Abstract
In the United States, there are well-known racial, ethnic, age, and sex differences in incarceration rates. Younger offenders are more likely to be sentenced to prison than are older offenders. Black and Hispanic rates of incarceration are six to eight times that of White offenders and males are 14 times as likely as women to be sentenced to prison. This research explores how the combined effects of race, ethnicity, age, and sex, net of legally relevant factors, influence the decision to incarcerate. We examine these effects across nine offense categories. The analysis is based on Florida felony conviction data for the years 2000 to 2006. We find that legally relevant factors significantly influence the incarceration decision. Young Black males are most disadvantaged at the incarceration decision.
Since the early 1980s, several states have implemented sentencing guidelines with the purpose of creating more certainty in sentencing outcomes, reducing unwarranted sentencing disparity, and increasing the overall transparency of the sentencing process (Stith and Cabranes 1998). Despite these efforts, there continues to be persistent race, ethnic, age, and sex disparities in the incarceration decision. In 2007, for example, men in the United States were incarcerated at a rate 14 times higher than that of women. Comparing males between the ages of 25 and 29, Blacks had an incarceration rate that was seven times that of Whites and Hispanics had a rate that was two times that of Whites (U.S. Department of Justice 2009).
The highly variable rates of incarceration across demographic categories have led many scholars to question the source of these disparities. Some have argued that the differences reflect differential involvement in crime (Cohen and Kluegel 1978; Hindelang 1978; Steffensmeier and Allan 1996). That is, the observed disparities in sentencing outcomes reflect differences in criminal involvement. For example, men are more involved in crime than are women and young Black and Hispanic males are engaged in crime more often than Whites of the same age. Other research suggests that the disparities point to the differential treatment that men and racial minorities experience during the sentencing decision because of the definitions of harm and danger that are unequally distributed across racial, ethnic, age, and sex categories (Kramer and Ulmer 2009; Spohn and Holleran 2000; Steffensmeier and Demuth 2006; Ulmer and Johnson 2004). 1 Zatz (1987) maintains that the biases that result from these definitions of harm and danger are not necessarily overt, but rather are systemic and socially conditioned by those defining who is most threatening in society.
The purpose of this research is to further explore these issues by examining the effects of race, Hispanic ethnicity, age, and sex on the decision to incarcerate. Using Florida felony conviction data between the years of 2000 and 2006 (N = 501,027), we specifically question whether young Black and Hispanic males pay an incarceration “penalty” for committing specific types of offenses. Steffensmeier, Ulmer, and Kramer (1998) were the first to recognize the importance of considering the additive effects of race, sex, and age on judicial decision making. To date, only a few studies have examined the combined effects of age, race/ethnicity, and gender on sentencing outcomes (see Kramer and Ulmer 2009; Spohn and Holleran 2000).
The current study extends prior research in three ways. First, we include Hispanics as well as Blacks in order to examine if young Hispanic males are just as likely as young Black males to be sentenced to jail and/or prison. Next, we empirically examine the effects for jail and prison across nine specific crime types (see also Holleran and Spohn 2004; Kramer and Ulmer 2009). Crime categories are highlighted here because prior studies have generally grouped them into three broad categories. Doing so, fails to more fully account for how crime types might differentially impact the decision to incarcerate (Kramer and Ulmer 2009; Spohn and Holleran 2000; Steffensmeier and Demuth 2000). Finally, we emphasize the importance of examining the effects for jail and prison separately. Holleran and Spohn (2004) note that combining jail and prison into one category increases the risk of measurement error because they are qualitatively different sentence types (see also Holleran and Spohn 2004). We therefore utilize a three category dependent variable to explore how the combined effects of race, ethnicity, age, and gender influence the incarceration decision.
Race, Ethnicity, Sex, Age, and Sentencing: Conceptual Issues
The relationship between criminal sentencing and the defendant’s race, ethnicity, sex, and age has most often been conceptualized using the “focal concerns” framework. This perspective emphasizes the relevance of both legal and extralegal considerations in the sentencing process. As developed by Steffensmeier (1980) and subsequently elaborated by a variety of researchers (e.g., Steffensmeier and Demuth 2001, 2006; Ulmer 1997; Ulmer and Johnson 2004), this perspective suggests that there are three primary concerns which may combine to influence judicial decision making. These are blameworthiness, protection of the community, and practical constraints and consequences.
