The Decision to Prosecute Drug Offenses and Homicides in Marion County, Indiana


Chapter 3

Data Analysis


Under Indiana Code felony charges can be brought against an individual for numerous violations, including murder, robbery, attack, vehicle manslaughter, distribution of a controlled substance, burglary, carjacking, battery, reckless homicide, involuntary manslaughter, theft, possession of a controlled substance, firearm violation, criminal recklessness, vehicle theft, resisting arrest, fleeing a police officer, and child molestation. Depending on severity, the charge is classified A, B, C, or D.[1]

Although the original study design of this report included homicides, because prosecution numbers and arrest numbers are highly coincident and some of the information pertinent to the prosecution decision process was unavailable, an analysis of homicide prosecution decisions would have rendered invalid and unreliable results. However, arrest and prosecution data on homicides are included in this chapter.

An analysis was conducted on the decision to prosecute in Marion County for individuals arrested in 1993 and 1997 for two types of charges: possession of narcotics or a controlled substance and dealing narcotics or a controlled substance. Drug offenses are the most numerous violations for arrest, prosecution, and incarceration in Marion County. For both types of charges, analyses were conducted to compare arrest rates with population rates by race, differences in arrests for different violations by race, and evaluation of race as a factor in the decision to prosecute.

In the criminal justice system, the prosecutor has broad discretion as to whom to prosecute. This decision can be influenced by factors such as “the strength of the case, the prosecution’s general deterrence value, the Government’s enforcement priorities, and the case’s relationship to the Government’s overall enforcement plan.”[2] A prosecutor’s decision to prosecute a specific defendant cannot be based upon arbitrary classifications, such as race or religion. The equal protection clause only prohibits disparate treatment, not disparate impact.[3] Direct or circumstantial evidence can constitute proof of intent of disparate treatment.[4] 

The courts should consider seven factors as circumstantial evidence of intentional discrimination.[5] The first factor involves disparities in the administration of the law. Evidence alone will not satisfy the equal protection claim’s discriminatory intent, if the pattern of discriminatory administration is not severe.[6] Other evidence could consist of (1) objectionable purposes underlying the official action; (2) sequence of events leading up to the decision; (3) deviations from ordinary procedures; (4) substantive departure from normal considerations in the decision making; (5) legislative or administrative history; and, (6) direct testimony by the decision maker.[7] 

In Wayte v. United States[8] the Supreme Court articulated the test for the selective prosecution defense. In order to prove a selective prosecution claim, the defendant must show that the government’s enforcement of a facially neutral law had a discriminatory effect, and the enforcement was motivated by a discriminatory purpose.[9] To prove the first prong, the defendant must prove that others similarly situated have not been prosecuted.[10] To establish the second part of the test, the defendant must prove selection based on an impermissible reason, such as race, or other impermissible classification.[11]

In United States v. Armstrong,[12] the Supreme Court held that the defendant’s study was insufficient to meet the threshold standard entitling the defendants to discovery on a claim of discriminatory prosecution.[13] The study showed that all crack cocaine cases brought by the federal prosecutor in the Central District of California during a specific time period were against African American defendants. The Court held that the defendants failed to demonstrate that similarly situated individuals of a different race were not prosecuted.[14] The defendant’s study did not show the existence of the essential elements of a selective prosecution claim, because it failed to identify individuals who were not African American, and could have been prosecuted for the offenses for which the respondents’ were charged, but were not so prosecuted.[15] Armstrong sets a standard that applies to state and local prosecutors, as well as federal prosecutors.[16]

Analysis of incarceration records for drug offenses shows that African Americans are prosecuted and incarcerated at a rate significantly higher than their percentage of the population. African Americans are 21 percent of the Marion County population, yet they are the majority of those incarcerated for drug offenses. Of those incarcerated for drug possession, 32.7 percent are white, while 62.6 percent are African American.[17] Similarly, 57 percent of inmates convicted of dealing drugs are African American, while 37 percent are white.[18]

DEMOGRAPHIC DATA

Demographic information was obtained from the 1990 census. The racial and ethnic characteristics of Marion County are set out in table 3.1. The county is essentially divided between two racial groups, whites and African Americans.[19]

Population

Whites are 76.7 percent of the total population; African Americans are 21.2 percent. Latinos and Asian Americans are the other 3.1 percent of the county’s population. Because the populations of Latinos and Asian Americans are so small relative to whites and African Americans, they were excluded as separate racial/ethnic groups in this study.


