Introduction

It has long been known that lower-class black males are disproportionately under some type of criminal justice system control. Some estimate that up to one-third of black men are under state control [27]. This trend to disproportionately incarcerate one of the most disadvantaged populations in the United States has resulted in the increased adoption and use of intermediate sanctions; including intensive supervision probation, boot camps, halfway houses, day reporting centers, and the like. Intermediate sanctions are intended primarily for offenders convicted of first-time, low-level offenses or drug offenses, both convictions for which disadvantaged blacks predominate [27]. Purportedly, the purpose is to put offenders for whom there is still hope back on the right path quickly and effectively. Intermediate sanctions also serve as a cheaper alternative than incarceration for this largely low-risk population. Given the monetary and political appeal of these sanctions, it is not surprising to find that their expected benefits typically are not supported by empirical research. As Welsh and Farrington ([31]: 159) note, “anecdotal evidence, program favorites of the month, and political ideology seemingly drive much of the crime policy agenda.” Programs created with little reflection in best practice research and with little evaluation result in diversion of offenders to programs that do not appropriately match their needs. This could cause offenders to exit programs prior to their completion.

In this article I argue that black males are more likely to drop out of boot camp than white males because they have been inappropriately diverted to this intermediate sanction. I frame my arguments within definitions of masculinity amongst low-class black males. I first assert that black males have been historically stereotyped as hypermasculine and are not considered to be a diverse group. Second, I describe how this perspective has been incorporated into the criminological literature and use Anderson’s [1] work to frame my arguments and analysis. Third, I discuss previous research on boot camps and apply the implications of that research to this thesis. Fourth, I analyze differences in boot camp drop out by race using data gathered by MacKenzie [16] from a boot camp in Florida. Finally, I discuss the implications of my findings and identify areas for future research.

Review of literature

Most research on corrections focuses mainly on recidivism. However, it is important to consider dropouts for similar reasons. First, analyzing who drops out can help to improve the correctional program. When there is a possibility that an offender can “fail” a corrections program, explanations of failure can uncover weakness in the program. Second, dropouts are costly. Identifying potential failures can save money by diverting failures to different programs. Finally, offenders can directly benefit by being placed in a program that will work best for them, which may also lead to lower re-offending levels by program participants.

To understand how black males may be less successful and potentially misplaced in boot camp, it is important to recognize how black men are viewed by mainstream white society.

Black masculinity

Specific ideas of masculinity vary by individual, but a general consensus exists on what is “manly.” Kilmartin [12] asserts that personality traits, accepted and prohibited activities/actions, and roles all have masculine forms. For example, men should have aggressive, dominant personalities. They should face problems head on and independently. They are typically husbands and fathers; people who support others financially and emotionally. Typically, variations on these acceptable forms are viewed negatively. This perspective on masculinity has, however, been shaped by whites (especially males) who reside in the upper strata of society and exercise hegemonic control over definitions of masculinity.

According to these definitions, black men are stereotyped as being hypermasculine. Hypermasculinity emphasizes exaggerated masculine traits [12] and physical domination over others. This negative ascription to black males can and has been functioning to unite whites and reinforce hegemonic white masculinity [10]. As such black men are not consistently afforded the benefits of being a male in society, whereas white men are [9].

The distinction of masculinity by race and ensuing social degradation of blacks began with slavery [32]. During slavery, black males were viewed as sex-crazed men that, should they be allowed to roam free, would attempt to rape white women. The association between racialized masculinity and violence continued throughout the Civil Rights Era when the visibly angry black man emerged and riots broke out in black neighborhoods. Presently this image is perpetuated in the media that, for some whites, serves as the only window into black life. Black males who successfully make it to television are typically hip-hop artists and sports figures—two venues where hypermasculinity abounds. Black males are also often displayed in both factual news and fictional entertainment as violent criminals.

This social perspective that lumps all black men together and typifies them as hypermasculine has permeated research. Researchers often fail to recognize black masculinity as its own topic instead choosing to view it as a subtopic that contrasts with a norm of “real” masculinity [23]. Negative consequences result when policies are based on research that does not recognize differences among black males. Criminological theories and research are not immune from this issue.