Blameworthiness focuses on the defendant’s culpability and the degree of injury associated with the crime. The objective of sentencing in this context is to ensure that the punishment fits the crime (Steffensmeier and Demuth 2006). Crime seriousness and the defendant’s prior criminal history are central considerations under this focal concern. The defendant’s role in the criminal event as leader, accomplice, follower, as well as whether the defendant has been victimized by others are also important factors in deciding the defendant’s blameworthiness (Steffensmeier et al. 1998).
Protection of the community is concerned with future criminality and the danger the defendant may pose to the community. Albonetti (1991) maintains that when judges are making assessments about the likelihood of future offending they generally have incomplete information and might therefore, rely on attributional stereotypes that are linked to offender characteristics (Bridges and Steen 1998). Steen, Engen, and Gainey (2005) demonstrated that young Black males are stereotyped and are therefore more likely to be seen as blameworthy. These attributions are potentially used as a kind of perceptual shorthand (Steffensmeier et al. 1998) that in turn helps to facilitate judicial decision making.
Finally, practical constraints and consequences can be concerns at both the organizational and individual level. Organizationally, judicial decision making may be influenced by the objective of ensuring that there is an orderly flow of cases through the court system. In addition, judges must consider the impact of their sentencing decision on jail and prison capacities (Ulmer 1995; Ulmer and Kramer 1996). At the individual level, the defendant’s “ability to do the time” and the costs that may be imposed on the families of defendants who are being sentenced are both relevant sentencing concerns (Steffensmeier and Demuth 2001).
Although there is general consensus in the sentencing literature that legally prescribed factors such as offense type, severity of offense, and prior record are typically the strongest predictors of punishment outcomes (Spohn 2000), the focal concerns perspective suggest that the sentencing process is not simply legalistic but one in which both legal and extra-legal considerations may affect the sentencing decision. For example, a review of presentence reports in the state of Washington led Bridges and Steen to conclude that minority offenders are often sanctioned more harshly because of the perception that they are “more threatening and therefore more deserving of punishment and control” (1988:556). Similarly, Steffensmeier et al. (1998) in interviews with Pennsylvania judges found that judges consistently viewed young Black men “as dangerous, committed to street life, and less reformable than women and older offenders” (787).
It is the case that quantitative sentencing data cannot directly demonstrate stereotyping as qualitative and social psychological data would be most appropriate. However, the quantitative patterns in sentencing data can provide more or less support by focal concerns and other attribution-based propositions. In the next section, we further explore these issues by focusing particular attention on the independent and joint effects of race, ethnicity, age, and sex on the decision to incarcerate. We also emphasize the importance of assessing their direct and combined effects across crime types.
Prior Research: Race, Hispanic Ethnicity, Age, Sex, and the Decision to Incarcerate
The direct effects of race, ethnicity, age, and sex on the decision to incarcerate have been well summarized in several comprehensive reviews (Kramer and Ulmer 2009; Pratt 1998). The empirical evidence suggests that Blacks and Hispanics are more disadvantaged than are Whites. Steffensmeier and Demuth (2001) found that Black defendants in Pennsylvania were significantly more likely to be incarcerated (jail or prison) than Whites for both drug and nondrug offenses (see also Steen et al. 2005). They also reported that not only were Hispanics more likely than Whites to be incarcerated, but their overall sentencing experience was harsher than that of Blacks. Although research has clearly demonstrated that Whites are most advantaged during the sentencing decision, Kramer and Ulmer (2009) have found that racial and ethnic disparities have declined since the implementation of sentencing guidelines in Pennsylvania.
Early reviews of the relationship between sex and the sentencing decision concluded that men are more disadvantaged than are women in part because men are more likely to be perceived as dangerous and threatening to the community (Griffin and Wooldridge, 2006; Nagel and Hagan 1982; Steffensmeier, Kramer, and Streifel 1993; Ulmer and Kramer 1996). Daly (1989) suggests that the judicial leniency shown toward women reflects their often perceived child care responsibilities that would be adversely affected by harsher criminal penalties (see also Daly 1987). Steffensmeier and Demuth (2006) also noted that relative leniency toward women may also be influenced by the focal concern of blameworthiness, inasmuch as judicial actors more often “view crimes committed by women as an outgrowth of their own victimization (246). One notable exception was reported by Koons-Witt (2002), whose analysis of sentencing both presentencing and postsentencing guidelines in Minnesota found that the defendant’s sex was not significantly related to the likelihood of incarceration at any of the study’s time points. 2
Age has not been the primary focus of sentencing research but prior studies have found that age is inversely related to the prospect of being incarcerated (Ulmer and Bradley 2006; Ulmer and Kramer 1996). Steffensmeier, Kramer, and Ulmer (1995) were the first to discover a curvilinear relationship between age and sentencing when they demonstrated that the probability of incarceration increased with age until the age of 30 and then declined at an increasing rate as defendants got older. The curvilinear effect was reproduced for different categories of crime but was strongest among violent offenders and weakest among those sentenced for drug crimes.