TABLE 3.1
Population of Whites and Blacks in Marion County

Population

Whites 76.7%
Blacks 21.2%
Other 3.1%

Source: 1990 U.S. census data.


Income 

Differences between African American households and white households are particularly notable in the lowest income categories. Nearly 60 percent of all African American households in Marion County have an income lower than $25,000 (in 1990 dollars), while 38 percent of white households in the county have an income level below $25,000.


TABLE 3.2
Household Income of Whites and Blacks in Marion County

Income below $25,000

Income below $10,000

Whites 38.3% 11.0%
Blacks 59.6% 27.2%

Source: 1990 U.S. census data (adjusted for inflation).


DATA ANALYSIS OF ARRESTS AND PROSECUTION: DRUG OFFENSES

Data on Drug Offenses

Data on arrests for dealing and for possession of narcotics and controlled substances were obtained from the Marion County Criminal Justice Information Agency for the years 1993 and 1997. The agency also provided prosecution decision data on all individuals prosecuted for drug offenses.

Arrest records were obtained for all drug-related offenses in 1993 and 1997 recorded by the Indianapolis Police Department and the Marion County Criminal Justice Information Agency. These data included such information as the name and address of the arrestee; location of the arrest; arrest charge; case identification number; and the race, age, gender, and education level of the arrestee.


TABLE 3.3
Arrests for Possession of Narcotics by Race, 1993 and 1997, Marion County

Number

Percent

Whites 2,388 32.8%
Blacks 4,849 66.6%
Other 45 0.6%

Source: Marion County Criminal Justice Information Agency.


Arrests for Possession of Narcotics

Arrest records for possession of narcotics by race and year are shown in table 3.3. In 1993 and 1997, 7,282 people were arrested in Marion County for narcotics possession. Of those, 4,849, or 66.6 percent, were African American. There were 2,388 whites arrested for possession of narcotics, 32.8 percent of such arrests.

The analysis divided possession of narcotics into three categories: (1) possession of cocaine, (2) possession of a controlled substance, and (3) possession of marijuana. The arrest numbers and percentages by race for the categories are shown in tables 3.4, 3.5, and 3.6, respectively. 

Arrests for Possession of Cocaine

In 1993 and 1997, 2,617 people were arrested in Marion County for cocaine possession. Of those, 2,287, or 87.4 percent, were African American. There were 330 whites arrested for possession of cocaine, 12.6 percent of such arrests.


TABLE 3.4
Arrests for Possession of Cocaine by Race, 1993 and 1997, Marion County

Number

Percent

Whites 330 12.6%
Blacks 2,287 87.4%
Other 0 0%

Source: Marion County Criminal Justice Information Agency.


Arrests for Possession of a Controlled Substance

In 1993 and 1997, 368 people were arrested in Marion County for possession of a controlled substance. Of those, 132, or 35.8 percent, were African American. There were 236 whites arrested for possession of a controlled substance, 63.2 percent of such arrests.


TABLE 3.5
Arrests for Possession of a Controlled Substance by Race, 1993 and 1997, Marion County

Number

Percent

Whites 226 63.2%
Blacks 132 35.8%
Other 0 0%

Source: Marion County Criminal Justice Information Agency.


Arrests for Possession of Marijuana

In 1993 and 1997, 4,297 people were arrested in Marion County for possession of marijuana. Of those, 2,430, or 56.6 percent, were African American. There were 1,822 whites arrested for possession of marijuana, 42.4 percent of such arrests.