Criminological literature

Group-level theories of criminality attempt to explain why certain segments of the population are more or less likely to commit crimes than others based on the social milieu of that group. These theories attribute one type of value system to the criminal group. For the immediate purposes, these theories can be grouped into three categories: theories that consider socioeconomic status, theories that consider race, and theories that combine race and socioeconomic status. Theories that consider socioeconomic status posit their arguments within a culture of poverty framework. This framework argues that being poor carries its own cultural values with it. In this value system, crime does not carry the same negative weight as it does in higher classes. In fact, crime can be viewed as a means of gaining status [7]. Hypermasculine behaviors are displayed through fighting and being unafraid to engage in fights [6]. Dominance can also be asserted by displaying a tough attitude [21], consistent with a hypermasculine system of values.

Theories in the second category consider only race and argue that racial minorities have their own subculture. The values in this subculture contrast with the values of the majority race in all subject areas, including crime. Wolfgang and Ferracuti ([33], see also [8]), for instance, argue that black culture is a subculture with especially violent tendencies. These violent tendencies are historical in nature and have become part of the value system of blacks. They assert that violence is typically the first resort to problem solving and is used often.

More recent theories couple poverty with race. This integration is somewhat natural due to the confluence of race and poverty and has served to provide a more comprehensive explanation of differential rates of crime amongst groups. Messerschmidt [20] considers the intersection of gender with race with class, concluding that lower-class minorities have a construction of masculinity that differs from that of the middle-class. The lower-class minority version of masculinity (which appears to be shared by both males and females) lends itself more toward specific types of street crimes and physical violence. He asserts that this group is forced to adopt these tendencies because they are the only means to success. Other traditional avenues, such as education, are blocked because of the group’s race and class. Messerschmidt recognizes different constructions of masculinity for different races, but fails to fully develop the idea of intra-racial heterogeneity.

Anderson [1] departs from past trends and recognizes that the low-class black population is not homogenous. He notes that there are two different types of families that live in low income, predominantly minority areas: “decent” and “street.” The street families are those that engage in deviant behavior and follow the “code of the streets.” The decent families abide by traditional middle class ideals. Decent families tend to work hard and attend religious services. Family members may not earn much money but are legitimately employed and/or enrolled in school. Street families in contrast, do not value hard work and often contain unemployed adults. These families are the creators and perpetuators of the “code of the street.”

The code of the street centers on “juice” or respect. Respect is earned by defending oneself in a physical fight and seeking out physical fights. The ability to appear capable of violence at any time and induce fear in others is seen as a pivotal part of this attitude. In addition, attempts to control others are exercised through means of physical violence. The “code of the streets” permeates the entire neighborhood. Decent boys are susceptible to and may find it necessary to engage in street behavior while not necessarily abiding by street values. Anderson [1] dubs their ability to transition from street to decent and vice-versa “code switching.” The boys that code switch do not feel as though the hypermasculine street values are a way of life, but are forced to switch in order to avoid physical harm and social stigmatization. Hence, decent boys are capable of changing roles, but it is not by choice. Once they return home hypermasculinity is devalued and returning to “decent” is desirable.

Regardless of the group or groups considered, these theories essentially have the same theme: hypermasculinity is the dominant type of masculinity for black males. Masculinity is displayed through acts of violence; and violence is the way that lower-class black males gain status and respect. Outside of the world of social science, the low-class distinction is often lost and all black males are viewed as a homogenous group. Again, the media serves to perpetuate this image.

What are the consequences of this stereotypical hypermasculinity for criminal justice processing? First, from this perspective, all black males have the ability to commit violent crime at a moments notice. Second, this perspective does not allow for the “decent” black male to exist. Young black men who commit crime cannot be viewed as “good kids who just got off track.” That distinction is reserved for young white men. This perspective has permeated into courtrooms and affects punishment decisions resulting in blacks being sentenced more harshly than whites [4, 25]. When faced with a “hardened offender” one punishment option is to place them in an institution where they will be met on the same hypermasculine playing field. The correctional boot camp is one such institution.

Boot camp

The purpose of boot camp incarceration is to rehabilitate offenders through strict discipline, swift punishment for rule infractions, and by instilling respect for authority. The attempt to achieve these goals is done through a quasi-military style regiment. Boot camps proliferated as a sanction for first-time, non-violent offenders during the mid-1980s. The length of incarceration in boot camps is typically 90 to 120 days [14]. Boot camps vary by jurisdiction in their program specifics but typically have a shared structure [14].

