Racial Disparity in Iowa Prisons: Possible Theoretical Explanations for Racial Differences Between Inmates on the Level of Service Inventory

Abstract

This article presents a test of several theoretically informed hypotheses that characterize differences between Whites and African Americans incarcerated in the Iowa prison system. The authors judge differences by comparing inmates’ responses on the Level of Service Inventory–Revised or LSI-R, which is a standardized risk/need assessment instrument used to classify Iowa inmates. The hypotheses are based on ideas found in theories of structural distributive justice, general strain theory (GST), and macro-structural explanations of crime. Iowa is an interesting case study because it ranks near the top in the United States in the proportion of Black to White prisoner disparity. This disparity serves as a lens that sharpens distinctions between the populations. The findings suggest that in comparison to White prisoners, African American inmates have higher total LSI-R scores than White inmates and that prior to incarceration African American prisoners had more difficulty finding work, were more likely to have an official record of violent crime, and were more likely to associate with people who were involved in crime than were White inmates. Additionally, the results suggest that in comparison to White inmates, African American prisoners were more likely to feel that their prison sentences were unfair and to act in ways that were indicative of this. These findings are consistent with explanations found in macro-structural theories of crime as well as concepts found in GST and structural distributive justice theory. The authors briefly discuss the implications of these findings.

Keywords

crime criminological theories race and corrections felon disenfranchisement prisoners rights

Introduction

For many years now, racial disparity in America’s prisons has been a challenging and yet fascinating issue for social science researchers and practitioners. The centrality of this topic has generated considerable discussion and debate. Much of this dialogue has centered on examining the correlates of disparity. Although research consistently indicates that African Americans are overrepresented in rates of victimization, arrests, and imprisonment (Blumstein, 1982; Federal Bureau of Investigation, 2008), currently there is not a clear consensus regarding the reasons for this. In criminology, explanations include descriptions that speculate disparity results from a subculture of violence, macro-structural inequalities, differential association, and learning, social control, strain, and negative emotions generated by injustice (Agnew, 1992, 2006; Akers, 1998; Blau & Blau, 1982; Burgess & Akers, 1966; Cloward & Ohlin, 1960; Hirschi, 1969; Krivo, Peterson, & Ruth, 2009; Merton, 1938; Peterson & Krivo, 2005; Sampson, 1987; Sampson &Wilson, 1995; Saperstein & Penner, 2010; Shaw & McKay, 1942; Wolfgang & Ferracuti, 1967).

The purpose of this research is to extend this conversation by examining a small slice of this subject matter—racial disparity in Iowa prisons. Iowa is an interesting case study because it ranks near the top in the United States in the proportion of Black to White prisoner disparity and yet has a very low African American population. The African American population in Iowa is about 2.8%, but African Americans comprise roughly 22% of the Iowa adult prison population. Accordingly, the ratio of African American to White prisoners in Iowa is among the highest in the country (The Sentencing Project, 2011; U.S. Bureau of Statistics, 2011). In what follows, we present and test several theoretically important hypotheses that characterize differences in the Iowa offender population. We judge differences by comparing inmates’ responses on the Level of Service Inventory–Revised or LSI-R (Andrews & Bonta, 1995, 2006), which is a standardized risk/need assessment instrument used to classify Iowa inmates. The hypotheses are based on ideas found in theories of social structural power and fairness, general strain theory (GST), and macro social-structural theories of crime. We believe our findings may serve as a small step along the road leading toward a better understanding racial disparity in prison in theoretically meaningful ways. In the next sections, we first present a brief literature review of the criminological theories that inform the hypotheses, then explicitly state the hypotheses, and follow up with a descriptive summary of the Iowa prison population and tests of the hypotheses.

Literature Review

Macro-Structural Explanations

The macro-structural approach specifies how specific features of community structure produce differential crime rates among subgroups (Byrne & Sampson, 1986; Sampson & Wilson, 1995). This perspective does not stress individual involvement in crime, but rather attempts to isolate the characteristics of communities that lead to criminality (Sampson & Wilson, 1995). The tradition traces its roots to the “ecological school” centered at the University of Chicago in the 1920s and 1930s (Park, McKenzie, & Burgess, 1925; Shaw & McKay, 1942). According to this view, group differences in crime are generated by a reduced sense of social cohesion and community organization. For a given community, low socioeconomic status, racial, and ethnic heterogeneity along with high residential mobility reduces social cohesion and thereby generates crime and violence (Sampson & Wilson, 1995; Shaw & McKay, 1942). Social order exists when individuals in a community exhibit a high degree of personal bonding with each other and the goals and values of conventional society. In poorer, more heterogeneous communities characterized by high levels of family disruption and resident turnover, social controls begin to break down. This failure results in decreased supervision of children as well as a reduced sense of connection with the community and its values. Crime and violence is the ultimate result.

Blau and Blau’s (1982) theory set the stage for contemporary structural explanations regarding the relationships between race, ethnicity, and crime. Their work mixed ideas found in the Chicago school with parts of Merton’s (1938) definitive treatise. The core idea of this work is that a disjunction between culturally defined goals and the distribution of socioeconomic resources results in diffuse feelings of frustration and alienation. In the United States, expressions of these emotions often take the form of violent crime because they are perceived as particularly illegitimate since they are frequently based on ascriptive and nonmeritocratic characteristics like race.

In 1987, Wilson posited that racial differences in crime and other social problems were rooted in the divergent community experiences of Whites and African Americans. Wilson argued that differences in White and Black offending stem from the extreme concentrations of disadvantage found in many African American neighborhoods. These pockets of disadvantage isolate African American residents from mainstream society and its connections to jobs and conventional role models. These same conditions also generate social disorganization, family disruption, and significant reductions in informal social controls generally found in other types of neighborhoods (Massey & Denton, 1993; Peterson & Krivo, 1993; Sampson, 1987; Shaw & McKay, 1942).

