Page 1

LEXSEE 41 STAN. L. REV. 121
Copyright (c) 1988 The Board of Trustees of Leland Stanford Junior University
Stanford Law Review
November, 1988
41 Stan. L. Rev. 121
LENGTH: 19998 words
COMMENT: Protecting the Innocent: A Response to the Bedau-Radelet Study
NAME: Stephen J. Markman * Paul G. Cassell **
BIO:
* Assistant Attorney General, Office of Legal Policy, United States Department of Justice. B.A., 1971,
Duke University; J.D., 1974, University of Cincinnati College of Law.
** Special Assistant United States Attorney, Eastern District of Virginia; former Associate Deputy Attorney
General, United States Department of Justice. B.A., 1981, Stanford University; J.D., 1984, Stanford University.
Special thanks to Peter Rient for his assistance. The views expressed in this article are the authors'; they do not
necessarily reflect the position of the Department of Justice.
SUMMARY:
... Given the fallibility of human judgments, the possibility exists that the use of capital punishment may result in the
execution of an innocent person. ... The use of this phrase -- and its variant, "proved to be innocent" -- clearly suggests a
high degree of certainty in the determination of innocence. ... Josephine Brown, the wife of victim Edgar Brown,
testified that shortly before the murder she had seen a brown Rambler heading towards their residence with a black man
driving it. ... The authors' claim that Appelgate was innocent, stripped of irrelevancies, rests on the incorrect assertion
of an earlier author that "virtually no evidence against Appelgate existed beyond Mrs. Creighton's unsupported word."
... There are two principal difficulties with the argument that we should not use capital punishment because of the risk
of executing an innocent person. ... The use of capital punishment entails some risk that an innocent person will be
executed. ... They assert that with the abolition of the mandatory death penalty, "society runs both the risk of executing
the innocent and the risk of recidivist murder, whereas it is only necessary that society run one or the other." ...
TEXT:
[*121] Given the fallibility of human judgments, the possibility exists that the use of capital punishment may result in
the execution of an innocent person. This terrible prospect raises the issue of whether the risk of error in administering
the death penalty is sufficiently high both to outweigh the potential benefits of capital punishment and to offend the
moral sensibilities that must support a free society's criminal justice system. Despite occasional claims that specific
individuals have been put to death for crimes they did not commit, the risk of executing the innocent has never been the
subject of thorough and rigorous empirical study. Recently, however, two opponents of the death penalty, Hugo Adam
Bedau and Michael L. Radelet, have published the results of "sustained and systematic" research purporting to show
that the use of capital punishment entails an intolerable risk of mistaken executions. n1 According to the authors, 350

Page 2
41 Stan. L. Rev. 121, *121

persons have been wrongly convicted of capital or "potentially capital" crimes in the United States during this century;
and twenty-three innocent persons have actually been executed. n2 These alarming conclusions cannot be taken at face
value. Not only is the Bedau-Radelet study severely flawed in critical respects, it wholly fails to demonstrate an
unacceptable risk of executing the innocent. To the contrary, it confirms -- as convincingly as may be possible -- the
view that the risk is too small to be a significant factor in the debate over the death penalty. n3 Because of these
shortcomings, [*122] the study deserves a response, lest it gain currency to the detriment of clear thinking on the
subject for years to come. n4
The response presented here proceeds along two lines. The first is directed at the relevance and validity of the
study itself. As the discussion in Part I indicates, the study suffers from a number of flaws. The most serious of these
are the authors' reliance on material irrelevant to the risk of wrongful executions and their method of determining
innocence. With respect to relevance, only about seven percent of the study deals with cases of allegedly erroneous
executions. Moreover, even as regards these cases, the authors' decision to include cases from the early part of this
century, long before the adoption of the extensive contemporary system of safeguards in the death penalty's
administration, skews their analysis of capital punishment under contemporary circumstances. With respect to
methodology, the authors' method is overly subjective. It is also one-sided in its description of the cases of alleged
error. Indeed, with respect to some of the allegedly mistaken executions, there appears to be little resemblance between
the authors' descriptions and the actual cases.
The second line of response, presented in Part II, rebuts the argument that capital punishment should be abolished
because the risk of erroneous execution is unacceptably high. n5 That argument fails for several reasons. First, no
sound reason exists for believing that there is currently an intolerable risk of executing an innocent person. Over the
past fifteen years, procedural protections have been adopted to reduce as much as possible the likelihood that error will
be committed or, if committed, that it will go undetected. More to the point, the authors present no credible evidence
that any innocent person has been executed during this period; and they do not claim that any individual now awaiting
execution is innocent. In addition, the argument undervalues the important reason why the great majority of Americans
and their legislators recently have determined that capital punishment should be reinstated, notwithstanding society's
inability to administer the death [*123] penalty with "godlike perspicacity." n6 That reason is to save lives. Through
a combination of deterrence, incapacitation, and the imposition of just punishment, the death penalty serves to protect a
vastly greater number of innocent lives than may be lost through its erroneous application. Finally, society would be
guilty of a self-destructive "failure of nerve" n7 if it were to forego the use of an appropriate and effective punishment
simply because it is not humanly possible to eliminate the risk of mistake entirely.
I. FLAWS IN THE BEDAU-RADELET STUDY
Starting with the unexceptionable proposition that "[f]ew errors made by government officials can compare with
the horror of executing a person wrongly convicted of a capital crime," n8 Bedau and Radelet have undertaken an
empirical study aimed at providing "a better understanding of the miscarriages of justice in capital or potentially capital
cases that have occurred in the United States during this century." n9 To this end, they present "350 cases in which
defendants convicted of capital or potentially capital crimes in this century, and in many instances sentenced to death,
have later been found to be innocent." n10 Twenty-three of these cases, they claim, resulted in the execution of an
innocent person. n11
In fact, the Bedau-Radelet study is too flawed to provide much support for the authors' argument against the death
penalty. A complete analysis of the study's shortcomings, including an examination of the validity of the authors'
conclusions with respect to each of the 350 cases presented, is outside the scope of this article. Nevertheless, even a
partial analysis of the study's methodology and conclusions is sufficient to expose the study's defects.
A. The Misleading Nature of the Study
As noted above, the authors assert that 350 Americans have been wrongly convicted of capital or "potentially

Page 3
41 Stan. L. Rev. 121, *123

capital" crimes during this century. n12 The category of "potentially capital" cases, however, consists in large measure
of cases in which the death penalty was not available or was not the sentence given. It includes cases of persons
convicted in states without capital punishment for the defendant's crime, convicted of lesser offenses such as
second-degree murder and [*124] voluntary manslaughter, and convicted of capital offenses but sentenced to
imprisonment rather than death. n13 The authors justify including these "potentially capital cases" on the ground that,
"except for some relatively adventitious factor," they might have culminated in a death sentence. n14 This approach
seriously misstates the magnitude of the problem of wrongful capital convictions. The authors have pointed to only 200
allegedly wrongful convictions for first-degree murder during this century. n15 In only 139 of these cases were the
defendants actually sentenced to death. n16 More to the point, the death sentence was carried out in only twenty-three
of the 350 cases cited by the authors. n17 Thus, strictly speaking, only 6.6 percent of the study is relevant to the issue
of wrongful executions. Moreover, the authors cite but a single allegedly erroneous execution during the past
twenty-five years -- that of James Adams. n18 A review of that case demonstrates, however, that Adams was
unquestionably guilty. n19 Thus, Bedau and Radelet have made no persuasive showing that anyone has been wrongly
executed since new capital punishment procedures were instituted in the wake of Furman v. Georgia. n20 In short,
what the authors have done could be compared to studying traffic deaths before the adoption of traffic signals.
The authors also mislead readers through their presentation of the [*125] data. They apparently seek to elicit
shock or alarm, rather than to promote reasoned consideration of the issues. Thus, with respect to their finding that an
innocent person was sentenced to death in 40 percent of the cases they present, they state: "This is error in capital cases
in the strictest sense of the term, at an average rate of three such errors every two years." n21 Yet the eye-catching
statistic is misleading. This is an alleged average annual rate of erroneous death sentences. But, since the discussion's
context is the human propensity to err in making capital judgments, the appropriate statistic would be the number of
erroneous death sentences divided by the total number of such sentences.
Also objectionable is the authors' emphasis on erroneous death sentences, when the more significant consideration
is the number of erroneous executions. Presumably, the issue is whether the magnitude of the risk of executing
innocent persons makes capital punishment unwise as a matter of policy. The size of this risk can only be evaluated,
however, by dividing the number of innocent persons who have been executed by the total number of persons executed.
n22 The number of persons wrongly convicted of capital crimes and sentenced to death is less significant for at least
two reasons: First, if these persons were not executed, their erroneous convictions and death sentences have no bearing
on the wisdom of executing persons properly convicted of capital crimes. Second, focusing on trial court dispositions
alone completely ignores the procedural protections for detecting erroneous convictions that are available through direct
appeal and collateral attack. n23
[*126] B. The Subjectivity of the Study
1. Determining innocence.
The overwhelming problem with the Bedau-Radelet study is the largely subjective nature of its methodology and
therefore of its conclusions. The linchpin of the study is the claim that the authors have identified "350 cases in which
defendants convicted of capital or potentially capital crimes in this century, and in many cases sentenced to death, have
later been found to be innocent." n24 This claim raises the critical question of the meaning of the phrase "found to be
innocent." The use of this phrase -- and its variant, "proved to be innocent" n25 -- clearly suggests a high degree of
certainty in the determination of innocence. The use of the passive voice, however, belies that the authors themselves
made many of the findings; and neither the standard they use to determine innocence nor the proof they offer to meet
that standard permits such assurance.
The authors' standard is simply their belief, arrived at after reviewing the information available to them regarding
each case, that "a majority of neutral observers, given the evidence at [the authors'] disposal, would judge the defendant
in question to be innocent." n26 The type of information upon which the authors rely includes the alleged subsequent
discovery that no capital crime was committed, n27 alleged acknowledgment of error by the state, n28 another

Page 4
41 Stan. L. Rev. 121, *126

person's confession, n29 a combination of lack of evidence against the defendant and the implication [*127] of
another person as the culprit, n30 a state official's opinion that the defendant was not guilty, n31 and the "weight of
the evidence" available to subsequent investigators. n32 More specifically, in the twenty-three cases involving
allegedly mistaken executions, the information upon which the authors rely to "prove" innocence consists of another
person's confession (six cases), n33 the implication of another person (three cases), n34 a state official's opinion (six
cases), n35 and "subsequent scholarly judgment" n36 that the defendant was innocent (eight cases). n37
The authors' dubious method of finding convicted individuals to have been innocent strikes the neutral observer
reading the authors' very brief discussion of wrongful executions. n38 There one reads the following:
The evidence for judgment of error in these cases will naturally interest many, but what we can report will satisfy
only a few. In the notorious cases, we have relied on the judgment of other scholars whose investigative work
convinces us that error did indeed occur. Nothing we can say here will cause the controversy that still surrounds each of
these cases to abate. In some of the more obscure cases, we have relied entirely on the opinions of officials whose
views we believe deserve to be taken seriously. In none of these cases, however, can we point to the implication of
another person or to the confession of the true killer, much less to any official action admitting the execution of an
innocent person. n39
[*128] Given this admission, a neutral observer is entitled to be skeptical that this study contributes to a better
understanding of the prevalence of error in the administration of the death penalty. The use of such questionable
methodology invites readers to suspend use of their critical faculties and simply accept the results of this study on faith
-- faith that in fact the authors have sufficient information to make a judgment in each of these cases and faith that they,
rather than the jurors and judges who considered the evidence admitted at trial, have made the correct judgment. Such
faith hardly seems justified, given Bedau and Radelet's tendency to rely upon sources that either do not support their
statements and conclusions or contradict them. n40 These errors and other aspects of the study give the impression that
it is more an argumentative tract than a fair-minded inquiry. n41
2. The Adams case.
Examination of individual cases in which the authors claim that an innocent person was executed heightens the
reader's skepticism. Consider, for example, the case of James Adams. According to the authors, Adams was
erroneously executed in 1984. This case is particularly important because it is the only alleged example of an innocent
person's execution since Furman v. Georgia. n42 The following is the authors' full description of the evidence for and
against Adams's guilt:
ADAMS, JAMES (black). 1974. Florida. Adams was convicted of first-degree murder, sentenced to death, and
executed in 1984. Witnesses located Adams' car at the time of the crime at the home of the victim, a white rancher.
Some of the victim's jewelry was found in the car trunk. Adams maintained his innocence, claiming that he had loaned
the car to his girlfriend. A witness identified Adams as driving the car away from the victim's home shortly after the
crime. This witness, however, was driving a large truck in the direction opposite to that of Adams' car, and probably
could not have had a good look at the driver. It was later discovered that this witness was angry with Adams for
allegedly dating his wife. A second witness heard a voice inside the victim's home at the time of the crime and saw
someone fleeing. He stated this voice was a woman's; the day after the crime he stated that the fleeing person was
positively not Adams. More importantly, a hair sample found clutched in the victim's hand, which in all likelihood had
come [*129] from the assailant, did not match Adams' hair. Much of this exculpatory information was not discovered
until the case was examined by a skilled investigator a month before Adams' execution. Governor Graham, however,
refused to grant even a short stay so that these questions could be resolved. n43
Our review of the record in the Adams case reveals that this description is seriously misleading. The authors have
misrepresented or excluded critical evidence of the defendant's guilt and have exaggerated the significance of
supposedly exculpatory evidence. In addition, there is no basis for their suggestion that Adams was denied a fair

Page 5
41 Stan. L. Rev. 121, *129

opportunity to establish his innocence.
Bedau and Radelet begin their defense of James Adams by claiming that one witness who identified Adams leaving
the scene of the crime "was driving a large truck in the direction opposite to that of Adams' car, and probably could not
have had a good look at the driver." The trial testimony, as recounted by the State in its response to Adams's second
federal habeas corpus petition, gives a different picture of the witness's identification:
Willie Orange . . . passed a car travelling in the opposite direction which was wobbling all across the road so much .
. . that Orange had to pull over as far to the right as he could. The car slowed before it reached him, and moved back to
the left side, travelling west. Orange identified the car as being a brown Rambler, and identified the driver as being the
defendant. He was positive of his identification of defendant as the driver. The next day he picked defendant out of the
lineup. n44
The article also states that "[i]t was later discovered that this witness was angry with Adams for allegedly seeing his
wife." n45 Defense counsel made this suggestion, however, in their opening statement at trial in 1974 and did not later
pursue it. n46
The Bedau-Radelet article also misrepresents the testimony of the second witness to which it refers. According to
the authors, this witness [*130] "heard a voice inside the victim's home at the time of the crime and saw someone
fleeing. He stated this voice was a woman's; the day after the crime he stated that the fleeing person was positively not
Adams." n47 Resort to the trial proceedings gives quite a different impression of this witness's testimony:
[Foy Hortman, the witness to whom Bedau and Radelet refer,] finally drove his truck toward the house in order to
find Mr. Brown [the murder victim] and saw a brown Rambler in front with mirrors on the outside and something
hanging on the mirror on the inside. As he approached the house, he thought he heard someone say, "In the name of
God, don't do it." He then listened for a short while, re-entered his truck and started to back out when he saw the door
that opens into the carport of the house open. The day was "real cloudy," and it rained a bit as Hortman had approached
the house. When he saw someone leaving the house, he stopped his truck and got out. The person leaving the house
was "real black," and had something sticking out in his front pocket. Hortman exchanged a few words with the man,
who continued walking away toward the Rambler. As soon as the man rounded the corner of the house Hortman heard
a door open and a car leave. The man Hortman saw appeared to be 30 to 35 years old, approximately 6 feet tall, n48
and Hortman noticed no mustache on him. Hortman said the lighting conditions underneath the carport area was "kind
of dark" since the sun "wasn't shining a bit." On cross-examination, Hortman testified that the voice that he heard from
inside the house did not sound like Mr. Brown's voice. On direct examination, he had testified that the voice sounded
"kind of like a woman's voice, kind of like strangling or something and I thought it was on the radio or T.V. or
something." On recross and direct examination, he stated that the voice sounded like "it was kind of strangled or
something," and that "if he couldn't hardly talk it would sound that way, if he had been beat up like he was." Hortman
also testified that he could not say that defendant was or was not the man that he saw, and that he did not identify
defendant in a lineup the next day. n49
As this testimony makes clear, the voice Hortman heard said, "In the name of God, don't do it." It was not seriously
disputed at trial that this must have been Mr. Brown's last plea for life, and therefore it is without consequence that the
voice sounded in certain respects like a woman's. Moreover, contrary to the authors' implication that the real killer was
a woman, n50 the witness made it clear that the person he saw leaving the scene of the crime was a man, even though
he was unable to identify the man positively as Adams.
Finally, and "[m]ore importantly," the article contends, "a hair sample [*131] found clutched in the victim's hand,
which in all likelihood had come from the assailant, did not match Adams's hair." n51 The article even suggests that
this evidence was newly discovered but that the Governor refused to delay the execution in 1984 to consider the issue.
The hair sample issue, however, actually was raised and litigated in 1984 before the Florida courts. They, like many
other state courts, have procedures for evaluating recently discovered evidence suggesting that a defendant may have