Age also has relevance for sentencing in terms of the focal concern of community protection. Portillos (1998) noted that the putative threat of Latinos is particularly associated with younger as opposed to older males (see also Campo-Flores 2005). Similarly, it is young Black males who are frequently perceived as “the dangerous class” since “the brunt of the stereotyping [as threat] falls most heavily on them” (Steffensmeier et al. 1998:769).
Despite numerous studies examining the direct effect of race, ethnicity, age, and gender on judicial decision making, few have assessed their combined influence on the likelihood of incarceration. Steffensmeier et al. (1998) were the first to assess the “punishment cost of being young, Black, and male.” Using Pennsylvania Guidelines data from 1989 to 1992, their main effects showed that the prospect of incarceration (jail or prison) was increased for Blacks, decreased for women, and curvilinear for age, with offenders over 50 and under 20 least likely to be incarcerated and those aged 21 to 29 most likely. This pattern was sustained for most race/sex pairings, with very young (18-20) and older men and women of both races less likely to receive an incarceration sentence than young (21-29) Black and White offenders of either sex (1998:776-7). The authors concluded “young Black men stood out as the offender group most at risk to receive the harshest sentences” (1998:787). Spohn and Holleran (2000) extended the analysis of this issue by including Hispanic ethnicity and employment status and found that young Black and Hispanic defendants were “consistently more likely than middle-aged White offenders to be sentenced to prison in each jurisdiction” (2000:301). 3
Kramer and Ulmer (2009) found similar results in their analyses of the sentencing decision (see also Kramer and Ulmer 2002; Ulmer et al 2007). In particular they demonstrate that young Black and Hispanic males have higher odds of incarceration and longer sentence lengths than do other racial, ethnic, and age categories. In addition, they found that while young Black males are disadvantaged, young Hispanic males receive more severe sentences than do their White or Black counterparts. Despite these differences the authors point out that these race and ethnic disparities in sentencing have declined across time with the implementation of sentencing guidelines.
Race, Ethnicity, Age, and Sex: Incarceration Penalty Across Crime Type
In the United States, there are well-known differences in perceptions of criminality (Gibbs 1988; Tittle and Curran 1988). Popular discourse about drugs and violence, for example, has become particularly associated with young Black and Hispanic males (Kennedy 1997; Loury 2003). Steffensmeier and Demuth observed “drug distribution and sales in particular are identified with Black dominated gangs on the one hand and with Hispanic traffickers on the other” (2001:152). Portillos notes that “the assumption is frequently made that if you are a young Latino, and especially a Latino male, you are a gun wielding, drug selling gang-banger, unless proven otherwise” (1998:156). These stereotypes about young Black and Hispanic males, coupled with the perception of stable future criminality reinforces the perception that young Black and Hispanic males are dangerous and in need of more severe punishment.
In their study of federal courts (2000) and the state courts in Pennsylvania (2001), Steffensmeier and Demuth found that both Blacks and Hispanics were more disadvantaged than Whites, with Hispanics having greater odds of incarceration than both Blacks and Whites. This pattern was substantially stronger in both studies for drug crimes than for nondrug crimes. Unnever and Hembroff’s (1988) found a similar pattern in their analysis of drug offenders in Miami.
Spohn and Cederblom’s (1991) examination of sentencing for violent crimes found that race influenced the incarceration decision for assault but not for more serious violent crimes such as rape, robbery, and murder. Using 1990 data from the State Courts Processing Statistics Program, Maxwell et al. (2003) found that there were no significant differences between Blacks and Whites in the decision to incarcerate for assault and sexual assault. Interestingly, they also report that Blacks were less likely than Whites to go to prison for murder. Among robbery defendants, both Blacks and Hispanics had greater odds of receiving a prison sentence. In addition, the effects for Hispanic ethnicity were inconsequential for murder, assault, and sexual assault.