The key distinguishing feature of boot camps is an emphasis on all things physical. Rule infractions have physical consequences. Participants can be required to do intense physical labor. Programs can incorporate group exercise into daily routine in order to build discipline. These physical challenges are not voluntary, but are demanded by people in authority. The pressure to live up to these demands is both social and practical. When an offender fails to meet physical challenges it indicates that he is weak which is stigmatizing. It can also be viewed as noncompliance and could lead to further corporal punishments or to ejection from the program. The emphasis on success through physical accomplishments and threats of physical punishment are key features of hypermasculinity.

Unlike prisoners in traditional facilities, boot camp inmates are required to successfully complete the boot camp program. Participation in the program can be either voluntary or mandatory depending on the state. For those states with mandatory participation offenders can fail to complete the program only by being administratively removed. In states with voluntary participation the inmate can be ejected or choose to leave. High drop out rates can make boot camps an especially costly correctional option [30]. Tonry and Lynch [28] estimate that one-third to one-half of all boot camp entrants fail to complete the program. In an evaluation of eight state boot camp programs MacKenzie and colleagues [17] found completion rates between 3% and 51%. Other studies have found drop out rates around 25% [3, 5]

Benda and colleagues [2] considered the question of what types of offenders would be more likely to successfully complete boot camp. They hypothesized that gang members would be more likely than non gang members to complete a boot camp program because they adhere to hypermasculine values while in the gang. He found support for this hypothesis; gang members were more likely to complete boot camp. But, a racial twist emerged: black males had lower odds of completing the program than white males. While this result was not significant, it was interesting because it contradicts the image of black males as hypermasculine.

It may then be the case that black offenders who are sentenced to boot camp are not hypermasculine. Statutory requirements typically restrict boot camp to low-level and first time offenders. Yet, judges may be more likely to send black offenders to boot camp because of their hardened criminal image even though some offenders could be placed on less restrictive supervision. Although some black offenders are properly placed, this other group is put into an environment that is inconsistent with their value system. As such, they are not likely to successfully complete the program.

From this brief literature review, I develop and test two hypotheses. These hypotheses are based mainly on results from the Benda et al. [2] study and Anderson’s conceptual framework.

  1. Hypothesis 1:

    Black males (because they are inappropriately diverted to boot camps) are more likely than whites to dropout of boot camp.

  2. Hypothesis 2:

    “Decent” black males are most likely, in comparison to whites and “street” black males, to drop out of boot camp.

Data and method

The data used to test these hypotheses were originally gathered by MacKenzie [16] as part of an assessment of the effectiveness of boot camps in preventing recidivism. The original study evaluated boot camp facilities in seven different states. I only consider data from the state of Florida because Florida had the highest number of dropouts amongst all of the states analyzed. The original sample contained offenders from five consecutive cohorts who began their sentence between January and September 1990. The sample consisted both of boot camp participants and prison parolees. I analyzed only the offenders participating in boot camp. This produced a sample of 102 offenders. Data were missing on less than 5% of the sample.

Florida’s boot camp is for youthful, male offenders. The program capacity was 100 offenders. MacKenzie ([15]: 60–61) notes that there were a high number of dropouts in the boot camp program and most were dismissed from the program due to discipline problems; offenders could not choose to leave the program. In addition boot camp participation was court ordered. She reports that inmates spent “a little less than 2 hours per day in counseling or educational programs…most of this time was devoted to rational emotive therapy” (MacKenzie [15]: 61).

Dependent variable

Dropping out of boot camp is the dependent variable. It is a dichotomous variable indicating whether or not the participant dropped out of boot camp (coded 1 = participant dropped out; 0 = participant did not drop out). Thirty-nine percent of the sample failed to complete the programFootnote 1.

Independent variables

The race of the offender is the primary variable of interest. I created a dichotomous variable to indicate whether the participant was black or not (coded 1 = participant is black; 0 = participant is white). Fifty-eight percent of the sample is blackFootnote 2.