TABLE 3.6
Arrests for Possession of Marijuana by Race, 1993 and 1997, Marion County

Number

Percent

Whites 2,430 56.6%
Blacks 1,822 42.4%
Other 45 1.0%

Source: Marion County Criminal Justice Information Agency.


Arrests for Dealing Narcotics

In 1993 and 1997, 1,892 people were arrested in Marion County for dealing narcotics. Of those, 1,384, or 73.2 percent, were African American. There were 508 whites arrested for dealing narcotics, 26.8 percent of such arrests.


TABLE 3.7
Arrests for Dealing Narcotics by Race, 1993 and 1997, Marion County

Number

Percent

Whites 508 26.8%
Blacks 1,384 73.2%

Source: Marion County Criminal Justice Information Agency.


Analysis

The first set of analyses are straightforward tests for independence among arrest rates, types of arrest rates, decision to prosecute rates, and population designed to show disparate impact assuming binomial distributions.

The second set of analyses is a binary logit regression, with the decision to charge an individual with a crime as the dependent variable and race specifically modeled as an independent variable. 

Tests for Independence 

Tests for independence are structured to determine relationships between sets of characteristics, for example racial groups and geographical region or religion, or as in this study, arrest rates. Observed frequencies are compared with expected frequencies, which are based on the assumption there are no between-group differences.[20]

African Americans are 66.6 percent of all arrests for drug possession and 21.2 percent of the county population. Employing the above, the expected range for the number of African Americans to be arrested for drug possession is between 1,381 and 1,645.[21] The observed (actual) arrest rate of African Americans for drug possession in the years 1993 and 1997 is 4,849. This disparity is depicted in figure 3.1.

African Americans are 73.2 percent of all arrests for drug dealing, though they compose just 21.2 percent of the county population. Employing the above test, the expected range for the number of African Americans arrested for drug dealing is between 242 and 344.[22] The actual arrest rate of African Americans for drug dealing in the years 1993 and 1997 is 1,384. This disparity is depicted in figure 3.2.

 

FIGURE 3.1
Expected Arrest Range for African Americans and Observed Arrests, Drug Possession



1,381            1,645                                  4,849

Expected arrest range                            Observed arrests
for possession


FIGURE 3.2
Expected Arrest Range for African Americans and Observed Arrests, Drug Dealing

 
 

242                344                                       1,384

Expected arrest range                             Observed arrests
for dealing


Differences in Arrest Rates for Different Crimes

The chi-square test is employed to test the hypothesis of independence between race, i.e., African American and white, and the type of drug possession arrest, i.e., possession of cocaine, possession of a controlled substance, and possession of marijuana. Following standard statistical procedures, table 3.8 displays a contingency table with the actual and expected (in parentheses) frequencies for the three types of drug arrests by race. 


TABLE 3.8
Contingency Table of Race and Type of Drug Possession Arrest

Cocaine possession

Controlled substance possession

Marijuana possession

Blacks 2,287
(1,753)
132
(247)
2,430
(2,849)
Whites 330
(864)
236
(121)
1,822
(1,403)
Total 2,617 368 4,252

Source: Marion County Criminal Justice Information Agency.


Calculating the chi-square at 926, a significant dependence is found between race and the type of drug possession offense for which an individual is arrested. Specifically, from table 3.8 it is observed that African Americans are significantly more likely to be arrested for drug possession and possession of cocaine (chi-square; at 0.05=5.99), whereas whites are significantly more likely to be arrested for possession of marijuana and controlled substances.

Logit Regression of the Decision to Charge

A logit regression was employed to test whether race was a contributing variable in a prosecutor’s decision to charge an individual. The decision to charge/not charge was set as the dichotomous dependent variable. The independent variables in the model are age, race, previous offenses (yes/no), public defender (yes/no), and year (1993/1997). 