Page 6
41 Stan. L. Rev. 121, *131

been erroneously convicted. In this Florida proceeding, the state pointed out that the hair sample analysis was made
known to defense counsel at trial in 1974. Defense counsel chose not to use it n52 because the sample was not
particularly probative. n53 It was a remnant of a sweeping of the ambulance and so could have come from another
source. n54
All of the evidence relating to the hair sample, along with several other issues, was placed before the state court
which denied Adams's motion to vacate sentence in an order dated April 25, 1984. n55 Adams apparently decided not
to pursue the hair sample issue. He did not even raise it when he appealed to the Florida Supreme Court on other
grounds. n56
It seems reasonable to conclude, therefore, that the article offers little evidence to support its contention that Adams
was in fact innocent of the murder. More tellingly, the article ignores the compelling evidence that convinced a
unanimous jury of his guilt beyond a reasonable doubt.
Josephine Brown, the wife of victim Edgar Brown, testified that shortly before the murder she had seen a brown
Rambler heading towards their residence with a black man driving it. When she returned home at approximately 10:50
a.m., she saw her husband lying on the floor in a pool of blood. He was still alive, and she attempted to treat [*132]
him. He was taken by ambulance to a hospital but died the next day. Edgar Brown's physician testified that he died
from blows to the head. A bent fireplace poker, stained with type O blood, the same type as the victim's, was found in
the house. n57
In addition to Willie Orange and Foy Hortman, whose testimony implicating Adams has been reviewed above,
three other witnesses testified that they saw Adams or his car at, or in the immediate vicinity of, the Brown ranch on the
morning of the murder. n58 In addition, on the afternoon of the murder, Adams took his brown Rambler to an auto
body shop and asked to have the car painted another color. The police spotted the car as the one that had been identified
leaving the scene of the crime and later arrested Adams driving a friend's Oldsmobile. When he was arrested, Adams
had nearly $ 200 in his pockets, mostly in ten-and twenty-dollar bills. One of the twenty-dollar bills was stained with
type O blood. When Adams was asked about the blood on the money, he said that it came from a cut on his finger. His
blood was type AB, however, while the victim's was type O. n59 Police found a pair of pants, a striped shirt, a pair of
eyeglasses, a blue suit, two rings, and a gold watch in the car's locked trunk. The pants and shirt both had type O blood
on them. The optician who made the glasses identified them as Edgar Brown's; the watch and rings were identified as
coming from Brown's house. Adams acknowledged that the clothing was his and claimed that the jewelry "was stuff
that he had acquired over a period of time." n60 Further, the owner of the construction company where Adams worked
testified that Adams had failed to show up for work on the day of the murder. n61 Finally, Adams's possession of the
money appeared suspicious. The construction company paid him only $ 2.50 an hour, and Adams's neighbor had made
two loans to him totalling $ 35 shortly before the murder. n62
Adams's alibi defense was that he had not driven down the road in front of the Brown ranch on the day of the
crime. He said that he went to a bar to have a beer at approximately 9:00 a.m., then went to Vivian Nickerson's home,
arriving shortly after 10:00 a.m. At approximately 10:15 a.m., he said, he lent his car to Nickerson, who left for
approximately thirty-five to forty-five minutes to buy cards. That afternoon, he added, he put some jewelry and some of
his clothes in the Oldsmobile's trunk. He denied, however, that he had ever seen the rings or watch [*133] identified
as coming from the victim's house. Contrary to his earlier statement to the police, he also denied owning the pants and
striped shirt on which the blood was found. Finally, he stated that he had never seen the victim's glasses before. n63
Adams's testimony was inconsistent with his earlier statement and was inadequate to explain his possession of the
property taken from the victim's home.
In addition, Adams's principal alibi witness contradicted him on the critical issue of his whereabouts at the time of
the crime. The witness, Vivian Nickerson, testified that Adams arrived at her home at approximately 11:15 a.m., at
least twenty minutes after the murder had been committed, and that Adams also had a "strange looking face" that day.
n64

Page 7
41 Stan. L. Rev. 121, *133

To summarize, consideration of all the evidence demonstrates beyond a reasonable doubt that Adams was guilty of
the crime for which he was executed. The authors are able to maintain that Adams was the victim of a "miscarriage of
justice" only by resorting to a distorted view of the evidence, derived solely from the defendant's brief before the
Clemency Board, and by ignoring the compelling evidence directly linking him to the murder. n65 Moreover, as is
shown by the following brief review of some of the other alleged examples of erroneous executions, the authors'
misrepresentation of the Adams case is not an aberration. They show similar disregard of the evidence in many other
cases.
3. Other alleged erroneous executions.
The authors state that "[c]ritics who seize on the weakest cases and generalize from them to the whole set will miss
the forest for a few trees." n66 Our examination reveals, however, that there are not enough "trees" to make up a copse,
let alone a "forest." In evaluating Bedau and Radelet's conclusions, we focus on cases in which the authors assert
[*134] that innocent people have actually been executed, because they are the most relevant to evaluating the
magnitude of the risk of wrongful executions. n67 We narrow the focus further to eleven of the remaining twenty-two
cases in which the authors claim that an innocent person was executed. We do so only because appellate opinions for
them set forth the facts proved at trial in detail sufficient to permit a neutral observer to assess the validity of the
authors' conclusions. n68 Examination of these opinions, supplemented in some instances by reference to the authors'
sources and to other materials, demonstrates the weakness of the authors' methods and conclusions. n69
EVERETT APPELGATE. Appelgate and a codefendant, Frances Creighton, were executed in 1936 for having
murdered Appelgate's wife with rat poison. The authors' claim that Appelgate was innocent, stripped of irrelevancies,
rests on the incorrect assertion of an earlier author that "virtually no evidence against Appelgate existed beyond Mrs.
Creighton's unsupported word." n70 Even if that were true, which it is not, it would not compel the conclusion that
Appelgate was innocent. More to the point, questions of credibility were for the jury to resolve. The proof at trial
certainly entitled the jury to credit the testimony of Creighton. She had admitted her guilt, but claimed that she had
acted at the behest and under the domination of Appelgate, who had threatened to cause trouble for her and her husband
by publicizing their past involvement in a case of suspected murder. As the New York Court of Appeals put it after
reviewing the evidence:
[T]he jury were justified in finding [Appelgate] guilty. He had reason [*135] to fear his wife. His motive for
disposing of her is apparent. Ruth [Creighton's daughter] and he were in love with each other. . . .
. . . Appelgate evidently wanted to marry Ruth. Marriage was talked over and Ruth was crying about it before Mrs.
Appelgate was in her grave. The conduct of the defendant showed that he had no fear of consequences and at least no
love or affection for his wife. She had threatened to expose him. He had committed a state's prison offense. The night
the rat poison was bought he went to the drug store with Mrs. Creighton. He admits giving her money for the price of
the purchase. He admits giving his wife the eggnog and taking a very long time to do it, an hour or more before she
could get it down. He kept feeding it to her. He declined to have a post mortem and was not at all interested in whether
a crime had been committed. n71
Creighton's admitted involvement in the crime permits only two possible explanations of the murder: Either
Creighton, acting alone, killed Mrs. Appelgate; or Creighton and Appelgate, acting in concert, killed her. Together with
other evidence, the proof related to motive amply supported the jury's acceptance of the latter hypothesis. Evidence
showed that Creighton disliked both Appelgate and his wife and wanted them to move out of her house. Those facts,
however, hardly supplied Creighton with a motive to murder Mrs. Appelgate, particularly considering that it was not
Mrs. Appelgate, but her husband, who was blackmailing Creighton. In contrast, Appelgate had two powerful motives
for wanting his wife dead: She had threatened to expose his unlawful sexual relationship with Creighton's minor
daughter Ruth, and she stood in the way of his marriage to Ruth.
In short, the authors have presented little to persuade a neutral observer that Appelgate was not guilty of murdering

Page 8
41 Stan. L. Rev. 121, *135

his wife. Beyond that, their claim that Appelgate was innocent is discredited by the fact that two of the sources they cite
to support Appelgate's innocence actually believed he was guilty. n72
[*136] SIE DAWSON. Dawson was executed in 1964 for murdering a two-year-old boy in the course of a lethal
assault on the boy's mother. The principal evidence against him was a confession which he repudiated at trial. He
claimed that he had confessed only to avoid being turned over to a mob and that the woman's husband had committed
the murders. Bedau and Radelet base their conclusion that Dawson was innocent on a belief that his confession was
false and that the husband was the true culprit, and on the fact that "[y]ears later, newspaper stories revived doubts that
had surrounded the conviction from the beginning." n73 Since the jury that convicted Dawson did not share these
doubts, the critical questions in reassessing his guilt or innocence are the truth of his confession and the plausibility of
his later claim that another person was the killer.
With respect to the first question, the authors tell us that Dawson had a low IQ and that he confessed after spending
more than a week in custody. What they do not tell us is that Dawson "stoutly maintained his innocence" for five days
after the alleged threats to turn him over to the mob, n74 that he admitted the murder only after learning that an
eyewitness to the crime was still alive, n75 that he repeated his confession twelve days later, n76 and that every detail
of the confession except the actual commission of the crime was corroborated by other witnesses' testimony. n77 As to
the claim that the woman's husband was the murderer, the authors also fail to mention that "other witnesses accounted
for the whereabouts of [the husband] on the fatal day with such minute detail that it would have been almost an
impossibility for him to have committed the criminal act." n78
[*137] STEPHEN GRZECHOWIAK AND MAX RYBARCZYK. These defendants were executed in 1930
following their convictions for a murder they committed together with Alexander Bogdanoff during a robbery. Bedau
and Radelet believe a neutral observer would find them not guilty because they maintained their innocence "[i]n their
final words" and because Bogdanoff exonerated them at trial and again shortly before his execution. n79 However, the
opinion of the New York Court of Appeals belies the claim of innocence. n80 It begins with the following passage:
Evidence which is not challenged establishes that on the 27th day of July, 1929, Ferdinand Fechter was killed in the
city of Buffalo, Erie County. His death was the result of wounds inflicted by robbers who escaped with a large amount
of money which Fechter was carrying from a bank to his home. On September 7, 1929, the defendant Rybarczyk was
arrested. Early the next day the other two defendants were arrested. At the police station they were questioned
separately and together. Two made written statements admitting that they had taken part in the robbery. The defendant
Rybarczyk made no written statement, but the evidence establishes that he admitted before a number of witnesses that
he had taken part in the robbery and that the written statement of Grzechowiak, made in his presence, was true. n81
Since the defendants confessed, the court was justified in observing that the "evidence . . . is clearly sufficient to
sustain a conviction." n82 On that point, even the dissenting justice -- who disagreed with the majority's [*138]
rejection of the defendants' belated challenge to the sufficiency of the indictment -- concurred: "The defendants
committed an atrocious murder, and are clearly guilty." n83 Bedau and Radelet offer no persuasive reason for
accepting the defendants' protestations to the contrary. n84
JOSEPH HILLSTROM ("JOE HILL"). Hillstrom was executed in 1915, having been convicted of shooting a
storekeeper to death. Bedau and Radelet conclude that he was innocent because "[t]he prosecution was based on
sketchy circumstantial evidence and was in part the result of collusion between the prosecution and the trial judge in an
atmosphere of anti-union hostility." n85 In fact, the evidence, which was reviewed in great detail by the Utah Supreme
Court, n86 admitted of no rational conclusion other than guilt.
Shortly before the shooting, a witness saw a man on a brightly lit street outside the victim's store. She testified that
he was similar to the defendant in size and features, that he had the same slim face, sharp nose, and large nostrils, and
that he had the same scar on the side of his face and neck. A witness to the shooting testified that the murderer was
about the same size as the defendant, had a similarly shaped head, and wore clothing like the defendant wore that night.