The result of these studies suggest that the effects of extra-legal factors such as race, ethnicity, age, and sex on the incarceration decision might be in part influenced by the severity of the offense. Kalven and Zeisel’s (1966) liberation hypothesis suggest that with more serious offenses, judicial decision making is more strongly determined by the seriousness of the crime. On the other hand, with less serious offenses judges are relatively less constrained and therefore, might rely on nonlegal factors when making the decision to incarcerate. This study underscores the importance of exploring crime type and how it influences the combined effects of race, ethnicity, age, and sex on the incarceration decision.
While prior studies have generally demonstrated that young Black and Hispanic males are more disadvantaged during the incarceration decision, they only examined these effects across broad categories (property, violent, and drug). Spohn and Holleran (2000) found that both young Black and Hispanic males were more likely to be sent to prison, with no crime-specific analyses. Kramer and Ulmer (2009) reported on the likelihood of prison or jail incarceration for young Black and Hispanic males, but they also did not examine crime-specific effects.
Like Kramer and Ulmer, we examine prison and jail outcomes separately. And as they along with Spohn and Holleran did, we consider the consequences of being young, male, and Hispanic as well as being young, male, and Black. However, while those two studies did not consider the contextual importance of crime type, we examine these results for nine specific crime categories. In short, we are the first study to consider whether young Black males and young Hispanic males are disadvantaged in sentencing to either jail or prison for an extended number of specific crimes. This affords the most complete and detailed examination of these questions to date, and it does so using more recent data from a new geographical context.
Summary of Expectations
The present analysis is undertaken with several expectations. First, Black, Hispanic, and male defendants will more likely be incarcerated than White and female defendants, net of legally relevant factors. Second, the effects for age will be curvilinear and we expect the effects of race and ethnicity will be stronger for males than for females. Next, across all nine offense categories young Black and Hispanic males will have greater odds of being sentenced to jail or prison. Moreover, we anticipate that while young Black and Hispanic males will be most disadvantaged across crime types, the odds of their incarceration will be greater for drug crimes or less serious offenses where judges are less constrained by the law (Crawford, Chiricos, and Kleck 1998). Finally, although we expect young Black and Hispanic males to be disadvantaged relative to White males, we do not anticipate similar disparities to be observed among females. Steffensmeier and Demuth (2006) found that there were essentially no differences in the incarceration and sentence length decisions handed out to White, Black, and Hispanic female defendants. They essentially argue that Black and Hispanic female defendants benefit from their status as women. In this study, we further explore this issue by assessing whether the incarceration decision among women varies across crime type. 4
Data and Methods
The data to address these issues are taken from the Florida Department of Corrections (DOC) Sentencing Guidelines database and its Offender-Based Information System (OBIS). The sample includes 501,027 defendants convicted and sentenced in the State of Florida from 2000 to 2006.
The dependent variable takes into account three decision categories jail, prison, and community supervision with community supervision as the reference category. As recommended by Holleran and Spohn (2004) and demonstrared by Kramer and Ulmer (2009), we empirically examine the effects of jail and prison separately as they are distinct punishment outcomes with those who are sentenced to jail serving less than 1 year. In Florida, the median jail sentence is approximately 90 days, as compared to the average prison term of 33 months.
Independent Variables
Table 1 presents descriptive statistics for all of the variables included in the analyses. This research focuses particular attention on the combined effects of age, sex, and race or ethnicity, and the anticipated incarceration penalty paid by young Black and young Hispanic males. Age at sentencing is measured both continuously (18 to 69 years) and dichotomously, as age groups (18-29 and 30-69). 5 Age-squared is included in the main effects model to account for the previously observed nonlinear relationship between age and sentence severity (Steffensmeier et al. 1998). Race, ethnicity, and sex are measured dichotomously as non-Hispanic Black (Black = 1), Hispanic (Hispanic = 1), and male (male = 1).
Descriptive Statistics for Variables Included in the Analysis
Note: The sample size for those sentenced to probation is 259,359, jail is 129,915, and prison is 111,753.
In addition to the primary factors described above, several variables are included in our models to control for defendant and offense characteristics often associated with sentencing severity. These include current offense, prior criminal record, and whether or not the case went to trial. While not shown in the descriptive statistics, we also control for sentencing year and the judicial circuit in which the defendant was convicted 6 .
