciding whether a policy does or does not work.

D. Categorizing Violence-Prevention Programs

This paper categorizes violence-prevention programs in the following way:

There is great variation in the types of programs instituted at different schools. Two-thirds of schools offer alternative schools or programs for disruptive students, and over 60 percent offer conflict-resolution and peer-mediation training, but these are only two sorts of programs out of literally hundreds. Unfortunately, evaluation of these programs has been slim. The Carnegie Council on Adolescent Development, after a survey of such programs, remarked that "it is impossible to state with conviction which types of violence prevention programs or intervention strategies reviewed are the most effective." Few violence prevention programs even collect evaluation data. In many programs, data collection is limited to measuring the attitudes of program participants, or measuring the number of services provided. Most programs, in fact, only aim at changing attitudes or social skills, though the relationship between these and actual violent behavior has not been firmly established. This has important implications for education policy. All evidence--or, rather, the lack thereof--points toward adopting a policy that does not mandate one sort of program across the board. Even where evidence exists, it is often inconclusive, and for good reason--programs will work in some places, but not in others, because schools and students are different.

We have found no evidence that any one anti-violence program works best. Instead, we have found the truism validated that a one-size-fits-all policy fits no one. The best way to reduce school violence--separating the programs that work from those that work less well, or are the results of the latest academic fads--seems to be to encourage different schools to innovate and try out different approaches, conduct proper evaluations and make the information available to parents as a marketing tool, and to subject schools to the discipline of competition to enhance both parental options and accountability.

Part 3

A Survey of Popular Methods

A. School Management

The first set of methods we address for dealing with school violence goes under the general term of "school management." These methods include everything related to discipline and punishment administered at the school site, the rules and regulations by which the school is managed, and the consequences of violating these rules.

1. The value of discipline

"Love is a boy, by poets styl'd; Then spare the rod and spoil the child," Samuel Butler wrote in 1664 in his poem Hudibras. The belief in discipline and punishment as an effective way to mold moral beings is, of course, older than the 17th century. "Foolishness is bound in the heart of a child; but the rod of correction shall drive it from him," the book of Proverbs tells us. The notion is, besides, intuitively plausible, and has produced tolerably good effects over the centuries.

While discipline and punishment have been--and continue to be--quite unpopular among academics, especially in the last 30 or so years, the practice itself is making a bit of a comeback. Educators on the front lines, parents, and politicians have observed the increase in violence at public schools since the 1960s, have observed the contemporary decrease in the belief in and use of discipline and punishment to maintain order, and have wondered whether there is not somewhat of a connection between the two.

Disorder occurs when many students do not recognize the legitimacy of school rules and violate them often, and when many students defy the authority of the enforcers of these rules, that is, teachers and staff. Disorder can take the form of students arriving late, students wandering the halls, or even graffiti and litter. All of these invite students to test the limits further; in fact, testing the rules becomes part of the fun. Students who are not stopped when they wear hats, litter, carry forbidden beepers, or write on walls, soon challenge more important rules, like "Thou shalt not assault other students." John Devine calls such a situation--where the school disciplinary structure yields whenever it is pushed--the "marshmallow effect."

Disciplinary methods, meant to enforce students' respect of school rules, take many forms. They include the drafting of behavior and discipline codes with strict penalties attached to the violation of their provisions. They include the use of suspension and expulsion as methods of keeping students in line--and, in the more violent schools, they include the use of the criminal justice system. In some schools and in some jurisdictions, the notion of discipline extends to the use of corporal punishment. In many cases, punishment consists of being sent to an "alternative education" center, such as a boot camp or a school for at-risk students. And in a few extreme cases, teachers have taken discipline into their own hands and, in classically American fashion, have successfully sued their unruly students.

These methods, aside from according with many people's conceptions of justice and fair treatment, may work sometimes. Zero-tolerance laws in Texas, according to the Texas Federation of Teachers, have been associated with a 6 percent decrease in threats of violence to students, a 33 percent decrease in threats of violence to teachers, a 10 percent decrease in assaults on students, and a 35 percent decrease in assaults on teachers.

At least six states have passed legislation to hold parents and guardians more responsible for students' behavior. In Alabama, the 1994 Safe School and Drug-free School Policy makes parents and guardians financially liable for property damage caused by their underage children. In Nevada, the Felonies Committed on School Property law "removes the limitation on the civil liability of parents for the delinquent act of a minor." In some teacher-vs.-student lawsuits, the students and their families have been fined, and the students have been expelled from school. Teachers' unions in Chicago, New York and Miami now urge teachers to sue when a student's behavior becomes intolerable. In New York City, the United Federation of Teachers reported that physical attacks on teachers and staff were down 23 percent from 1993 to 1994. The union attributed the change to the extensive support it provides teachers, including those suing students.

2. Questions of effectiveness

But regardless of any successes of disciplinary methods, important questions remain: How strict should the rules be, and how stiff the punishments? Is punishment always a last resort, or should it sometimes be the first? If it is not the first, what should be tried before? If it is the last, how does one know when one has exhausted all other possibilities? How important is consistency and predictability in punishment, as opposed to case-specific, individually determined punishments?

By all available accounts, the effectiveness of punishment and the circumstances in which it should be administered depend on the characteristics of the student, the facts of the disruption at hand, the nature of the school, and the educational philosophy of its teachers and administrators; there is no one answer that would apply across the board.

There is little evidence, for instance, that zero-tolerance laws have systematically decreased misbehavior. Moreover, in their reaching after consistency, the codes may, in some cases, sacrifice fair treatment. In Wichita, Kan., a 16-year-old junior-high-school student was expelled for having a paintball gun in his car. A seven-year-old second grader in Kingstown, R.I., was suspended for four days for showing a pocketknife at recess. An 11-year-old girl in South Carolina was suspended and arrested for taking a kitchen knife to school so she could cut her chicken. (Officials only found out about the knife because the girl asked her teacher whether she could use it.) A six-year-old in Pawtucket, R.I., was suspended for ten days for bringing a butter knife to school to cut his cookies.

Similarly, while we may approve of strict treatment and criminal punishment for youths who commit criminal acts, the juvenile-justice system is not very effective. Juvenile courts often only intervene after serious violence occurs. According to a recent study of juvenile courts, less than one-third of youths accused of violent acts stay in custody; the rest are put on probation or set free. Only 3 percent are tried in adult courts, and even those are often given light punishments, as judges, who routinely see older, more dangerous defendants, are more likely to put children on probation.

It is not only the criminal justice system that is not as effective as one might hope. Alternative schools, which are often used as an alternative to relying on law enforcement, have been criticized for being "little more than a watered-down version of the traditional school program, where students are warehoused rather than educated, [where] there is little to distinguish these alternatives from traditional schools." Community college officials criticize the college-as-alternative-setting theory on the grounds that it transfers the problem from high schools to colleges and, moreover, makes high schools look better because a student who participates in a college-based program is counted not as a "dropout," but as a "transfer." Another drawback of such programs is that even good programs have high recidivism rates--often 70 percent or higher. Even "boot camps," which concentrate on military discipline, have high repeat-arrest rates. One successful program, Associated Marine Institutes, which runs 35 programs in 8 states, many involving youths in marine environmental projects, has repeat-arrest rates under 50 percent, but this is still very high.

And many educators believe that far from being an effective deterrent to violence, many punitive methods, such as corporal punishment, can in fact be counterproductive. Critics charge that "violence breeds violence"; corporal punishment teaches children that violence is an acceptable way to compel behavior, and makes them more likely to be violent themselves. Corporal punishment is often misdirected--while most violence is in higher grades, much corporal punishment occurs at primary and intermediate levels, and is more rarely used against bigger students who might retaliate. Corporal punishment, instead of being used as a last resort, is often used as a first punishment for nonviolent and minor misbehaviors. Some studies have found that eliminating corporal punishment in a school does not increase misbehavior. Corporal punishment can also, depending on its frequency, duration, and intensity, induce post-traumatic stress disorder in its victims, and the victims themselves may show an increase in absenteeism, apathy, and vandalism. , At least one critic has brought up the possible sexual implications to the hitting of teenage girls by male principals.