To indicate whether or not the offender was “decent,” I included whether or not the offender was employed and whether or not the offender was enrolled in school. These variables were self-reported as part of a survey instrument administered immediately prior to entry into the boot camp. They are both dichotomous variables (coded 1 = employed/enrolled in school; 0 = unemployed/not enrolled in school). Forty-five percent of the sample was employed at the time of the offense and 10% of offenders were both enrolled in school and employed at the time of arrest. While income was measured, over half of the data was missing on this variable; hence it was not included in the present analysis.

I added two control variables from official records that may play a role in whether or not the offender drops out of boot camp. First, I included whether or not the offender was sentenced for a violent or non violent offense. The type of offense could play a role in the likelihood of dropping out of boot camp for numerous reasons (e.g. these people are more likely to engage in anti-social behavior or, given program’s hypermasculine environment, violent offenders may be more able to adapt to the program). This variable is dichotomous (coded 1 = violent offense; 0 = non-violent offense). Seventy percent of the sample committed a nonviolent offense. Second, I included the age at which the offender was sentenced. This variable was added to control for the impact that age could have on the likelihood of dropping out. It is reasonable to expect that an older offender would be more compliant with institutional rules [29]. This variable is continuous. The mean age was 18.8 years and age had a near normal distribution.

Three items from the survey instrument were included as control variables to address the impact that offenders attitudes may have in their prospects of dropping out of boot camp. Respondents were asked to rate their agreement with a series of statements. The responses ranged from one to five and were on a Likert-type scale with one indicating strong disagreement and five indicating strong agreement. I included three statements to control for any predetermined decision the offender made about the program. These pre-program entry attitudes may indicate whether the participant had already determined that he was going to fail or succeed. The three statements are: “There is nothing in this place that will help me;” “The programs in this place will never help me in any way;” “I believe I am tough enough to handle this place.” Surprisingly, the correlation between the first two statements is below 0.5. The mean scores for these variables and the other variables are displayed in Table 1.

Table 1 Descriptive statistics of the variables used to model the likelihood of dropping out of boot camp

I used a logit model to address the research question. Logit models are appropriate for dichotomous dependent variables. All results are reported in odds ratios as they are more interpretable and accessible for the research question.

Results

I display four models that assess the impact of being black on dropping out of boot camp (see Table 2). The results are reported as odds ratios. The first three models are insignificant as indicated by the chi-square statistic. Model one considers only the effect of the being black on dropping out. Being black did not significantly affect the odds of dropping out of boot camp. Model two adds the independent variables without considering any interactions. Again, being black was not significant, but positive responses to the question about being tough enough for the institution decreased the odds of being a boot camp dropout. The third model reports the interactions of the indicators of decency with being black. Neither of these interactions were significant. In addition, a model (not displayed) that used a scale from the responses to “nothing will help me” and “the programs will not help me” was tested. The addition of this scaled variable was not significant.

Table 2 Odds ratios from logit models on the dependent variable of whether the offender dropped out of boot camp

The final model was significant (χ 2 = 22.42, p < 0.01). This model includes the interaction of the toughness survey item, the only significant variable from Model 3, with being black. Because of collinearity between the interaction term and being black, the tough variable was first centered about its mean. This sufficiently decreased the variance inflation factor (VIF) of the interaction and its components. The interactions from model three were not included in the final model in order to maintain parsimony.

The final model showed that having a stronger belief in one’s toughness (or at least responding more affirmatively to that question) significantly decreased the odds of dropping out of boot camp by approximately 75%. When, however, the race of the offender was interacted with that response, the results were drastically different. Black offenders had increased odds of dropping out of boot camp. In fact, each increase in agreement with the toughness statement increases the odds of a black male dropping out by a factor of 3.6.

Discussion and conclusion

Turning specifically to the research hypotheses, neither was supported. Being a “decent” black male (as measured through employment or school) had no effect on the odds of dropping out of boot camp. Also, being a black male had no direct effect on dropping out of boot camp. However, the significant interaction between one’s confidence that he is tough enough to complete the program and being black warrants further discussion.

