.-_ -..-...-...- -- .--....” ...”.
14’t~l)rll;l I‘\’ I!J!JO

DEATH PENALTY
SENTENCING
Research Indicates
Pattern of Racial
Disparities

.
I
---~

United States
General Accounting Office
Washington, D.C. 20548

GAO

General Government Division

B-236876
February 26, 1990
The Honorable Joseph R. Biden, Jr.
Chairman, Committee on the Judiciary
United States Senate
The Honorable Strom Thurmond
Ranking Minority Member, Committee
on the Judiciary
United States Senate
The Honorable Edward M. Kennedy
Chairman, Subcommitteeon Immigration
and Refugee Affairs
Committee on the Judiciary
United States Senate
The Honorable Jack Brooks
Chairman, Committee on the Judiciary
House of Representatives
The Honorable Hamilton Fish, Jr.
Ranking Minority Member, Committee
on the Judiciary
House of Representatives
The Anti-Drug Abuse Act of 1988 (Public Law 100-690) requires us to
study capital sentencing procedures to determine if the race of either
the victim or the defendant influences the likelihood that defendants
will be sentencedto death. We did an evaluation synthesis-a review
and critique of existing research- on this subject to fulfill the mandate.
This report provides a summary of our findings and a discussion of our
approach and data limitations.
A

Approach

An evaluation synthesis is a critical integration of findings from existing
empirical research on a given topic- in this case death penalty sentencing after the Furman decision.’First, we identified and collected all
potentially relevant studies done at national, state, and local levels from
‘In Furman v. Georgia, 408 U.S. 238 (1972), the Supreme Court found unconstitutional death
sentences imposed under state statutes which allowed juries to impose these sentences in w arbitrary
or capricious manner. In response to this decision, states adopted new statues that addressed the
concerns raised by the Court.

Page 1

GAO/GGD-90-57Racial Di~~paritiesin Sentencing

both published and unpublished sources.Computer-generatedbibliographic searchesand manual reviews of the bibliographies of studies
that we obtained contributed to our list of potentially relevant material.
We also surveyed 21 criminal justice researchersand directors of relevant organizations whose work relates to death penalty sentencing to
identify additional research. We screenedmore than 200 annotated citations and referencesto determine relevance to our review. We excluded
studies that (1) were based primarily on data collected prior to the
Furman decision and (2) did not examine race as a factor that might
influence death penalty sentencing.From this initial screening we
obtained 63 studies that we determined to be relevant.
We then reviewed each of the 63 studies to determine both appropriateness and overall quality of the research. We excluded studies that did
not contain empirical data or were duplicative (a few researcherspublished several articles, with the most current including data and findings
cited in earlier versions). Twenty-eight studies remained after this
assessment.The information included in these studies forms the basis
for our findings.
Next, we rated the 28 studies according to research quality. Two social
scienceanalysts independently rated each study in five dimensions:(1)
study design, (2) sampling, (3) measurement,(4) data collection, and (6)
analysis techniques, A rating for overall quality was also given. A third
analyst reviewed the raters’assessmentsto ensure consistency. In addition, a statistician reviewed the studies that used specialized analytic
techniques to assesswhether the techniques were applied correctly and
whether the analyses fully supported the researchers’conclusions.
Finally, we extracted all relevant information on the relationship of race
to death penalty sentencing from each of the studies. This information
was compared and contrasted across studies to identify similarities and
differences in the findings.
Evaluation synthesis has benefits and limitations, The major benefit is
that evidence from multiple studies can provide greater support for a
finding than evidence from an individual study. The major limitation is
that this approach dependson the quantity and quality of the design
and methodology of available studies and the comprehensivenessof
their reporting. In this case,the body of research concerning discrimination in death penalty sentencing is both of sufficient quality and quantity to warrant the evaluation synthesis approach.

Page 2

GAO/GGD-90-57Racial Disparities in Sentencing

We evaluated 28 studies which were done by 21 sets of researchers.”
The studies covered homicide casesfor different time periods through
1988, many states that have the death penalty, and different geographic
regions of the country. In three instances, two or more articles were generated from a single database, with each article focusing on a different
aspect of the sentencing process. A few researchers used data from
other studies in their analyses. Overall, the 28 studies constitute 23 different data sets.

Des ription of the
Stu 4 ies

We rated almost half of the studies as high or medium quality; the
remainder were rated as low. It is important to evaluate research quality for two reasons: (1) the results of the synthesis should be based on a
sufficient number of medium or high quality studies; and (2) it is important to note differences in studies’findings, if any, by the quality of the
studies, By quality we mean the strength of the design and the rigor of
the analytic technique that leads to a level of confidence we have in the
study findings. We judged a study to be high quality if it
was characterized by a sound design that analyzed homicide cases
throughout the sentencing prqcess;
. included legally relevant variables (aggravating and mitigating circumstances); and
. used statistical analysis techniques to control for variables that correlate with race and/or capital sentencing.