3. The civil-rights revolution

Moreover, the disciplinary options available to schools have been restricted in recent decades by the civil-rights revolution. While this may be bad news from the point of view of public school administrators interested in adopting punitive measures, it is also a necessary consequence of compulsory education and mandated attendance at specific schools. When the government provides a service, it is also obligated to provide the service fairly, and assure safeguards against abuses of power. Private schools are provided voluntarily, using private money, and are chosen, and so are not subject to due process restrictions; private schools can, by and large, contract with whomever they like on whatever terms they like. But due process considerations must be considered for all government services--whether it be the disbursing of Social Security checks, the awarding of driver's licenses, or the choosing of contractors. The fact that education is compulsory and that attendance at a particular school is assigned makes the burden on the government all the greater. It is not by accident that public schools have a hard time suspending and expelling students. The alternative--government-run schools that punish left and right and expel students frivolously--would be even worse. This may also be one of many reasons why public schools generally have a worse record of violence than private schools.

Suspension, expulsion, and corporal punishment are some of the disciplinary measures subject to legal limits. Public embarrassment has been successfully challenged in court. So has grade reduction, once used routinely as retaliation for disciplinary infractions; some courts have treated grades as a constitutionally protected "property interest." Dress codes and locker searches have been challenged as well. (These are discussed later in the paper.) School officials are also potentially liable for civil damages. Administrators are now increasingly wary of disciplining students.

In short, while punishment is often necessary to maintain order, it is not always effective. And even where it might be effective, a public school may be restricted from using it, often for excellent reasons.

B. Environmental Modification

1. Security-related solutions

While some violence-prevention strategies focus on disciplinary measures to deter and punish school crime and violence, other strategies focus on changing the school environment. Some of these methods, such as metal detectors, security guards, and video cameras, try to improve behavior by enhancing security. Other methods try to indirectly influence violence rates by changing the general attitudes of the students--these range from dress codes to changing school size to sponsoring after-school activities to changing the culture so that employers demand good performance in high school.

Why metal-detectors, video cameras, and security guards may be effective is obvious enough. The following anecdotes illustrate:

If surveillance were free and totally effective, perhaps one could unqualifiedly endorse such methods across the board, though students may find such a police-like atmosphere unsettling. Certainly, one's educational philosophy, including one's view of privacy, in part determines how desirable these methods are, even when they work. Leaving such concerns aside for the moment, though, we may note that because surveillance is both expensive and not fully effective, there are enough reservations to be expressed on empirical grounds alone:

. Metal detectors can only detect metal, and not even all of that. Hand-held "wands" are more often used than walk-through detectors; while they are less expensive (on average, $115 versus $2,500), they are also less effective. Lost time is also a high cost; since it would take hours to screen every student, many schools don't check everyone. Some New York schools only screen one student in nine, though at less crowded times they have been known to scan one in three to five. Even with partial scanning, long waits and bottlenecks are common, and often detract from the educational process; students sometimes come to their first class half an hour late.

. Video cameras may reduce violence, though they are only as good as the people doing the surveillance. If--for example, because of tight budgets--no one is available to actually watch the screens, and if this becomes known, video cameras might lose their deterrent value. Unmonitored cameras are said to be one of the least-effective deterrents to robberies in banks and convenience stores, and areas with expensive and easily removed computer equipment could make the schools more attractive to burglars. Some schools do put up "placebo cameras" to create the illusion of surveillance, but even if these cameras have some deterrent effect, they could create liability issues for the school. The cameras may create the illusion of security, and a student attacked within "view" of such a camera could claim he reasonably expected security to come to his aid. Similar concerns apply to cameras that are working but are unmonitored.

Moreover, cameras cost money; the Huntsville system cost $1.7 million and required licensing by the Federal Communications Commission (because it used microwave-based cameras). Less sophisticated systems cost less, but there are also maintenance and personnel costs to consider, as well as the costs of keeping the videotapes in a secure location, possibly off-site. Black and white cameras are cheaper than color cameras, but are also less useful in identifying students. Hand-held cameras are cheaper, but require extra labor and potentially put their operators at risk.

John Devine has described many of the inadequacies of security guards in many lower-tier New York schools. Because of the school security bureaucracy, principals often find it difficult to fire bad guards. More importantly, guards cannot be everywhere, and excessive reliance on security guards may lull other participants in the school system into a sense that violence prevention is not their responsibility. Devine points out "the gradual withdrawal of teachers, over the past several decades, from the responsibility for Schoolwide discipline, when the union contract removed this function from their job descriptions or reduced it," and notes that in some ways, this withdrawal of teachers (and their replacement by guards) may have exacerbated disorder, as teachers no longer even try to prevent violence.

The same sorts of advantages and disadvantages may be listed for just about any form of security-based method--such as searching (or removing) lockers, requiring book-bags to be clear (or banning them), prohibiting overcoats and large bags, mandating picture IDs, fencing in campuses, and so on. While security measures may reduce violence, they have obvious limitations, in that no school can be made truly secure, just as society as a whole cannot be made truly secure through security-based measures alone. The absolute number, as well as the density, of students, requires a large commitment of surveillance and policing resources; moreover, large buildings have as many as 50 exits that have to be unlocked from the inside for quick escape in case of fire. Violent incidents rose 20 percent in 1992-93 in District of Columbia public schools, even though tougher security and a new closed-campus lunch policy were in place at the time. To do everything desirable in terms of increased security, says C.W. Burruss of the Dallas school district, "you're talking megabucks." Megabucks, most schools do not have.

2. Indirect behavior-based solutions

Indirect, behavioral solutions do not try to deal with the crimes themselves, but rather aim to create a "social" environment that, without reference to violence itself, will, as a pleasant side effect, produce fewer crimes. These methods are varied. Dress codes and uniform requirements try to change behavior on an individual level, on the theory that children wearing uniforms will be better behaved. On a larger level, after-school, extracurricular programs have been suggested. On a still larger level, the size of the schools themselves may affect the probability of crime.

In recent years, several hundred schools around the country, including some in such urban areas as Miami, Baltimore, Detroit, Milwaukee, Dallas and Louisville, have begun to require or encourage students to wear uniforms. School uniforms have two justifications. One is to reduce violence, by decreasing the probability that students will carry concealed weapons, fight over clothing jealousy, be victims of robbery and assaults because of their expensive jackets or shoes, or be victims of gang violence because they are wearing the colors or clothing associated with a gang. The other justification of uniforms is to modify student behavior, allegedly by fostering school pride, creating fewer clothes-related distractions, improving the atmosphere, increasing attendance, and thus encouraging a better learning environment.

Dress codes seem to have played a role in reducing violence in some districts. In Long Beach, crime decreased 36 percent in the year following implementation; fights dropped 51 percent, sex offenses 74 percent, weapons offenses 50 percent, assault and battery offenses 34 percent, and vandalism 18 percent. Similar results hold in Seattle, Norfolk, Va., and other cities where mandatory uniform policies (usually with opt-out provisions) were established. Many observers point anecdotally to the fact that parochial schools often have uniforms and also have low violence rates.

On the other hand, there has been little scientific study on the effectiveness of uniform policies; what little there has been has been at the elementary level, which is not where the violence is. A Harvard report suggests that since uniforms are generally voluntary, they may encourage discrimination against students who choose not to wear them, perhaps by students but also by teachers in their disciplinary actions. In Clarke Street Elementary School in Milwaukee, a uniform policy didn't work because there was no neighborhood vendor to sell uniforms to parents, and because many parents could not afford the uniforms. Also, since gang identity consists of more than merely colors, there is the danger that a uniform policy may create a false sense of security. Some analysts also believe that to be effective in changing student behavior, the uniform also has to be supported by the students themselves. At the Florence B. Price Elementary School on Chicago's South Side, the teachers, in solidarity with the students, wear forest green every Wednesday to show that they support the uniform policy. Such symbolic acts may not always be enough to instill respect for the uniform among the youth, who are notoriously blasé in such matters. Moreover, one educator's statement that "if everyone is dressed alike, they will feel equal" strikes an ominous chord with some, who feel that having children wear uniforms to avoid competition sidesteps the need to teach children to respect diversity among their peers.