Page 9
41 Stan. L. Rev. 121, *138

Following the murder, during which the victim shot the killer in the chest, two witnesses saw a man, whose appearance
and voice were similar to the defendant's, run from the scene of the crime. In addition, the proof showed that the
defendant sought medical attention for a fresh bullet wound in his chest on the night of the murder. The bullet was fired
from a gun of the same caliber as that used by the victim to shoot his assailant. The defendant told the doctor that the
wound resulted from a fight over a [*139] woman and asked him not to reveal the matter to others. Finally, when he
was arrested several days later, the defendant admitted having thrown away his gun shortly after leaving the doctor's
home. n87
While it is true that no witness positively identified the defendant as the murderer, no such proof was necessary in
light of the defendant's unexplained gunshot wound which, as the Utah Supreme Court aptly observed, was "quite as
much a distinguishing mark as though [the murderer] in the assault had one of his ears chopped off." n88 One
concurring judge stated: "[T]he jury were authorized to conclude -- indeed, it is not easy to perceive how rational men
could have arrived at any other conclusion -- that the appellant was, in fact, shot in Morrison's store at the time of the
homicide." n89 Maintaining the defendant's innocence is difficult in the face of this and the other evidence
summarized above. n90
MAURICE MAYS. The authors rely on two arguments to support their belief that this defendant was wrongly
convicted and executed in 1926 for murdering a woman in her bedroom during a robbery: (1) "The conviction rested on
the testimony of a police officer who had disliked Mays for years and on the testimony of an eyewitness who never got
a clear look at the killer," and (2) the "real killer" confessed several years after Mays was executed. n91 Neither
argument can withstand scrutiny.
The Tennessee Supreme Court opinion gives a more complete and accurate account of the evidence presented at
May's second trial for the murder. n92 A witness who had been sleeping in the same room as [*140] the victim, and
who had no motive to identify Mays falsely, testified that Mays was the man who fired the fatal shot. The lighting
conditions in the room were good, and she was positive of her identification because she spoke with him after the
shooting in a successful effort to persuade him not to assault her but to take money instead. Shortly after the crime she
described the murderer to the police with such specificity that they immediately realized that Mays was the man. They
went promptly to his house, where they found him, pretending to be asleep, in possession of a .38-caliber revolver, the
same type of gun used in the murder. It appeared that one bullet had recently been fired from the pistol. In addition,
Mays was wearing clothing and shoes the condition of which indicated that he had recently been in the vicinity of the
house in which the victim was killed. Police then took Mays before the eyewitness who, with no prompting by the
police, immediately, positively, and repeatedly identified him as the killer. Mays's defense was an alibi, but he failed to
account for his whereabouts at the time the murder was committed. n93 He also introduced testimony of a witness who
claimed that the arresting officer was out to get Mays, but the officer denied this. The appellate court noted that the
issue was immaterial since there was no evidence that the officer tried to influence the eyewitness's identification of
Mays in any way. n94
Finally, the notion that the person who later confessed to the murder was the true killer is baseless. According to
Bedau and Radelet, this person was a white woman, who claimed to have dressed up as a black man and said she
committed the murder because the victim was having an affair with her husband. n95 Despite their recognition that
confessions of third parties are often "false or dubious" and "must be evaluated with extreme caution," n96 the authors
provide no justification for believing this bizarre confession to be true. Although they do note that the authorities never
accepted it as true, n97 they fail to state the reasons for this official reaction: The woman's account of the murder
contained serious discrepancies, and her husband explained to the police that she had confessed falsely because she was
temporarily demented. n98 Given these facts, as well as the confession's implausibility in light of evidence which
convinced twenty-four jurors at two trials that Mays was guilty beyond a reasonable doubt, the woman's confession
provides no sound basis for crediting the authors' claim that Mays was innocent.
[*141] WILLIE MCGEE. McGee, a black man convicted by three separate juries of raping a white woman in
1945, was executed in 1951. n99 According to Bedau and Radelet, "[t]he chief evidence against McGee was a coerced

Page 10
41 Stan. L. Rev. 121, *141

confession that he gave after being held incommunicado for thirty-two days after his arrest." n100 The authors
apparently believe that this confession was false, because "the victim's husband and her two children, asleep in the next
room, never heard any commotion from the alleged attack," and because a journalist's subsequent investigation
"revealed that the victim had been consorting with McGee for four years and was angry at his efforts to terminate their
relationship." n101 They thus imply that she accused him of rape to make him pay for rejecting her.
Resort to the opinion of the Mississippi Supreme Court, which unanimously affirmed McGee's third conviction,
n102 shows that these allegations provide no basis for concluding that the jury convicted McGee mistakenly. First, with
respect to the confession, McGee made "quite a number of admissions of guilt of raping the woman," the first of which
occurred as he was being taken to, and while he was at, the crime scene the day after the rape. n103 At that time,
he stated he raped his victim; pointed out the house . . . where the event occurred; pointed out where he parked his
truck after seeing her lying on the bed, through the windows, before the light was extinguished; and other details -- all
corroborated by the physical facts and oral testimony of witnesses. n104
According to the appellate court, "[t]he proof overwhelmingly support[ed] the State's contention that these
statements were voluntarily made without duress or promise." n105
The opinion is also more informative than Bedau and Radelet with respect to the question of consent. It appears
that the victim did not alert her husband to the attack because
The woman was terrified, and resisted as best she could, even calling to her husband, who was on a sleeping porch
at the back of the house, [*142] and who could not be awakened, but ceased her resistance when [McGee] threatened
to cut her throat, and more poignantly to her, the throat of her ill infant child. n106
McGee's oral statement to the police following his arrest also belies consent. He implied that he had decided on the
spur of the moment to attack a woman who was a stranger to him. n107 Other proof that the victim did not know him
also existed. n108
ALBERT SANDERS. Sanders, a cab driver, was executed in 1918 for having participated with Fisher Brooks in
the robbery-murder of a female passenger the year before. The authors support their belief in Sanders's innocence only
with Brooks's testimony that Sanders was not involved in the crime, a claim that both men repeated before they were
hanged. n109 But the authors present no information that would lead a neutral observer to credit these claims. Indeed,
they omit from the description of the case several critical facts relevant to Brooks's credibility. First, they fail to
mention that Brooks's exonerating testimony was given after he himself had been convicted at a separate trial, and,
therefore, when he had nothing to lose by clearing Sanders. n110 Second, they overlook Brooks's admission at
Sanders's trial that he had told so many different stories about the murder that he was unable to recall what he had said
previously. n111 Third, they omit other facts making the exoneration [*143] unbelievable, including Brooks's
admission at Sanders's trial that he had previously told the authorities that Sanders had participated with him in the
commission of the crime. n112
But the case against Sanders consisted of far more than discrediting Brooks. It included testimony by a fellow
prisoner of Sanders and Brooks who overheard a conversation between the two men in which Sanders acknowledged his
guilt. n113 It also included persuasive circumstantial evidence of Sanders's complicity. The proof showed that Sanders
knew the victim had money, that the victim was stabbed, that Sanders had blood on his fingernails when he was arrested
shortly after the crime, that he had cut pieces from, and then hidden, the clothes he had worn on the night of the murder,
that bloodstains were found on those pieces and on his shoes, that he and Brooks were seen in close proximity shortly
before and after the murder, and that Sanders's alibi defense was concocted after the event. n114
CHARLES LOUIS TUCKER. Tucker was executed in 1906, having been convicted of stabbing a young woman to
death during a robbery. Although Bedau and Radelet disparage the evidence against Tucker as "circumstantial only,"
n115 the Massachusetts Supreme Judicial Court opinion and one of the authors' cited sources show that the jury was

Page 11
41 Stan. L. Rev. 121, *143

fully warranted in finding him guilty. n116 Proof of guilt included Tucker's bloodstained knife, which he first denied
owning but later admitted having broken into pieces in order to avert police suspicion, and which fit the wounds on the
victim's body and tears in her corset; a slip of paper bearing Tucker's handwriting, found near the body; and a pin
belonging to the victim, found concealed in Tucker's coat pocket. Furthermore, [*144] the unemployed Tucker had a
ten dollar bill shortly after the murderer took a similar bank note from the victim's purse. n117 In his defense, Tucker
relied on an alibi; but he did not testify himself, n118 and his alibi witness was apparently impeached with prior
inconsistent statements. n119
Bedau and Radelet do not explicitly contend that there was insufficient evidence of Tucker's guilt. n120 Their
position apparently is that the circumstantial evidence of guilt is outweighed by the fact that "[m]ore than 100,000
Massachusetts residents signed petitions on behalf of clemency" and that "[a]mong those convinced of [Tucker's]
innocence was the county medical examiner (who lost his job because of his stand) and a clergyman who said a witness
had told him she had perjured herself at the original trial." n121 The clemency petitions were not, however, the product
of legitimate doubts as to Tucker's guilt. They were the result of an irresponsible campaign by partisan journalists
whose reports of the case included only small parts of the state's evidence, falsely leading the public to believe that there
was not sufficient evidence to convict Tucker. n122 Finally, neither Bedau and Radelet, nor the sources they cite,
provide sufficient basis for crediting the medical examiner's and the clergyman's views instead of the conclusions of the
jurors who convicted Tucker, the judges who presided at his trial, and the Governor who rejected the clemency petitions
after a personal review of the case. n123
[*145] The foregoing discussion makes three pertinent points regarding Bedau and Radelet's study. First, the
authors have employed an unacceptable standard for determining innocence. It gives no weight at all to the considered
judgment of the juries and judges who decided and reviewed the cases. Second, in the cases we have reviewed, the
authors have consistently presented incomplete and misleading accounts of the evidence. Third, in each of the cases in
which a reasonably complete account of the facts is readily available, the authors' claim that the defendant was later
"found" or "proven" to be innocent is unconvincing. These critical flaws seriously diminish the value of the study to the
debate over the significance of the risk of erroneous execution.
II. FLAWS IN THE ARGUMENT FOR ABOLITION
Even if Bedau and Radelet's study could be taken at face value, it would not provide a rational basis for rejecting
capital punishment. The danger of executing the innocent, as well as the uniquely irremediable nature of such a
mistake, has long been recognized by death penalty proponents, n124 as well as by opponents. n125 Opponents
contend that society should eschew the use of the death penalty entirely because of this danger. Therefore, a response to
the study would be incomplete if it did not address the merits of that position.
There are two principal difficulties with the argument that we should not use capital punishment because of the risk
of executing an innocent person. First, it gives excessive weight to the slight risk of erroneous executions. This is
particularly so in light of new procedural protections governing imposition of the death penalty, for Bedau and Radelet
cannot provide even one credible example of an erroneous execution in the past fifteen years. Second, the argument
fails to consider the countervailing benefits of capital punishment. A fair assessment of risk and benefits demonstrates
that society has decided reasonably to make the death penalty available for certain offenses. n126
[*146] A. The Risk of Erroneous Execution
The use of capital punishment entails some risk that an innocent person will be executed. n127 Proponents of the
death penalty have commonly taken the position that this risk is not great. n128 The Bedau-Radelet study confirms this
view. After "sustained and systematic" research, n129 they point to only twenty-three out of more than 7000
executions as erroneous. Their judgments of error are unconvincing with respect to the twelve cases in which sufficient
facts are available to test them. Their methodology makes their conclusion in the other eleven cases suspect as well.
n130 Assuming, however, for the sake of argument that as many as twenty-three innocent persons have been executed,

Page 12
41 Stan. L. Rev. 121, *146

the rate of error would be only about one-third of one percent over the past eighty-seven years. n131 Moreover, even
accepting the authors' claims, this miniscule rate of error has been reduced by more than one-third since 1943. n132
But, say Bedau and Radelet, "[e]valuating the argument against the death penalty based on the fact that innocent
defendants have been and will be executed requires some care." n133 This is so, they add, because "[o]ur total of
twenty-three wrongful executions is not an estimate, and it cannot serve as a basis for a reasonable estimate, of the total
number of wrongful executions in the United States during this century." n134 The true number is far greater, they
hint; n135 but "any attempt to calculate the odds of executing the innocent . . . is doomed to fail." n136 It [*147]
must be remembered, however, that Bedau and Radelet are the ones who argue that existing law should be changed.
n137 Accordingly, they bear the burden of justifying their argument by producing credible evidence that the use of
capital punishment entails a significant risk of executing the innocent. This they have failed to do, and their assertion
that the task is impossible is no substitute. Responsible social policy should be based on the best information available,
not -- as the authors apparently would have it -- on speculation as to what the information would show if it were
available.
B. Safeguards Against Erroneous Execution
The safeguards that exist to protect criminal defendants, especially those charged with capital crimes, lessen the
risk of erroneous execution. In particular, a number of statutory protections have arisen since the Supreme Court's
decision in Furman v. Georgia. n138 Although the specific provisions of statutes governing capital punishment vary
among jurisdictions, these statutes generally share a number of features, most of which are intended to safeguard against
erroneous imposition of the death penalty. The important provisions of a typical death penalty statute, and of related
law, in the United States are as follows. n139
First, the statutes limit imposition of capital punishment to the most serious offenses against society -- intentional
crimes that involve the taking of human life. n140
Second, the defendant must be notified in advance of the state's intention to seek the death penalty and of the
aggravating factors the state proposes to prove as justifying a sentence of death, n141 so that the defendant will have a
sufficient opportunity to make ready a defense to the charge and to prepare for the post-trial penalty hearing if found
guilty.
Third, if the trial results in a guilty verdict, a post-verdict hearing must be held to determine the penalty. The
hearing is conducted before a jury if the defendant wishes, and he or she is entitled to continued representation by
counsel. n142
Fourth, in order to justify the death penalty's imposition, the state must prove beyond a reasonable doubt the
existence of at least one [*148] statutorily specified aggravating factor that is not outweighed by any mitigating
factors. n143 Aggravating factors specified in the statutes generally include commission of the offense by use of
torture, n144 in a manner that created a great risk of death to others, n145 by a person who had previously been
convicted of a serious violent crime, n146 for money, n147 and against certain designated public officials. n148
Typical mitigating factors include the defendant's youth, n149 impaired mental capacity, n150 commission of the
offense under unusual and substantial duress, n151 and playing the role of a relatively minor subordinate to the
principal offender in the crime's commission. n152
Fifth, the jury must unanimously determine, on the basis of its findings concerning aggravating and mitigating
factors, whether the death penalty is justified, or whether a lesser penalty should be imposed instead. n153 The judge
may not impose a death sentence unless the jury recommends it. n154
Sixth, the state's highest court automatically reviews all death sentences to ensure that there is an adequate legal
basis for imposing the death penalty, that the sentence was not imposed as a result of passion or prejudice, and that the
sentence is not excessive or disproportionate to the penalty imposed in similar cases. n155 In some cases, direct review

Page 13
41 Stan. L. Rev. 121, *148

is also available in the Supreme Court. n156
Seventh, the defendant may also seek a new trial on the basis of newly discovered evidence that she or he was not
guilty. n157 In addition, both state and federal laws allow a person whose capital conviction has been upheld by the
highest court of the state to obtain collateral review in the state courts, in the lower federal courts, and in the Supreme
Court. n158 After all judicial remedies have been exhausted, a condemned defendant may seek executive clemency.
n159
An example of the use of these extraordinary procedural protections in capital cases is the Adams case discussed
above. n160 Adams was [*149] convicted in 1974, and the judge followed the jury's recommendation by imposing a
death sentence. n161 The judgment and sentence were affirmed by the Florida Supreme Court, n162 and the United
States Supreme Court denied certiorari. n163 Pursuant to the rule of Gardner v. Florida, n164 the Florida Supreme
Court directed the trial judge to state whether he had imposed the death sentence on the basis of any information not
known to petitioner. The trial judge responded that he had not. The Florida Supreme Court denied relief on that basis,
n165 and the United States Supreme Court denied certiorari. n166 After the Florida Supreme Court resolved a dispute
concerning the appointment of counsel for Adams's post-conviction relief motion, n167 the trial court denied Adams's
first motion for post-conviction relief. The Florida Supreme Court affirmed. n168
Adams then sought federal habeas corpus relief in the United States District Court for the Southern District of
Florida. In a lengthy order, the court denied the petition. n169 The United States Court of Appeals for the Eleventh
Circuit affirmed, n170 and the United States Supreme Court refused to grant certiorari n171 and denied rehearing.
n172
In 1984, Adams filed a second motion for post-conviction relief which the state trial court denied. n173 On May 2,
1984, the Florida Supreme Court affirmed the denial in a written opinion. n174 The court also denied Adams's petition
for habeas corpus and his application for a stay of execution. n175
Adams then filed a second petition for habeas corpus relief in the federal district court, raising issues that had
already been litigated in his [*150] 1983 federal habeas corpus petition, which was also denied. n176 An Eleventh
Circuit panel granted a stay of execution pending appeal, but the Supreme Court vacated the stay. n177
Meanwhile, on May 1, 1984, Adams filed an application for executive clemency with the Florida Governor's
Office. The Clemency Board reviewed that application and denied it following oral presentations by representatives for
Adams and for the state. n178
Thus, over a 10-year period, various aspects of Adams's conviction and sentence were reviewed on at least twelve
separate occasions by numerous state and federal judges, as well as by the Governor's Clemency Board. This history,
typical of death penalty cases, n179 shows that procedural safeguards are adequate to prevent erroneous executions.
n180 It also exposes as baseless Bedau and Radelet's suggestion that Adams was the victim of a rush to judgment.
n181
C. The Effect of Safeguards in Recent Years
According to Bedau and Radelet, ninety-four defendants have been the subjects of "miscarriages of justice" in
capital or potentially capital cases since 1960. n182 However, only twenty-six of these persons were sentenced to
death. n183 Only two were executed, and the authors give unconvincing reasons for concluding that these executions
were erroneous. n184 There is, in short, no persuasive evidence that any innocent person has been put to death in more
than twenty-five years. Equally significant, Bedau and Radelet do not include in their catalogue any individual on death
row at the end of 1986. Although they seem to think that some of these persons were innocent, n185 they identify only
two such persons. n186 One of these had his conviction reversed and sentence [*151] vacated in 1987; n187 the
other was plainly guilty and was executed in 1987. n188

