Death Penalty Knowledge, Opinion, And Revenge: A Test Of The

University of Central Florida University of Central Florida

STARS STARS

Electronic Theses and Dissertations

2007

Death Penalty Knowledge, Opinion, And Revenge: A Test Of The Death Penalty Knowledge, Opinion, And Revenge: A Test Of The

Marshall Hypotheses In A Time Of Flux Marshall Hypotheses In A Time Of Flux

Gavin Lee

University of Central Florida

Part of the Criminology and Criminal Justice Commons

Find similar works at: https://stars.library.ucf.edu/etd

University of Central Florida Libraries http://library.ucf.edu

This Masters Thesis (Open Access) is brought to you for free and open access by STARS. It has been accepted for

inclusion in Electronic Theses and Dissertations by an authorized administrator of STARS. For more information,

please contact ST[email protected].

STARS Citation STARS Citation

Lee, Gavin, "Death Penalty Knowledge, Opinion, And Revenge: A Test Of The Marshall Hypotheses In A

Time Of Flux" (2007).

Electronic Theses and Dissertations

. 3237.

https://stars.library.ucf.edu/etd/3237

DEATH PENALTY KNOWLEDGE, OPINION, AND REVENGE:

A TEST OF THE MARSHALL HYPOTHESES IN A TIME OF FLUX

GAVIN LEE

B.S. Kaplan University, 2006

A thesis submitted in partial fulfillment of the requirements

for the degree of Master of Science

in the Department of Criminal Justice and Legal Studies

in the College of Health and Public Affairs

at the University of Central Florida

Orlando, Florida

Summer Term

2007

iii

ABSTRACT

This thesis tests the three hypotheses derived from the written opinion of Justice

Thurgood Marshall in Furman v Georgia in 1972. Subjects completed questionnaires at the

beginning and the end of the fall 2006 semester. Experimental group subjects were enrolled in a

death penalty class, while control group subjects were enrolled in another criminal justice class.

The death penalty class was the experimental stimulus. Findings provided strong support for the

first and third hypotheses, i.e., subjects were generally lacking in death penalty knowledge

before the experimental stimulus, and death penalty proponents who scored “high” on a

retribution index did not change their death penalty opinions despite exposure to death penalty

knowledge. Marshall’s second hypothesis--that death penalty knowledge and death penalty

support were inversely related--was not supported by the data. Two serendipitous findings were

that death penalty proponents who scored “low” on a retribution index also did not change their

death penalty opinions after becoming more informed about the subject, and that death penalty

knowledge did not alter subjects’ initial retributive positions. Suggestions for future research are

provided.

This thesis is dedicated to my children, Sebastian and Caitlin, and my astonishing wife, Miranda.

They have all endured my absence for the past 10 months. Without Miranda’s unyielding and

continuing support of my endeavors none of my achievements, within academia and beyond,

would have been possible. Thank you my love.

ACKNOWLEDGEMENTS

I would like to thank Dr. Brandon Applegate and Dr. Mark Lanier for their willingness to

form my thesis committee, and for their positive and insightful input throughout the process.

Special thanks are due to Professor Robert M. Bohm, my thesis Chair, invaluable advisor,

and mentor without equal. Professor Bohm has given me the knowledge, tools, and desire to

move on from this work and to enter our discipline with great enthusiasm to continue

researching, and eventually, to pass on this knowledge to my own students.

TABLE OF CONTENTS

LIST OF TABLES.......................................................................................................................viii

CHAPTER ONE: INTRODUCTION............................................................................................. 1

Furman v. Georgia and the Marshall Hypotheses...................................................................... 1

Testing the Marshall Hypotheses............................................................................................ 4

CHAPTER TWO: LITERATURE REVIEW................................................................................. 7

Results of Previous Studies......................................................................................................... 7

Hypothesis I: The American people are largely ignorant about the administration of capital

punishment.............................................................................................................................. 7

Hypothesis II: An informed American public would not support the death penalty.............. 9

Hypothesis III: If the primary reason people support capital punishment is based on the

principle of retribution, then increased knowledge about the subject would make little or no

difference to their death penalty opinions............................................................................. 15