Uniform policies have on occasion sparked legal disputes, including in Long Beach, Oakland, and Phoenix. The Supreme Court has held that schools can adopt reasonable dress codes and hair-length requirements that do not restrict political expression, but it has not ruled specifically on uniforms. The lawsuits have typically been unsuccessful for the plaintiffs or have been settled, but they can still be costly for the school. Generally, for a school uniform policy to pass muster, it should provide financial assistance for parents who might have problems affording the uniforms, and it should avoid restricting students in their political, religious, or other expression. The financial assistance may prove to be a substantial burden on some school districts, and may make a uniform policy inadvisable for many schools.

Some schools have also tried to decrease juvenile violence generally, and school violence indirectly, by offering school-based after-school activities. Some such programs merely give students a place to go, and try to promote camaraderie through clubs, sports leagues, camps, and other after-school programs. Under the Beacons Initiative in New York City, 37 schools stay open seven days a week from early morning until late evening, to provide "one stop shopping" services such as counseling, tutoring, recreational activities, vocational training, and a safe place for kids to hang out. It is unclear, however, how effective physical fitness programs are in reducing bad behavior like weapon-carrying or substance abuse. In some cases, those most interested in the activities may be those who need them the least. Some programs actually try to provide psychological services for potentially unstable students, like victims or observers of violent events. Examples include foster care programs for abused youth, respite day care for short-term reaction to problems, and crisis-management services to deal with a violent event. Such programs may help break the "cycle of violence," but these programs are rarely evaluated.

School size has also been offered as an explanation of school violence rates, with small schools being more personal, more conducive to learning, less likely to foster anonymity, and therefore more human and less violent. The empirical evidence is mixed, though. The Safe School Study of the late 1970s did indeed find that large schools have greater property loss through burglary, theft, and vandalism, and also have slightly more violence. But the authors of the study explained that larger buildings with more expensive equipment and more students provide more opportunity for loss, and per-capita property loss from large schools is not higher than in small schools. On the other hand, the proportion of students victimized is indeed higher in larger schools, perhaps because of the greater anonymity in large schools. It also found that the more students each teacher teaches, the greater the amount of school violence, perhaps because students develop fewer personal relationships with teachers. The study concluded that crowding--the size of the school population in relation to school capacity--was a greater problem, though, than size itself. Another study, from the Department of Education, reported no difference in worry about crime or in actual victimization for students at larger schools.

Even if smaller schools do help learning and reduce violence, reducing school size or class size may not be the most cost-effective method. In New York, some small, "model" schools were created by draining surrounding schools of their best students. John Devine, a small-school believer, says that creating small schools in isolation exacerbates the overall problems at the larger schools. A more thorough effort "would mean an expenditure of funds larger than any present-day politician would deem remotely reasonable or even imaginable." Similar cost problems plague the movement toward reducing class sizes in California, where the state government offers monetary incentives to schools that reduce class sizes. Many schools are now using spaces (auditoriums, libraries, cafeterias) that would otherwise go to other uses, hiring teachers that would otherwise be considered marginally qualified, and spending money that would otherwise have been spent on higher grades. Reducing class size is an idea that may work, but, like reducing school size, is an expensive idea if done for the sole purpose of reducing school crime, and is therefore not appropriate everywhere.

C. Educational and Curriculum-Based

Other strategies--apparently the most popular among academics, who dislike punishment as a way of dealing with school crime and violence, and favor addressing "root causes"--mostly involve new educational programs to improve student and teacher conflict-resolution skills, prevent or discourage gang membership, or to enhance students' self-esteem through new curricula.

1. Individual conflict resolution

We use the term "conflict resolution program" to lump together an assortment of violence-prevention programs. What makes them similar is their shared reliance on education instead of discipline as a way of preventing violence. From "Just say no to violence" (i.e., violence prevention) to "Can't we all just get along?" (i.e., conflict resolution) is, educationally, a short step.

A number of school systems have reported positive results from conflict-resolution programs. In New York, the "Resolving Conflict Creatively Program" (RCCP), established in 1985, teaches conflict resolution and peer mediation. The K-12 curriculum focuses on preventing violence, resolving conflicts, and avoiding bias. Teaching strategies include role-playing, interviewing, group discussion, and brainstorming. RCCP coaches teachers in this new style of classroom management, which involves sharing power with students and thus helping them deal with their own disputes.

In a 1988-89 evaluation of three community school districts where this program was implemented, 67 percent of teachers observed less student name-calling and fewer verbal put-downs, 89 percent of teachers believed the mediation program had helped students take more responsibility for solving their own problems, and 71 percent of teachers reported that students were less violent. The test results of fourth, fifth, and sixth grade participants showed that they learned key concepts of conflict resolution and could apply them in hypothetical situations. Over 98 percent of respondents in the five schools said that the program gave children an important tool for dealing with conflicts. The report concluded that RCCP was exemplary and that participants' assessments were extremely positive. Teachers believed that children's attitudes had changed for the better as a result of RCCP.

Unfortunately, what is remarkable about such "success stories" is how little actual success they seem to show. Despite the popularity of programs like RCCP (such approaches have been dubbed "the most effective way of intervening" in violence problems), many such programs lack proof that they significantly reduce school crime. The RCCP study, for instance, produced percentages on many variables, none of which were actual measures of school violence. A review of three popular violence prevention curriculums--Violence Prevention Curriculum for Adolescents, Washington [D.C.] Community Violence Prevention Program, and Positive Adolescents Choices Training--found no evidence of long-term changes in violent behavior or reduced risk of victimization. A main function of such programs is often to give the impression that school officials and politicians are doing something--anything--about the problem. Another study, after reviewing the existing research on violence prevention, concluded that many schools are engaged in well-intentioned efforts without any evidence that the programs will work, and--worse--that some programs actually influence relatively non-violent students to be more violence-prone. Yet another study surveyed 51 violence prevention programs around the country, including RCCP, and concluded that much more research needed to be done. Of the 51 programs, 30 percent conducted no evaluation, or had outdated or unavailable data. Another 10 percent collected no data aside from the number of people served. Another 16 percent did participant evaluations; 21 percent did outcome evaluations--but most of these evaluations were merely "before" and "after" measurements of participant attitudes and knowledge, using unvalidated measures with no control-group comparisons. "In short, there have been only a handful of programs that have been evaluated at a level approaching rigorous experimental design. None would meet the most rigorous methodologic standards of outcome evaluation." That schools adopt these programs without valid effectiveness information may merely indicate that they use these programs as a last resort. But why do they continue with the programs without any evidence that they work? If solid evaluations are useful to school administrators, it may seem surprising that these programs are so poorly evaluated. We can only conclude that schools do not evaluate these programs because they do not have to; from the administration's point of view, there is no significant loss in attendance or funding from a program that does not work well.

Moreover, even on educational-theory grounds, a conflict-resolution approach is not for everyone. Some schools have almost no violence, and manage to get by without muttering a word about violence anywhere in their curriculum, so whether such a program is even desirable in the first place depends on how much of a problem already exists, and how effective a curriculum change is expected to be in the conditions at hand. Other commentators believe that violence-prevention programs, which do nothing but talk about violence, are inherently limited in that they are often adopted as a substitute for actually stopping students in the act of violence, and moreover, inadvertently teach students that violence is a normal state of affairs to be adapted to, instead of being an aberrant situation to be reversed. At any rate, researchers seem to be unanimous that violence-prevention or conflict-resolution programs only work if properly done, and since there is currently no universal consensus on what constitutes doing violence prevention properly, we have every reason to expect empirical results of such programs to be highly mixed.