Page 14
41 Stan. L. Rev. 121, *151

The likely explanation for the absence of errors in capital cases during the past quarter century is the greater care
taken by the courts to assure the correct resolution of such cases and, particularly, the pains-taking reviews that occur in
cases in which the death sentence is actually imposed. The point is illustrated by the very few "miscarriages of justice"
claimed by Bedau and Radelet to have occurred in the decade after the Supreme Court upheld the constitutionality of
the death penalty. n189 Out of approximately 50,000 murder convictions during the period from 1977 to 1986, n190
the authors point to only thirty-one persons who, they claim, were wrongly convicted of capital offenses or potential
capital crimes. n191 Even if one accepts their claim that all of these convictions were mistaken, the authors' accounts
of these cases demonstrate that current post-conviction procedures work reasonably well in discovering and correcting
errors. After all, in each of these cases the mistake was discovered. Furthermore, in more than three-fourths of the
cases the error was discovered within four years of conviction, resulting [*152] in the defendant's acquittal or release
from prison, or in the dismissal of charges; and in eight of the cases the error was discovered within one year. n192
D. The Benefits of Capital Punishment
A thorough inquiry into the significance of the risk of executing an innocent person in the debate over capital
punishment should attempt to evaluate capital punishment's benefits. Bedau and Radelet simply despair of this effort,
concluding that "since there is no consensus in our society over the weight to be assigned to each variable in [the death
penalty] equation, we are doomed to disagree over how to compute it." n193 However, an assessment of the costs and
benefits leads to the conclusion that the minute risk of executing an innocent person is substantially outweighed by the
protection that capital punishment affords to society through incapacitation, deterrence, and just punishment.
1. Incapacitation.
The authors' assertion that we are "doomed to disagree" over the weight to be assigned variables in a capital
punishment cost-benefit evaluation is incorrect. We can compare the number of lives of innocent persons saved or lost
through the imposition and the abolition of the penalty. As a cost, Bedau and Radelet claim that twenty-three innocent
persons have been executed in this century. Even accepting this figure as an accurate representation of the penalty's
cost, n194 the number of innocent persons whose lives have been saved through the penalty's incapacitative effects
outweighs it.
Capital sentences, when carried out, save innocent lives by permanently incapacitating murderers. Some persons
who commit capital homicide will slay other innocent persons. n195 The death penalty is the most effective means of
preventing such killers from repeating their crimes. The next most serious penalty, life imprisonment without
possibility of parole, prevents murderers from committing some crimes but does not prevent them from murdering in
the prison environment. n196 Furthermore, [*153] some killers who commit capital murder eventually obtain parole.
n197
While it is impossible to determine precisely how many innocent lives the execution of convicted murderers has
saved, the available data suggest that the number is not insignificant. Of the roughly 52,000 state prison inmates
serving time for murder in 1984, an estimated 810 had previously been convicted of murder and had killed 821 persons
following those convictions. n198 Executing each of these inmates following the first murder conviction would have
saved 821 lives over the last several decades. Of course, as Bedau and Radelet point out, n199 only a fraction of
convicted murderers receive the death penalty. Therefore, the number of innocent lives saved would be substantially
smaller, although still more than the number of "innocent" persons Bedau and Radelet claim have been executed in this
century. n200
Bedau and Radelet's response to this arithmetic is inadequate. They assert that with the abolition of the mandatory
death penalty, "society runs both the risk of executing the innocent and the risk of recidivist murder, whereas it is only
necessary that society run one or the other." n201 Because of what they believe to be a "low incidence of recidivism,"
n202 they urge that society not run any risk of executing an innocent [*154] person. This response is curious because
it postulates that society has only two extreme choices: executing all murderers or executing none. In fact, society has

Page 15
41 Stan. L. Rev. 121, *154

many intermediate options, such as executing murderers only after a careful assessment of the aggravating and
mitigating factors, n203 or executing murderers after a finding of "dangerousness." n204 Because the middle course
approaches chosen by society serve important goals besides incapacitation (e.g., satisfying constitutional requirements,
allowing the possibility of mercy), it is hardly objectionable that they do not focus solely on maximizing the
incapacitative benefit of capital punishment.
2. Deterrence.
While the innocent lives saved through the incapacitative effect of capital punishment are important, the penalty
has also saved innocent lives through its general deterrent effect. Logic supports the conclusion that the death penalty is
the most effective deterrent for some kinds of murders -- those that require reflection and forethought by persons of
reasonable intelligence and unimpaired mental faculties. It stands to reason that capital punishment deters such persons
more than the next most serious penalty, life imprisonment without parole. As the Supreme Court observed in Gregg v.
Georgia, "There are carefully contemplated murders, such as the murder for hire, where the possible penalty of death
may well enter into the cold calculus that precedes the decision to act." n205
Statistical studies also suggest that capital sentences, like other criminal sanctions, have a deterrent effect. To be
sure, some statistical surveys, often conducted by opponents of the death penalty, have found no such effect. A detailed
review undertaken by the Department of Justice concluded, however, that few, if any, of these studies relied on rigorous
methodologies or adequately controlled for many variables that affect the homicide rate in the jurisdictions under
consideration. n206 Moreover, [*155] it appears to be common ground in the deterrence literature that a statistically
valid study should account not only for the response of criminals to penalties imposed by an outside authority (the
so-called "demand" for crime) but also for an outside authority's response to changes in crime (the so-called "supply"
for crime). n207 The "supply-demand" econometric studies that have been done to date support the conclusion that the
death penalty deters homicide.
The most recent substantial econometric study was performed by Professor Stephen K. Layson of the University of
North Carolina at Greensboro, who analyzed data for the United States from 1936 to 1977. n208 Layson concluded
that increases in the probability of execution reduced the homicide rate. Specifically, Layson found that, on average,
each execution deterred approximately eighteen murders. Layson's study of the United States data is consistent with his
earlier study concerning the deterrent effect of capital punishment in Canada n209 and with important empirical work
in this area by Isaac Ehrlich n210 and other scholars. n211 These econometric studies are buttressed by a growing
body of literature demonstrating that punishment has a deterrent effect on crime in a wide variety of settings, n212
including "cohort" studies using data on particular individuals rather than aggregate crime rates. n213 Indeed, the
premise that more severe penalties deter more serious crimes is fundamental to our criminal justice system.
If Layson's empirical work (which is strongly supported by deterrence theory) is correct, we can estimate that the
death penalty has deterred [*156] roughly 125,000 murders in this country in this century. n214 This figure dwarfs
the twenty-three innocent persons Bedau and Radelet claim have been executed in the same time period. More
important, it demonstrates rather starkly that under any realistic risk assessment the presence of capital punishment
saves more innocent lives than it jeopardizes. n215
Bedau and Radelet also assert that because no innocent defendant ever "consents" to running the risk of an
erroneous execution, the death penalty should not be imposed despite its deterrent effects. n216 Accepting this
reasoning would cast in doubt the legitimacy of a host of socially useful governmental enterprises. For example, one
could argue that the nation's prisons should be closed because innocent defendants have not consented to the risk of
erroneous imprisonment. The consent notion has not been adopted as a guiding principle for government action in the
criminal justice area because of its theoretical boundlessness. n217 Instead, our society's choices with respect to
sentencing and other justice-related matters are based on the benefits and hazards of the various options available.
n218

Page 16
41 Stan. L. Rev. 121, *156

[*157] 3. Just Punishment.
Through the imposition of just punishment, civilized society expresses its outrage and sense of revulsion toward
those who, by contravening its laws, have not only inflicted injury upon discrete individuals, but also weakened the
bonds that hold communities together. Certain crimes constitute such outrageous violations of human and moral values
that they demand retribution. It was to control the natural human impulse to seek revenge and, more broadly, to give
expression to deeply held views that some conduct deserves punishment, that criminal laws, administered by the state,
were established. The rule of law does not eliminate feelings of outrage but does provide controlled channels for
expressing such feelings. People can rely on society to sanction criminal conduct and to carry out deserved punishment.
The law's acceptability and effectiveness as a substitute for self-help depends, however, on the degree to which society's
members perceive the law as actually providing just punishment for particularly serious criminal offenses. The
intentional and unjustified taking of an innocent person's life is such an offense. As the Senate Judiciary Committee has
recognized, "[m]urder does not simply differ in magnitude from extortion or burglary or property destruction offenses;
it differs in kind. Its punishment ought to also differ in kind. It must acknowledge the inviolability and dignity of
innocent human life. It must, in short, be proportionate." n219 Determining what sanction is proportionate and,
therefore, what constitutes just punishment for committing certain types of murder is admittedly a subjective judgment.
Nevertheless, when there is widespread public agreement that the death penalty is a just punishment for certain kinds of
murder -- as there is in this country today n220 -- and when a jury acting under standards meeting constitutional
requirements determines that a particular person has killed another [*158] under circumstances for which the
legislature has found that death is the appropriate penalty, the resulting judgment is no less "just" because its validity
cannot be verified objectively.
The death penalty's retributive function thus vindicates the fundamental moral principle that a criminal should
receive her or his just deserts. Even if capital punishment had no incapacitative or deterrent utility, its use would be
justified on this basis alone. But as discussed above, n221 there is ample evidence that the death penalty protects
innocent lives through incapacitation and deterrence. Equally important, through the provision of just punishment,
capital punishment also affirms the sanctity of human life and thereby protects it. Walter Berns has traced this process.
n222 To summarize his explanation: We punish in part because we are angry at what the criminal has done and want to
pay him back. The law respects and acknowledges the rightness of our anger (which expresses caring for others) when
it punishes the object of that anger. The law thus validates and promotes the law-abiding person's abhorrence of
criminality. In short, retribution has a value-reinforcing effect. It deters crime not only by instilling fear of punishment,
but also, through a process of rewarding the anger aroused by the sight of crime, by praising law-abidingness.
Of course, in order to protect lives through the operation of this value-reinforcing effect, retribution must be
appropriate to the offense committed. "So the question becomes," as Berns puts it, "how do we pay back those who are
the objects of great anger because they have committed great crimes against us?" n223 The answer Berns gives is
worth quoting at some length:
The purpose of the criminal law is not merely to control behavior -- a tyrant can do that -- but also to promote
respect for that which should be respected, especially the lives, the moral integrity, and even the property of others. In a
country whose principles forbid it to preach, the criminal law is one of the few available institutions through which it
can make a moral statement and, thereby, hope to promote this respect. To be successful, what it says -- and it makes
this moral statement when it punishes -- must be appropriate to the offense and, therefore, to what has been offended. If
human life is to be held in awe, the law forbidding the taking of it must be held in awe; and the only way it can be made
to be awful or awe inspiring is to entitle it to inflict the penalty of death. n224
[*159] One could argue that the execution of innocent defendants undermines the death penalty's capacity to
provide the incapacitation, deterrence, and just punishment necessary to protect the innocent lives of others. With
respect to incapacitation, for example, if an innocent person is executed for murder, the true culprit is likely to remain at
large, free to kill again. Likewise, the threat of capital punishment theoretically may not deter persons from murdering
if they perceive the threat to be directed indiscriminately at the innocent as well as at the guilty. n225 Infliction of the

Page 17
41 Stan. L. Rev. 121, *159

death penalty upon innocent persons is also clearly not just punishment.
The force of these arguments depends, however, on the frequency with which persons are executed for capital
offenses they did not commit and, to the extent that the arguments relate to deterrence and just punishment, on the
degree of public knowledge of the number of mistaken executions. If such miscarriages of justice are relatively
common, then capital punishment -- in addition to taking the lives of innocent defendants -- might well be ineffective in
protecting innocent life generally. On the other hand, if such occurrences are unusual, then these consequences are
unlikely. In fact, erroneous executions appear to be extremely rare if not nonexistent. The risk of executing the
innocent thus has little detrimental effect on the death penalty's protection of innocent lives.
E. The Duty of Society
The foregoing discussion of the manner in which the incapacitative, deterrent, and retributive functions of capital
punishment protect innocent life demonstrates the reasonableness of concluding that the death penalty's benefits
outweigh the chance of a mistaken execution. Given this judgment, it is imperative that society demonstrate the
courage of its convictions. Unless it is prepared to forsake murder victims, society must demonstrate its self-confidence
and steadfastness of purpose by imposing and carrying out the death penalty in cases that reasonably meet the law's
stringent requirements. The minimal but unavoidable risk of error in the administration of capital punishment must not
be allowed to induce a "failure of nerve" n226 that would paralyze society [*160] from taking the steps necessary to
protect its citizens.
Ultimately, the Bedau-Radelet study offers little to opponents of capital punishment. The eighty-four percent of
Americans who support the death penalty do so not because they believe its administration is perfect, but because they
believe it to be prudently administered in a manner consistent with the society's interest in justice and the protection of
the innocent. Nothing presented by Bedau and Radelet undermines that conviction. n227
Legal Topics:
For related research and practice materials, see the following legal topics:
Criminal Law & ProcedureCriminal OffensesHomicideMurderCapital MurderPenaltiesCriminal Law &
ProcedureCriminal OffensesHomicideMurderFirst-Degree MurderPenaltiesCriminal Law & ProcedureDefensesGeneral
Overview
FOOTNOTES:

n1. See Bedau & Radelet, Miscarriages of Justice in Potentially Capital Cases, 40 STAN. L. REV. 21, 27,
78-81 (1987).
n2. Id. at 23, 36.
n3. As discussed in detail below, even if one were to concur in the authors' claim that 23 innocent persons
have been executed during this century, that number is only about one-third of 1% of the total number of
executions that have taken place since 1900. See note 22 infra and accompanying text. Of course, even the
execution of one innocent person would be a tragedy and is hardly insignificant in that sense. But so far as the
debate over the death penalty is concerned, the question that must be addressed is whether the risk of such
tragedies is significant either in absolute terms or in terms of the balance between the costs and the benefits of
capital punishment. The authors of this article believe that the risk is insignificant, in the sense that it is
statistically minuscule and also in the sense that it does not affect the validity of our society's considered