Critique of Previous Studies ..................................................................................................... 17

Sampling............................................................................................................................... 17

Experimental Stimuli................................................................................................................ 21

Measures of Death Penalty Opinion......................................................................................... 24

Measures of Death Penalty Knowledge.................................................................................... 28

Measures of Retribution............................................................................................................ 32

CHAPTER THREE: METHODOLOGY ..................................................................................... 36

Experimental Design................................................................................................................. 36

vii

Measurement Procedures.......................................................................................................... 38

CHAPTER FOUR: RESULTS ..................................................................................................... 41

Hypothesis I: The American people are largely ignorant about the administration of capital

punishment................................................................................................................................ 41

Hypothesis II: An informed American public would not support the death penalty................ 42

Hypothesis III: If the primary reason people support capital punishment is based on the

principle of retribution, then increased knowledge about the subject would make little or no

difference to their death penalty opinions................................................................................. 50

CHAPTER FIVE: CONCLUSION............................................................................................... 54

APPENDIX A: CCJ 4361-0001: DEATH PENALTY FALL, 2006............................................ 59

APPENDIX B: DEATH PENALTY OPINION AND KNOWLEDGE SURVEY...................... 64

LIST OF REFERENCES.............................................................................................................. 68

viii

LIST OF TABLES

Table 1 Studies that Tested All or Part of Marshall’s Hypotheses................................................. 6

Table 2 Sample Sizes of Previous Tests of the Marshall Hypotheses.......................................... 20

Table 3 Demographic Characteristics of Experimental and Comparison Groups on Pretest and

Posttest (in percentages) ............................................................................................................... 37

Table 4 Comparison of Pretest Percentages and Means of Experimental and Comparison Groups

on the Knowledge Index............................................................................................................... 42

Table 5 Comparison of means: Death penalty support for some people convicted of first-degree

murder........................................................................................................................................... 44

Table 6 Experimental Group: Concrete Items............................................................................. 45

Table 7 Comparison Group Concrete Items ................................................................................. 47

Table 8 Comparison of means: LWOP as an alternative to death penalty ................................... 49

Table 9 Comparison of means: LWOP + as a death penalty alternative...................................... 50

Table 10 Experimental Group “Low” Retributivists: Pretest v. Posttest...................................... 52

Table 11 Experimental Group “High” Retributivists: Pretest v. Posttest..................................... 53

CHAPTER ONE: INTRODUCTION

The purpose of this thesis is to test “The Marshall Hypotheses.” These hypotheses were

postulated by Justice Thurgood Marshall in his decision in the case of Furman v. Georgia (408

U.S. 238, 1972):

1. The American people are largely ignorant about the administration of capital punishment.

2. An informed American public would not support the death penalty.

3. If the primary reason people support capital punishment is based on the principle of

retribution, then increased knowledge about the subject would make little or no

difference to their death penalty opinions.

This research is important because the death penalty in the United States appears to be

driven by public opinion. Indeed, Bohm (2003) states that it would be very difficult to sustain the

penalty where the public does not support it. Unfortunately, many studies (described later) have

found that public opinion about the death penalty is not only poorly informed, but often

misinformed. The problem, as Justice Marshall recognized, is that the public’s decision about

whether an individual lives or dies, that is, its choice about the death penalty, should be “a

knowledgeable choice” (Furman v. Georgia, 1972 at 362, fn.145).

Furman v. Georgia and the Marshall Hypotheses

In 1972, the United States Supreme Court (the Court) decided three cases regarding the

constitutionality of the death penalty as it was applied at the time. The cases were Furman v.

Georgia (408 U.S. 238) consolidated with Jackson v. Georgia (No. 69-5-30) and Branch v.

Texas (No. 69-5031). The Branch and Furman cases dealt with the death penalty’s

constitutionality with respect to murder, whereas the Jackson case dealt with the same penalty,

but for the crime of rape (in the absence of murder).