2. Peer-group programs

Paramount, Calif., was one of the first cities to include a course in gang prevention in the school curriculum. The city has a serious gang problem, with multigenerational Hispanic gangs, a gang of immigrant youths, a Crip clique, and several tagger groups. Since the early 1980s, over 9,000 students in the second, fifth, and seventh grades have taken a 15-hour course called "Alternatives to Gang Membership" (ATGM). ATGM, which has been widely replicated in Southern California, seeks to reduce gang membership and activity by teaching students the harmful consequences of a gang lifestyle, how to not participate in it, and how to choose positive alternatives. ATGM tries to reach students early; the second grade program is taught in ten weekly 40-minute lessons, the fifth grade program is taught in 15 weekly 55-minute lessons, and the seventh grade follow-up program consists of eight biweekly lessons which expand on previous topics, such as peer pressure and drug abuse. The program also focuses on self-esteem, higher education and career opportunities, and uses guest speakers. Every year, ATGM holds about 50 bilingual neighborhood gang-education and gang-prevention meetings with parents and residents, at schools, churches, parks, community centers and private residences--to educate them about gangs. Program staff also contact individual students and their families, and meet one-on-one with at-risk students referred to them by teachers.

ATGM evaluations have typically used "before" and "after" participant questionnaires. These evaluations, and staff opinions, have suggested that the program was effective. Fifth graders who had neither positive nor negative feelings toward gangs before the program tended to have a negative attitude after the program. Of fifth graders from the original 1982-83 group, 90 percent said two years later that the program had helped them avoid gangs. The same students gave the same responses two years later. In February 1993, working with the Los Angeles County Sheriff's Department, city officials matched 3,612 names of ATGM participants with a listing of identified gang members; 96 percent were not gang members.

The trouble with such "success stories," of course, is the same as the trouble with conflict-resolution programs. We must take student questionnaires about attitudes with a large shaker of salt; it is, indeed, questionable whether they mean anything at all. The statistic that 96 percent of ATGM participants were not gang members also does not tell us much unless we can know how many would have become gang members if they had not been offered ATGM. While even the ATGM figures do not clearly show that the program is a success, other studies are even more pessimistic. Patrick Tolan and Nancy Guerra, who have reviewed the literature on peer-group interventions, conclude that "there is little evidence that this type of approach is effective in reducing antisocial or violent behavior, and some programs have demonstrated negative effects." Empirical studies of peer-mediation programs are "almost nonexistent." There is a relation between gang involvement and antisocial behavior, but most studies of gang-prevention programs--which have tried to decrease gang recruitment or to channel gang members to better community activities--either have flaws in their methodology, or suggest that the programs are ineffective. In one study, 800 members of four gangs were given athletic and social events and academic tutoring. Because these activities made the gang members spend more time together, criminal behavior increased. An experiment that tried to provide services to gang members without increasing their time together reduced criminal activity, but such programs are hard to devise.

Other gang-prevention efforts are mostly harmless but are also amusingly simplistic. One report, Working Together to Erase Gangs in Our Schools, from the National Consortium on Alternatives for Youth at Risk, tells teachers how to identify gang members and gang activity at their school. Bloods call each other "Blood" and Crips call each other "Cuz"; Latino gangs call gang members "cholo" while black gangs say "let's bail" for "let's leave." Teachers are told to watch out for caps and jackets with sports logos such as that of the L.A. Raiders, colored shoelaces, sagging pants worn low around the hips, tattoos, and hand signals. All this while warning teachers to "eliminate any preconceived notions you may have about gangs." Another author suggests watching out for students with beepers, and for "informal social groups" with unusual names, like "Females Simply Chillin'" or "Kappa Phi Nasty." A naïve teacher reading this report and accurately observing the behavior of today's high-school students would be forced to conclude that everyone must belong to a gang.

Similarly disappointing results have often been found for many drug-prevention programs, for instance the Drug Abuse Resistance Education (DARE) program. Several studies have found either that DARE had no effects on students' drug or alcohol use, that the effects were short-lived, or, for some groups, that it even slightly increased their tendency to use drugs or alcohol.

Part 4

Structural Considerations

So far, we have dealt with actual school-violence prevention methods--the different means that schools use to reduce the incidence of violence. Our general conclusion has been unsurprising--different methods work in different schools; no method clearly works in all cases. There are too many variables, most of them difficult to quantify, and all of them changing over time. To a hypothetical education planner trying to predict violence rates corresponding to different anti-violence programs, we may remark, as Yoda did to Luke Skywalker in The Empire Strikes Back, "Hard to see. Always in motion is the future."

Any policy that strives to impose a particular school-violence prevention method on many different schools is unlikely to be the best solution to school violence. This realization leads us to a more basic question--what policies can we adopt that will encourage schools to adopt the most appropriate anti-violence methods for their needs?

A. How Public and Private Schools Differ

1. Incentives

We have already suggested that decentralization is desirable--that centralized regulations are likely to end up micromanaging decisions that are best left to the schools themselves, since they tend to be more aware of their own communities, problems, and constraints. But decentralization and a simple "ability to be more aware" is not enough. Al Shanker has pointed out that when New York City schools were decentralized in 1968 and decisionmaking authority brought closer to the neighborhood level, the result, in many cases, was corruption, and board members who were ignorant of many important aspects of their schools. As a result, Shanker says, New York schools have recently become partly recentralized.

What is required is that decisionmaking authority go to people who are more able to know what will work in their case, and that these people operate within an institutional structure that gives them an incentive to actually find out what will work, and act on that knowledge. This means that a mechanism must be in place through which those who run schools are rewarded for making good decisions and punished for making poor decisions.

The institutional setting of private schools provides some lessons. Private schools have a better record at keeping violence down than public schools. Private schools are usually smaller and less bureaucratic. They are often more academically challenging, so that to the extent violence is perpetrated by unmotivated students faced with undemanding course offerings, private schools offer advantages over public schools. They often offer stronger accountability to parents and students, since their survival depends on performance and meeting parental and student expectations. Moreover, the voluntary nature of attendance at these schools gives them greater latitude to set rules and "contract" with students to abide by them. Some evidence suggests that by competing with public schools, private schools also force quality (including safety) up in public schools.

Choice matters, even in non-private schools. Yvonne Chan, principal at the Vaughn Learning Center--one of the first schools in California to be awarded a charter, in 1993, under California's charter-school legislation--described the revolutionary effects of choice, and the pride in having a school that the administrators, community, and students can feel to be "their own":

Because of the racial-ethnic problem [at the school before the charter]--my predecessor was pulled out of school because of death threats--the best they could do for me . . . was to give me three security guards . . . . Then [after getting the charter], we have to get those insurance's. If you're a vendor, like Prudential or [CIGNA], God, will you sell workman's' comp and liability to a school like Vaughn in the ghetto with all the vandalism and graffiti and theft? . . . But guess what? Right now, we have no theft, no nothing. Everybody takes ownership of this school.

In its first year, discipline referrals dropped from 500 to 100 a year. Likewise, public schools that use private contractors to manage them may be better able to enhance accountability and reduce violence by making achievement of these goals a contract renewal condition. For example, public schools now managed by the Edison Project use contracts with each student and their parents to set goals and evaluate student performance.

2. Doing what the government can't

Public schools, by contrast, labor under a host of legislative and judicial restrictions on discipline and punishment. Yet many of these restrictions exist for an excellent reason--to prevent abuse of government power and discriminatory provision of mandated government benefits. In a private context, where parents' choice of school is entirely voluntary, and where parents can contract with the school for any policy imaginable (as long, of course, as it is legal), these constraints naturally (and correctly) do not apply.

This is good news for advocates of the disciplinarian model--private schools often keep violence down through strict and uniform regulations. Researchers like James Coleman find that private-school discipline, while less legalistic than in public schools, is both perceived as fairer by students and (possibly as a result) more effective. (James Coleman also reports that private-school sophomores do, on average, two more hours of homework per week than their public-school counterparts, which may contribute to keeping them out of trouble, at least out of school and perhaps in school too.) But making students better able to attend private schools would also be good news for advocates of the non-disciplinarian model, as non-disciplinarian private schools are also widespread, and parents would be able to choose whatever private school suited their vision of what their children's education should look like.