The defendant’s current offense is measured in terms of their primary offense type and its seriousness. Primary offense type is measured using nine dummy variables, based on the FBI’ s UCR index crime categories of murder/manslaughter, sexual battery, robbery, aggravated assault/battery, burglary, grand larceny, and motor vehicle theft as well as drug possession and drug sale. Current offense seriousness is measured using the sum of Florida sentencing guideline points associated with the offense for which the defendant was convicted. This measure is continuous, ranging from 2.4 to 999.4 points, and includes points for any additional offenses, use of a weapon, and victim injury.
Early reviews of the sentencing literature (Kleck 1981; Peterson and Hagan 1984) emphasized the importance of controlling for prior record in the estimation of race effects on the sentencing decision. For this analysis, prior record is measured in three ways: (a) the sentencing guidelines points associated with the defendant’s prior record, (b) whether or not the defendant has ever been sentenced to prison in Florida (YES = 1), and (c) whether or not the defendant has ever had his/her sentence to community supervision revoked in Florida (YES = 1). Collectively, these measures capture the information on prior record afforded the sentencing judge and should, therefore, account for the influence of that factor on the sentencing decision. 7
Florida’s sentencing guidelines scoring system arguably provides one of the most comprehensive and discriminating measures of crime seriousness available for sentencing research (Burton et al. 2004). For our purposes, crime seriousness is measured by the sum of the sentencing guidelines points assigned to the most serious offense at conviction, additional offenses, extent of victim injury or sexual conduct, legal status and community sanction violations, and enhancement points for motor vehicle theft, use of a firearm, street gang involvement and domestic violence. By controlling for the recommended guideline sentence we are also accounting for the presumptive disposition (Engen and Gainey 2000). A score greater than 44 points is a guideline recommended prison sentence. We use this as a dichotomous measure of crime severity which is called “scored to prison” (YES = 1).
Finally, we include an additional measure related to the sentencing event, whether or not the case went to trial (YES = 1). Prior research has shown that offenders who choose trial over plea agreements receive harsher sentences because they are not seen as remorseful for their crimes.
Analysis
In the analyses that follow, we estimate multinomial logistic regression models which examine the likelihood of receiving two incarceration decisions—jail and prison. 8 Following the estimation of a main effects model, we then examine the joint effects of age, race or ethnicity, and sex on the probability of receiving a sentence to jail or prison. We further estimate the same joint effects for defendants across nine offense categories in order to comparatively assess the incarceration penalty that young Black and Hispanic males may pay when sentenced for specific types of crime.
Findings
Table 2 presents the results of the multinomial logistic regression model which measures the independent effects of the defendant’s age, race, ethnicity, and sex on the decision to sentence a defendant to jail or prison (community supervision is the reference category). Also included in the model are dummy variables for the primary offense type, sentence year, and judicial circuit (results are not presented in table but are available upon request). The coefficients in each model are presented as odds ratios and nonsignificant results are designated with an asterisk.
Odds Ratios Multinomial Logistic Estimates of Post Incarcerationa
a Reference category is community supervision (i.e., probation, community control/house arrest). Models include dummy variables for sentence year and judicial circuit in which sentenced. Models only include felony convictions.
* Coefficients are not statistically significant at p ≤ .001.
The findings for the main effects of race, ethnicity, and sex are consistent with prior research with Blacks having a higher probability of being sentenced to jail and prison. In general, males are more likely to receive a sentence to jail and prison than are women. Although, Hispanics are less likely than Whites to receive to a jail sentence, they are more likely than Whites to receive a prison sentence. In fact, even after controlling for legally relevant factors, being Hispanic increases the odds of a prison sentence by 23 percent. Black defendants on the other hand are generally more disadvantaged as they are more likely than Whites to receive a jail and prison sentence, with the odds of a jail sentence increasing by 47 percent and the odds of receiving a prison sentence increasing by 24 percent. The results also indicate a nonlinear relationship between age at sentencing and the decision to incarcerate, with the log odds increasing then rapidly declining. 9
The results in Table 2 also demonstrate that the odds of receiving a jail or prison sentence are greater with higher levels of crime seriousness and prior record points. Having prior prison commitments along with prior violations of community supervision increased the likelihood of incarceration. Although receiving a score to prison reduced the odds of going to jail, defendants who scored to prison were five times more likely to be sentenced to prison. Being convicted by trial rather than plea bargain also increased the likelihood of incarceration.