Page 18
41 Stan. L. Rev. 121, *160

judgment that the benefits of capital punishment outweigh the costs.
n4. A response to the Bedau-Radelet study is necessary for practical reasons as well. According to the
authors, defense attorneys have been attempting, in some cases successfully, to use the study's findings during
the penalty phase of capital cases to buttress their arguments for imprisonment rather than death. See Bedau and
Radelet, supra note 1, at 83.
n5. See Bedau & Radelet, supra note 1, at 78-81.
n6. W. BERNS, FOR CAPITAL PUNISHMENT: CRIME AND THE MORALITY OF THE DEATH
PENALTY 178 (1979).
n7. See van den Haag, In Defense of the Death Penalty: A Legal-Practical-Moral Analysis, 14 CRIM. L.
BULL. 51, 67 (1978).
n8. Bedau & Radelet, supra note 1, at 22.
n9. Id. at 26.
n10. Id. at 23-24 (footnote omitted).
n11. Id. at 36, 72-73 & Tables 2, 10.
n12. Id. at 23-24.
n13. Id. at 31-33.
n14. Id. at 33. The arbitrary nature of the authors' methodology is evident from their unwillingness to carry
through on their own definition of potentially capital crimes. While the authors do include all homicides that
could not or did not result in the imposition of the death sentence, they only include rapes if the defendant was
sentenced to death. See id. at 33-34. This suggests that the authors are more than a little uncomfortable with the
idea that the risk factor for executing an innocent person should include cases in which the defendant could have
been but was not sentenced to death and cases in which the death penalty for the crime was unavailable in the
particular jurisdiction.
In addition, the authors chose not to include robbery, kidnapping, burglary, and treason -- crimes for which
a capital sentence has been permissible punishment. See id. at 33-34. Again, this omission suggests that the
authors are unwilling to accept the implications of their definition of the relevant offense pool. Moreover, many
of the reasons the authors give for excluding these crimes -- for example, that they are less relevant given
modern constitutional holdings and are not typically thought of as capital crimes -- would apply equally to many
of the "potentially capital cases" which they do include. See id. The exclusion of treason is especially peculiar
because it is a quintessential capital offense and was included in an earlier draft of the authors' study. The
reason for the exclusion may be to avoid the criticism that would result from their declaring the innocence of

Page 19
41 Stan. L. Rev. 121, *160

Ethel and Julius Rosenberg, whom they categorized earlier as having been executed erroneously despite
acknowledging that "the most recent reinvestigations of the 1953 Rosenberg case . . . report new discoveries that
tend to confirm the prosecution's case (at least against Julius Rosenberg). . . ." Bedau & Radelet, Miscarriages of
Justice in Potentially Capital Cases 16-17, Appendix I at 42-43 (first draft Oct. 15, 1985) (on file with the
Stanford Law Review) (citing R. RADOSH & J. MILTON, THE ROSENBERG FILE: A SEARCH FOR THE
TRUTH (1983)).
n15. Bedau & Radelet, supra note 1, at 36, Table 1.
n16. Id. at 36, Table 2.
n17. Id.
n18. Id. at 72-73, Table 10.
n19. See text accompanying notes 42-65 infra.
n20. 408 U.S. 238 (1972).
n21. Bedau & Radelet, supra note 1, at 36.
n22. As the authors recognize, their own figures show this risk to be exceedingly small--the 23 allegedly
wrongful executions they identify constitute only 0.33% of the approximately 7000 executions that have taken
place in this country since 1900. See id. at 73.
n23. Moreover, elimination of capital punishment would have no effect on the number of wrongful
convictions for otherwise capital offenses.
The authors contest our assertion that the bulk of their work is beside the point. They note that we do not
dispute their claim that a number of wrongful death sentences would have been carried out had it not been for
interventions having nothing to do with normal judicial review procedures or statutory safeguards. See Bedau &
Radelet, The Myth of Infallibility: A Reply to Markman and Cassell, 41 STAN. L. REV. 161, 163-64 (1988).
Since these death sentences were not carried out, it is unnecessary to address the contentions that the defendants
involved were innocent and that the criminal justice system cannot take credit for sparing their lives. Even so,
we certainly do not concede these defendants' innocence, for we have not examined their cases. See text
following note 11 supra. What we can say without hesitation is that if they were innocent, their vindication was
possible only because the system permits all manner of post-conviction interventions.
Indeed the authors' refusal to give due weight to the influence of post-conviction procedures is curious in
light of their concession that "the criminal justice system is arguably shown to have been nearly as effective as it
could be" when a wrongful conviction is averted at trial. Bedau & Radelet, supra note 1, at 45. The same may
be said when an erroneous conviction is reversed on appeal, on collateral attack, or even when a pardon is
granted. Since post-conviction review is such an important feature of the criminal justice system, one cannot
judge the system fairly without taking it into account.

Page 20
41 Stan. L. Rev. 121, *160

n24. Bedau & Radelet, supra note 1, at 23-24 (footnote omitted).
n25. See, e.g., id. at 38.
n26. Id. at 47.
n27. Id. at 42. Subsequent discovery of an alleged victim alive does, of course, provide objective proof of
the innocence of a defendant convicted of murdering that person. The authors, however, claim this has occurred
in only seven cases, id. at 64-65 & Table 7; in none of those cases was the defendant executed.
n28. The circumstances that the authors regard as signifying official acknowledgment of error include an
award of indemnity, executive pardon, and reversal of the conviction on appeal followed by acquittal on retrial
or dismissal of the charges. See id. at 48-52. The authors also consider a pardon to be a good indicator of
innocence, although insufficient by itself. Id. at 50.
The hazards of relying on such circumstances as evidence of innocence are obvious and are well illustrated
by the case of Lem Woon. Woon was convicted of murder in Oregon in 1908 and sentenced to death; but the
sentence was commuted to life imprisonment, and he was subsequently pardoned on condition that he be
deported. See id. at 171. The deportation condition plainly belies any inference that Woon was innocent. More
importantly, pardon was used at that time in Oregon as a substitute for parole, as Bedau has previously pointed
out, there being no other way of shortening a life sentence for murder. See Bedau, Capital Punishment in
Oregon, 45 OR. L. REV. 1, 31 (1965). Moreover, since Bedau did not suggest in his earlier article that Woon
was innocent, see id. at 33-34, his inclusion among those now claimed to have been wrongly convicted is
puzzling.
n29. See Bedau & Radelet, supra note 1, at 52. The authors recognize that "[m]any such confessions are
false or dubious, and thus they must be evaluated with extreme caution." Id. As their discussion of the Mays case
indicates, however, id. at 52, 144-45, they are willing to accept an implausible confession of a person other than
the defendant whose innocence they assert. See notes 95-98 infra and accompanying text.
n30. See Bedau & Radelet, supra note 1, at 53. The validity of this basis for concluding that a defendant
was innocent depends on the correctness of the authors' appraisal of the weight of the evidence against the
defendant, as well as on the plausibility of the hypothetical case against the other person. As the discussion
below of the Adams, Appelgate, and Dawson cases indicates, there is no reason to credit the authors' judgments
that these defendants did not commit the crimes for which they were executed. See notes 43-78 infra and
accompanying text.
n31. See Bedau & Radelet, supra note 1, at 53. These officials are a county medical examiner and two
prison wardens. Id. at 53, 164; see also id. at 173 (Coding Schedule), 49, Table 5 (Coding Schedule's
description). Since the authors do not explain why the opinions of these individuals are entitled to such weight,
it is difficult to understand why they should be persuasive.
n32. Id. at 53-54. This heading includes "the unshaken conviction by the defense attorney and others who
studied the case" that the defendant was innocent and the persistence of "widespread belief in the innocence of
the executed defendant." Id. But the authors claim that the best evidence of innocence in cases in which the

Page 21
41 Stan. L. Rev. 121, *160

defendant was executed consists of a variety of defects in the original prosecution that were discovered later
through private investigation. Id. at 54.
n33. See id. at 117, 173 (Garner), 121, 175 (Grzechowiak and Rybarczyk), 144-45, 174 (Mays), 158, 173
(Sanders), 161, 173 (Shumway).
n34. See id. at 91, 178 (Adams), 92, 175 (Appelgate), 109, 177 (Dawson).
n35. See id. at 93, 173 (Bambrick), 95-96, 173 (Becker and Cirofici), 158-59, 176 (Sberna), 164, 173
(Tucker), 170, 175 (Wing).
n36. Id. at 49, Table 5.
n37. See id. at 91-92, 176 (Anderson), 106-07, 175 (Collins), 124-25, 175 (Hauptmann), 125-26, 173 (Hill),
136-37, 174 (Lamble), 145-46, 176 (McGee), 157-58, 174 (Sacco and Vanzetti).
n38. See id. at 72-75.
n39. Id. at 74 (footnote omitted). Also interesting is the authors' concession that the weight of their
evidence of innocence ranges from "conclusive" to "slight." Id. at 48. They do not explain why slight evidence
of innocence should convince a neutral observer to disregard a unanimous jury verdict of guilty beyond a
reasonable doubt.
n40. See, e.g., notes 72, 108, 116, 123 infra.
n41. This impression is reinforced by the authors' statements that they both strongly support abolition of the
death penalty. See Bedan & Radelet, supra note 1, at 80-81, 89-90. It is also strengthened by the authors' claim
that among those wrongly executed were Sacco, Vanzetti, Hillstrom, and McGee, defendants whose convictions
were politically controversial. See note 69 and text accompanying notes 85-90, 99-108 infra.
n42. 408 U.S. 238 (1972). It was for this reason, and because the court records were most readily available,
that we selected the Adams case for careful scrutiny. We chose the case before we had an opportunity to
examine the strength of the evidence for or against guilt.
n43. Bedau & Radelet, supra note 1, at 91. For this description of the case, the authors cite only the
defendant's Application for Executive Clemency. See id. at 91 n.350 (citing Application for Executive
Clemency, In re James Adams (May 1, 1984) (on file with the Stanford Law Review)).
n44. Response to Petition for Writ of Habeas Corpus at 8, Adams v. Wainwright, 709 F.2d 1443 (11th Cir.
1983) (No. 82-5595) (citing trial transcript at 375-77, 382-83).
n45. Bedau & Radelet, supra note 1, at 91.

Page 22
41 Stan. L. Rev. 121, *160

n46. In his opening argument, defense counsel stated, "We intend to show that Willie Orange knew Mrs.
Adams, James Adams' wife, that Willie Orange knew James Adams, and that Willie Orange knew that James
Adams' wife -- that James Adams was going out with his wife." Application for Executive Clemency, supra note
43, at 25 (citing trial transcript at 833). However, defense counsel did not produce evidence to support this
statement at trial, id., and Adams did not raise this claim in his Motion to Vacate Judgment and Sentence, served
on opposing counsel on April 24, 1984, Response to Motion to Vacate Judgment and Sentence, State v. Adams
(St. Lucie County, Fla. filed Apr. 25, 1984) (No. 73-284-CF) although newly discovered evidence could have
been raised in support of the motion. See Hollman v. State, 371 So. 2d 482, 484-85 (Fla. 1979).
n47. Bedau & Radelet, supra note 1, at 91.
n48. Adams testified at trial that he was 35 years old and was 5 feet, 9 inches tall. See Response to Petition
for Writ of Habeas Corpus, supra note 44, at 8 (citing trial transcript at 887).
n49. Id. at 8 (citing trial transcript at 353-54, 355-58, 361, 363, 365, 366-68, 371).
n50. See Bedau & Radelet, supra note 1, at 91.
n51. Id.
n52. See Response to Motion to Vacate Judgment and Sentence, supra note 46, at 11. Counsel did,
however, explore the possible origin of the hair in cross-examination of the detective who had delivered it to a
laboratory for testing. Id. at 13.
n53. The state's request for analysis of the sample described the hair as "[h]air removed from Mr. Brown's
hand, by his wife, while in the ambulance enroute to the hospital. This hair was thrown on the floor of the
ambulance, and the ambulance was cleaned out and the hair thrown in the trash can." Motion to Vacate
Judgment and Sentence at 2, State v. Adams, No. 73-284-CF (St. Lucie County, Fla., filed Apr. 1984), quoted
in Response to Motion to Vacate Judgment and Sentence, supra note 46, at 12.
n54. Note 53 supra. Even assuming that the hair had been found in Mr. Brown's hand, there was another
plausible source. Elease Smith, a black woman who lived on Brown's property, had helped Josephine Brown,
the victim's widow, treat the victim's wounds. In particular, she helped Mrs. Brown wash off the victim's face.
See id. at 9 (citing trial transcript at 480-81, 483, 485); see also Response to Motion to Vacate Judgment and
Sentence, supra note 46, at 13 ("the trial transcript also suggests that, if the hair was indeed in the victim's hand,
its source might have been Elease Smith, the black lady who was helping Mrs. Brown aid her husband").
n55. Order, State v. Adams, No. 73-284-CF (St. Lucie County, Fla. filed Apr. 25, 1984).
n56. See Adams v. State, 449 So. 2d 819, 820 (Fla. 1984) (listing grounds for relief proffered by Adams).
n57. See Response to Petition for Writ of Habeas Corpus, supra note 44, at 5, 9, 11 (citing trial transcript at
267-97, 778-79, 491-92).

Page 23
41 Stan. L. Rev. 121, *160

n58. See id. at 6 (citing trial transcript at 322-28, 339-47, 398-408).
n59. The victim's wife and son testified that Edgar Brown carried large amounts of money with him, usually
in $ 20 bills, and that he had a sizeable sum with him on the day of the murder. See id. at 7-11 (citing trial
transcript excerpts).
n60. Id. at 12, 15 (citing trial transcript at 609, 612-15, 648-49, 682, 817-18, 878, 879, 730-31, 807-11).
n61. Id. at 13 (citing trial transcript at 671).
n62. See id. (citing trial transcript at 626, 639, 671, 675).
n63. See id. at 14 (citing trial transcript at 843, 856, 857-70, 875-79).
n64. See id. at 16 (citing trial transcript at 959). A witness for the state testified on rebuttal that the distance
from the Nickerson residence to the Brown ranch was approximately 12 miles and that the travel time at 55
miles per hour was 15 minutes. See id. at 16-17 (citing trial transcript at 991-93).
n65. In their article, Bedau and Radelet state that "if new evidence were to show that we have erroneously
included a case in which the defendant was truly guilty, we would of course have no objection to deleting it
from our catalogue." Bedau & Radelet, supra note 1, at 48. On the basis of an independent review of the Adams
case, the authors of this article have requested that Bedau and Radelet delete this case from their list. See Letter
from Assistant Attorney General Stephen J. Markman to Professor Hugo Bedau (Nov. 17, 1987) (on file with the
Stanford Law Review). Bedau has written in reply, "During the months ahead, my colleague and I will give
close attention to reviewing the information about Adams that you have supplied us. You may be sure that if it
or other subsequent evidence causes us to change our mind, you will be duly informed." Letter from Hugo Adam
Bedau to Stephen J. Markman (Dec. 10, 1987) (on file with the Stanford Law Review). Bedau and Radelet
maintain, however, that "nothing [in this response] removes the very serious doubts in the Adams case, or in any
other case included in our catalogue." Bedau & Radelet, supra note 23, at 163.
n66. Bedau & Radelet, supra note 1, at 48.
n67. See notes 12-23 supra and accompanying text.
n68. The authors cite these opinions in footnotes to their descriptions of the cases, but they invariably fail to
acknowledge the weight of the evidence of guilt that these opinions show was produced at trial.
n69. We discuss nine of the eleven cases below. The cases of Sacco and Vanzetti are not discussed here
only because adequate treatment of the evidence supporting their guilt would expand this response inordinately.
Suffice it to say that the evidence produced at trial, which was reviewed in meticulous detail by the
Massachusetts Supreme Judicial Court, see Commonwealth v. Sacco, 255 Mass. 369, 151 N.E. 839 (1926), was
plainly sufficient to support the verdict. Furthermore, the principal argument offered by the authors to show the
defendants' innocence -- the claim that another man confessed to the crime, see Bedau & Radelet, supra note 1,