There are many other things the government cannot do. The government cannot indoctrinate children with any particular brand of religion-based morality. But the connection between violence and moral values is not accidental. Many believe that truly addressing problems of violence depends on inculcating a sense of moral values in children. And while morality is possible without religion, many people derive their morality from religion. Many parents also believe that morality aside, religious schools also provide structure to children who lack structure in their lives. Government-run schools, again, for excellent reasons, are forbidden from using religion to inculcate moral values--but many parents find morality more acceptable for their children, and many students find it more compelling as a personal guide, if it is religiously based. This is yet another reason why one might expect private schools, particularly religious schools, to do a better job at controlling violence.

For reasons related to discrimination law, the government cannot run same-sex schools; on the other hand, many private schools, including Catholic schools, have been same-sex. That boys are generally more violent than girls is well-known and not surprising (though this is becoming less true, at least in public schools). Same-sex education may not reduce violence appreciably for boys, though it may reduce relationship-related violence, and some have suggested that it may reduce violence by giving boys from single-parent homes "healthy male role models," thereby helping to break "the cycle of welfare and intergenerational illegitimacy." By removing boys, same-sex education may also reduce violence substantially for girls.

Also, for obvious reasons related to discrimination law, the government cannot run same-race schools. Some educators and parents, though, believe that all-black schools can provide significant benefits to black children, particularly by exposing black boys from fatherless families to positive black male role models they can identify with. More generally, to the extent that a black community may share certain cultural characteristics (much as ethnic communities do), such schools may succeed by being more in tune with community values and prompting greater parental involvement and student interest. Several cities, including Baltimore, Milwaukee, Minneapolis, Seattle, Cleveland, Portland, Ore., and Camden, N.H., have opened schools with Afrocentric curricula. In Detroit, Malcolm X Academy, a public school, strives to be all-black and all-male and has an Afrocentric curriculum. Students are taught Swahili and refer to male and female instructors as "Baba" and "Mama" (Swahili for "father" and "mother"). The school sports a red, black, and green "African" flag, and displays pictures of Malcolm X, Thurgood Marshall, and other prominent blacks. It features books on black history and literature, and emphasizes the contributions of blacks to math and science. (Teachers at Malcolm X also enforce a strict dress code, and also are free to spank unruly children.) Though 75 percent of its students are raised in single-parent households and more than 60 percent are poor enough to get free lunches, Malcolm X students have higher scores on standardized tests, higher GPAs, and better attendance rates than district norms--and, more interestingly for our present purposes, have low violence rates.

Since the school is government-run, it has been desegregated by court order, along with Detroit's two other black-male academies. Still, it is mostly black because of its location, and still almost all-male because the community has rallied around the school and few girls have applied. The school has made enemies of the American Civil Liberties Union and the National Organization for Women, but the principal denies that his school is segregationist. "Why shouldn't our children learn about their origins, too?" asks principal Clifford Watson. Actually, Watson's critics have a point; the school is indeed segregationist, but no more so than, say, a Jewish school, of which there are many.

Government-run schools should be restricted from endorsing this brand of racialism, just as Judaeocentric curricula are inappropriate in public schools, regardless of Jews' needs for positive role models. But if this type of school truly offers educational benefits for some, as Watson and the parents of his students believe, it should be allowed and encouraged--except, of course, without government funding.

B. Doing the Numbers

As private school enrollment began rising after more than a decade of decline--in Florida, for instance, combined private-school enrollment in Dade, Broward, and Palm Beach counties rose by 10 percent in 1994--private schools came to experience many of the same problems as public schools, including crowding, discipline, and drugs. But this increase in private-school enrollment has come about because of parents' dissatisfaction with the crowding, discipline, and drug problems at public schools. "I think a lot of people right now are afraid to send their kids to public school," said Edward Gilgenast, headmaster of the Admiral Farragut Academy in St. Petersburg. And these problems are still significantly smaller at private schools. In the words of Sister Noreen Werner, schools superintendent for the Archdiocese of Miami, "we have the same problems they do; we just have them in less numbers." Seventy percent of respondents to a national poll felt that private schools did a better job keeping out drugs and violence; 6 percent thought public schools did a better job.

Relevant statistics, culled from different studies, on the performance of public and private schools, are shown in Table 3. On the availability of drugs, the prevalence of violence and property offenses, the extent to which students avoid places at school or fear attacks, private schools are consistently shown to be safer places to be than public schools. While victimization in general is lower in private schools than in public schools, physical attacks are lowest by the largest amount. Private-school teachers are also more positive about their students than are public-school teachers; private-school students are more positive about their classmates than are public-school students; and private-school administrators are more likely to give their schools high marks than are public-school administrators. The comparisons between assigned public schools and chosen public schools also indicate that while choice--whether attendance is voluntary--and responsibility--who has ultimate control and who must bear the costs of bad behavior--are always important, ownership--whether a school is public or private--is also important.

C. Religious Schools

It is often asserted that private schools do well because they can expel whomever they like; thus, they can weed out the most difficult-to-educate students--like those with emotional or physical handicaps--foisting them on the public school system. Religious schools, though--particularly Catholic schools--have a legendary reputation for educating the difficult-to-educate. On average, Catholic schools expel less than 1 percent of their students, and suspend less than 3 percent of them. Cardinal John O'Connor of New York City, responding to a long-standing challenge by the American Federation of Teachers' Al Shanker, even offered to enroll 5 percent of the city's most difficult-to-educate students in parochial schools for Fall 1996. Mayor Rudolph Giuliani accepted the offer, originally floating the prospect of using vouchers to fund the transfers. The money must now come from private sources, because of concerns about the separation of church and state. (At any rate, public schools already do not accept everyone. Nationwide, more than 100,000 difficult-to-educate students--students with physical handicaps, learning disabilities, emotional troubles, or involvement with the juvenile-justice system--are already enrolled in private secular and religious schools at taxpayer expense.)

Table 3: Selected Statistics on Public and Private Schools

Private Public

Students who...

Say their school have too much drugs and violence 22% 48%

Say drugs are available at their school 52% 70%

Were victimized at school 7% 9%

. Violent offense 1% 2%

. Property offense 6% 8%

Avoid places at school 3% 6%

Fear an attack at school 13% 22%

Private Public (chosen) Public (assigned)

Students who...

Know of the occurrence of victimization 45% 71% 73%

Witnessed victimization 32% 54% 58%

Worried about victimization 13% 27% 26%

Were actually victimized 7% 10% 12%

. Were bullied 5% 8% 9%

. Were physically attacked 1% 4% 4%

. Were robbed < 0.5% 1% 1%

Catholic Other private Public

Students who...

. Talk back to teachers 29% 27% 51%

. Disobey instructions 20% 17% 39%

Administrators who...

. Think student absenteeism... 15.2% 13.8% 56.6%

. Think cutting classes... 4.6% 0% 37.0%

. Think verbal abuse of teachers... 4.7% 5.3% 9.6%

. Think drug and alcohol use... 26.2% 18.0% 48.5%

. Think vandalism of school property... 13.8% 11.7% 24.5%

...is a serious or moderate problem.

Total private Catholic Other religious Non- sectarian Public

Teachers who...

Think student misbehavior and substance abuse

. interferes with education 16% 12% 15% 23% 38%

Think student tardiness or cutting classes

. interferes with education 15% 10% 17% 18% 52%

Think student attitudes reduce their chances

. for success 31% 26% 28% 42% 61%

Note: Because these numbers come from different studies, they may not all agree exactly. Sources: Jean Johnson and Steve Farkas, Getting By: What American Teenagers Really Think About Their Schools (New York: Public Agenda, 1997), p. 42. Lisa D. Bastian and Bruce M. Taylor, School Crime: A National Crime Victimization Survey Report, U.S. Department of Justice, September 1991, NCJ-131645, pp. 2, 4, and 11. Mary Jo Nolin, Elizabeth Davies, and Kathryn Chandler, Student Victimization at School, National Center for Education Statistics, NCES-95-204, October 1995, pp. 7-8. Peter Benson and Marilyn Miles McMillen, Private Schools in the United States, National Center for Education Statistics, February 1991, NCES-91-054, pp. 97-99.