Consistent with this thesis, Sampson and Wilson (1995) argued that neighborhood disadvantage predicts violence. Although this is true for both Whites and Blacks, it is especially relevant for poor Black communities because these neighborhoods are more likely to be characterized by extreme poverty, public housing, and social disorder. While the causes of Black crime are not unique, the contextual neighborhood basis for race and crime is especially important for African Americans because macro-structural factors function to concentrate Black poverty and family disruption in urban inner-city areas making for stark living conditions in many of these isolated communities. Sampson and Wilson claim that it is nearly impossible to reproduce the circumstances under which African Americans often live in White communities because even the most disadvantaged White communities have less deprivation than the typical disadvantaged African American community (p. 42). Consequently, White communities are less likely to demonstrate adaptations that reduce social control (Krivo & Peterson, 2005). In short, Sampson and Wilson held that macro-social patterns of residential inequality produce concentrated pockets of the truly disadvantaged. This social isolation generates structural barriers that undermine social organization and crime control. And this leads to community-level violence. Racial segregation and isolation of Blacks also differentially exposes them to structural conditions that result in cultural adaptations congruent with violence (Sampson & Bean, 2006). These cultural adaptations take the form of value systems and attitudes that provide a basis for the tolerance of crime and deviance (Sampson & Wilson, 1995, p. 50). It is important to note that Sampson and Wilson were not making a “culture of violence” argument. Instead, they suggested that ghetto-specific practices (including violence) are learned by precept and imitation and that they do not become internalized and so are not necessarily enduring. Consequently, these customs do not influence behavior irrespective of environment, but rather are driven by macro-structural forces.

Several studies have tested principal features of Sampson and Wilson’s (1995) argument. Results from these studies have generally been supportive. For example, McNulty (2001), Krivo and Peterson (1996), and Shihadeh and Shrum (2004) demonstrated that the effects of structural variables in predicting violent crime is consistent across a broad range of cities and neighborhoods using a wide variety of research methodologies and forms of data (see also, Krivo & Peterson, 2000; Ousey, 1999; Parker & Johns, 2002; Parker & Pruitt, 2000).

GST

The principal claim of GST is that people are pressured into crime by the various strains they experience. Strain is an event or condition that people dislike (Agnew, 2006, p. 3). It can be broadly defined as the loss of something valuable, mistreatment by others, or as blocked goals. The theory is an elaboration of earlier strain theories including Merton’s (1938), Cohen’s (1955), and Cloward and Ohlin’s (1960) theories. However, these early theories defined strain more narrowly than GST, and mainly focused on strain generated by blocked goals. The core argument of the early theories was that Americans desire success—defined as money and status—but also share a cultural belief that success should be achieved legitimately through hard work. However, for some classes of people, like the poor and minority members, there is discontinuity between these goals and the means to achieve them. For these segments of the population, the legitimate means for achieving success are often blocked and this impasse results in strain. Frustration or strain is an aversive experience that people must deal with. According to these early theories, people often chose crime as one way to cope with strain.

The concept of strain was defined more broadly in GST. In addition to classifying strain as the result of failure to attain goals, strain results from personal mistreatment and loss. Additionally, the theory distinguished between objective strains and subjective strain and between strains that are personally experienced and those that are anticipated or vicarious.

As noted, GST argues that strain causes people to experience negative emotions which create pressure for corrective action and that some people are better able to deal with strain than are others. Consequently, people poorly equipped to deal with strain sometimes turn to crime as a way of dealing with the negative emotions produced by it. GST outlines several reasons why strains may lead to crime. First, certain negative emotional states can be conducive to crime. Anger is a good example. Research has demonstrated that people experience anger when they think they have been treated unfairly (Jasso, 1980). Simply put, anger generates intense pressure for corrective action (Agnew, 2006). As we will describe in more detail below, this pressure often takes the form of justice restoring attempts (Jasso, 1980; Markovsky, 1985). Anger also reduces a person’s ability to cope with the situation in a legal or legitimate way. This is because angry people often fly off the handle and so are less able to judge the situation accurately, are less able to effectively communicate with others, and often care less about the costs of crime than do less angry people (Agnew, 2006,p. 33). Research suggests that demographic differences among groups are an important factor in determining how people cope with strain. Some groups tend to experience strains that are more likely to generate anger or be perceived as unjust than are others. These types of strains are conducive to coping mechanisms associated with crime. For example, Agnew (2006) and Broidy and Agnew (1997) used GST to explain gender differences in offending rates. These authors argue that males are more likely than females to be exposed to high magnitude strains like harsh discipline, negative school experiences, and abusive peer relationships that create pressure for criminal activity. GST asserts that these same types of pressures are partially responsible for differences in offending rates between Blacks and Whites (Asseline, 2009).

A second reason why strain leads to crime is that strain reduces elements of social control while it simultaneously increases the probability of the social learning of crime. These ideas rely heavily on Hirschi’s (1969), Sutherland and Cressey’s (1978), and Akers’ (1998) theories. According to GST, when strains are recurrent or chronic in nature, they often reduce individuals’ ability to resist crime because they damage relationships with conventional others, such as parents, teachers, and friends and also may reduce investments in conventional goals and activities. Chronic strain may also decrease a person’s belief that crime is wrong (Agnew, 2006, p. 42). GST is grounded in the metatheoretical assumption that modern society consists of conflicting norms and definitions of appropriate behavior and that these definitions of behavior are learned through contact with others. Criminal activity depends then, in part, on increased association with others who share definitions tolerant of criminal behavior. The more a person has contact with others who hold tolerant definitions of crime, the more likely the person is to engage in criminal activity (Agnew, 2006). The basic idea is that chronic strain generates reduced social control, and this in turn increases the chances that people will come into contact with members of groups that hold tolerant definitions of crime. GST argues that the primary reason why African Americans have higher rates of offending than Whites is because Blacks are more likely to experience these chronic strains (Agnew, 2006).

Research investigating the relationship between strain and crime has generally supported GST’s theoretical claims. For example, support has been found for the connection between crime and strain (Aseltine, Gore, & Gordo, 2000; Baron, 2004, 2007, 2008; Froggio & Agnew, 2006) and anger and crime (Mazerolle & Piquero, 1998; Piquero & Sealock, 2000); and the impact of objective and subjective strains on criminal behavior (Agnew & White, 1992; Lagrange & Silverman, 1999; Mazerolle & Maahs, 2000).