The case identification number from the arrest data was matched with case identification numbers from prosecution decision data. However, the data match rate between the listed arrest case identification number and the listed prosecution case identification number was only 67 percent. The analysis proceeded assuming unmatched cases were distributed between groups at expected rates. 

Individual cases with multiple arrests were condensed to one entry, and an additional independent variable—multiple charges—was computed. In addition, an interaction variable was created using race and age. 

Education was dropped as a considered independent variable because of the high number of cases for which the data was missing. Similarly, the income of the arrestee was omitted from the analysis because the data analysis indicated problems with the reliability and validity of the data.

The inclusion of the variable year (1993 or 1997) was to isolate differences in the prosecution decision between the Newman (1997) and Modisett (1993) administrations. Because of theorized differences in the charges, two separate regressions were computed: analysis of the decision to prosecute dealing narcotics and the decision to prosecute possession of narcotics.

Results: Decision to Charge In Narcotics Dealing 

The race of the person arrested was not found to be a significant contributing factor in the decision to prosecute narcotics dealing arrests, holding other variables constant.[23] Although there was a positive relation between race, i.e., African American, and the decision to prosecute, the relationship was too small to confidently conclude that race was a contributing factor in the decision to prosecute. The only variable found to have a significant contribution was previous arrests. None of the other variables (age, gender, public defender, year, or race) was found to significantly contribute to the decision to prosecute. Of particular interest is the finding that year (1993 or 1997) is not a contributing factor, which implies that the two prosecutors have both followed similar policies in the decision to prosecute. 

HOMICIDES

Data on arrests for homicides were obtained from the Marion County Criminal Justice Information Agency for the years 1993 and 1997. The agency also provided prosecution decision data on all individuals prosecuted for homicides. 

Arrests

Arrest and prosecution records on all homicides in 1993 and 1997 were obtained from the Indianapolis Police Department and the Marion County Sheriff as recorded by the Marion County Criminal Justice Information Agency. The data included such information as the name and address of the arrestee; location of the arrest; arrest charge; case identification number; the race, age, gender, and education level of the arrestee; and type of charge by the prosecutor’s office, i.e., aggravated assault, involuntary manslaughter, and murder.

Though African Americans were disproportionately arrested and prosecuted relative to their population in Marion County (see table 3.9), the arrest and prosecution rates of African Americans were coincident with the African American victimization homicide rates in the county.

Further analysis showed a high coincidence between rates of arrest and rates of prosecution, more than 90 percent. This produced a mulitcollinearity in the data and precluded further analysis with logit regression to test whether race was a contributing variable in a prosecutor’s decision to charge an individual.


TABLE 3.9
Prosecution for Homicides, 1993 and 1997, Marion County

Number

Percent

Whites 133 24.6%
Blacks 405 75.0%
Other 2 0.4%

Source: Marion County Criminal Justice Information Agency.


Death Penalty Prosecution

There is a perception in the minority community that race plays a role in the death penalty charging decision. For the death penalty to be sought, the defendant must be charged with murder and there must be one of a list of statutory aggravating circumstances present in connection with that murder. Those aggravators may include intentional killing during the course of a felony, the murder of a person under the age of 12 or over the age of 65, or other factors making the murder a more heinous and aggravated act.

If such a factor is present, the decision whether to charge the death penalty is completely discretionary with the prosecutor. During the course of the prosecution, however, the prosecutor still has the prerogative to offer a plea bargain to the defendant for a penalty less than death, which is generally a plea to life without parole.

Monica Foster of Hammerle, Foster, Allen & Long-Sharp testified that in the past six years, death penalty charges have been filed by the prosecutor in 15 cases. One of those cases was dismissed, leaving 14 individuals facing death penalty charges. Of the remaining 14 defendants, six defendants were white and eight were African American. Of the six white defendants charged with the death penalty, the prosecutor offered a plea bargain for life in five cases; in only one case where the defendant was white was no plea for life offered. Of the eight cases with African American defendants, seven defendants were charged with the death penalty, and only one was offered a plea for a penalty less than death.[24] 

Scott C. Newman, Marion County prosecutor, said Monica Foster’s assertions are false. Newman reported to the Advisory Committee that 22 death penalty cases have been handled by the Newman administration in the Marion County Prosecutor’s Office during the years 1995–2000.