For students from comparable backgrounds, absenteeism, disciplinary problems, threats to teachers, and rates of violence among students are lower among Catholic-school students. Many parents choose religious schools for reasons quite unrelated to religion; "Our school is free of drugs, free of violence and free of sex," says Sulaiman Alfraih, principal of the boys' school at the Islamic Saudi Academy in Washington, D.C. "Regardless of their ideology, the parents love to see their kids in a very safe, clean environment." "Our schools have a sense of order," says Sister Catherine McNamee, president of the National Catholic Education Association. "Parents feel their children are safe, especially in urban areas, and they will develop a sense of moral values."

Many Catholic schools used to have uniforms, though many today merely have a general dress code (i.e., no baggy jeans or nose rings). Most have few guards or security gadgets, and overwhelmingly, they do not incorporate violence prevention as such into the curriculum. Naturally, they have (often voluntary) school prayer, and strict behavior codes. Legendary Catholic school discipline (i.e., being rapped on the knuckles by a menacing nun) is more lax today than it once was. The success of Catholic schools is mostly attributed to such factors as "high expectations, firm discipline, academic rigor, and a sense of community." One writer describes the typical Catholic-school approach to discipline, in the person of Brother Greg (a pseudonym):

Coming off a 20th reunion gathering, a class of '76 graduate recalled, "Brother wanted you to learn. He knew his stuff. He had a sense of humor. He respected you. But if you played the badass or mouthed off or hassled other kids, no matter how big you were, he would take off his shirt, show his martial arts thing, and, if you pressed it, kick your butt but good." In four years, how many butts did you actually see him kick? "None--but that's the point. We knew he could and would--and had! We also knew he cared, and that he didn't play favorites. White or black. Jock or not. Going to college or back to [the local bar]."

One 15-year-old, who has attended both a Catholic school and a lower-tier New York public school, puts the matter quite clearly: "It's like here [in the public schools], the teachers . . . don' say anythin' when you miss their class or mess up your homework; the nuns, they make you look stupid and feel bad kind of like my mom treats me." The Jesuits, renowned among Catholics (and in the outside world) for their quality educational system:

were not afraid to confront students who failed to uphold their responsibilities . . . . When you did something right, you got immediate positive reinforcement . . . . What Jesuit teachers feared more than any arrogant student was a fellow teacher (Jesuit or lay) who didn't know how to control a class . . . . Fear was certainly not considered the best method for motivating students in these highly competitive schools, but neither was it disdained . . . . My graduate students today stare at me in disbelief when I relate to them how, as late as the 1960s, as a young priest-housemaster living in the student dormitory of a Jesuit university, I was doing midnight bedchecks, bailing students out of the local precinct when they got in trouble with the law, communicating with their parents, and checking to see if they were going to daily Mass.

When teachers were asked to rate aspects of their school climate, Catholic school teachers gave their schools generally higher marks than public-school teachers, but the difference was greatest in teacher assessment of student behavior (see Table 4).

Table 4: Percent of Teachers Reporting Positive School Climate in Public and Catholic High Schools: 1984

Factor Public Catholic

Principal leadership 50 59

Staff cooperation 52 68

Student behavior 39 73

Teacher control 66 81

Teacher morale 74 85

Source: Peter Benson and Marilyn Miles McMillen, Private Schools in the United States: A Statistical Profile, With Comparisons to Public Schools, U.S. Department of Education, Office of Educational Research and Improvement, NCES 91-054, February 1991, p. 115, figure 5-5.

D. Catholic School Principals Speak

1. Public and Catholic schools in Los Angeles

The Los Angeles Unified School District has its own police department, which has been in existence since 1948. The LAUSD Police Department has about 280 sworn personnel--one chief, three assistant chiefs, four lieutenants, 25 sergeants, 18 detectives, six senior police officers, and about 223 police officers. The police department serves about 58,394 regular employees, 811,713 students (in school year 1995-96), and 899 schools and centers spanning an area covering 708 square miles. Table 5 shows LAUSD crime statistics for years 1990-91 through 1995-96, with offenses ranging from assault with a deadly weapon and homicide to property crimes and trespassing.

Table 5: Violence in Los Angeles Public Schools

1990-91 1991-92 92-93 93-94 94-95 95-96

Assault 104 121 132 119 99 *

Assault with a Deadly Weapon 458 483 399 308 292 226

Battery 874 776 741 629 686 n/a

Chemical Substance Offenses 248 259 384 665 959 1471

Crimes Against Property 7396 7905 7215 6676 5449 5441

Destructive Devices 178 176 108 52 111 93

Extortion 2 5 4 6 3 *

Homicide 1 1 2 0 1 3

Loitering/Trespassing 286 216 150 142 385 733

Possession of Weapons 1305 1403 1325 1032 1018 416

Robbery 475 433 451 401 461 422

Sex Offenses 404 429 409 427 477 93

District enrollment (thousands) 790 800 811 792 795 812

Note: n/a = not available. Assault reporting rules changed in 1995-96; see battery. Extortion reporting rules changed in 1995-96; see robbery. Source: Los Angeles Unified School District home page, http://www.lausd.k12.ca.us/police/crimstat/

According to the California Safe Schools Assessment 1995-96 annual report to the legislature, Los Angeles County had an enrollment of 1,511,054 (including the 811,713 in LAUSD). The financial loss to the county due to crime-related incidents (mostly property crimes) was just under $12 million. California public schools also invest a large amount of resources into violence-prevention programs: $7.2 million statewide for the School Violence Reduction Grant Program; $10 million for eight or more sites (in a three-year demonstration grant) for the Targeted Truancy and Public Safety Grant Program; $50,000 for each school that applies for the School Community Violence Prevention Grant Program; $8,000 for each applying school for the Conflict Resolution and Youth Mediation Grant Program; $5,000 for each of 100 schools (plus a district matching fund) for Safe School Plan Implementation Grants; $3 million statewide for the Gang Risk Intervention Program; and $4.03 per pupil (a federal fund entitlement) for Title IV Safe and Drug Free Schools and Communities.

Catholic school enrollment is 2.6 million nationwide. Minority students account for nearly one-fourth of the total, and a rising percentage (now 13.2 percent) of the students are not Catholic. On average, Catholic schools expel less than 1 percent of their students, and suspend less than 3 percent of them. Most Catholic principals (84 percent) say that "discipline is a strong emphasis."

Total enrollment in Los Angeles County parochial schools is 93,200. There are 207 elementary schools (K-8), four middle schools, and 45 high schools (9-12). According to Sister Mary Joanne, research analyst for the Los Angeles Archdiocese, there is no need for a formal violence tracking system in Los Angeles-area Catholic schools because the number of incidents is so small. The Catholic school principals we interviewed reported less than one incident per year that would require police involvement. "There have been no incidents on school property in the last two years, although there has been violence in the community which has affected the children," according to John Quarry, principal of San Mogul Elementary. And as Margaret Nadeau, principal of St. Malachy Elementary, said, "I have worked here for six years and we have only called the police one time when outside gangs were causing trouble on our street corner. In fact, the police do not even know where we are located." When the police are called at Los Angeles Catholic schools, it is usually because of a disturbance from outsiders who come on or near the school campus.

Sister Mary Joanne confirmed our findings from the interviews with local principals and said that the low rates of violence are generalizable to all Catholic schools in Los Angeles County.

2. Ten Los Angeles Catholic schools

We interviewed ten Catholic school principals at K-8 schools in East and South-Central Los Angeles. The student populations at these schools were 100 percent minority. At Santa Teresita Elementary School, all 274 students are Hispanic; at Santa Isabel, all 304 students are Hispanic; and at St. Lawrence of Brindisi, 60 percent of students are Black and 40 percent are Hispanic. These Catholic schools also have a high student-teacher ratio. The smallest student-teacher ratio was 28-to-1, and most schools had a ratio closer to 35-to-1. (A Wall Street Journal editorial once remarked that Mrs. Roman, an eighth-grade teacher at New York's Our Lady Queen of Angels, "manages a class of 46, a number that would send most public school teachers on strike.")