Interestingly, even after controlling for crime seriousness, there is considerable variation in the likelihood of a jail or prison sentence across the primary offense categories (grand larceny is the reference category). Drug crimes have the largest effect on the odds of receiving a jail sentence with drug possession increasing the odds by 48 percent and drug sale/trafficking by 22 percent. Murder/manslaughter, sexual battery, and robbery offenses are all less likely to receive a jail sentence.
The results are slightly different for prison incarceration. First, defendants convicted of drug crimes are less likely to be sentenced to prison, with drug sale/traffic exerting no significant influence on the odds of receiving a prison sentence. Robbery and murder/manslaughter are the most likely to receive a prison sentence with the odds for murder/manslaughter increasing by 50 percent and the odds for robbery increasing by 64 percent, when compared to property crimes. Although motor vehicle theft increased the odds of receiving a sentence to jail it did not significantly influence the likelihood of receiving a prison sentence.
Table 3 presents the odds of being sentenced to jail or prison for 11 age, race, ethnicity, and sex-group categories, with Black males aged 18-29 as the reference category. 10 Collectively, the findings support our hypothesis that young Black males, in particular, pay an incarceration penalty, even after controlling for legally relevant factors. Specifically, when compared to all other race–ethnic-age-gender categories there is only one instance in which young Black males are less likely than others to be sentenced to jail. Moreover, the effects for defendants sentenced to prison are even more compelling. That is, in every comparative instance, young Black males have a higher probability of being sentenced to prison than all other groups, even after controlling for legally relevant factors. Overall, among both younger and older defendants sentenced to jail, the lower odds ratios are found for Hispanic males and females. This suggests that they are less likely than either Whites or Blacks to receive a jail sentence. However, for prison sentences, White males and females, in both contexts have the lowest odds ratios for incarceration and on that account are consistently more advantaged than Blacks and Hispanics in this regard.
Odds Ratios, Multinomial Logistic Estimates of Postconviction Incarceration for Age, Race, Hispanic Ethnicity, and Sex Categories
a Reference category
Models include all controls from Table 2.
* Logit coefficient not statistically significant at p ≤ .001.
The next set of analyses examine whether young Black and Hispanic males have a higher probability of incarceration across nine offense types. The only prior research to examine the crime-specific effects of age, race, and sex on the decision to incarcerate was reported by Steffensmeier et al. (1998). In their analysis across three broad crime categories, they found that the disadvantage for young Black males was similar across violent, property, and drug offenses. We extend this analysis by including Hispanics as well as females and by examining jail and prison separately for nine specific crime types.
Table 4 shows the odds of receiving a sentence to jail for 11 categories of defendants clustered by age, sex, race, and Hispanic ethnicity. As with Table 3, the estimates are based on full models using all of the variables shown in Table 2, but for parsimony only the age, sex, and race/ethnicity results are shown. Because estimates based on small subsamples may be unreliable, we follow the recommendation of Cohen (1990) and do not report results for particular age, sex, race/ethnicity clusters with n’s less than 30. For both jail and prison outcomes, these small subsamples only include female defendants for the crimes of murder and sexual assault.
Odds Ratios, Multinomial Logistic Estimates for Postconviction Jail Incarceration for Specific Crimes by Age, Race, Hispanic Ethnicity, and Sex.
a Reference category.
b Categories are not included in models as cell sizes are inadequate for valid and robust estimates (see Cohen 1990).
* Logit coefficient not statistically significant at p ≤ .001.
Models include all controls from Table 2.
For jail incarceration there are no significant age, sex, race/ethnicity effects for the very serious crimes of murder and sexual battery. For robbery, the only significant results involve the greater likelihood of incarceration for older White or Black males as compared to the reference category of young Black males. One possible explanation for this unanticipated outcome is suggested by the mean prior record scores for these older defendants which are substantially and significantly (p < .0001) higher than prior record scores for their younger counterparts. Older White males sentenced to jail for robbery have prior record scores that average 44 percent higher than young Black males and this difference is 266 percent when older Black males are considered. Similar differences are found using prior prison commitments as a measure of prior record. In these contexts, it would appear that prior record as well as other legal controls, trump any consequences of being young, male, and minority. There are two other instances in which older Black males are more likely to be sent to jail than their younger counterparts (drug sale/trafficking and burglary), and in these instances mean prior record scores for the older defendants were, respectively, more than two and three times higher (p < .0001) than for Black males aged 18-29.