Structural Power and Distributive Justice

As noted, GST theorizes that strain which is perceived as unjust produces strong emotional reactions such as anger and frustration and that these types of emotions generate intense pressure for corrective action. This action often takes the form of violent crime. GST argues that certain characteristics of strains influence perceptions of injustice. People tend to perceive strain as unjust when (i) they think it is undeserved, such as when a person believes that it did not result from his or her own behavior; (ii) when a person perceives the strain as excessive or overly harsh; (iii) when the strain violates strongly held social norms and values; or (iv) when victims perceive it as very different from their past treatment in like circumstances or as different from the treatment of similar others in parallel circumstances (Agnew, 2006, p. 64).

These arguments dovetail with concepts found in distributive justice theories. These conceptually related theories investigate how people evaluate the fairness of allocations of benefits and burdens (Hegtvedt, 2005; Hegtvedt & Markovsky, 1995). People who perceive incongruence between actual and expected outcomes experience emotional reactions (Markovsky, 1988; Turner & Stets, 2005). Under fairly robust conditions, under-reward vis-à-vis expected outcomes generates feelings of anger or resentment. These emotional reactions have important consequences for behavior, with more intensely felt emotions increasing the probability of justice-restoring behavior (Homans, 1961; Jasso 1980; Markovsky, 1985; Younts &Mueller, 2001).

Berger, Zelditch, Anderson, and Cohen’s (1972) status value theory embedded interpersonal justice evaluations in larger social contexts by emphasizing the importance of referential standards in justice evaluations. A referential structure consists of an actor’s expectations about the associations of social characteristics and reward levels. An actor’s input and reward levels are viewed in terms of their status or prestige values rather than in the purely economic terms as in micro-exchange theories. Many social characteristics are associated with symbolic prestige or status value. This status value is linked with normative expectations for reward. People evaluate justice by cognitively assessing the rewards they receive with the amount an abstract generalized other person with equal status would receive in like circumstances.

Jasso (1980) and Markovsky (1985) combined elements of justice and status value theories to unpack some of the complexities of justice evaluations and reactions (Hegtvedt & Markovsky, 1995). Central to these theories are mathematical justice evaluation functions that calculate a logarithmic transformation of the ratio of an actor’s actual rewards to a reward standard. The transformation ensures that a person’s experience of injustice is a function of the degree to which a person cares about the evaluation. Barnum, Markovsky, and Richardson (2010) built on these ideas by incorporating the effects of social structure into the justice evaluations functions. According to this model, the perceptions of injustice and social structure work in tandem to create pressure for justice-restoring attempts including crime.

The LSI-R

The LSI-R is a survey-based instrument designed to assist correctional personnel in the assessment of an offender’s risks and needs. The LIS-R is administered to every incarcerated offender in Iowa upon admission to one of the state’s correctional facilities.¹ It is primarily used for programming and intervention. The instrument is a quantitative survey of the attitudes and attributes of offenders, including background information. Its purpose is to identify the major and minor risk factors that have been empirically shown to affect needs and risks. These factors represent reasonable targets for predicting future criminal activity and are useful in implementing intervention strategies that can be used by correctional staff for inmate treatment planning (Andrews & Bonta, 1995, 2006).

The content of the LSI-R is based primarily on the social learning perspective for criminal behavior. The principal claim of this perspective is that the same learning processes that produce conventional behavior also generate deviant behavior (Akers & Jensen, in press). The theory consists of four major dimensions: differential associations, definitions, differential reinforcement, and imitation (Akers, 1992; Akers & Sellers, 2004). Differential association refers to direct interaction with people who behave in a specific way or who express norms, attitudes, and values that is congruent with the behavior. Associations that occur in primary groups and those that happen early in life, or that occur often and frequently have the most impact on behavior (Akers, 1998; Burgess & Akers, 1966). Definitions refer to a person’s justifications and attitudes that define a specific behavior as right or wrong (Akers & Jensen, in press). Definitions are used to rationalize past behavior or anticipated actions. Differential reinforcement refers to anticipated or actual awards and punishments that result from a given act. The greater the value, frequency, and probability of reward for the act, the more likely it will be repeated (Akers & Jensen, 2003). Finally, imitation refers to behaving in a similar fashion as others. It generally occurs after direct observation of a similarly situated other (Akers & Jensen, in press).

The LSI-R survey is comprised of 54 items each answered as either a yes or a no, or given a score on a sliding scale ranging from 0 to 3 (Andrews & Bonta, 1995). The anchor score 3 indicates a satisfactory situation with no need for improvement, the opposite anchor, 0 is a score suggesting a very unsatisfactory situation that is clearly in need of improvement. The LSI-R is scored using a binary scoring format. Questions that are answered with a yes (or a 0 on the sliding scale questions) are coded as a 1. All other answers are scored as 0. The number of questions coded as 1 are simply added together to form LSI-R total score. Higher scores are indicative of more risk/need. The questionnaire items are grouped into 10 subcomponents which represent areas of potential risk.² Correctional staff generally administers the survey using a paper-and-pencil format from a standardized interview template. The staff also generally uses both the interview responses and the information from the inmate’s personnel file to complete the form.

The psychometric properties of the LSI-R are strong. Research demonstrates that the instrument has good interrater and test–retest reliability (Andrews, 1982; Bonta & Andrews, 1993), internal consistency and validity (Andrews, 1982) including face (Bonta, Motiuk, & Ker, 1985), convergent and divergent-construct validity (Andrews, Bonta, Motiuk, & Robinson, 1984; Bonta & Motiuk, 1990) as well as reasonable generalizability (Shields, 1991, 1993).

Hypotheses

In this section, we present a number of hypotheses that are derivable from the theories outlined above. Macro-structural formulations argue that concentrations of extreme of disadvantage form in some urban areas. Functionally, these pockets of privation create structural barriers that isolate residents from mainstream society and its corresponding connections to jobs and employment. As noted, African Americans are more likely to reside in disadvantaged areas than are members of other groups so it follows then, that when all else is equal, African Americans should experience more difficulty obtaining employment than other people. Consequently:

Hypothesis 1.1: Net of other factors, African Americans incarcerated in Iowa Prisons are more likely to report a preincarceration history of unemployment than are White inmates.

Hypothesis 1.2: Net of other factors, African Americans incarcerated in Iowa prisons are more likely to report that they have relied on social assistance during their lives than are White inmates.