The Catholic schools we contacted do not have student mediation and conflict resolution programs, metal detectors or security guards, locker searchers or small class sizes. They manage to maintain discipline without many of the popular public school methods for preventing violence. Our interviews identified three sorts of strategies Catholic schools use to promote order, maintain discipline, and avoid violence--assertive discipline, contact with parents, and a strong sense of moral values.

3. Assertive discipline

Amity Schlaes wrote of New York's Our Lady Queen of Angels that it "enjoys another, giant advantage not shared by its public counterparts: the freedom to demand civilized behavior from its students. A blue school handbook lays out a stern line: `Self-discipline is the Christian ideal which all students are encouraged to achieve.' The `Rules of general behavior' include `polite greeting to each other' and `holding doors and stepping back to let adults pass first.'"

All of the schools we contacted have a clear and consistent discipline policy. Public schools have zero tolerance for bringing weapons to school; Catholic schools have zero tolerance for

misbehaving. All types of misconduct carry serious consequences, so student misbehavior never gets to the point where students are carrying weapons.

According to the principal of San Miguel Elementary School, "the number one component to prevent violence is a very strong, assertive discipline program. Although teachers have autonomy to find the best way to control their classrooms, they consistently enforce a set of rules that all students are made aware of." At St. Lawrence, students are given a handbook at the beginning of the year, and the teachers review the handbook with students at mid-year to remind them of appropriate behavior. At this school, the punishment associated with different types of misconduct becomes harsher as the year progresses.

At Holy Cross Middle School, Sister Daniel Therese Flynn explains that the policies in the student handbook are strictly enforced. "We do not deviate. We have complete consistency in applying our policies. If students throw punches, for example, they are both suspended. There is no determination of who is at fault. There is not one predator and one victim--we do not act as a court of law so as to divide students into groups. There is no arbitration; everyone gets the same penalties."

Ms. Collins of St. Gregory pointed out that when children learn to respond to discipline in the first grade and the child stays in the Catholic schools for eight years, a sense of self-control becomes ingrained in the child. "We teach children self-discipline," explains John Quarry of San Miguel Elementary, who has expelled only two children in 12 years.

At Santa Isabel the principal, Sister Joanne Marie, pointed out that all teachers present a unified front of consistency. "We have zero problems because we emphasize that misconduct is just not permitted. When two seventh-grade boys were caught smoking marijuana before school, we took it very seriously. We made a heavy-duty big deal. They know, the other students know, their parents know--we set an example--this behavior is not tolerated."

In Catholic schools, students know the exact consequences for their behavior. At St. Thomas the Apostle school, for example, there is a very specific process leading up to student expulsion. If students receive three pink slips, they are put on probation. If they receive three more, they are asked to withdraw from the school. School principal Dan Horn explains that "it rarely gets to this point. The kids know the policy and they have a sense of shame when they receive a pink slip because they know it is serious. The student's parents are contacted even before the first pink slip is issued. Before a student is asked to withdraw, every effort has been made to work with the student and parent and we even recommend outside counseling. The last automatic probation was for an eighth-grade boy who continued to verbally and sexually harass a female student."

4. Contact with parents

Catholic schools keep in close contact with parents through letters or phone calls. Catholic school teachers call between 28 and 35 parents on a regular basis. At Holy Cross Middle School, the teachers maintain constant contact with parents to report positive and negative student conduct. Santa Isabel has mandatory parent meetings; every Tuesday, students take home a progress report detailing the student's behavior, which the parents must sign.

And at Santa Teresita Elementary, when two seventh grade boys grabbed a note from a seventh grade girl, they received a detention slip just for "the nonsense" and because they took someone else's property. Sister Mary Virginia, the school principal, told the boys to have their parents call her at home that night. One boy did, but the other did not. Sister Mary Virginia called the second boy's home at 9:30 that night; his parents had been told nothing about the incident or his detention. The principal made sure the parents were aware of the incident.

5. Moral values

All of the school principals we talked to stressed the importance of explicit moral values in maintaining a safe and positive environment in their schools.

As St. Malachy's principal, Margaret Nadeau, explained, "Catholic schools have the moral advantage; we live by the Ten Commandments and install a strong sense of right and wrong in our children. We talk about values and teach the children to respect their teachers and each other. Our teachers demand respect. Children cannot live without a framework. We spell out our expectations and the children appreciate this--they appreciate the safe environment."

Similarly, Holy Cross Middle School's Sister Daniel Theresa says, "We make youngsters aware that they have a moral obligation to behave. Their parents are sacrificing their time and money to send them to this school." And at Santa Isabel, they emphasize "saying kind things rather than unkind." At St. Thomas the Apostle school, Dan Horn explains that the faculty has a strong philosophy of respect and dignity. "Beyond just academics, we care for the students. And both students and teachers share in that philosophy."

E. Compulsory Schooling

To be most effective, choice in education may need to go further than merely allowing parents to choose which school their child goes to. There is an interesting case to be made that compulsory schooling laws themselves exacerbate school violence problems, and that repealing or softening such laws, at least at the high-school level, would alleviate school violence, would improve the quality of education, would not flood the streets with delinquents, and would not appreciably increase crime in society at large.

The costs of compulsory schooling are twofold. First, public schools find it difficult to expel troublesome students. When a troublesome student attends class, he can make education difficult for the willing students; when he doesn't attend class, as is often the case, he blurs the line between intruders and students, making it harder to maintain order. An anonymous ninth- and tenth-grade teacher at a large high school in New Jersey puts the problem this way:

You normally can't kick [a low-performing, disruptive kid] out of the class. School administrators want to keep that kid in the classroom because they say whatever he gleans might help him next year when he takes the class again. But that isn't what happens. For short periods of time, you can remove him from the class--which takes time and energy--and place him in what's called the "restriction room," a sort of detention hall held during regular school hours. But you still have to make assignments for him and follow up on those, both of which take time away from actively working students. And when he comes back to the class, he is usually just as disruptive, and likely to drag marginal students down to his level.

Second, compulsory schooling may not even benefit the dropout. Compulsory schooling laws are often called "compulsory education" laws, but they are more accurately called "compulsory enrollment" laws. For unwilling, disaffected students, who have not chosen their school and who feel like prisoners, enrollment does not equal education. Such students are hostile, do not respect authority, and do not feel their education is worthwhile, and the higher the age of compulsory education, the more such students there are.

It is no coincidence that many academically or artistically selective schools--such as Boston Latin School, the Bronx High School of Science, Aviation High School, and the Murry Bertraum High School for Business Careers--are both safe and academically meritorious. They are entirely chosen, and have a critical mass of willing students. Thus, Aviation High School, for instance, serves both plane lovers like then-17-year-old Bridgette Miles and students like then-junior Pastora Rivas who was "looking for a place where there wasn't going to be a fight every day." According to Toby, schools can ensure such a critical mass:

by insisting that educational achievement is the primary mission of schools. Such a policy implies that the small minority of high school students who lack the slightest interest in learning anything except how to drive their teachers into another profession would have to mend their ways in order to remain enrolled. Taking high school education seriously means that it is not enough for a youngster to be on the high school rolls and show up occasionally. Dropout prevention is not an end in itself; a youngster who does not pay attention in class and do homework ought to drop out.

What would happen if schools really allowed and even encouraged potential dropouts to drop out? Since most children are ruled by their parents, most children, even unwilling ones, would still go to school. The data from different states with different ages of compulsory attendance confirms that the vast majority of students would still attend school. Table 6 compares the percentages of students in 1970 that attended school until ages 14, 15, 16, 17, and 18, in two groups of states--the five states that compelled attendance to age 15 or under at the time, and the four states that compelled attendance to age 18.