The question before the Court was: Does “the imposition and carrying out of the death

penalty in these cases constitute cruel and unusual punishment in violation of the Eighth and

Fourteenth Amendments” (Furman v. Georgia, 1972 at 239). In one of the lengthiest Supreme

Court decisions in its history, the Court held, by a 5-4 ruling, that the death penalty was indeed

unconstitutional. Every justice wrote an opinion. That the justices were so forceful and singular

in their views, and so narrowly split, tended to show that the Supreme Court was a microcosm of

the powerfully held and diverse views of the general public. Justices Douglas, Stewart and White

wrote that the death penalty was indeed unconstitutional, but not irredeemably so. In contrast, the

remainder of the majority, Justices Brennan and Marshall, declared the death penalty to be

unconstitutional in all cases. The Court’s rationale for the death penalty’s unconstitutionality was

that the penalty was carried out in an arbitrary and capricious manner, with many of the justices

citing racial concerns as particularly troublesome (Furman v. Georgia, 1972). The outcome of

the Court’s ruling was the voiding of all existing death penalty statutes, and the commutation of

the sentences of all condemned inmates to either life without the possibility of parole or, in some

instances, life with the possibility of parole.

In Marshall’s Furman opinion, he wrote that if the American public’s death penalty

opinion was going to be probative, it was a constitutional necessity that the public should be

knowledgeable of the death penalty and that this knowledge should be discerned (Furman v.

Georgia, 1972 at 362, fn. 145). Justice Marshall explained, “The question with which we must

deal is not whether a substantial proportion of American citizens would today, if polled, opine

that capital punishment is barbarously cruel, but whether they would find it to be so in the light

of all information presently available” (Furman v. Georgia, 1972 at 362). In clarification,

Marshall further remarked, “This is not to suggest that with respect to this test of

unconstitutionality people are required to act rationally; they are not. With respect to this

judgment, a violation of the Eighth Amendment is totally dependant on the predictable,

subjective, emotional reactions of informed citizens” (my emphasis) (Furman v. Georgia, 1972

at 362).

As noted, in his first hypothesis Marshall presumed that most Americans are not

knowledgeable about the death penalty, which begs the question what Americans must know if

they are to be considered informed about the death penalty. In Furman, he suggested the

following:

that the death penalty is no more effective a deterrent than life imprisonment, that

convicted murderers are rarely executed but usually sentenced to a term in prison; that

convicted murderers usually are model prisoners, and that they almost always become

law abiding citizens upon their release from prison; that the costs of executing a capital

offender exceed the costs of imprisoning him for life; that while in prison, a convict

under sentence of death performs none of the useful functions that life prisoners perform;

that no attempt is made in the sentencing process to ferret out likely recidivists for

execution; and that the death penalty may actually stimulate criminal activity (Furman v.

Georgia, 1972 at 364, footnotes omitted).

Marshall believed, as suggested in his second hypothesis, that “[t]his information would surely

convince average citizens that the death penalty was unwise” (Furman v. Georgia, 1972 at 364).

Regarding his third hypothesis, Marshall recognized that even if Americans knew the

aforementioned information, they still may not consider the death penalty morally reprehensible

and oppose it. The reason, according to Marshall, is that many Americans support the death

penalty primarily for retributive reasons, and retributivists are less likely to respond to

knowledge-based claims. On the subject of retribution, Marshall wrote, “Retaliation, vengeance,

and retribution have been roundly condemned as intolerable aspirations for a government in a

free society” (Furman v. Georgia, 1972 at 343). He went on to posit, “The history of the Eighth

Amendment supports only the conclusion that retribution for its own sake is improper” (Furman

v. Georgia, 1972 at 345) and that “no one has ever seriously advanced retribution as a legitimate

goal of our society” (Furman v. Georgia, 1972 at 363). Marshall’s view of retribution, which he

equated with purposeless vengeance, was so negative that he maintained:

I cannot believe that at this stage in our history, the American people would ever

knowingly support purposeless vengeance. Thus, I believe that the great mass of citizens

would conclude on the basis of the material already considered that the death penalty is

immoral and therefore unconstitutional (Furman v. Georgia, 1972 at 363-364).

It is clear from his previous remark he was using the precedent set by the Court’s decision in

Trop v. Dulles (356 U.S. 86, 1958 at 102) that a punishment must be judged in terms of “the

evolving standards of decency that mark the progress of a maturing society.”