Since African Americans are more likely to live in areas characterized by extreme poverty and disadvantage they face macro-level structural barriers that are unique to their experience. These challenges result in significant differences between Whites and Blacks in terms of the hardships they can expect to face. In fact, Sampson and Wilson (1995) claim that even the most disadvantaged White neighborhoods have less deprivation than the typical disadvantaged African American community. And as noted, GST argues that these types of hardships should be thought of as strains that generate powerful negative emotions like anger and feelings of injustice. Moreover, the legacy of racism in America increases the potency of these strains because it increases the odds that African Americans will perceive these barriers as undeserved and illegitimate. Consequently, we believe African Americans will be more sensitive to feelings of injustice stemming from their life chances than will Whites. Therefore:

Hypothesis 2.1: Net of other factors, African Americans incarcerated in Iowa prisons are more likely to feel that their sentences are unfair than are White Iowa inmates.

Hypothesis 2.2: Net of other factors, African Americans in Iowa prisons are more likely to behave in ways that are indicative of people that feel their sentence is unfair than are White Iowa inmates.

As noted, GST and distributive justice theories each suggest that strain which is perceived as unjust generates high levels of anger and intense pressure for corrective action (sometime called justice-restoring attempts). Both these traditions also argue that when justice restoring attempts are predicated on anger they often take the form of crime. As outlined above, Blacks are more likely to perceive the barriers they face as illegitimate than are members of other groups. Consequently, African Americans should be more likely to experience a strong sense of injustice and anger as a result of macro-structural roadblocks than will others. And so, Blacks should be more likely to engage in criminal justice-restoring attempts than members of other groups. It follows then:

Hypothesis 3.1: Net of other factors, African Americans in Iowa prisons are more likely to have a history of violent criminal activity than are White Iowa prisoners.

Hypothesis 3.2: Net of other factors, African American inmates are more likely to have an extensive criminal history than are White Iowa inmates.

GST also posits that chronic strain generates reduced social control. This takes the form of weak emotional bonds with conventional others and institutions. The absence of these relationships increases the likelihood for crime because under such circumstances people have less of a stake in conformity. In particular, such people tend not to care about the opinions and beliefs of family and friends. They also have less to lose in terms of jobs and relationships if they violate the law. As we have seen the macro-social approach suggests these types of chronic strain are more likely for Blacks than Whites. If these arguments are true, then we would expect the following:

Hypothesis 4.1: Net of other factors, African American offenders’ ties to conventional others will be weaker than White Iowa inmates.

Hypotheses 4.2: Net of other factors, African American offenders’ ties to conventional institutions will be weaker than White Iowa inmates.

Finally, GST argues that reduced social control contributes to the social learning of crime. This occurs because people with weak ties to conventional others are more likely to come into contact with criminals and people who hold tolerant definitions of crime. Additionally, the macro-structural approach argues that African Americans are more likely to live in residential pockets distinguished by high levels of social disorganization and crime. This increases the odds that African Americans will come into contact with others who engage in criminal activity. Accordingly:

Hypothesis 5.1: Net of other factors, African Americans in Iowa prisons will have had more preincarceration close contact with others who hold tolerant definitions of crime than will White Iowa inmates.

As will be explained in detail below, we use specific questionnaire items from the LSI-R to test these hypotheses. The tests of the hypotheses are predicated on two fundamental assumptions found in Sampson and Wilson’s (1995) theory. First, that macro-structural barriers generate clusters of truly disadvantaged. These clusters are often centered in neighborhoods comprised mainly of African American residents. And second, that as a consequence even the most disadvantaged White neighborhoods have less deprivation than the typical underprivileged African American community. It is important to note that the effects of deprivation impacts all the residents in these areas—criminals and noncriminals alike. Therefore, to the degree that these assumptions are true, we reason that: (i) some criminal behavior is a consequence of macro-structural barriers, and (ii) in general African Americans are more likely to encounter these barriers than are Whites, (iii) therefore, we expect the differences between African Americans and White inmates in our sample to be at least partially explained by these macro-structural forces and their correlates.³ This means, of course, that these explanations are useful in understanding differences between inmate groups in our sample, even though the sample itself is limited in that it consists solely of imprisoned felons.

Data and Methodology

Descriptive Information for Participants

The data for this study were provided by the Iowa Department of Corrections. These data include information about all individuals who were incarcerated in Iowa’s nine correctional facilities in 2005. At that time, over 90% of inmates incarcerated were male, roughly 75% were White and approximately 24% were minority members. The overwhelming total of minority members was African American (22%). Native Americans accounted for roughly 1.6% of the population and Asian and Pacific Islanders for the remaining 0.9%. Table 1 provides demographic and summary information for the incarcerated offenders.

Table 1.

Demographic and Summary Information for all Iowa Prison Inmates in 2005

Race	Male	Female	Total	%
White	5,897	559	6,456	75
Black	1,735	171	1,906	22
Native	116	22	138	1.6
Asian	77	5	82	.9
Total	7,825	757	8,582	100
Percentage	91	9	100

Most of the male offenders incarcerated in Iowa committed violent offenses—this is true for all races. Table 2 shows a breakdown of the percentage of male offenders in the Iowa system for White and Black offenders by type of crime, and also the Black-to-White odds ratios for each offense type. The results show that African Americans were more likely to be incarcerated for violent crimes and property crimes than were Whites, but Whites were more likely to be in prison for drug offenses and other miscellaneous crimes (these types of crimes include offenses like drunk driving, animal cruelty, etc.). In the Iowa prison population, being an African American increased the odds of being incarcerated for a violent offense by 25%, and increased the odds by 14% of being incarcerated for a property crime. In this same population, being White increased the odds of being imprisoned for a drug offense by 37% and increased the odds by 22% of being incarcerated for other types of crimes.

Table 2.

The Percentage and Number of Male Inmates Incarcerated in Iowa by Type of Crime Committed as well as the Odds Ratios Comparing Black to White Male Inmates by Type of Crime

	Percentage and number by crime				Black–White odds ratio
Race	Violent	Drug	Property	Other	Violent	Drug	Property	Other
White	42.7% (2,519)	27.13% (1,600)	18.6% (1,097)	11.5% (681)	0.79	1.37	0.87	1.22
Black	48.3% (837)	21.3% (370)	20.1% (358)	9.6% (167)	1.25	0.72	1.14	0.81

Note. The values in parentheses indicate the number of inmates incarcerated.