Table 6: White Males Enrolled in School by Age and Compulsory Attendance Requirement, 1970 (percentage) (Compulsory Attendance Required by State Law)

Age To Age 15 or Under (five states) To Age 18 (four states) Difference

14 94.6% 97.1% 2.5%

15 93.7% 96.5% 2.8%

16 90.2% 94.9% 4.7%

17 85.8% 90.1% 4.3%

18 70.3% 71.3% 1.0%

Note: The five states with compulsory attendance until age 15 or under are Arkansas, Louisiana, Maine, Mississippi, and Washington. The four states with compulsory attendance until age 18 are Hawaii, Ohio, Oregon, and Utah. Source: Jackson Toby, "The Schools," in Crime, ed. James Q. Wilson and Joan Petersilla (1995), chp. 7, p. 24, citing United States Bureau of the Census, Census of Population 1 (Washington, D.C.: U.S. Government Printing Office, 1973), ch. D, parts 5, 13, 20, 21, 37, 39, 46, 49.

For all years, the percentages of students that stay enrolled in school are similar. Since enrollment is an overestimate of attendance, the differences in attendance should be even smaller. And since attendance is an overestimate of learning, the true differences should be even smaller than that (and while "learning" cannot be measured directly, we cannot rule out the possibility that the lower-compulsory-attendance states might come out ahead under such a comparison).

School violence would also decrease for three reasons. First, those who don't want to be there, who disproportionately exhibit delinquent behavior, would leave. Second, since schools, freed from the requirement to take all comers regardless of behavior, would be able to maintain higher standards, incorrigible students who do not want to drop out would be expelled. Third, once schools enforce higher standards, individual students' behavior would probably improve. "Making schools tougher academically, with substantial amounts of homework, might have the paradoxical effect of persuading a higher proportion of families to encourage their children to choose of their own volition to try to learn . . . . Keeping internal dropouts in school is an empty victory." As the crude comparison in Table 7 indicates, higher ages of compulsory attendance seem to be associated with higher rates of secondary-school crime. The interesting variable in the table is the right-hand column, which calculates the difference between secondary-school crime and elementary-school crime. As the age of compulsory school attendance rises, so does this difference.

Table 7: Referral of School Crimes to the Police by Age of Compulsory School Attendance in the State, 1974-75

Age of Compulsory School Attendance In ElementarySchools In Secondary Schools Difference

[sterling] 15 (AR, LA, ME, MS, WA) 3.1 8.0 4.9

16 (36 states + DC) 3.2 10.5 7.3

17 (NV, NM, PA, TX, VA) 3.8 11.6 7.8

18 (HI, OH, OR, UT) 4.8 20.1 15.3

Note: Because there are so few states in the 18 age group, extreme values for one of them greatly influence the average. Hawaii, for example, had by far the highest rate of school crime on both the elementary and secondary levels. If Hawaii were excluded from the average and the remaining eight states with compulsory ages of school attendance of 17 or higher were averaged, the result would be 4.0 for elementary schools and 11.4 for secondary schools, with a difference of 7.4. Source: Jackson Toby, "The Schools," in Crime, ed. James Q. Wilson and Joan Petersilla (1995), ch. 7, p. 25, citing United States Department of Health, Education, and Welfare, Violent Schools -- Safe Schools: The Safe School Study Report to the Congress (Washington, D.C.: U.S. Government Printing Office, 1978), p. B-6.

Would dropouts increase the crime rate in the outside world? Probably not. Intuitively, one can observe that juvenile arrest do not increase much during summer vacation, even though students, including violent ones, are not in school. One can also observe that many students who eventually drop out already spend a lot of time outside of school, since their nonattendance rates are high. Quantitative studies support this intuition, and suggest that while dropping out may be a symptom of larger problems, it is not itself the problem, and in fact, forcing dropouts to stay in school will likely be counterproductive, both for the school and the would-be dropout. While dropouts do indeed have high delinquency rates, dropping out is not the cause of delinquency. Dropouts generally adopt antisocial behaviors while still in school, often as a result of experiences in the school itself. Once they drop out, their delinquency does not increase (see Figure 2); according to one study, dropping out actually decreases the dropouts' rate of delinquent behavior and the likelihood of official police contact.

Figure 2: Cross-Time Delinquency Score for Three

Levels of Education

High school dropouts

Stayins who were "primarily students" after high school

Stayins who were not "primarily students" after high school

Notes: Lines connect means for only those participating in all data collections. Circles indicate means based on all Time 1 respondents, regardless of further participation, who could be classified into analysis groups. Ordinates are scaled to show time 1 grand mean +/-1 standard deviation.

Source: Bachman, Green, Wirtanen, Youth in Transition, p. 124.

Thus, only a small minority of students are likely to drop out, and these are possibly the students that ought to drop out in any event; these dropouts would not appreciably increase violence in society at large. As schools became able to be more demanding, other would-be dropouts might conclude that education was valuable and worthwhile. Schools may well be safer for the other students as well, increasing the value of the education for well-behaved students, and possibly slowing down the flight of students from public schools.

Voluntary high schools may account for some of the successes of the Japanese educational system; Japanese high schools are voluntary, and can therefore be selective and demand hard work from willing students. Ninety-four percent of Japanese junior high school graduates attend high school, and 90 percent of them complete it. As high school attendance becomes more selective and voluntary, higher academic and behavioral standards seep into junior high schools, where students know that their acceptance into the high school of their choice depends on how they do in junior high. Japanese junior high schools (which are compulsory) are more violent than Japanese senior high schools, even though most junior high students are too busy preparing for high school admission exams to break the rules.

Schools might benefit not only by allowing students to drop out, but also by encouraging adult dropouts to return to school--not in special adult classes, but together with children. In Chicago, DuSable High School, which allows adults in regular classrooms, has found that "returning students," "urban Rip Van Winkles," act as "unofficial teacher's deputies" and provide a "calm and wisdom" that reinforces the power of teachers. Children are often ashamed to misbehave because their adult relatives, or other adults they know (who "don't have time for no foolishness"), might see what they are doing. As 17-year-old senior Alex Lee remarked, "I don't want to be cursing and acting silly around them. I got respect for old people. Some of them are 40, 45 years old." For those children who misbehave anyway, adults can also be extra disciplinary aids. Once, the principal of DuSable, Charles Mingo, thought he saw a girl beating a boy in the hall. It turned out to be a mother disciplining her son, who was about to skip gym class. "She popped him right there in the hall and marched him off to the gym." (On the other hand, while sending teenagers to school with middle-aged adults may seem intuitively appealing, sending them to school with college-age students is perhaps another story.)

In the words of the anonymous New Jersey teacher:

[In the real world, dropouts] could ponder their choices without draining time and resources from other kids who want to learn. Teachers would have more time to teach, and principals would have the opportunity to meet responsible students instead of dealing with the same problem kids over and over. I'm sure some of the dropouts would do well in the work world, especially those who got into a trade that emphasized experience over book learning. I'm equally sure that others would come back to school with their attitudes adjusted . . . . Compulsory education laws obscure the fact that most students would choose to be in school anyway--and that choice is a major motivator in learning. Perhaps more important, in the end, such laws make willing students pay the freight on unwilling ones. And those charges are pretty steep.

Part 5

Conclusion

Our conclusion is threefold.

A preferred public policy solution to school violence, therefore, lies not in changing the individual acts of individual schools, but rather in creating an educational environment relying less on centralized, government-run, compulsory approaches, and more on localized, voluntary ones, including private-school options.

Part 6

About the Authors

Alexander Volokh has been a policy analyst with the Reason Public Policy Institute (RPPI) since 1994. Reason Foundation has published numerous studies on education innovations, charter schools, and contract services, including Alternative Teacher Organizaitons: Evolution of Professional Associations, Charter School Innovations: Keys to Effective Charter Reform, and Using Contractors to Cook, Clean, and Drive the Bus.

Lisa Snell is an assistant policy analyst at the RPPI and works closely with RPPI's Education Studies Program. She is the co-editor of Privatization Watch, a monthly newsletter published by the RPPI's Privatization Center, and has written numerous articles on education and privatization.