For aggravated assault and motor vehicle theft, most jail outcomes show no age, sex, race/ethnicity effects, but where significant results are found, young Black males are disadvantaged. For burglary (with one exception noted above) and grand larceny, the most common result is that young Black males are the most likely group to be sentenced to prison. For drug possession and drug sale/trafficking, the only exception to the pattern of young Black males being significantly more likely to go to jail involves older Black males (see discussion above). For these drug crimes, the young Black male disadvantage is both consistent and relatively strong.
There is no significant difference in the odds of jail incarceration for young Hispanic males as compared to young Black males for murder, sexual assault, robbery, and aggravated assault. For the remaining crimes, young Hispanic males are significantly less likely to receive a jail sentence than young Black males. It should also be noted that for the seven crimes with adequate n-sizes involving female defendants, a consistent pattern is observed. For both younger and older defendants, Black females have a greater likelihood of being sent to jail than either Hispanics or Whites for each of the seven crimes. For example, for robbery and aggravated assaulted White and Hispanic females appear to be treated more similarly than Black females. However, for grand larceny, motor vehicle theft, drug possession, and drug sale/traffic, Black, White, and Hispanic females are generally less likely to be sent to jail than are young Black males. These patterns provide some evidence that women are treated more leniently than are males as every category that has statistically significant relevance they are less likely to be sentenced to jail.
Table 5 repeats the foregoing analyses with the results of models predicting the odds of receiving a prison sentence. As with jail sentencing, there are no significant age, sex, race/ethnicity effects for murder and sexual battery. For robbery, the young Black male disadvantage is only found in comparison with young White males, and young White, Black, and Hispanic females. For robbery, older White males are more likely to be imprisoned than young Black males. As with similar anomalous results in the jail analysis, it can be noted that prior record points for White males aged 30-69 average more than three times higher (p < .0001) than those for Black males aged 18-29. For motor vehicle theft, the young Black male disadvantage is again limited to comparisons with older Black males, young Black females, and older White females. All other coefficients are not significant.
Odds Ratios, Multinomial Logistic Estimates for Postconviction Prison Incarceration for Specific Crimes by Age, Race, Hispanic Ethnicity, and Sex
a Reference category.
b Categories are not included in models as cell sizes are inadequate for valid and robust estimates (see Cohen 1990).
* Logit coefficient not statistically significant at p ≤ .001.
Models include all controls from Table 2.
For the remaining crimes, there is more of the hypothesized pattern to be seen in the estimated odds of a prison sentence. That is, the likelihood of prison incarceration is significantly greater for young Black males in, 8 out of 11 comparisons for burglary, 9 out of 11 for drug possession, and 10 out of 11 involving aggravated assault and drug sale/trafficking. For the seven crime categories not involving drugs, there is no significant difference in the odds of being sentenced to prison for young Hispanic males as compared to young Black males. However, for drug possession and for drug sale/trafficking, young Hispanic males are significantly advantaged relative to Black males aged 18-29. Finally, unlike the pattern displayed for jail, Black woman are not consistently disadvantaged relative to Whites and Hispanics in sentencing to prison.
Conclusions
This study has examined whether there is young Black or Hispanic male “incarceration penalty” relative to the sentencing patterns experienced by other demographic groups. We focused in particular on young Black and Hispanic males because they are most often associated with symbols of crime, drugs, and violence in American society (Kennedy 1997; Steen et al., 2005). To the extent that these symbols of criminality influence judicial decision making, young Black and Hispanic males would potentially be more disadvantaged in sentencing outcomes.
This is the first study to examine the crime-specific effects of being young, male, and either Black or Hispanic on sentencing to both jail and prison. Prior related work either did not involve Hispanics, made use of an incarceration measure combining jail and prison and was limited to three broad crime categories (Steffensmeier et al. 1998) or was not crime specific in its analyses of prison sentencing (Spohn and Holleran 2000) or jail and prison outcomes (Kramer and Ulmer (2009). By including Hispanics, examining jail and prison separately and disaggregating our analysis by nine specific crime types, we provide the most comprehensive assessment of this research question to date.