During that period, 11 capital defendants were African American, and 11 capital defendants were white. Among capital cases in which the defendant was African American, he said, eight cases have been disposed of as follows: 

Among capital cases in which the defendant was white, he said, the cases have been disposed of as follows: 

LIMITATIONS OF THE STUDY

Data Limitations

Valid and reliable research can be hampered by data quality and completeness. As discussed earlier, the case identification number from the arrest data matched the case identification numbers from prosecution decision data at a rate of only 67 percent. 

There are also issues regarding the quality of the data with respect to race. There are issues about who is African American and Latino, and this variable in this study comes from the arresting officer’s opinion of the race/ethnicity of the person arrested. Finally, there may be variables or information pertinent to the explanation of the data but not included in the model because of unavailability.

Similarly, other factors may come into play that are difficult to model explicitly. For example, education may be a significant factor affecting criminal justice equity in that it may play a role in how well a person can articulate the facts surrounding the case and thereby assist his or her lawyer in presenting a case. Another factor may be prior records. For example, the number of prior criminal events may influence the criminal justice system, e.g., whether a person receives bail, which in turn can affect whether or not a person has access to a better defense, which in turn may affect the case in court.

Limitations in Analysis of Differences between Groups

The differences between groups in arrest data in this section are highly significant, but there are issues about race being highly related to other factors, which might in part explain why a particular racial group is treated the way that it is. Some of these explanations may be legally legitimate in the sense of we accept that explanation and live with it. Some of it may be illegitimate socially and legally in the sense that we need to do something about that explanation. Regardless, race is often a variable related to resources that may have an impact on equality in the criminal justice system, e.g., resources determine what kind of lawyer a person has, how well the lawyer argues, and how long the lawyer can afford to keep that case in court.



[1] Ind. Code Anno. §§ 35-50-2-4, 35-50-2-5 to 35-50-2-7 (MB, LEXIS 1999).

[2] Wayte v. United States, 470 U.S. 598, 607 (1985).

[3] Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U.S. 252, 265 (1977).

[4] Id. at 266.

[5] Id. at 266–68.

[6] Id. at 266.

[7] Id. at 268.

[8] Wayte, 470 U.S. at 598.

[9] Id. at 268.

[10] Id. at 269.

[11] Id. at 609–10.

[12] 517 U.S. 456 (1996).

[13] Id. at 470.

[14] Id.

[15] Id.

[16] See also Reno v. American-Arab Anti-Discrimination Comm., 525 U.S. 471 (1999); Wade v. Unites States, 504 U.S. 181 (1992); Hunter v. Underwood, U.S. 222 (1985).

[17] 1995 statistics from Indiana Department of Corrections.

[18] Ibid.

[19] See footnote 3, chapter 1.

[20] Probabilities associated with binomial experiments are readily obtainable from the formula b(x; n,p) where the mean (µ) = n*p, n being the population and p the population proportion, and the standard deviation (s) = (n·p·q)˝

[21] This assumes a tolerance range based on 3 standard deviations.

[22] Ibid.

[23] The coefficient of the race variable was 0.39, but it was not significant (p=0.087). The chi-square value was 8.12 with 5 degrees of freedom.

[24] Monica Foster, testimony before the Indiana Advisory Committee to the U.S. Commission on Civil Rights, fact-finding meeting, Indianapolis, IN, Jan. 29, 1999, transcript, pp. 116–22.

[25] Letter of Scott C. Newman to Constance M. Davis, regional director, Midwestern Regional Office, U.S. Commission on Civil Rights, May 12, 2000.