Page 24
41 Stan. L. Rev. 121, *160

at 158 -- was not even raised on appeal from the denial of the defendants' motions for a new trial. Moreover, as
the authors recognize, the most recent work on the subject, by a long-time student of the case, presents recently
discovered evidence of Sacco's and Vanzetti's guilt. See id. at 158 n.818 (citing F. RUSSELL, SACCO AND
VANZETTI: THE CASE RESOLVED (1986)).
n70. Bedau & Radelt, supra note 1, at 92 (quoting W. BROWN, THEY DIED IN THE CHAIR 103 (1963)).
In fact, Appelgate himself supplied telling evidence of his own guilt. He admitted the facts that gave him a
powerful motive for killing his wife, he admitted having assisted in the purchase of the poison, he admitted
having prepared the poisoned eggnog and having spent two or three hours getting his wife to drink it, and he
refused to permit an autopsy until persuaded to do so by the authorities. See People v. Creighton, 271 N.Y. 263,
272-77, 2 N.E.2d 650, 654-56 (1936). Moreover, the jury that found Appelgate guilty was instructed that "you
cannot convict the defendant Appelgate upon [Creighton's] testimony alone, unless you find that her testimony is
corroborated by other evidence which tends to connect Appelgate with the commission of the crime." Id. at 281,
2 N.E.2d at 658.
n71. See Creighton, 271 N.Y. at 280-81, 2 N.E.2d at 658 (1936). The "state's prison offense" referred to by
the court was second degree rape, committed on numerous occasions against Creighton's minor daughter Ruth,
with whom Appelgate had been living together "almost as man and wife." Id. at 269, 2 N.E.2d at 653. The threat
to "expose" Appelgate mentioned by the court was established by evidence that, following a fight a few weeks
before the murder during which her husband slapped her, Mrs. Appelgate said: "[I]f you ever do that again I will
tell something that will put you where you belong." Id. at 268, 2 N.E.2d at 652.
n72. See D. KILGALLEN, MURDER ONE 190-91, 230 (1967) (Appelgate "very nearly got away" with the
murder); L. LAWES, MEET THE MURDERER 334-35 (1940) ("Frankly, I do not doubt the culpability" of
Appelgate), cited in Bedau & Radelet, supra note 1, at 92 n.362. The authors also state, incorrectly, that the
Governor of New York denied a clemency petition despite having "doubts about Appelgate's guilt." Id. at 92.
One of the authors' sources cited for support wrote that the governor did not "question [Appelgate's] guilt. He
took keen interest in the proceedings and made every possible effort to see that justice was meted out." L.
LAWES, supra, at 334-35. Thus, even though they rely on no evidence other than that presented at trial, Bedau
and Radelet believe that a neutral observer would find Appelgate innocent when a unanimous jury, a unanimous
Court of Appeals, the Governor, and two of their three secondary sources reached the opposite conclusion.
n73. Bedau & Radelet, supra note 1, at 109.
n75. Id. at 421 (Drew, J., dissenting). The authors assert, falsely, that there were no eyewitnesses to the
crime. Bedau & Radelet, supra note 1, at 109. In fact, as the Florida Supreme Court opinion indicates, there
was an eyewitness -- the victim's 4-year-old brother, Donnie, who had been beaten and left to die at the crime
scene. See Dawson, 139 So. 2d at 410. When found a day later, Donnie told his father, the police chief, and a
family friend that Sie Dawson had committed the murder. See St. Petersburg Times, Sept. 24, 1977, at 12A, col.
1. The prosecution did not place Donnie on the stand, presumably because of his age; and the witnesses who
had heard him identify Dawson as the killer could not testify as to his statements because of the hearsay rules.
Seventeen years later, when interviewed by the St. Petersburg Times about the details of the crime, he explained
how Sie Dawson had brutally killed his mother with a hammer. See id. at col. 5.
n76. See Dawson, 139 So. 2d at 418 (Thornal, J., concurring).

Page 25
41 Stan. L. Rev. 121, *160

n77. Id. at 414.
n78. Id. In this connection, Judge Thornal observed that "the circumstantial evidence is completely
consistent with the details revealed by the confession, whereas it is not equally consistent with the details
described by Dawson from the witness stand when he undertook to repudiate his confession." Id.
Material contained in the clemency files of the Florida governors is also relevant to the claim that the
woman's husband was the murderer. Although the authors refer the reader to these files, see Bedau & Radelet,
supra note 1, at 109 n.478, they neglect to mention that the materials they contain provide no support for their
position. Those materials indicate that investigations by both the Parole Commission and a representative of the
florida Governor's Office found nothing to substantiate the defense claim that the husband had committed the
crime. See Memorandum from William A. Norris, Jr. to Governor Bryant re: Sie Dawson 2 (May 1, 1964) (on
file with the Stanford Law Review); Memorandum from Al Healy to Bill Norris re: Sie Dawson (Apr. 29, 1964)
(on file with the Stanford Law Review). The materials also reveal that "intensive investigation at the time [of the
crime] failed to develo[p] any indication that Alvin Clayton [the husband] was implicated in the murders."
Memorandum from Al Healy, supra, at 2.
Finally, if there is any need for additional refutation of the authors' claim that Dawson was innocent, Bedau
himself supplies it. He indicated in 1982 that the Dawson case "remain[ed] in the limbo of uncertainty" because
"[t]he original news story [regarding Dawson's supposed innocence] merely reported allegations and was
inconclusive; no subsequent inquiry known to me has established whether Dawson was really innocent." Bedau,
Miscarriages of Justice and the Death Penalty, in THE DEATH PENALTY IN AMERICA, 236-37 (H. Bedau
ed. 1982) (citing to the same newspaper articles as are cited in Stanford Law Review article).
n79. Bedau & Radelet, supra note 1, at 121.
n80. See People v. Bogdanoff, 254 N.Y. 16, 171 N.E. 890 (1930).
n81. Id. at 19, 171 N.E. at 891. The confessions were reliable. A year after his conviction, Rybarczyk told
the governor's clemency investigator, see note 84 infra, that the police did not use any third-degree methods to
obtain the confessions. See Buffalo Evening News, July 3, 1930, at 1, col. 8, at 3, col. 6. Indeed Rybarczyk was
apparently motivated to confess to the murder by a jailhouse meeting with his mother. See id., Nov. 5, 1929, at
3, col. 1. Moreover, Rybarczyk's confession included an account of another mobster slaying that he had
observed. The truthfulness of that account, and of his confession in the Fechter case, was confirmed after his
conviction, when police found the body of the apparent victim in circumstances that fit Rybarczyk's description.
See id., Nov. 5, 1929, at 1, col. 6.
n82. Bogdanoff, 254 N.Y. at 32, 171 N.E. at 896. The confessions were corroborated by other evidence of
the defendants' guilt. At least three witnesses with good views of the crime scene positively identified
Grzechowiak, Rybarczyk, or both, as participants in the murder. See Buffalo Evening News, Oct. 24, 1929, at 1,
col. 6, at 3, col. 1; id., Oct. 25, 1929 at 1, col. 1.
n83. Bogdanoff, 254 N.Y. at 40, 171 N.E. at 899 (Crane, J., dissenting, on the grounds of the majority's
rejection of the challenge to the sufficiency of the indictment).
n84. See Bedau & Radelet, supra note 1, at 121. The authors' reliance on the defendants' exoneration by

Page 26
41 Stan. L. Rev. 121, *160

their codefendant Bogdanoff is misplaced. The lengthy police investigation that preceded the arrests of the three
men included infiltration of the gang, a wiretap, and around-the-clock surveillance of the Grzechowiak residence
for several weeks. It showed that Grzechowiak and Rybarczyk were members of Buffalo's notorious "Blue
Ribbon Gang," headed by Bogdanoft. See Buffalo Evening News, Nov. 6, 1929, at 2, col. 1. Bogdanoff had
boasted to an undercover agent that the gang was wanted for the Fechter robbery-murder. Id., July 3, 1930, at 3,
col. 8. The only reason Bedau and Radelet suggest for believing Bogdanoff is his refusal to name "his true
accomplices." Bedau & Radelet, supra note 1, at 121. In fact, after the Governor refused to consider clemency
for Grzechowiak and Rybarczyk without this information, Bogdanoff claimed that his accomplices had not been
members of his gang but three gunmen from Chicago whose names, addresses, and descriptions he furnished to
the authorities. See Buffalo Evening News, July 5, 1930, at 3, col. 1. The Governor's legal advisor and others
investigated Bogdanoff's story; but "[n]othing was brought forth to substantiate the claim, which was regarded
as a last attempt to save Rybarczyk and Grzechowiak." Id., July 18, 1930, at 3, col. 3. Police also viewed the
claim as inconsistent with the known facts of the case. See N.Y. Times, July 17, 1930, at 14, col. 3. See
generally id., July 18, 1930, at 3, col. 3.
n85. Bedau & Radelet, supra note 1, at 125.
n86. See State v. Hillstrom, 46 Utah 341, 150 P. 935 (1915).
n87. Id. at 349, 150 P. at 939.
n88. Id. at 356, 150 P. at 941.
n89. Id. at 371, 150 P. at 947 (Frick, J., concurring); see also id. at 373, 150 P. at 948 (McCarty, J.,
concurring) (the evidence that the defendant was one of the perpetrators of the crime was "about as conclusive
on that point as though the witnesses had positively identified the defendant as the taller of the two men who
were at and in the immediate vicinity of the crime just before and immediately after it was committed").
n90. Indeed, Bedau and Radelet cite a work of fiction to support their conclusion that Hillstrom was not
guilty. Bedau & Radelet, supra note 1, at 126 n.588 (citing "generally" W. STEGNER, JOE HILL: A
BIOGRAPHICAL NOVEL (1969) (first published under the title THE PREACHER AND THE SLAVE)). The
novel's foreword explicitly explains that it "is fiction, with fiction's perogatives and none of history's limiting
obligations. . . . Joe Hill as he appears here -- let me repeat it -- is an act of the imagination." W. STEGNER,
supra, at 13-14 (emphasis added). It is more than a little troubling that Bedau and Radelet use a work of fiction
as one of their two sources for the proposition that Hillstrom was innocent. More disturbing yet, Bedau and
Radelet cite the Stegner novel "generally" to support Hillstrom's innocence despite the fact that the author's view
was to the contrary. See Stegner, Correspondence: Joe Hill, NEW REPUBLIC, Feb. 9, 1948, at 39; see also
Stegner, Joe Hill: The Wobblies Troubadour, NEW REPUBLIC, Jan. 5, 1948, at 20.
n91. Bedau & Radelet, supra note 1, at 144. The authors also rely on the fact that Mays maintained his
innocence up to the time he was executed, although they recognize that a defendant's "own protestations" are
"not enough" by themselves to prove innocence. Id. at 48. As the analysis of the Mays case shows, however,
the arguments made to support the defendant's protestations of innocence are unavailing.
n92. May's previous conviction of the crime had been overturned because the judge rather than the jury had

Page 27
41 Stan. L. Rev. 121, *160

imposed the death penalty, an error that had nothing to do with the defendant's guilt or innocence. See Mays v.
State, 143 Tenn. 443, 226 S.W. 233 (1920).
n93. Mays v. State, 145 Tenn. 118, 123-32, 238 S.W. 1096, 1097-1101 (1921).
n94. Id. at 133-35, 238 S.W. at 1101.
n95. Bedau & Radelet, supra note 1, at 52, 145.
n96. Id. at 52.
n97. Id. at 145.
n98. Edgerton, A Case of Prejudice: Maurice Mays and the Knoxville Race Riot of 1919, S. EXPOSURE,
July/Aug. 1983, at 64. Edgerton also reports another salient fact Bedau and Radelet omit: The Governor denied
clemency only after his own review of the case and that of an independent commission convinced him that the
verdict was correct. (The commission also suggested that Mays might have been insane.) See id. at 63.
n99. McGee's first conviction, by an all-white jury in the county where the crime was committed, was
reversed because the trial court had erroneously denied a change of venue. McGee v. State, 200 Miss. 592, 26
So. 2d 680 (1946). His second conviction, in a different county but also by an all-white jury, was overturned
because blacks were excluded on account of their race from the grand jury that indicted him. McGee v. State,
203 Miss. 592, 33 So. 2d 843 (1948) (en banc). He was indicted again in the original county, this time by a
grand jury with three black members and was found guilty a third time. McGee v. State, 40 So. 2d 160 (1949)
(en banc), cert. denied, 338 U.S. 805 (1949).
n100. Bedau & Radelet, supra note 1, at 145.
n101. Id. at 145-46.
n102. McGee v. State, 40 So. 2d 160.
n103. Id. at 169.
n104. Id.
n105. Id. The court also rejected McGee's claim that his subsequent written confession had been coerced by
daily and nightly beatings, in part on the ground that "[t]he very extravagance of such extreme assertions
furnishes, in large measure, their own refutation, averting the credibility of reasonable men." Id. at 170.
n106. Id. at 171. Moreover, as the court pointed out, the fact of rape was corroborated not only by McGee's

Page 28
41 Stan. L. Rev. 121, *160

admissions, but also by the victim's husband and a neighbor who testified that the victim was hysterical
immediately after the attack, by her physician who testified that voluntary sexual intercourse would not likely
have produced the condition of her sexual organs, and by physical signs of entry into her bedroom through a
window. Id.
n107. See text accompanying notes 103-105 supra; see also McGee, 40 So. 2d at 170-71. McGee's failure to
testify at trial, McGee, 40 So. 2d at 170, also undermines the consent claim. Nothing in the court's opinion
suggests that McGee was deterred from taking the stand by a prior conviction or the like. To the contrary, the
court described him as an "intelligent" man who "could read and write, and had a responsible job, delivering
goods and collecting therefor, in the employ of the Laurel Wholesale Grocery Company." Id.
n108. The authors also claim that "[i]nvestigation by journalist Carl Rowan revealed that the victim had
been consorting with McGee for four years and was angry at his efforts to terminate their relationship" and that
"local blacks were too intimidated to give this evidence in court." Bedau & Radelet, supra note 1, at 146.
Rowan's investigation consisted of a few conversations with blacks in the town where the crime had occurred.
See C. ROWAN, SOUTH OF FREEDOM 174, 177-85 (1952). He concluded that "the evidence of this
so-called clandestine romance was no more ironbound than was the evidence to the contrary," Id. at 185, and he
questioned how the hypothesis of consent on the night in question could "jibe with testimony of a woman
defense witness that McGee was with her at the time of the alleged attack." Id. at 186. Indeed, it appears that
Rowan had deep-seated doubts about McGee's innocence, contrary to what the authors imply, and was
convinced only of the injustice of subjecting him to the death penalty for an offense that had never been thought
to warrant that punishment when committed by a white. See id. at 176, 191-92.
n109. See Bedau & Radelet, supra note 1, at 158.
n110. See Sanders v. State, 202 Ala. 37, 39, 79 So. 375, 377 (1918). Bedau and Radelet state incorrectly that
Sanders was convicted "with Fisher Brooks" and that "[b]oth men were sentenced to death and executed in
1918." Bedau & Radelet, supra note 1, at 158. In fact, the two men were convicted separately; and Brooks was
executed in 1917, almost a full year before Sanders. See Sanders, 202 Ala. at 38, 79 So. at 376.
n111. See Mobile Register, July 21, 1917, at 1, col. 1, at 9, col. 1 ("'I've just told so many stories I don't
know what I have told'").
n112. See id. at 9, col. 1. At Sanders's trial, Brooks claimed that he had acted alone in committing the
murder. Testimony from a night watchman contradicted this point. He saw and heard a screaming woman in a
taxi (like Sanders, Brooks was a taxi driver) with two men, on the road to the bridge from which the victim's
body was dumped into a river. Brooks also claimed that he stopped the taxi to throw the body from the bridge
into the river. But a guard at the bridge testified that he heard a splash as the taxi drove over the bridge with
increasing speed. Blood was found on the west railing, the opposite side from the steering wheel in the taxi as it
drove over the bridge. See id., July 20, 1917, at 1, col. 7, at 9, col. 3; id., July 21, 1917, at 1, col. 1, at 9, col. 2.
n113. See Sanders v. State, 202 Ala. at 38, 79 So. at 376. According to a local newspaper account of the
trial, the witness said he overheard Brooks assure Sanders that Brooks's testimony would "help" Sanders, to
which the latter replied, "'I'm guilty as .'" Mobile Register, July 21, 1917, at 1, col. 1.