It is likely that offenders interpreted “I am tough enough to handle this place” to mean that they could withstand the rigors of boot camp both mentally and physically. It makes intuitive sense that one’s belief to be successful will result in a higher probability of success. This is the basic premise of a self-fulfilling prophecy. Why then, does this hold true for white males but not blacks? There are four possible scenarios that could explain the result of the interaction, some methodological. First, black offenders may have experienced respondent fatigue which would affect how they responded to the survey (e.g., circled “strongly agree” for every answer). However, a brief exploration of the data reveals that this is not, in fact, the case. Second, black offenders could have been intentionally dishonest on the survey. There is some evidence that race affects how honestly respondents answer survey questions (see [11]). Third, black offenders may have believed that they could tolerate boot camp, but after experiencing it, they discovered they were unable to. Finally, black offenders may actually be “tough enough” for boot camp but they do not complete the boot camp for other reasons not related to their ability to mentally or physically endure the situation. These last three explanations deserve discussion because they return to ideas of how masculinity, hypermasculinity, and being a black male intersect.

Intentional dishonesty is when the offender answers affirmatively when he either did not know his prospects or did not believe he would actually complete the program. Regardless of the offender’s precise assessment of the situation, by strongly agreeing with the statement that he is “tough enough to handle boot camp” when he does not believe so, he is attempting to appear as though he is tough in general. In this scenario, an application of Miller’s [21] focal concern of toughness is also appropriate. According to Miller, toughness is a hypermasculine attitude that dominates one’s psyche. The male must appear as though he is emotionally and physically strong at all times. Respondents to the survey may have felt the need to feel tough even though the questionnaire was self-administered and reviewed by researchers with whom the offender would not see beyond the study. Hence, the propensity toward failure with increased confidence is really the result of an attempt to save face and not any real attitude about the program.

Black offenders may be more apt to believe that they can handle anything. Since it is likely that they are coming from dangerous communities, they may feel that they have good survival skills in any environment. In this case, the respondents would be code switchers. They fail in the hypermasculine boot camp because they cannot adapt to having to constantly play the street role. This argument has limited support by the directional (but not significant) relationship between the employed non-whites and boot camp failure. However, given that employment is positively related to dropping out without the race interaction suggests a general misfit between boot camp and offenders who are “more decent” regardless of race.

It is also possible that black offenders are failing boot camp for numerous other reasons unrelated to the belief that they are tough enough to handle the institution. In other words, this result could simply be correlated with other unmeasured factors that actually lead to dropping out. Other possible explanations could include violating institutional rules, not complying with the boot camp program, or having treatment issues necessitating transfer to another facility. Unfortunately, the data do not allow for a thorough investigation of these alternatives.

Contrary to the findings of Benda and colleagues [2], the interactions between being black and indicators of decency were not significant. Because it was impossible to explore whether offenders were gang affiliated, the test was weak. Similarly, the data did not allow a decomposition of race into different ethnic groups. Therefore, in addition to African American, the black and white category may include other ethnicities. This aspect further weakens the study. All future criminological research should be sure to include ethnicity as a factor due to the many unique circumstances surrounding different ethnicities (see for example [13, 18, 19, 22, 24]).

Future research should have several objectives. First, data should be collected for the purpose of addressing questions about boot camp failure. In particular, “exit” type interviews should be conducted with drop outs. This will allow researchers to compare variables collected before and during camp but also learn from the failures. Such information can be used to better assign clients and tailor programs to lessen the likelihood of boot camp drop out. Second, efforts should be made to quantify the distinction between decent and street families. This will not only help for the immediate purposes of studying what impacts dropping out of boot camp but will also provide further empirical support (see [26]), or lack thereof, for Anderson’s theory. Finally, boot camps should be studied contextually, across different states and different courts. This will determine how much variation in who drops out can be attributed to differences across jurisdictions.

Whether the boot camp option is an appropriate sanction for black males remains unclear. The results showed some differences between black males’ and white males’ success at boot camp but most coefficients are not significant. These results may be a function of the small sample size. Directionally, at least, blacks appear more apt to drop out of boot camps than whites. It is important to recognize that black males may have correctional needs that are different from their white counterparts simply because of the many consequences of growing up black in a white dominated society. Slight program modifications could change this dichotomy and make boot camp a reasonable alternative to prison for all inmates. If rehabilitation is goal of a given correctional program, then every effort should be made to place the right people in the right programs. This is better practice in the long-run for both economic and social welfare purposes. Benda and colleagues’ [2] call for a classification system that could help determine whether offenders will be successful at programs is valid and supported by this analysis. Such a system cannot, however, be developed without more evidence and replicable results on risk factors associated with program non-completion.