.

We judged a study as medium quality if we found it to be lacking in one
or more of the above characteristics. However, the medium quality studies generally were more similar to high quality studies than to low quality studies. Low quality studies typically had weak or flawed designs,
relied on less reliable statistical analysis, and were simplistic in interpretation of the data. Studies published before 1986 comprised a larger proportion of lower quality studies than those published subsequently. This
coincides with the relatively recent development and use of a more
sophisticated statistical technique appropriate for use with data such as
those in death penalty studies.

Limitations of the
Studies
a

the design and analysis of the research. We identified three major limitations among these studies: (1) the threat of sample selection bias, (2) the
problem of omitted variables, and (3) the small sample sizes.
“Appendix I includes a list of the studies we used in the synthesis.

Page 3

GAO/GGD99-57 Racial Dieparities in Sentencing

0236870

-_.

Sampleselection bias implies that the casesunder consideration are not
representative of all the casesof interest. The crim inal justice system is
characterized by discretionary processesof selection at different points
in the system. Racial factors may influence decisionsat different stages
of the process.A study that considered only whether persons convicted
were sentencedto death was especially prone to the biasing effect of
sample selection. Racial factors may have influenced decisionsearlier in
the process,such as whether the prosecutor requested that an offender
be charged with capital murder. This discretion exercised early in the
processmay have the effect of concealing(masking) race effects if analysis is lim ited only to the later stages.
We found sample selection bias in more than half of the low quality
studies; these studies typically analyzed only those casesin which the
defendant was convicted of capital murder or received the death penalty. Studies that included all reported homicides and followed the disposition of these defendants from initial charge through subsequent
stagesof the judicial process are not likely to have been affected by this
bias. More than two-thirds of the studies we rated high or medium quality picked up casesprior to conviction and followed these casesthrough
the judicial process.
Another lim itation is the problem of omitted variables. This lim itation is
especially important in studies examining racial discrimination. This is
becausethe effect of race is considered the residual-after all relevant
and important variables have been controlled, the effect that remains,
the residual, is interpreted to be racial disparity. Omitting relevant variables can affect results by failing to reduce the residual appropriately,
thus enhancing the perceived racial disparity. Omitted variables in
death penalty research are potentially of two types: (1) variables that
were known and were believed to be correlated with race or the death
penalty and (2) variables that were not known and may be correlated
with race or the death penalty outcome.
Several of the higher quality studies controlled for many variables. For
example, one high quality study controlled for more than 200 variables.
Only a few variables are shown to be highly explanatory. Most of these
are controlled for in the better quality studies. However, there are variables such as strength of evidence or socioeconomicstatus of the victim
and defendant which are difficult to measure or obtain. If there are
important omitted variables (either becausethey are difficult to measure or becausethey are unknown), other explanations for the differencesin death penalty outcomescannot be excluded. But for another
Page 4

GAO/GGD-90-57 Racial Disparities in Sentencing

variable to influence the existing disparity it would have to (1) be
jointly correlated with both race and the death penalty outcome and (2)
operate independently of the factors already included in the analysis.
A third limitation relates to the consequencesof the small sample sizes
in the analyses of death penalty imposition. The imposition of the death
penalty is a relatively rare event. As such, in most studies there were
very few casesat the end of the process-the sentencing and imposition
stages.The small sample size places limits on the usefulness of statistical techniques for analysis at these final stages and thus limits the rigor
of analyses at these stages.
While the severity of the limitations varied, as reflected in the studies’
ratings, these limitations do not preclude a meaningful analysis of the
studies. We have considered quality in evaluating the studies and arriving at our findings.

Fifidings

Our synthesis of the 28 studies shows a pattern of evidence indicating
racial disparities in the charging, sentencing, and imposition of the
death penalty after the Furman decision.
In 82 percent of the studies, race of victim was found to influence the
likelihood of being charged with capital murder or receiving the death
penalty, i.e., those who murdered whites were found to be more likely to
be sentencedto death than those who murdered blacks.”This finding
was remarkably consistent across data sets, states, data collection methods, and analytic techniques. The finding held for high, medium, and
low quality studies.
The race of victim influence was found at all stagesof the criminal justice system process,although there were variations among studies as to
whether there was a race of victim influence at specific stages.The evidence for the race of victim influence was stronger for the earlier stages
of the judicial process (e.g., prosecutorial decision to charge defendant
with a capital offense, decision to proceed to trial rather than plea bargain) than in later stages.This was becausethe earlier stages were comprised of larger samples allowing for more rigorous analyses. However,

“When we refer to a finding of racial disparities at the sentencing and imposition stages we are, in
fact, including disparities that occurred in earlier stages of the judicial process, e.g., charging and
decision to proceed to trial.