Testing the Marshall Hypotheses

The bold statements made by Justice Marshall in such a landmark case drew almost

immediate attention and study from many social scientists due to its far reaching scope and

originality of thought. A literature review revealed seventeen studies that tested all or part of the

Marshall hypotheses. Table 1 shows the studies and what part of the Marshall hypotheses they

tested. However, there were problems with the testability of the hypotheses from the start: What

constitutes knowledge, and how is it to be measured? What constitutes death penalty opinion,

and how is it to be measured? What constitutes retributive reasons for death penalty support, and

how are they to be measured? In addition, previous studies of the Marshall Hypotheses faced

problems with sampling method, the time and context in which the data were gathered, and the

stimuli used to make subjects more knowledgeable. The review of this literature examines the

previous studies in light of these problems and reports the methods they employed, their

findings, and their relative strengths and weaknesses.

Table 1

Studies that Tested All or Part of Marshall’s Hypotheses

HYPOTHESES

STUDY

H1 H2 H3

Sarat and Vidmar (1976) Y Y Y

Lord, Ross and Lepper (1979) Y Y N

Vidmar and Dittenhoffer (1981) Y Y N

Ellsworth and Ross (1983) Y Y N

Bohm (1989) Y Y N

Bohm (1990) Y Y N

Bohm et al. (1990) Y Y N

Bohm et al. (1991) Y Y Y

Bohm and Vogel (1991) Y Y N

Bohm et al. (1993) Y Y N

Sandys (1995) Y Y N

Wright, Bohm and Jamieson (1995) Y Y N

Bohm and Vogel (2004) Y Y Y

Clarke, Lambert and Whitt (2000-2001) Y Y N

Lambert and Clarke (2001) Y Y N

Cochran and Chamlin (2005) Y Y Y

Cochran, Sanders and Chamlin (2006) Y Y Y

Note: Y= yes; N= no.

CHAPTER TWO: LITERATURE REVIEW

Results of Previous Studies

Hypothesis I: The American people are largely ignorant about the administration of capital

punishment

At the outset, it is important to note, as Bohm (2003, p. 260) has, that “[w]hat Marshall

fails to stipulate is how much information a citizen must know in order to be informed.” Bohm

(2003, p. 260) asks, “Must one know all of it or will 50 to 60 percent suffice?” “Without setting a

standard,” Bohm asserts, “Marshall leaves unanswered the key question of what it means to be

informed or knowledgeable about the death penalty” (p.260). In the first test of the Marshall

hypotheses, Sarat and Vidmar (1976) found that, in response to six death penalty-related items, a

mean of 46 percent was answered correctly. The authors concluded, “few persons in our sample

could be labeled ‘informed’ about the death penalty” (p.171), and further, “subjects knew little

about the death penalty, particularly its effectiveness” (p.195).

Ellsworth and Ross (1983) also found their subjects ignorant about the death penalty.

They used nine death penalty-related items to test their subjects’ knowledge. There was only one

item of the nine that more than 50 percent answered correctly. They concluded, “Our findings of

high levels of ignorance, often willingly admitted, are consistent with the findings of Sarat and

Vidmar [1976]” (p.144).

Bohm, Clark and Aveni (1990) found that their experimental-group subjects initially

answered correctly only 50 percent of 14 knowledge items, while comparison-group subjects

answered correctly only 39 percent of the knowledge items. Using a different sample, Bohm,

Clark and Aveni (1991) found that their subjects initially answered correctly, on average, only 52

percent of 14 knowledge items. Whether these results represent being knowledgeable, the

authors write, “[W]e choose to regard such a score [52%, 50%, or 39%] as indicative of being

relatively uninformed” (p.369).

Wright, Bohm and Jamieson (1995) used 17 knowledge items to measure death penalty

knowledge. Their subjects answered correctly an average of only 47 percent of the items. They

found that, with two exceptions, “no…knowledge item was answered correctly by more than

61.4% of all subjects, and only seven of the 17 knowledge items were correctly answered by

more than half the subjects” (p.64). The exceptions were that 78 percent of their subjects

correctly knew or guessed that “poor people who commit murder are more likely than rich

people to be sentenced to death,” and 71 percent of their subjects correctly knew or guessed that

“for the same crime, men have been more likely to be executed than women” (p.64). Wright et

al. concluded that the “students had some knowledge about the death penalty, but they cannot be

considered well informed” (p.64).