Page 29
41 Stan. L. Rev. 121, *160

n114. See Mobile Register, July 20, 1917, at 1, col. 7, at 9, cols. 3-4; id., July 21, 1917, at 9, cols. 1-3; id.,
July 22, 1917, at 1, col. 1.
n115. Bedau & Radelet, supra note 1, at 164.
n116. See Commonwealth v. Tucker, 189 Mass. 457, 76 N.E. 127 (1905); E. PEARSON, MASTERPIECES
OF MURDER 157-79 (1963). Indeed, the bill of exceptions which the defendant presented to the appellate
court conceded that "there was evidence for the jury on all issues presented to them." Tucker, 189 Mass. at 467,
76 N.E. at 130. Moreover, Pearson's chapter which the authors cite generally to support their discussion of the
case, see Bedau & Radelet, supra note 1, at 164 n.869, does not suggest that Tucker was innocent. In fact, it
actually implies the opposite conclusion. See E. PEARSON, supra, at 168 ("even without [the handwriting
evidence] the Government's case was very strong"); id. at 171 (the Governor's rejection of the defendant's
petition for clemency was "conscientious and admirable").
n117. Tucker, 189 Mass. at 465-76, 76 N.E. 129-34; E. PEARSON, supra note 116, at 164-68.
n118. Tucker's defense counsel later reported that Tucker had decided not to testify after being told by
counsel that he should testify if he were innocent but not if he were guilty. E. PEARSON, supra note 116, at
178-79.
n119. See Tucker, 189 Mass. at 479, 484, 76 N.E. at 135, 137; see also E. PEARSON, supra note 116, at
170 (in denying Tucker's motion for a new trial, the trial judges described most of the evidence produced on his
behalf as untrue). Moreover, it appears that Tucker's alibi witnesses could not account for Tucker's whereabouts
at the time of the murder. See Hewett, Shreds of Evidence, BOSTON MAG., Nov. 1983, at 133, 134.
n120. See Bedau & Radelet, supra note 1, at 164. As noted above, see note 116 supra, Tucker admitted on
appeal that the case was strong enough to go to the jury. Beyond that, in denying Tucker's motion for a new
trial, the two judges who presided at the trial expressed the view that "'a verdict of acquittal would have been a
failure of justice.'" E. PEARSON, supra note 116, at 170. The Governor concurred. His review of the case led
him to state, "I am compelled to an undoubting belief in his guilt." Id. at 175.
n121. Bedau & Radelet, supra note 1, at 164.
n122. See E. PEARSON, supra note 116, at 169-72; see also F. RUSSELL, SACCO AND VANZETTI:
THE CASE RESOLVED 203-16 (1986) (discussing ill-founded public protest on behalf of condemned
prisoners generally and of Tucker in particular). The true nature of the agitation on Tucker's behalf is
demonstrated by the fact that "[t]he names of entire families, including, in some instances, children in the cradle,
were solemnly set down at the foot of these documents." E. PEARSON, supra note 116, at 170.
n123. The medical examiner, who testified at trial that he thought the victim had been killed with a knife of
the type owned by Tucker, claimed at the clemency hearing that such a knife could not have been the murder
weapon, and that the state had deliberately framed Tucker for the murder. See Hewett, supra note 119, at 134,
136. The trial judges, however, at the Governor's request, reviewed the evidence presented at the clemency
hearing. They reported that they could not find "that there would now be any material change in the testimony

Page 30
41 Stan. L. Rev. 121, *160

of the medical experts." E. PEARSON, supra note 116, at 173. The authors are also mistaken about the trial
witness's alleged admission of perjury to a clergyman. The witness made no such admission. What the
clergyman reported was a third person's statement contradicting 136. This double hearsay would not convince a
neutral observer of Tucker's innocence without proof that the witness lied and that true testimony would have
raised a reasonable doubt about Tucker's guilt.
n124. See, e.g., W. BERNS, supra note 6, at 178; van den Haag, The Death Penalty Once More, 18 U.C.
DAVIS L. REV. 957, 967 (1985).
n125. See, e.g., C. BLACK, CAPITAL PUNISHMENT: THE INEVITABILITY OF CAPRICE AND
MISTAKE 22-30 (2d ed. 1981); Pollak, The Errors of Justice, in CAPITAL PUNISHMENT 207 (T. Sellin ed.
1967).
n126. The laws of 37 states, as well as certain federal laws, permit imposition of the death penalty. See U.S.
DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS, CAPITAL PUNISHMENT, 1986, at 3 (1987)
[hereinafter CAPITAL PUNISHMENT, 1986].
n127. See, e.g., SENATE COMM. ON THE JUDICIARY, REPORT ON S. 1765, ESTABLISHING
CONSTITUTIONAL PROCEDURES FOR THE IMPOSITION OF CAPITAL PUNISHMENT, S. REP. NO.
251, 98th Cong., 1st Sess. 14 (1983) [hereinafter "SENATE REPORT"].
n128. W. BERNS, supra note 6, at 178 (erroneous executions are unlikely); SENATE REPORT, at 14 (the
risk is "minimal"); van den Haag, supra note 124, at 967 (miscarriages of justice are "rare"). Indeed, Bedau
himself has observed that it is "false sentimentality to argue that the death penalty ought to be abolished because
of the abstract possibility that an innocent person might be executed, when the record fails to disclose that such
cases occur." Bedau, The Death Penalty in America, FED. PROBATION, June 1971, at 32, 36. Bedau now
asserts that "[t]he publication of the present research, shows the reality of such mistakes to be a virtual
certainty." Bedau & Radelet, supra note 1, at 26. However, in our judgment, the study does not show with
"virtual certainty" that mistaken executions have occurred in this century; and so Bedau's earlier assessment
remains essentially accurate. See also Kaplan, Administering Capital Punishment, 36 U. FLA. L. REV. 177,
186 ("It is true that in the modern history of the death penalty, the last thirty-some years, probably no innocent
person has been executed").
n129. Bedau & Radelet, supra note 1, at 27.
n130. See text accompanying notes 42-123 supra. The remaining 11 allegedly erroneous executions predate
modern times. Only two occurred during the past 50 years, and both of these took place more than 40 years ago.
See Bedau & Radelet, supra note 1, at 73, Table 10 (Sberba 1938; Anderson 1945).
n131. See note 22, supra.
n132. Bedau and Radelet claim that between 1900 and 1942 there were 19 erroneous executions out of an
estimated 5229 total executions -- a rate of error of 0.36%. Bedau & Radelet, supra note 1, at 73. Between
1943 and 1985, they claim, there were four mistaken executions out of a total of 1863 -- a rate of error of 0.22%.

Page 31
41 Stan. L. Rev. 121, *160

Id.
n133. Id. at 78.
n134. Id. at 84.
n135. Id.; see also id. at 29.
n136. Id. at 78.
n137. See id. at 86-90.
n138. 408 U.S. 238 (1972).
n139. In addition, it can be argued that the various forms of prejudice which Bedau and Radelet claim or
imply contributed to the execution of allegedly innocent persons have been subtantially reduced, if not virtually
eliminated. See Bedau & Radelet, supra note 1, at 125 (anti-union hostility in the Hillstrom case), 144-45 (racial
prejudice in the Mays case), 145-46 (racial prejudice in the McGee case), 157-58 (anti-communist prejudice in
the Sacco and Vanzetti case).
n140. See, e.g., ALA. CODE § 13A-5-40 (1982 & Supp. 1987).
n141. See, e.g., N.J. REV. STAT. § 2C:11-3c(2)(e) (Supp. 1988).
n142. See, e.g., ALA. CODE §§ 13A-5-45, 54 (1982).
n143. See, e.g., MISS. CODE ANN. § 99-19-101(3) (Supp. 1987).
n144. See, e.g., GA. CODE ANN. § 17-10-30(b)(7) (1982).
n145. See, e.g., id. § 17-10-30(b)(3).
n146. See, e.g., LA. CODE CRIM. PROC. ANN. art. 905.4A(3) (West Supp. 1988).
n147. See, e.g., id. art. 905.4A(5).
n148. See, e.g., GA. CODE ANN. §§ 17-10-30(b)(5), (8) (1982).
n149. See, e.g., LA. CODE CRIM. PROC. ANN. art. 905.5(f) (West 1984).

Page 32
41 Stan. L. Rev. 121, *160

n150. See, e.g., FLA. STAT. § 921.141(6)(f) (1985).
n151. See, e.g., id. § 921.141(6)(e).
n152. See, e.g., id. § 921.141(6)(d).
n153. See, e.g., LA. CODE CRIM. PROC. ANN. art. 905.6 (West 1984).
n154. See, e.g., id. art. 905.8.
n155. See, e.g., MISS. CODE ANN. §§ 99-19-105(1)-(3) (1987).
n156. See 28 U.S.C. § 1257(2) (1970).
n157. See, e.g., TEX. CRIM. PROC. CODE ANN. art. 40.03(6) (Vernon 1979). 28 U.S.C. § 2254 (1982).
n158. See, e.g., LA. CODE CRIM. PROC. ANN. arts. 924, 924.1, 930.3 (West 1981).
n159. See, e.g., ARIZ. CONST. art. 5, § 5; ARIZ. REV. STAT. ANN. §§ 31-402A, 31-443 (1986).
n160. See text accompanying notes 42-65 supra.
n161. Florida law differs from the typical death penalty statute in that it permits the judge to impose the
death penalty after considering the jury's recommendation. FLA. STAT. § 921.141(2), (3) (West 1985). In the
Adams case, the judge found numerous aggravating factors to support imposition of the death penalty. For
example, at the time of the murder Adams had escaped from the Tennessee prison system where he was serving
a 99-year sentence for rape. Adams v. State, 341 So. 2d 765, 769 (Fla. 1976).
n162. Adams, 341 So. 2d at 765.
n163. Adams v. Florida, 434 U.S. 878 (1977).
n164. 430 U.S. 349 (1977).
n165. Adams v. State, 355 So. 2d 1205, 1206 (Fla. 1978).
n166. Adams v. Florida, 439 U.S. 947 (1978).
n167. Adams v. State, 380 So. 2d 421 (Fla. 1980).

Page 33
41 Stan. L. Rev. 121, *160

n168. Adams v. State, 380 So. 2d 423 (Fla. 1980). At that time, Adams was also one of 123 Florida death
row inmates who unsuccessfully sought relief in Brown v. Wainwright. 392 So. 2d 1327 (Fla.), cert. denied, 454
U.S. 1000 (1981).
n169. See Adams v. Wainwright, 709 F.2d 1443, 1445 (11th Cir. 1983).
n170. Id.
n171. Adams v. Wainwright, 464 U.S. 1063 (1984).
n172. Adams v. Wainwright, 465 U.S. 1074 (1984).
n173. See Adams v. State, 449 So. 2d 819, 820 (Fla. 1984).
n174. Id. at 820-21. Adams did not raise a claim that he was innocent before the Florida Supreme Court.
Nor did he report to the Court any "newly discovered" evidence of his innocence even though such evidence
could, in proper circumstances, be the basis for relief under Florida law. See Hallman v. State, 371 So. 2d 482,
484-85 (Fla. 1979).
n175. Adams, 449 So. 2d at 820-21.
n176. See Adams v. Wainwright, 734 F.2d 511, 511-12 (11th Cir. 1984) (noting the petition's successive
nature and its denial by the district court).
n177. Wainwright v. Adams, 466 U.S. 964 (1984).
n178. Interview with Andrea Hillyer, Florida Attorney General's Office (March 1987).
n179. See, e.g., Darden v. Wainwright, 473 U.S. 928, 929 (1985) (Burger, C.J., dissenting from grant of
certiorari) ("In the 12 years since petitioner was convicted of murder and sentenced to death, the issues now
raised in the petition for certiorari have been considered by this Court four times . . . and have been passed upon
no fewer than 95 times by federal and state court judges."). See generally Who Is on Trial? Conflicts Between
the Federal and State Judicial Systems in Criminal Cases: Hearing Before the Subcomm. on Gov't Information,
Justice, and Agriculture of the House Comm. on Gov't Operations, 100th Cong., 2d Sess. 25-27 (1988)
(statement of Paul Cassell, Associate Deputy Attorney General).
n180. See Bedau & Radelet, supra note 1, at 70-71, 84.
n181. See id. at 91. It should also be noted that in 1986 the average elapsed time from sentencing to
execution in death penalty cases was more than seven years. See CAPITAL PUNISHMENT, 1986, supra note
126, at 9.

Page 34
41 Stan. L. Rev. 121, *160

n182. Bedau & Radelet, supra note 1. at 38, Table 4.
n183. Id.
n184. See notes 42-43, 73-78 supra and accompanying texts (discussion of the Adams and Dawson cases).
n185. Bedau & Radelet, supra note 1, at 38-39.
n186. See id. at 29-31 n.40 (cases of John Henry Knapp and Earl Johnson). The authors also refer to seven
other "recent capital cases involving inmates released in the first six months of 1987 from death rows . . .
because of evidence raising doubts about their guilt," id.; but according to the information supplied by the
authors, six of these individuals (Peek, Cobb, Williams, McManus, Brown, Ramos) were no longer under a
sentence of death at the end of 1986. Their convictions and death sentences had been vacated before then. The
same is true of the seventh individual, Hill, whose case was actually decided in 1985, not 1987. See Hill v.
State, 473 So. 2d 1253 (Fla. 1985).
n187. Bedau & Radelet, supra note 1, at 30 n.40 (case of John Henry Knapp).
n188. Id. (case of Earl Johnson). The authors claim that "substantial doubts remain" about Johnson's guilt,
but -- once again -- the appellate opinions in the case show that Johnson was certainly guilty. The elderly victim
of the assault that preceded the police officer's murder positively identified Johnson as her assailant the next day
and at trial. She was unable to make a positive identification earlier because "she was in a state of shock and
was confused due to the viciousness of the beating on her head and body as well as the attempt upon her life and
safety." Johnson v. State, 416 So. 2d 383, 387 (Miss. 1982). Moreover, Johnson's claim that his confession was
false appears to have been an afterthought. See Johnson v. State, 508 So. 2d 1126, 1127 (Miss. 1987). He did
not attack its admissibility on direct appeal, see Johnson v. State 416 So. 2d 383, 387 (Miss. 1982), in his state
application for a writ of error coram nobis, see Johnson v. Thigpen, 449 So. 2d 1207 (Miss. 1984), or in his
federal petition for habeas corpus. See Johnson v. Thigpen, 623 F. Supp. 1121 (S.D. Miss. 1985), aff'd, 806 F.2d
1243 (5th Cir. 1986), cert. denied, 107 S. Ct. 1618 (1987). In any event, Johnson's confession was corroborated
in at least three important respects: it led to the recovery of the gun that the murderer took from the officer and
used to kill him, Johnson v. State, 416 So. 2d at 385-86; witnesses placed a vehicle like that owned by Johnson
at the crime scene at the time in question, Johnson v. Thigpen, 623 F. Supp at 1126; and Johnson admitted to a
trustee at the jail where he was being held that he had shot the officer, Johnson v. Thigpen, 806 F.2d at 1252;
Johnson v. State, 416 So. 2d at 387.
n189. See, e.g., Gregg v. Georgia, 428 U.S. 153 (1976).
n190. We derived the 50,000 figure by multiplying the approximately 200,000 murder arrests between 1977
and 1986 by the 25% conviction rate for that crime during the preceding few years. See U.S. DEP'T OF
JUSTICE, FEDERAL BUREAU OF INVESTIGATION, CRIME IN THE UNITED STATES, 1986, at 41,
Table 1 (1987).
n191. Bedau & Radelet, supra note 1, at 178-79. Only five of the 31 cases cited resulted in imposition of a
death sentence. In none of these was the sentence actually carried out. Id.; see also notes 185-188 supra and