Page 6

GAO/GGD-90-67Racial Disparities in Sentencing

5230870

decisionsmade at every stage of the process necessarily affect an individual’s likelihood of being sentencedto death.
Legally relevant variables, such as aggravating circumstances,were
influential but did not explain fully the racial disparities researchers
found. In the high or medium quality studies, researchersused appropriate statistical techniques to control for legally relevant factors, e.g.,
prior crim inal record, culpability level, heinousnessof the crime, and
number of victims. The analysesshow that after controlling statistically
for legally relevant variables and other factors thought to influence
death penalty sentencing(e.g., region, jurisdiction), differences remain
in the likelihood of receiving the death penalty based on race of victim .
The evidence for the influence of the race of defendant on death penalty
outcomeswas equivocal. Although more than half of the studies found
that race of defendant influenced the likelihood of being charged with a
capital crime or receiving the death penalty,4 the relationship between
race of defendant and outcome varied across studies. For example,
sometimesthe race of defendant interacted with another factor. In one
study researchersfound that in rural areas black defendants were more
likely to receive death sentences,and in urban areas white defendants
were more likely to receive death sentences.In a few studies, analyses
revealed that the black defendant/white victim combination was the
most likely to receive the death penalty. However, the extent to which
the finding was influenced by race of victim rather than race of defendant was unclear.
Finally, more than three-fourths of the studies that identified a race of
defendant effect found that black defendants were more likely to
receive the death penalty. However, the remaining studies found that
white defendants were more likely to be sentencedto death.
To summarize,the synthesis supports a strong race of victim influence.
The race of offender influence is not as clear cut and varies across a
number of dimensions.Although there are lim itations to the studies’
methodologies,they are of sufficient quality to support the synthesis
findings.

“About two-thirds of these studies were of high or medium quality.

Page 6

GAO/GGD-90-57 R&al

Disparities in

Sentencing

We are sending copies of this report to cognizant congressionalcommittees, the Attorney General, and other interested parties.
Major contributors to this report are listed in appendix II. Pleasecall me
at 276-8389 if you have any questions.

Lowell Dodge
Director, Administration
of Justice Issues

Page 7

GAO/GGD-90-57Racial Disparities in Sentencing

7
I

3

C&tents

Letter

1

Apbendix I
Lise of Studies

10

Ap$endix II
Major Contributoxy to
This Report

13

Page 8

GAO/GGDBM7 Racial Diaparltles in Sentencing

Page 9

GAO/GGD9O-67Racial Diaparltiea in Sentencing

Appendix I

I&t of Studies

Arkin, Stephen. “Discrimination and Arbitrariness in Capital Punishment: An Analysis of Post-Furman Murder Casesin Dade County, Florida, 1973-1976.”Stanford Law Review, Vol. 33 (November 1980) 76101.
Baldus, David, GeorgeWoodworth, and Charles Pulaski. Equal Justice
and the Death Penalty: A Legal and Empirical Analysis. Boston: Northeastern University Press, 1990.
Barnett, Arnold. “Some Distribution Patterns for the Georgia Death Sentence.” University of California Davis Law Review, Vol. 18, No. 4 (Summer 1986) 1327-1374.
Berk, Richard and Joseph Lower-y. “Factors Affecting Death Penalty
Decisionsin Mississippi,” (unpublished manuscript, June 1986).
Bienen, Leigh et al. “The Reimposition of Capital Punishment in New
Jersey: The Role of Prosecutorial Discretion.” Rutgers Law Review (Fall
1988) 27-372.
Bowers, William. “The Pervasivenessof Arbitrariness and Discrimination Under Post-Furman Capital Statutes.” Journal of Criminal Law &
Criminology, Vol74. No. 3 (Fall 1983) 1067-1100.
Bowers, William and Glenn Pierce. “Arbitrariness and Discrimination
under Post-Furman Capital Statutes.” Crime and Delinquency (October
1980) 663-636.
Ekland-Olson,Sheldon. “Structured Discretion, Racial Bias and the
Death Penalty: The First DecadeAfter Furman in Texas.” Social Science
Quarterly, Vol. 69 (December 1988) 863-873.
Foley, Linda. “Florida after the Furman Decision: The Effect of Extralegal Factors on the Processingof Capital Offense Cases.”Behavioral Sciences& the Law, Vol. 6, No. 4 (Autumn 1987) 467-465. Foley, Linda and Richard Powell. “The Discretion of Prosecutors,
Judges, and Juries in Capital Cases.” Criminal Justice Review Vol. 7, No.
2 (1982) 16-22.
Gross, Samuel and Robert Mauro. “Patterns of Death: An Analysis of
Racial Disparities in Capital Sentencingand Homicide Victimization.”
Stanford Law Review, Vol. 37 (1984) 27-153.
Page 10