Cochran, Sanders and Chamlin (2006) only report the amount of knowledge gained as a

result of participating in a death penalty class. They measured knowledge on a pretest at the

beginning of the class and on a posttest at the end of the class. They report, “The average

difference in these scores was a 45.4 percentage point improvement . . . . The percent of change

in knowledge gain was also substantial (203%). That is, on average these students doubled their

time 1 score on the final exam. In terms of their relative knowledge gains, the students, on

average, improved their time 1 scores by 71% of what they maximally could have at time 2” (p.

213). These findings indicate that Cochran et al.’s subjects initially were not well informed about

the death penalty (also see Cochran and Chamlin, 2005). Other studies of the Marshall

hypotheses did not test hypothesis one.

In sum, seven studies directly tested Marshall’s first hypothesis. These analyses led

Bohm et al. (1991, p. 361) to claim, “A majority of Americans have taken a very strong position

on an issue about which they are substantially uninformed.” It appears that Marshall’s first

hypothesis is supported by the available data.

Hypothesis II: An informed American public would not support the death penalty

The literature review revealed two studies that strongly supported Marshall’s second

hypothesis, nine studies that provided some support, and six studies that seem to suggest that

Marshall was wrong.

Studies that Strongly Support Marshall’s Second Hypothesis

Vidmar and Dittenhoffer (1981) and Sandys (1995) found a strong inverse relationship

between knowledge about the death penalty and support for it, confirming Marshall’s contention.

In Vidmar and Dittenhoffer’s 1981 study, support of the death penalty decreased from 48 percent

to 24 percent, opposition increased from 33 percent to 71 percent, and undecideds decreased

from 19 percent to 5 percent, when subjects became more knowledgeable about capital

punishment. Sandys’ 1995 study found that support of the death penalty decreased from 56

percent to 22 percent, opposition increased from 35 percent to 65 percent, and undecideds

increased from 9 percent to 13 percent, again as subjects’ capital punishment knowledge

increased. The Vidmar and Dittenhoffer (1981) and Sandys (1995) studies are unique in the

literature because they are the only studies to find a majority of subjects opposed to the death

penalty after exposure to knowledge about it. A problem with both studies, however, is that they

were based on extremely small samples: Vidmar and Dittenhoffer (1981) had 21 experimental

group subjects, and Sandys (1995) had 23 experimental group subjects. Other problems with

these studies are addressed later.

Studies that Provide Some Support for Marshall’s Second Hypothesis

Sarat and Vidmar (1976), Bohm (1989), Bohm, Clark and Aveni (1991), Bohm and

Vogel (1991), Wright, Bohm and Jamieson (1995), Clarke, Lambert and Whitt (2000-2001),

Lambert and Clarke (2001), Cochran and Chamlin (2005), and Cochran, Sanders and Chamlin

(2006) all found some degree of support for the second hypothesis, but none of them reported a

majority of subjects opposed to the death penalty after becoming informed about it. However, in

a few studies a majority of some subgroups were opposed to the death penalty after becoming

more informed. Sarat and Vidmar (1976) found that after exposure to knowledge about the death

penalty support declined by 20 percentage points (from 62% to 42%), opposition increased by 11

percentage points (from 27% to 38%), and undecideds increased by 11 percentage points (from

10% to 21%).

Following exposure to knowledge, Bohm (1989) discovered that death penalty support

fell 26 percentage points (from 82% to 56%), opposition increased 24 percentage points (from

18% to 42%), and undecideds increased by 2 percentage points (from 0% to 2%). The racial

change was an interesting factor. While white support decreased from 88 percent to 72 percent

and opposition increased from 12 percent to 29 percent, black support decreased from 72 percent

to 28 percent and opposition increased from 28 percent to 67 percent (undecideds are not

reported). Two other interesting findings of this study were that females were the group most

resistant to change (support decreased from 74% to 65% and opposition increased from 26% to

35%), and no female was ever undecided.