Measures for LSI-R Summaries

The total LSI-R score values in Table 3 are on average higher for African Americans than White inmates. The mean African American LSI-R total equals 32.96, while Whites’ average LSI-R total is 31.07. An independent sample t test shows this difference is significant (t = 8.03, df = 5,945, p < .001). t Tests also show significant differences in eight subcategories. However, caution should be used in interpreting these results because the sample size is quite large. Consequently, these results may be due more to the precision of the parameter estimates resulting from statistical power than from substantive differences in the population. Cohen’s d is a standardized effect size measure of the difference between two means. It can be used to evaluate whether observed differences are substantively meaningful or important. Table 3 reports Cohen’s d values for the significant variables. The effect sizes for all but four variables are quite small and likely inconsequential (d < .2). However, the effect sizes for the total LSI-R score (d = .25), criminal history (d = .33), criminal friends (d = .34), and personality and emotional problems (d = .25) are moderate in size. These larger effect sizes suggest the results are meaningful and also imply that African Americans’ higher mean total LSI-R score stems in large part from the criminal history and criminal friends subcategories. This in turn indicates that on average when compared to White prisoners, African Americans incarcerated in Iowa prisons tend to have more extensive criminal histories, and are more likely to have preincarceration friends and associates who were involved in criminal activity. The information in Table 3 also reports that White offenders in the Iowa system are more likely to score meaningfully higher than Blacks on the personality and emotional problems subcategory. This suggests that on average White offenders imprisoned in Iowa are more likely than Black offenders to have mental health issues.

Table 3.

Mean LSI-R Subcategory and Totals for White and Black Males and Cohen’s d Statistic for Effect Sizes

Race	LSI-R subcategory
	Hist.^*	Emp.^*	Fin.	Fam^*	Res.	Rec.^*	Friend.^*	Drug^*	Per.^*	Att.^*	Total^*
White	6.31	5.94	1.12	4.39	1.92	1.77	2.96	6.75	2.32	3.23	31.07
Black	7.00	6.27	1.08	4.19	1.98	1.66	3.31	6.50	1.95	3.06	32.96
Cohen’s d	.33	.14	—	−.12	—	−.17	.34	−.19	−.25	−.13	.25

^* p < .001 for independent samples t tests between Blacks and Whites.

Note. The abbreviations used in the table refer to LSI-R subcategories. These are Criminal history, education/employment, finances, family/marital conditions, accommodation, leisure and recreation, companions, alcohol/drug problems, emotional/personal issues, and attitude/orientation.

Overall, these findings support several hypotheses. Results from Cohen’s d statistic are consistent with Hypotheses 3.1 and 3.2 (when compared to Whites, African Americans have more extensive criminal histories) and Hypothesis 5.1 (when compared to Whites, African Americans have more antisocial companions). However, using the LSI-R subcategory totals to evaluate the hypotheses is somewhat problematic. The instrument was designed to assess prisoner needs/risks for classification. It was not created to test our hypotheses. Consequently, LSI-R subcategories include questionnaire items that are not directly relevant for tests of the hypothesis. For example, the LSI-R criminal history subcategory in addition to asking about a client’s preincarceration criminal history includes questions about the client’s punishment history for institutional misconduct and escape history. These concepts are not relevant for testing Hypotheses 3.1 and 3.2. To address this problem, we use individual questionnaire items found in LSI-R subcategories to test hypotheses.

Measures for Tests of Hypotheses

As noted, the LSI-R consists of 54 questionnaire items each answered as either a yes or a no, or given a score on a sliding scale ranging from 0 to 3 (Andrews & Bonta, 1995). The no answers and the anchor score 3 each indicates a satisfactory situation with no need for improvement. The yes scores and the opposite anchor, 0 are scores suggesting a very unsatisfactory situation that is clearly in need of improvement. Questions that are answered with a yes, or a 0 on the sliding scale, are coded as a 1. All other answers are scored as 0. Nine binary response questions were used to test the hypotheses. These questions were chosen because of all the LSI-R questionnaire items, they most closely measure the concepts expressed in the corresponding hypotheses. A complete list of these hypotheses is found in Appendix A.

Results for Logistic Regression

Logistic regression is used for tests of each of the nine hypotheses. Race is the principal independent variable used in each model. The other variables are included for control. As noted in previous sections, prior research resulting from GST and macro-structural explanations strongly suggests that offending rates vary by age, gender, level of education, and the seriousness of the offense. Moreover, findings from macro-structural theories make clear offending rates also vary by place of residence (urban vs. nonurban). Accordingly, each model includes these control variables in the analysis.⁴ Iowa counties with cities having populations of at least 38,000 in 2005 were coded as urban areas (19 of Iowa’s 99 counties were urban).⁵ For education, the analysis compared inmates with a high school diploma, general equivalency diploma or better to those with less education. All forcible felonies were considered serious offenses. Other crimes were coded as nonserious offenses. The Table 4 gives the odds ratio coefficients resulting from logistic regression analyses.

Table 4.

The Odds Ratio Coefficients for Logistic Regression Analyses of the Hypotheses

Measures	The odds ratio coefficients for the hypotheses
	Hypothesis 1.1	Hypothesis 1.2	Hypothesis 2.1	Hypothesis 2.2	Hypothesis 3.1	Hypothesis 3.2	Hypothesis 4.1	Hypothesis 4.2	Hypothesis 5.1
Race	.60^*	.72^*	.68^*	.75^*	.32^*	.731^*	1.01	.99	.58^*
Gender	1.27^*	.74^*	1.54^*	1.63^*	3.13^*	1.50^*	.60^*	.61^*	1.34^*
Age	.95^*	1.01^*	1.01^*	.98^*	.99	1.05^*	1.03^*	1.03^*	.99
Seriousness	.51^*	.75^*	1.62^*	.79^*	5.87^*	.303^*	1.35^*	1.32^*	1.20^*
Urban	1.1^*	.93	.98	1.11	1.01	1.18	.96	.95	1.18^*
Education	.57^*	.86	.86^*	.77^*	.74^*	.65^*	1.11	1.10	.63^*
Log likelihood	−4100	−3983	−4486	−4528	−2855	−3949	−.4365	−.4436	−4527
Pseudo R ²	.083	.001	019	.017	.112	.083	.028	.030	.023

Note. Race, gender, seriousness, urban, and education are dummy variables. White inmates, male inmates, inmates incarcerated for forcible felonies, inmates arrested in urban areas, and inmates who have at least a high school education (including general equivalency diploma) are coded as 1 in the analyses.