Overall, the findings provide support for the focal concerns contention that judicial decision-making is most significantly influenced by legally relevant factors such as crime seriousness and prior criminal offending. The results also suggest that sentencing outcomes are apparently related at least in part to factors other than those directly associated with the seriousness of the crime and the defendant’s prior criminal history. Primary among these factors is the defendant’s age, race, Hispanic ethnicity, and sex which are consistent with the premise that stereotypes and generalized perceptions of who is most dangerous in society may exist in the American justice system.
When the combined effects of age, sex, and either race or ethnicity are examined, it is most often the case that for the most serious crimes of murder, sexual assault, and robbery, there are no significant differences among demographic groups in the likelihood of being sentenced either to jail or prison. This is also true for aggravated assault when it comes to jail sentences. In several instances, older White and older Black male defendants appear to be disadvantaged relative to young Black males and for these contexts it was shown that prior record levels of the older defendants were substantially and significantly higher than those for young Black males aged 18-29, likely driving the sentencing outcome.
However, for less serious property and especially drug crimes, a different pattern obtains. For these crimes, there is a strong tendency for young Black males to be disadvantaged relative to other groups in the decision to sentence defendants either to jail or prison. For the harsher prison sentence outcome, there is no significant difference in the disadvantage of young Hispanic males as compared to their young Black counterparts for seven of the nine crimes. But where there is a significant difference between these young male minorities, it is young Blacks who are significantly disadvantaged with regard to prison sentences for drug crimes.
This pattern of findings provide some support for Kalven and Zeisel’s (1966) liberation hypothesis which argues that with more severe offenses the sentencing outcome is more strongly constrained by the seriousness of crime, which places greater limits on judicial discretion. On the other hand, with less serious offenses judges are hypothesized to be relatively less constrained by the type of offense and the defendant’s prior criminal history. This in turn potentially gives judges greater discretion which increases the probability that extra-legal factors such as race, will influence the sentencing decision. This may explain why so few significant age, sex race/ethnicity effects are found for murder, sexual battery, and robbery, and why those effects are more readily apparent for less serious crimes such as larceny and drug offenses.
As with any research, the present undertaking is not without limitations. First, we lack information on charging decisions which rest almost entirely in the hands of the prosecutor. Although we do not have complete information on how prosecutors use their discretion, we do know that this is likely to have a significant impact on the sentencing decision (see Kramer and Ulmer 2009). Next, we also lack information on the defendant’s employment and socioeconomic status as well as familial responsibilities. Each of these characteristics are likely to be correlated with race, ethnicity, gender, and age (see Ulmer 1997). Despite these considerations, the US Sentencing Commission has concluded that factors such as employment and familial status should not be relevant to the sentencing decision and to the extent that they do influence sentencing outcomes undermines “a fair system of justice and makes penalties disproportionate to the severity of the crime” (Kramer and Ulmer 2009:115).
Beyond providing partial and crime-specific confirmation of the incarceration penalty hypothesis, this research also highlights the relevance of the additive nature of race, ethnicity, sex, and age on the sentencing decision. Scholars who examine the direct effects of these factors on sentencing risk overlooking how they may jointly create a disadvantage for young Black or Hispanic males in the justice system. The young Black male disadvantage demonstrated here, especially for less serious crimes, undoubtedly contributes to the growing Black/White disparity in incarceration rates. Currently, there are 3,161 Black male prisoners per 100,000 Black males in the United States. Among White males there are only 723 per 100,000 (U.S. Department of Justice 2009). Although the Hispanic male prison populations are not as large as Black males, they are still incarcerated at a rate double that of White males (1,632 per 100,000 Hispanic males).
Finally, in the date we have analyzed here over 46 percent of the young male and 50 percent of older male defendants were convicted primarily of drug offenses. As noted, the young Black male incarceration disadvantage was most consistently apparent for drug possession and for drug sales/trafficking. Notwithstanding the debatable “dangerousness” of a drug offender and the even more debatable racial and ethnic distribution of drug use in the United States, the racial and ethnic disparities of prison populations seems likely to continue its exponential growth unless our current policies toward the enforcement of drug laws and the current pattern of sentencing for drug offenders are reevaluated.
Footnotes
Notes
Acknowledgment
The authors would like to thank Eric Stewart, Gary Kleck , Brian Stults and the anonymous reviewers for their insightful comments on an earlier draft of this manuscript.
The authors declared no potential conflicts of interests with respect to the authorship and/or publication of this article.
The authors received no financial support for the research and/or authorship of this article.