Bohm et al. (1991) employed four measures of death penalty opinion: 1) death penalty

for some people convicted of first-degree murder, 2) death penalty for all people convicted of

first degree murder, 3) whether one could convict if served on a capital jury, and 4) whether one

could pull the lever to instigate an execution. After becoming more informed about the death

penalty subjects opposed to the death penalty for some persons increased from 28.0 percent to

49.5 percent, subjects opposed to the death penalty for all persons increased from 28.3 percent to

46.6 percent, persons who could not convict increased from 5.4 percent to 21.5 percent and

persons who could not pull the lever increased from 37.6 percent to 40.9 percent. The last change

was the only one not statistically significant at p < 0.05. Bohm et al. (1991) did not report

changes in support of the death penalty or changes in undecideds.

Bohm and Vogel (1991) saw death penalty support decrease from 85.5 percent to 57

percent, opposition increase from 11.5 percent to 33 percent, and undecideds increased from 3

percent to 10 percent, after exposure to death penalty knowledge. Similarly, Wright et al. (1995)

found a decrease in support and an increase in opposition after becoming more informed about

the death penalty. Support declined from 79 percent to 56 percent and opposition increased from

13 percent to 36 percent. Those with no opinion remained constant at 8 percent.

Clarke et al. (2000-2001) also found some change in death penalty opinion after subjects

became more informed about the penalty. Death penalty support decreased from 60 percent to 55

percent, death penalty opposition increased from 28 percent to 32 percent, and the percentage of

those who declared themselves to be “uncertain” moved from 12 percent to 13 percent. The

authors concluded, “The reduction in support for the death penalty…was not the major reduction

originally postulated by Justice Marshall when he wrote in Furman that persons who ‘were fully

informed as to the purposes of the penalty and its liabilities would find the penalty shocking,

unjust and unacceptable’” (Clarke et al., 2000-2001, p.337, fn.127). In another study, Lambert

and Clarke (2001) specifically addressed whether knowledge of deterrence and knowledge of

innocence affected death penalty opinions. They found that knowledge about deterrence did not

change the death penalty opinions of 55.6 percent of their subjects but changed somewhat the

death penalty opinions of 43.6 percent of their subjects and changed a great deal the death

penalty opinions of 0.8 percent of their subjects. Similarly, knowledge about innocence did not

change the death penalty opinions of 55.6 percent of their subjects but changed somewhat the

death penalty opinions of 34.7 percent of their subjects and changed a great deal the death

penalty opinions of 9.8 percent of their subjects.

Cochran and Chamlin (2005) and Cochran et al. (2006) showed identical results. (The

2006 sample included the 2005 sample.) After exposure to death penalty knowledge, death

penalty support decreased by 14 percentage points (from 64% to 50%), opposition increased by

27 percentage points (from 19% to 46%), and undecideds decreased by 13 percentage points

(from 17% to 4%). However, they also found that one-third of their subjects did not change their

death penalty opinions, and 13 percent of their subjects increased their level of death penalty

support after learning about the death penalty (Cochran et al., 2006). Though some may take

issue, the researchers concluded, “[I]mparting knowledge about the death penalty does not work

as well as Justice Marshall argued” (Cochran et al., 2006, p. 214).

Studies that Do Not Support Marshall’s Second Hypothesis

Six studies failed to support Marshall’s second hypothesis: Lord, Ross, and Lepper

(1979), Ellsworth and Ross (1983), Bohm, (1990), Bohm et al. (1990), Bohm, Vogel, and Maisto

(1993), and Bohm and Vogel (2004). Contrary to Marshall’s second hypothesis, Lord et al.

(1979) did not find that death penalty knowledge reduced death penalty support; rather

knowledge about the death penalty strengthened previously held death penalty opinions. Lord et

al. explained, “The net effect of exposing proponents and opponents of capital punishment to

identical evidence…was to increase further the gap between their views”—to “polarize” their

views (p. 2105). They noted, “Data relevant to a belief are not processed impartially” (p. 2099)

but, instead, are assimilated in a biased manner. They concluded:

Judgments about the validity, reliability, relevance, and sometimes even the meaning of

proffered evidence are biased by the apparent consistency of that evidence with the

perceiver’s theories and expectations. Thus individuals will dismiss and discount

empirical evidence that contradicts their initial views and will derive support from

evidence, of no greater probativeness, that seem consistent with their views.