^*p < .01.

Results from logistic regression support seven of the nine hypotheses. Macro-structural concepts form the core sources of Hypotheses 1.1 and 1.2. Logistic regression odds ratios show that net of control variables, the odds are 1.66 times greater that African American prisoners in the Iowa system have been chronically unemployed and 1.38 times greater that they have received public assistance than White inmates. Hypotheses 2.1 and 2.2 are derived from GST and macro-structural explanations. Results show that when compared to White prisoners, African American inmates are 1.47 times more likely to deny the fairness of their sentences and 1.33 times more likely to object to their prison classifications. Hypotheses 3.1 and 3.2 stem from GST, justice theories, and macro-social explanations. Results here indicate that when compared to White inmates, the odds are 3.13 times greater that Black prisoners will have an official record of assault and violence and 1.36 times greater that they will have a record of three or more prior convictions. Finally, Hypothesis 5.1 originates from concepts found in GST and macro-social structural explanations. Tests of this hypothesis reveal that in comparison to White prisoners, the odds are 1.72 times greater that African American inmates in Iowa prisons will have only a few preincarceration friends not involved in criminal activity. However, the results for 4.1 and 4.2 are not significant. This suggests that when looking at preincarceration work and school histories, African American prisoners are no more likely to have disliked work or school or the officials associated with these institutions than are White inmates.

Discussion

The results from the analyses suggest that African American and White prisoners in Iowa differed in meaningful ways along several important outcome measures of the LSI-R. First, on average Black prisoners’ total LSI-R scores were higher than White prisoners’ total scores. Recalling that higher scores are indicative of more need/risk, this finding suggests that all else being equal, Blacks in the Iowa system may be more likely to be placed in higher security classifications and undergo more intense supervision than White inmates. Second, the effect sizes from these tests indicate that much of the disparity in the total scores stems from the history, criminal friends, and the personality and emotional problems subcategories of the inventory. These findings indicate that a significant portion of the variance in the total LSI-R scores is a product of preincarceration differences between the racial groups in terms of criminal behavior and friendships. We believe these differences may be attributable to dissimilarities between the groups in terms of macro-structural factors resulting from childhood neighborhood conditions. These structural barriers may generate strain and frustration and this may ultimately lead to crime and violence. Third, the tests of hypotheses using logistic regression and individual LSI-R questionnaire items as dependent variables are consistent with most of the claims. All the hypotheses except 4.1 and 4.2 are supported. These results provide evidence that prior to incarceration and in comparison to White inmates, African American prisoners had more difficulty finding work, were more likely to have an official record of violent crime, and were more likely to associate with people who were involved in crime. Additionally, the results suggest that in comparison to White inmates, African American prisoners were more likely to feel that their prison sentences were unfair and to act in ways that were indicative of this. For each of the tests, the odds ratios from logistic regressions were modest to moderate in size. The effects were strongest for hypotheses regarding preincarceration history of violent crime and criminal friends and weakest for hypotheses about pre imprisonment work history. All told, the findings suggest that a prior history of violence is the most important predictor of differences between White and Black inmates in our analysis.

Finally, the results provide little evidence that African American inmates’ ties to conventional institutions (or others) are weaker than are White prisoners’ ties. This finding disconfirms Hypotheses 4.1 and 4.2. Future work should be conducted using a wider range of outcome measures to check if this outcome is genuine or due to the choice of questionnaire items that were selected to test the hypotheses. When considered as a whole, the results from the analyses fit theoretical ideas stemming from social disorganization, strain, and differential association theories better than they do explanations grounded in social control theory.

Limitations

We acknowledge certain limitations. First, the values for pseudo R ² in each of the logistic regression analyses were small, indicating that the independent variables used in the models did not explain much of the variance in the data. We ran several diagnostic models to investigate the small R ². The diagnostic models built-in many more independent variables (including information about birth state, housing conditions, leisure activities, current family status, financial status, drug and alcohol problems, and attitudes) and although the pseudo R ² increased somewhat (R ² ≈ .20), the odds ratio effect sizes for race remained essentially unchanged. Consequently, although some of the unexplained variance was eliminated by these much larger models (some included over 50 variables), the substantive results remained essentially the same. It seems likely that the uniform nature of the data explains the low R ² values. The data set consisted entirely of offenders, so it was not possible to compare LSI-R scores for offenders to those for noncriminals. It seems probable that this restricted the variance in the independent variables and limited R ² values.

Second, the variance in the odds ratio effect sizes from logistic regression may in part be an artifact of the LSI-R itself. Metatheoretically, the development of this instrument relied heavily on ideas and concepts from the social learning perspective of criminal behavior. Accordingly, many LSI-R questionnaire items were intended to measure social learning correlates, including differential association, differential reinforcement, and imitation. These correlates are similar to concepts found in our theoretical hypotheses. These include concepts regarding previous criminal history (which is similar to differential reinforcement in social learning theory) and concepts regarding criminal friendships (which is similar to differential association in social learning theory). Simply put, the LSI-R may be more sensitive at detecting social learning outcomes than other effects and this may in part account for the variance in effect sizes found in the analyses.

Finally, to a certain degree our data were incomplete. Although the data contained information about where each offender was arrested, the data did not include information regarding the inmate’s place of residence or information about where the inmate was reared. Consequently, it was not possible to directly test whether African American prisoners truly were raised in neighborhoods affected by higher levels of deprivation than were White inmates (an essential assumption underlying our hypotheses). Future work should directly measure and include this information, so this assumption can be explicitly tested.