GAO/GGJMO-57Racial Disparities in Sentencing

Appendix I
List of Studies

Keil, Thomas and Gennaro Vito. “Race and the Death Penalty in Kentucky Murder Trials: An Analysis of Post-GreggOutcomes.”Forthcoming in Justice Quarterly.
Keil, Thomas and Gennaro Vito. “Race, Homicide Severity, and Application of the Death Penalty: A Consideration of the Barnett Scale.”Criminology, Vol. 27, No. 3 (1989) 511-535.
Kleck, Gary. “Racial Discrimination in Criminal Sentencing:A Critical
Evaluation of the Evidence with Additional Evidence on the Death Penalty.” American SociologicalReview, Vol. 46 (1981) 783805.
Klein, Stephen, Allen Abrahamse, and John Rolph. “Racial Equity in
Prosecutor Requestsfor the Death Penalty.” Unpublished manuscript,
The Rand Corporation (1987).
Klein, Stephen. “Relationship of Offender and Victim Raceto Death Penalty Sentencesin California.” Unpublished manuscript, The Rand Corporation (1989).
Klemm, Margaret F. “The Determinants of Capital Sentencingin Louisiana, 19751984.” Dissertation, University of New Orleans (1986).
Lewis, Peter, Henry Mannle, and Harold Vetter. “A Post-Furman Profile
of Florida’s Condemned-AQuestion of Discrimination in Terms of Race
of the Victim and a Comment on Spenkelink v. Wainwright.” Stetson
Law Review, Vol. IX, No. 1 (1979) l-46.
Murphy, Elizabeth. “The Application of the Death Penalty in Cook
County.” Illinois Bar Journal, Vol. 93 (1984) 90-95.
Nakell, Barry and Kenneth Hardy. The Arbitrariness of the Death Penalty. Philadelphia: Temple University Press, 1987.
Paternoster, Raymond and Ann Marie Kazyaka. “The Administration of
the Death Penalty in South Carolina: Experiences Over the First Few
Years.” South Carolina Law Review, Vol. 39, No. 2 (1988) 245-414.
Radelet, Michael and Margaret Vandiver. “The Florida Supreme Court
and Death Penalty Appeals.” Journal of Criminal Law and Criminology,
Vol. 74, No. 3 (1983) 913-926.

Page 11

GAO/GGD-90-57Racial Disparities in Sentencing

Appendix I
List of Studies

Radelet, Michael and Glenn Pierce. “Race and Prosecutorial Discretion in
Homicide Cases.”Law and Society Review, Vol. 19, No. 4, (1985) 687621.
Radelet, Michael. “Racial Characteristics and the Imposition of the
Death Penalty.” American SociologicalReview, Vol. 46, (1981) 918-927.
Riedel, Marc. “Discrimination in the Imposition of the Death Penalty: A
Comparison of the Characteristics of Offenders SentencedPre-Furman
and Post-Furman.” Temple Law Quarterly, Vol. 49, No. 2 (1976) 261287.
Smith, Dwayne M. “Patterns of Discrimination in Assessmentsof the
Death Penalty: The Caseof Louisiana.” Journal of Criminal Justice, Vol.
15 (1987) 279-286.
Vito, Gennaro and Thomas Keil. “Capital Sentencingin Kentucky: An
Analysis of the Factors Influencing Decision Making in the Post-Gregg
Period.” The Journal of Criminal Law & Criminology, Vol. 79, No. 2
(Summer 1988) 483-508.
Zeisel. Hans. “Race Bias in the Administration of the Death Penalty: The
Florida Experience.” Harvard Law Review, Vol. 95, No. 2 (December
1981) 456-468.

Page 12

GAO/GGD-90-57Racial Disparities in Sentencing

.

I Appeidix II

M$jor Contributors to This &port

; Government
General
Division, Washington,
D.Ci

(131810)

Harriet C. Ganson, Analyst-in-Charge
Lisa Cassady, Social ScienceAnalyst
Jam es L. Frem m ing, Consultant Douglas M . Sloane, Statistical Consultant

Page 13

(hi 1.twrs hrg!,

‘I’tw first, fivv
$2.00 t~wll.

Mwylatrcl

copies

20877

of tvtc11 report.

are free. Atldit,iotlal

‘I’tit~rt~ is a 25% discount,
singlth acltlress.

OJI

Ortlt~rs

by cash or by check

ollt

t.0

must, tw prepaid
t.tlP

SJl~Jc~riJlt,~JJ<lt~tlt.

orders

Of

for

hClJIJl~JltS.

100 or more

or

copit

copitts

money

order

art’

rnaihi

nlacit!

to a

_------