(Lord et al., 1979, p. 2099)

Ellsworth and Ross (1983) and Bohm et al. (1990) likewise reported that “biased

assimilation” and “attitude polarization” could account for the results in their studies. (In this

thesis, the concepts attitude and opinion are used interchangeably). Ellsworth and Ross (1983)

claimed, “The picture that emerges is one of an emotionally based attitude, tempered by a sense

of social desirability” (p. 152). Bohm et al. (1990) surmised, “Thus reasons provided for death

penalty opinions appear to be primarily rationalizations or justifications for emotionally based

opinions, and we ought not to expect them to be particularly responsive to reasoned persuasion”

(p. 184).

Bohm (1990) also failed to find support for Marshall’s second hypothesis. In this study

Bohm asked subjects to publicly announce their death penalty positions at the beginning of each

death penalty class (more about the death penalty class later). He discovered that at the end of the

semester (after exposure to information about the death penalty) death penalty support decreased

only 5 percentage points (from 62% to 57%), and death penalty opposition increased only 5

percentage points (from 31% to 36%). By contrast, death penalty support among comparison

group subjects decreased by 12 percentage points (from 67% to 55%), and death penalty

opposition among comparison group subjects increased by 14 percentage points (from 22% to

36%). Thus, there was more change among comparison group subjects, who did not receive

death penalty information, than there was among subjects who received a semester’s worth of

death penalty information. Bohm (1990) surmised that having to publicly commit to a death

penalty position inhibited either death penalty opinion change or the voicing of death penalty

opinion change.

Finally, Bohm et al. (1993) and Bohm and Vogel (2004) conducted panel studies two-to-

three years and more than ten years, respectively, after subjects completed a semester-long death

penalty class. Both these studies showed that subjects’ support for the death penalty decreased

between the beginning of the death penalty class and its end. However, follow-up data revealed a

“rebound” in death penalty support. That is, after two-to-three years or more than ten years later,

subjects’ level of death penalty support had returned to, or exceeded, their initial level of death

penalty support. The researchers concluded that “classroom knowledge” may not prove

especially effective in changing death penalty opinions in the long term. Nevertheless, as Bohm

and Vogel (2004, p. 325) cautioned:

This does not mean that death penalty opinions are intransigent. Opinions do change, as

evidenced by the more than sixty-five year history of public opinion polls on the death

penalty in the United States. This study only indicated that, under one condition [a

semester-long death penalty class], most death penalty opinions (and reasons for

opinions) might be impervious to long-term change. If Justice Marshall had in mind a

stimulus like the one employed in this study, then his belief that death penalty opinions

could be changed substantially might be wrong.

Hypothesis III: If the primary reason people support capital punishment is based on the principle

of retribution, then increased knowledge about the subject would make little or no difference to

their death penalty opinions

Only five of the studies tested Marshall’s third hypothesis. Of those that did, all but one

(Cochran and Chamlin, 2005) found at least some support for it (Sarat and Vidmar, 1976, Bohm

et al., 1991, Bohm and Vogel, 2004, Cochran et al., 2006).

Summarizing their findings, Sarat and Vidmar (1976, p. 194) wrote:

retributiveness is more important in differentiating among supporters and opponents of

capital punishment than is any of the kinds of information contained in the three

experimental conditions…retributive motives are highly correlated with the extent of

change in death penalty support…persons low in retribution, persons whose level of

support for the death penalty was initially quite low…showed a further alteration in their

positions…[t]he effect of information among respondents scoring high on the measure of

retribution was…uniformly minimal…[t]he broad pattern of results thus appears to

confirm Marshall’s suspicion that even an informed public opinion might not reject the

death penalty to the extent that initial support for it is grounded in a desire for vengeance

and retribution against those who commit capital crimes.

Bohm et al. (1991) found that those who were highly retributivist did not change their

initial death penalty support very much (from 81.8% to 80%), whereas those who were less

retributivist did change their opinions to a greater extent (23.3% in support to 6.7% in support).