Conclusions

The study tested several theoretical hypotheses using prisoners’ scores on The LSI-R, which is an instrument used by the Iowa Department of Corrections for offender programming and classification. The hypotheses are based on ideas found in theories of social structural power and fairness, GST, and macro social-structural theories of crime. The findings support seven of the nine hypotheses. Specifically, the results suggest that on average, African American prisoners have higher total LSI-R scores than White inmates and that net of certain important control variables, African American prisoners had more difficulty finding work, were more likely to have an official record of violent crime, and were more likely to associate with people who were involved in crime than were White inmates. Additionally, the results suggest that in comparison to White inmates, African American prisoners were more likely to feel that their prison sentences were unfair and to act in ways that were indicative of this. These findings are consistent with explanations found in macro-structural theories of crime as well as concepts found in GST and distributive justice theory.

The results from this study suggest at least two key areas for future research. Each centers on inmate misconduct. First, as noted, African American inmates in the Iowa system tend to be placed in higher security classifications and undergo more intense supervision than White inmates. This is because Blacks’ LSI-R total scores are on average higher than Whites’ scores. Future research should investigate whether the differences in security classifications and supervision interact with race to affect institutional misconduct. Results from this line of research may have implications for understanding the dynamics of race and prisonization and how these processes affect institutional misconduct. Second, future work should more conclusively investigate whether differences between White and Black inmates stem in part from the extreme deprivation found in some predominately African American neighborhoods. Specifically, whether variables such as childhood family structure and preincarceration neighborhood conditions result in behavioral styles that are imported into prison. We believe a potentially fruitful avenue of exploration would be a study of the association between race and neighborhood violence. The goal of this research would be to understand the degree to which neighborhood conditions create an environment where violence is learned by precept and imitation in order to recognize the extent to which prisoners bring these behavioral styles into prison with them when incarcerated. We believe the present study may serve as an important first step along the path to increased understanding of complex institutional phenomena like these.

Footnotes

Appendix A

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47–87.

Agnew

(2006). Pressured into crime: An overview of general strain theory. Los Angeles, CA: Roxbury.

Agnew

White

(1992). An empirical test of general strain theory. Criminology, 30, 475–499.

Akers

(1992). Drugs, alcohol, and society: Social structure, process and policy. Belmont, CA: Wadsworth.

Akers

(1998). Social learning and social structure: A general theory of crime and deviance. Boston, MA: Northeastern University Press.

Akers

Jensen

.(Eds.). (2003). Advances in criminological theory (Vol. 11). New Brunswick, NJ: Transaction.

Akers

Jensen

. (2006). The empirical status of social learning theory of crime and deviance. In Cullen

Wright

Blevins

(Eds.), Taking Stock: The status of criminological theory. Advances in Criminological Theory (Vol. 15, pp. 37–76).

Akers

Sellers

(2004). Criminological theories: Introduction, evaluation, and application (4th ed.). Los Angeles, CA: Roxbury.

Andrews

(1982). The level of service inventory (LSI): The first follow up. Toronto: Ontario Ministry of Correctional Services.

10.

Andrews

Bonta

(1995). The level of service inventory–Revised. Toronto, ON: Multi-Health Systems.

11.

Andrews

Bonta

(2006). The psychology of criminal conduct (4th ed.). Newark, NJ: LexisNexis.

12.

Andrews

Bonta

Motiuk

Robinson

(1984). Some psychometrics of practical risk/needs assessment in corrections. Paper presented at the 92nd Annual Convention of the American Psychological Association, Toronto, ON.

13.

Aseltine

Gore

Gordon

(2000). Life stress, anger and anxiety and delinquency: An empirical test of general strain theory. Journal of Health and Social Behavior, 41, 256–275.

14.

Asseline

. (2009). Using general strain theory to understand drug and alcohol use in Canada: An examination of how strain, its conditioning variables and gender are interrelated. Retrieved October 22, 2011, from http://qspace.library.queensu.ca/bitstream/1974/5131/1/Asselin_Nicolas_R%20P_2009%20August_Master%20of%20Arts.pdf

15.

Barnum

Markovsky

Richardson

(2010). Structural power, fairness perceptions, and deviant acts A paper presented at the Annual Meetings of the American Sociological Association. Atlanta, GA.

16.

Baron

(2004). General strain, street youth and crime: A test of Agnew’s revised theory. Criminology, 42, 457–483.

17.

Baron

(2007). Street youth, gender, financial strain and crime: Exploring Broidy and Agnew’s extension to general strain theory. Deviant Behavior, 28, 278–302.

18.

Baron

(2008). Street youth, unemployment, and crime: Is it that simple? Using general strain theory to untangle the relationship. Journal of Crime and Criminal Justice, 50, 399–434.

19.

Berger

Zelditch

Jr Anderson

Cohen

B. P.

(1972). Structural aspects of distributive justice: A status-value formulation. In Berger

Zelditch

Anderson

(Eds.), Sociological theories in progress (Vol. 2, pp. 119–246). Boston, MA: Houghton Mifflin.

20.

Blau

(1982). The cost of inequality: Metropolitan structure and violent crime. American Sociological Review, 47, 114–129.

21.

Blumstein

(1982). On the racial disproportionality of United States' prison populations. Journal of Criminal Law and Criminology, 73, 1259–1281.

22.

Bonta

Andrews

(1993). The level of service inventory: An overview. IARCA Journal, 5, 6–8.

23.

Bonta

Motiuk

(1990). Classification to correctional halfway houses: A quasi-experimental evaluation. Criminology, 28, 497–506.

24.

Bonta

Motiuk

Ker

(1985). The Level of Supervision Inventory (LSI) among incarcerated offenders. Report #1. Toronto: Ministry of Correctional Services, Ontario.

25.

Broidy

Agnew

(1997). Gender and crime: A general strain theory perspective. Journal of Research in Crime and Delinquency, 34, 275–306.

26.

Burgess

Akers

(1966). A differential association-reinforcement theory of criminal behavior. Social Problems, 14, 128–147.

27.

Byrne

Sampson

(1986). Key issues in the social ecology of crime. In Byrne

Sampson

(Eds.), The social ecology of crime (pp. 25–46). New York, NY: Springer-Verlag.

28.

Cloward

Ohlin

(1960). Delinquency and opportunity: A theory of delinquent gangs. Glencoe, IL: Free Press.

29.

Cohen

(1955). Delinquent boys: The culture of the gang. Glencoe, IL: The Free Press.

30.

Federal Bureau of Investigation. (2008). Crime in the United States (2008). Washington, DC: U.S. Government Printing Office.

31.