Page 35
41 Stan. L. Rev. 121, *160

accompanying text. But see id. at 29 n.40 (nine new cases in which the death sentence may have been
wrongfully imposed, with one resulting in an execution).
n192. See Bedau & Radelet, supra note 1, at 96 (Beeman), 161 (L. Smith), 112-13 (Fay), 125 (Hicks), 170
(S. Wilson), 112 (Estes), 155 (Reynolds), 93 (Bachelor), 146 (McIntosh), 91 (Amado), 156 (Robertson), 129-30
(Jaramillo), 151-52 (Petree), 156 (Robinson), 103 (Carden), 129 (James & Peterson), 126-27 (Holbrook), 126-27
(Rucker), 103 (Carter), 113 (Ferber), 100-01 (A. Brown), 149 (Norwood), 143 (Malloy), 163-64 (Torres).
n193. Bedau & Radelet, supra note 1, at 78. Despite this caution, Bedau and Radelet do not hesitate in
announcing their view that the costs outweigh the benefits and that "abolition is the better policy." Id. at 80-81.
n194. The authors' protestations that their figures "understate" the numbers of innocent persons executed are
unconvincing. See notes 133-136 supra and accompanying text.
n195. See, e.g., Tison v. Arizona, 107 S. Ct. 1676, 1677 (1987) (defendants aided the prison escape of their
father who was serving a life sentence for murdering a prison guard and who then proceeded to kill four more
innocent persons).
n196. "At least five federal prison officers have been killed since December 1982, and the inmates in at
least three of the incidents were already serving life sentences for murder." W. Weld & P. Cassell, Report to the
Deputy Attorney General on Capital Punishment and the Sentencing Commission 28 (Feb. 13, 1987).
n197. See SENATE COMM. ON THE JUDICIARY, ESTABLISHING CONSTITUTIONAL
PROCEDURES FOR THE IMPOSITION OF CAPITAL PUNISHMENT, S. REP. NO. 251, 98th Cong., 1st
Sess. 12 (1983). A recent study of recidivism among young parolees revealed that 6% of young adults paroled
in 1978 after having been convicted of murder were arrested again for murder within the following six years.
See U.S. DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS, RECIDIVISM OF YOUNG PAROLEES
4, Table 4 (1987) [hereinafter RECIDIVISM OF YOUNG PAROLEES].
n198. See Memorandum from Lawrence A. Greenfeld to Steven R. Schlesinger 2 (Dec. 18, 1985) (on file
with the Stanford Law Review). Of the 1405 on death row at the end of 1984, an estimated 132 had prior murder
convictions. Id. at 1. These figures were based on a Bureau of Justice Statistics 1979 survey of 11,397 state
prison inmates, weighted and adjusted for the 1984 prison population. See U.S. DEP'T. OF JUSTICE,
BUREAU OF JUSTICE STATISTICS, PROFILE OF STATE PRISON INMATES, 1986, at 7 (1988); U.S.
DEP'T OF JUSTICE, BUREAU OF JUSTICE STATISTICS, CAPITAL PUNISHMENT, 1984, at 7, Table 9
(1985) [hereinafter CAPITAL PUNISHMENT, 1984] (of 1405 persons serving death sentences, 104 had been
previously convicted for homicide).
n199. See Bedau & Radelet, supra note 1, at 80.
n200. For instance, one might derive a rough estimate by multiplying 810 convicts by 3.1% -- the
proportion of convicted murderers who had been sentenced to death out of those admitted to prison for homicide
in 1983. See CAPITAL PUNISHMENT, 1984, supra note 198, at 9, Table A-1. This calculation yields a figure
of approximately 24 lives saved through the use of capital punishment over the last several decades. This

Page 36
41 Stan. L. Rev. 121, *160

estimate may be conservative. Since the death penalty is reserved for the most aggravated murders, it seems fair
to conclude that the perpetrators pose a greater risk of subsequent homicide than the population of convicted
murderers generally. See, e.g., TEX. CODE CRIM. PROC. ANN. art. 37.071(b)(2) (Vernon 1981 & Supp.
1988) (requiring a finding that "there is a probability that the defendant would commit criminal acts of violence
that would constitute a continuing threat to society"). Moreover, the 23 allegedly wrongful executions span this
entire century while 24 lives saved through incapacitation covers only the last few decades.
n201. Bedau & Radelet, supra note 1, at 80.
n202. Id. at 81. But see RECIDIVISM OF YOUNG PAROLEES, supra note 197, at 4, Table 4. The data
cited reveals that of 11,404 persons originally convicted of "willful homicide" and released during 1965 to 1974,
34 were returned to prison for commission of a subsequent criminal homicide during the first year alone. See
also H. Bedau, Recidivism, Parole, and Deterrence, in THE DEATH PENALTY IN AMERICA 173, 175 (3d
ed. 1982) (cited in Bedau & Radelet, supra note 1, at 79 n.296).
n203. See, e.g., GA. CODE ANN. § 17-10-30 (1982).
n204. See, e.g., TEX. CODE CRIM. PROC. ANN. art. 37.071(b)(2) (Vernon 1981 & Supp. 1988).
n205. 428 U.S. 153, 186 (1976) (plurality opinion). Anecdotal evidence supports this view. For instance,
according to the Attorney General of Kansas, one of the contributing factors leading to the 1935 reenactment of
the death penalty in Kansas for first-degree murder was the spate of deliberate killings committed in Kansas by
criminals who had previously committed such crimes in surrounding states where their punishment, if captured,
could have been the death penalty. These criminals admitted having chosen Kansas as the site of their crimes
solely for the purpose of avoiding a death sentence in the event that they were captured. Report of the Royal
Commission on Capital Punishment 1949-53, at 375 in 7 REPORTS OF COMMISSIONERS, INSPECTORS,
AND OTHERS 677 (1952-1953). For more anecdotal examples, see W. Weld & P. Cassell, supra note 196, at
app. C.
n206. See W. Weld & P. Cassell, supra note 196, at 15-19.
n207. See id. at 19 (explaining supply-demand econometric model of deterrence).
n208. See Layson, Homicide and Deterrence: A Reexamination of the United States Time-Series Evidence,
52 S. ECON. J. 68, 75, 80 (1984); Layson, United States Time-Series Homicide Regressions with Adaptive
Expectations, 62 BULL. N.Y. ACAD. MED. 589 (1986).
n209. See Layson, Homicide and Deterrence: Another View of the Canadian Time-Series Evidence, 16
CAN. J. ECON. 52 (1983).
n210. See Ehrlich, Capital Punishment and Deterrence: Some Further Thoughts and Additional Evidence,
85 J. POL. ECON. 741 (1977). This study is a response to criticisms of an earlier study by Ehrlich. See Ehrlich,
The Deterrent Effect of Capital Punishment: A Question of Life and Death, 65 AM. ECON. REV. 397 (1975).

Page 37
41 Stan. L. Rev. 121, *160

n211. See, e.g., Wolpin, Capital Punishment and Homicide in England: A Summary of Results, 68 AM.
ECON. REV. 422 (1978); Phillips & Ray, Evidence on the Identification and Causality Dispute About the Death
Penalty, in APPLIED TIME SERIES ANALYSIS 313 (1982). These studies are particularly informative since
Bedau and Radelet argue in their reply to this article that nothing we cite addresses the issue of whether the
death penalty has a marginal deterrent effect over long-term imprisonment. Bedau & Radelet, supra note 23, at
168-69. But Wolpin, as well as Phillips and Ray, controlled for imprisonment length and observed a unique
deterrent effect attributable to death sentences.
n212. See, e.g., Bartel, Women and Crime: An Economic Analysis, 17 ECON. INQUIRY 29 (1979); Block,
Nold & Sidak, The Deterrent Effect of Antitrust Enforcement, 89 J. POL. ECON. 429 (1981); Corman, Criminal
Deterrence in New York: The Relationship Between Court Activities and Crime, 19 ECON. INQUIRY 476
(1981).
n213. See, e.g., Viscusi, The Risks and Rewards of Criminal Activity: A Comprehensive Test of Criminal
Deterrence, 4 J. LAB. ECON. 317 (1986); Witte, Estimating the Economic Model of Crime with Individual
Data, 94 Q.J. ECON. 57 (1980). But see Myers, Estimating the Economic Model of Crime: Employment Versus
Punishment Effects, Q.J. ECON. 157 (1983) (challenging Witte's conclusions).
n214. Between 1900 and 1985, 7092 persons were lawfully executed. See Bedau & Radelet, supra note 1,
at 73. We multiplied Layson's estimate of the execution/homicide tradeoff (18) by 7092 to get 125,000.
Layson's data extends from 1933 to 1977. See Layson, supra note 208, at 70. We assume in our equation that
the tradeoff he calculates would apply at the same level for the entire period from 1900 to 1985.
n215. For instance, even if one were to contend that there was only a small (say 25%) chance that Layson's
figures were correct, then one would have to weigh 31,250 lives (25% x 125,000) in any risk assessment
equation.
Of course, to complete the calculation, one must consider the possibility of a counter-intuitive
"brutalization" effect (i.e., an increase in the homicide rate) attributable to society's use of the death penalty. See
Kaplan, The Problem of Capital Punishment, 1983 U. ILL. L. REV. 555, 560-61; see also Bedau & Radelet,
supra note 1, at 80. The anecdotal and statistical support for such an argument is extremely thin. See W. Weld
& P. Cassell, supra note 196, at 15-16. Moreover, Layson's equations demonstrate that any "brutalization"
effect is outweighed by capital punishment's deterrent effect. See id. at 16 n.17.
n216. Bedau & Radelet, supra note 1, at 79.
n217. As Ernest van den Haag has observed: "'All human activities -- building houses, driving a car, playing
golf or football -- cause innocent people to suffer wrongful death, but we don't give them up because on the
whole we feel there's a net gain.'" N.Y. Times, Nov. 14, 1985, at A19, col. 1. Similarly, van den Haag argues,
capital punishment is justified even though a few innocent persons may be executed, because it produces "a net
gain in justice . . . being done." Id. Bedau and Radelet claim that van den Haag's comparison is misleading, see
Bedau & Radelet, supra note 1, at 79, but it appears that they have misunderstood his position. Van den Haag's
argument is not that those who participate in such activities as building houses or driving cars consent to their
exposure to risk, but that such activities inevitably create risks for others who are not involved in those activities.
Society tolerates these risks because, on balance, the consequences of these pursuits are more desirable than not.
Nor is it fair of Bedeau and Radelet to criticize van den Haag's comparison on the ground that "the intention of

Page 38
41 Stan. L. Rev. 121, *160

capital punishment is to kill the convicted, whereas this is not the intention of the practices to which van den
Haag draws a parallel." Id. In truth, it is no more "the intention of capital punishment" to kill the innocent than it
is the intention of builders or drivers to do so. In either case, a mistake that causes the death of an innocent
person is an unintended cost of the activity.
n218. According to the authors, "[g]iven the irremediability of the death penalty and the availability of an
adequate alternative punishment, the execution of any defendant is unnecessary and cannot be justified even if
there were substantial benefits of capital punishment." Bedau & Radelet, supra note 1, at 79. One flaw in this
reasoning is that it takes as a premise the very question at issue -- "the availability of an adequate alternative
punishment." Beyond that, the conclusion that even substantial benefits cannot justify the execution of actual
murderers suggests that the authors' objection to capital punishment is not really based on the risk of executing
the innocent. They would apparently oppose the death penalty even if it could never result in a wrongful
execution and even though it would deter an indefinite number of prospective murders. See id. The authors
suggest that supporters of capital punishment "might see the issue differently" if they were "the innocent
defendants facing the executioner." Id. But, by the same token, opponents of capital punishment might see the
issue differently if they were the innocent victims of a kidnapper who had not yet decided whether to kill the
hostages. Cf. F. CARRINGTON, NEITHER CRUEL NOR UNUSUAL 99 (1978) (robbery victim was shot and
left for dead by assailants who had decided not to leave a witness because a murder would make no difference in
the "time" they would "do" if caught).
n219. SENATE COMM. ON THE JUDICIARY, supra note 197, at 13.
n220. Media General/Associated Press published the most recent national poll regarding the death penalty
based on interviews conducted from November 7-14, 1986. That poll found that 84% of Americans favor the
death penalty for especially brutal murders, and only 11% opose capital punishment entirely. The poll also
revealed that support for the death penalty crossed all racial, educational, economic, and regional lines. Media
General/Associated Press Poll: Death Penalty, Nov. 7-14 (1986), reprinted in W. Weld & P. Cassell, supra note
196, at app. B, at 1, 8.
n221. See notes 194-215 supra and accompanying text.
n222. Capital Punishment: Hearings Before the Comm. on the Judiciary, 97th Cong., 1st Sess., 246-47
(1981) (testimony of Walter Berns).
n223. Berns, Defending the Death Penalty, 26 CRIME & DELINQ. 503, 509 (1980).
n224. Id. at 511. Lesser penalties are inadequate to achieve this purpose and may, indeed, demean the value
of human life. As the husband of a murder victim stated following the recent convictions of several defendants
for his wife's brutal slaying in the District of Columbia: "I hate that D.C. doesn't have the death penalty. It's a lot
of people who think they can kill somebody and get away with it. Now we got to feed those prisoners. They'll
be sitting up looking at TV and having fun." Wash. Post, Dec. 19, 1985, at A39, col. 4.
n225. See E. van den Haag, Must the American Criminal Justice System Be Impotent? 2 (Sept. 19, 1985)
(White Paper prepared for the Washington Institute for Values in Public Policy) ("The sanctions of the criminal
justice system are effective only if inflicted upon the guilty and not on the innocent. If innocents were as liable

Page 39
41 Stan. L. Rev. 121, *160

to be punished as guilty persons, punishment would not be a reason for avoiding crime.").
n226. See van den Haag, supra note 7, at 68:
The irrevocability of a verdict of death is contrary to the modern spirit that likes to pretend that nothing ever
is definitive, that everything is open-ended, that doubts must always be entertained and revisions made. Such an
attitude may be proper for inquiring philosophers and scientists. But not for courts. They can evade decisions
on life and death only by giving up th eir paramount duties: to do justice, to secure the lives of the citizens, and
to vindicate the norms society holds inviolable.
Id.
n227. This conclusion is only strengthened by the authors' reply to our critique. See Bedau & Radelet,
supra note 23. As proponents of the view that capital punishment should be abolished because it entails an
unacceptable risk of erroneous execution, see Bedau & Radelet, supra note 1, at 80-81, they have the burden of
persuasion. Yet they now admit that neither they nor any previous researchers have proved any executed
defendant to be innocent. See Bedau & Radelet, supra note 23, at 164. They also now concede that the risk of
executing the innocent is relatively unimportant to the death penalty debate. See id. at 165-66.