Froggio

Agnew

(2006). The relationship between crime and “objective” versus “subjective” strains. Journal of Criminal Justice, 35, 81–87.

32.

Hegtvedt

(2005). Distributive justice. In Ritzer

(Ed.), Encyclopedia of social theory (pp.206–210). Thousand Oaks, CA: Sage.

33.

Hegtvedt

Markovsky

(1995). Justice and injustice. In Cook

K. S.

Fine

G. A.

House

(Eds.), Sociological perspectives on social psychology (pp. 257–280). Boston, MA: Allyn & Bacon.

34.

Hirschi

(1969). The causes of delinquency. Berkley: University of California.

35.

Homans

(1961). Social behavior. New York, NY: Harcourt, Brace & World.

36.

Jasso

(1980). A new theory of distributive justice. American Sociological Review, 45, 3–32.

37.

Krivo

Peterson

(1996). Extremely disadvantaged neighborhoods and urban crime. Social Forces, 75, 619–650.

38.

Krivo

Peterson

(2000). The structural context of homicide: Accounting for racial differences in process. American Sociological Review, 65, 547–559.

39.

Krivo

Peterson

(2005). Macro-structural analysis of race and violent crime: Recent lessons and new directions from research. Annual Review of Sociology, 33, 331–356.

40.

Krivo

Peterson

Ruth

(2009). Segregation, racial structure and neighborhood violent crime. American Journal of Sociology, 114, 1765–1802.

41.

LaGrange

Silverman

(1999). Low self-control and opportunity: Testing the general theory of crime as an explanation for gender differences in delinquency. Criminology, 37, 41–71.

42.

Markovsky

(1985). Toward a multilevel distributive justice theory. American Sociological Review, 50, 822–839.

43.

Markovsky

(1988). Injustice and arousal. Social Justice Research, 2, 223–233.

44.

Massey

Denton

(1993). American apartheid: Segregation and the making of the underclass. Cambridge, MA: Harvard University Press.

45.

Mazerolle

Maahs

(2000). General strain and delinquency: An alternative examination of conditioning influences. Justice Quarterly, 17, 753–778.

46.

Mazerolle

Piquero

(1998). Linking exposure to strain and anger: An investigation of deviant adaptations. Journal of Criminal Justice, 26, 195–211.

47.

McNulty

(2001). Assessing the race-violence relationship at the macro level: The assumption of racial invariance and the problem of restricted distributions. Criminology, 39, 467–490.

48.

Merton

(1938). Social structure and anomie. American Sociological Review, 3, 672–782.

49.

Ousey

(1999). Homicide, structural factors, and the racial invariance assumption. Criminology, 37, 405–425.

50.

Park

McKenzie

Burgess

(1925). The city: Suggestions for the study of human nature in the urban environment. Chicago, IL: University of Chicago Press.

51.

Parker

Johns

(2002). Urban disadvantage and types of race-specific homicide: Assessing the diversity in family structures in the urban context. Journal of Research in Crime and Delinquency, 39, 277–303.

52.

Parker

Pruitt

(2000). Poverty, poverty concentration, and homicide. Social Science Quarterly, 81, 555–570.

53.

Peterson

Krivo

(1993). Racial segregation and black urban homicide. Social Forces, 71, 1001–1026.

54.

Peterson

Krivo

(2005). Macrostructural analyses of race, ethnicity, and violent crime: Recent lessons and new directions for research. Annual Review of Sociology, 31, 331–356.

55.

Piquero

Sealock

(2000). Generalizing general strain theory: An examination of an offending population. Justice Quarterly, 17, 449–484.

56.

Sampson

(1987). Urban black violence: The effect of male joblessness and family disruption. American Journal of Sociology, 93, 348–382.

57.

Sampson

Bean

(2006). Cultural mechanisms and killing fields: A revised theory of community-level racial inequality. In Peterson

Krivo

Hagan

(Eds.), The many colors of crime: Inequalities of race, ethnicity and crime in America (pp. 8–36). New York: New York University Press.

58.

Sampson

Wilson

(1995). Toward a theory of race, crime, and urban inequality. In Hagan

Peterson

(Eds.), Crime and inequality (pp. 37–54). Stanford, CA: Stanford University Press.

59.

Saperstein

Penner

(2010). The race of a criminal record: How incarceration colors racial perceptions. Social Problems, 57, 92–113.

60.

Shaw

McKay

(1942). Juvenile delinquency and urban areas. Chicago, IL: University of Chicago Press.

61.

Shields

(1993). The use of young offender-level of service inventory with adolescents. IARCA Journal, 5, 10–26.

62.

Shields

(1991). The young offender-level of service inventory as a predictor of recidivism. A paper presented at the annual meeting of the Canadian Psychological Association, Calgary Alberta.

63.

Shihadeh

Shrum

(2004). Serious crime in urban neighborhoods: Is there a race effect? Social Spectrum, 24, 507–533.

64.

Sutherland

Cressey

(1978). Criminology (10th ed.). Philadelphia, PA: J.B. Lippincott.

65.

The Sentencing Project: Research and Advocacy for Reform. (n.d.). Retrieved October 22, 2011, from http://www.sentencingproject.org/map/map.cfm

66.

Turner

Stets

(2005). The sociology of emotions. Cambridge, MA: Cambridge University Press.

67.

United States Bureau of Statistics. (n. d.). Retrieved October 22, 2011, from http://www.ojp.usdoj.gov

68.

Wilson

(1987). The truly disadvantaged. Chicago, IL: University of Chicago.

69.

Wolfgang

Ferracuti

(1967). The subculture of violence towards an integrated theory in criminology. London, England: Tavistock.

70.

Younts

Mueller

(2001). Justice processes: Specifying the mediating role of perceptions of distributive justice. American Sociological Review, 66, 125–145.

Racial Disparity in Iowa Prisons: Possible Theoretical Explanations for Racial Differences Between Inmates on the Level of Service Inventory–Revised

Abstract

Keywords

Introduction

Literature Review

Macro-Structural Explanations

GST

Structural Power and Distributive Justice

The LSI-R

Hypotheses

Data and Methodology

Descriptive Information for Participants

Measures for LSI-R Summaries

Measures for Tests of Hypotheses

Results for Logistic Regression

Discussion

Limitations

Conclusions

Footnotes

Appendix A

Notes

References