In Bohm and Vogel’s (2004) panel study, the effect of retribution and knowledge on

death penalty opinions were ascertained at four points in time. As Marshall predicted, subjects

who agreed with the statement, “Society has a right to get revenge when a very serious crime like

murder has been committed,” changed their death penalty opinions little over time. However,

subjects who disagreed with the statement also did not change their death penalty opinions very

much over time. What Bohm and Vogel (2004) found surprising was that death penalty

knowledge had little effect on death penalty opinions regardless of subjects’ belief about

retribution.

Likewise, Cochran et al. (2006) reported, “Those students who maintained their level of

death penalty support between time 1 and time 2 were also the most retributive; similarly those

who maintained their time 1 level of opposition to capital punishment were the least retributive”

(p. 223). Thus, these findings not only support Marshall’s third hypothesis about highly

retributive subjects, but they also show that death penalty opponents who are not retributivists

are unlikely to change their death penalty opinions after becoming more knowledgeable about

the subject either.

The one study that did not provide support for Marshall’s third hypothesis was Cochran

and Chamlin (2005). They observed that “gains in knowledge, when objectively assessed, did

significantly decrease support for capital punishment among those who espoused retribution” (p.

580).

Critique of Previous Studies

Sampling

Experimental / Quasi-Experimental Design

Of all the studies that tested the Marshall hypotheses, only Sarat and Vidmar (1976) used

a true experimental design. Sarat and Vidmar randomly selected a final sample of 181 subjects

from Amherst, Massachusetts and randomly assigned them to experimental and control groups.

All of the other studies are quasi-experimental, because they either do not employ random

selection and assignment of subjects to experimental and comparison groups or do not use

comparison groups at all.

Geographical and Temporal Aspects

Sarat and Vidmar (1976), Lord et al. (1979), Vidmar and Dittenhoffer (1981), and

Ellsworth and Ross (1983) all used data gathered at a time, or in a place where there were no

executions being carried out, so in a sense the death penalty was an abstract concept. Data in

Sarat and Vidmar’s study were collected in Amherst, Massachusetts in 1975. There had not been

an execution in the United States since 1968. Although Massachusetts did not abolish its death

penalty until 1984, it had not executed anyone since 1947 (Bedau, 1997). In addition,

Massachusetts is considered an especially liberal area of the United States and therefore it

probably is not representative of the United States as a whole (The Bay Area Center for Voting

Research, nd).

Lord et al. (1979) did not report when their data were collected from Stanford University

students in California. However, by 1979, the publication year of their research, there could not

have been more than three executions nationwide in more than a decade. The last execution in

California was in 1967. Also, although Stanford University draws its students from all over the

United States and the world, it is not clear how geographically representative a sample of

Stanford University students is.

Vidmar and Dittenhoffer (1981) collected their data in Canada; they did not indicate

when their data were collected. Canada abolished its death penalty in 1976 and had not executed

anyone since 1962 (Bedau, 1997). Therefore their findings may not be particularly relevant to the

United States

Ellsworth and Ross (1983) collected their data in the San Francisco Bay area during the

summer and fall of 1974. Like the data in the Sarat and Vidmar (1976) study, the data in the

Ellsworth and Ross (1983) study were collected in a liberal area of the country (The Bay Area

Center for Voting Research, nd) at a time when executions were not being conducted. However,

unlike the Sarat and Vidmar (1976) study, Ellsworth and Ross took pains to reduce any liberal

bias in their sample. Specifically, they used “a stratified random sampling procedure based on

the precinct voting records of… [four] towns in the 1972 California referendum on the

restoration of capital punishment” (p. 123). Two of the towns tended to support capital

punishment in the 1972 referendum, and the two other towns tended to oppose capital

punishment in the same referendum. To obtain an even more representative sample, each town

was divided into five groups, “according to the degree of support for capital punishment they had

shown in 1972” (p. 123). From each group, within each town, 25 households were randomly

sampled.

All the remaining studies collected data from students at universities in the United States

at a time when the death penalty was a salient issue, that is, when people were being executed.

Bohm (1989), Bohm (1990), Bohm et al. (1990), Bohm et al. (1991), Bohm and Vogel (1991),

Bohm et al. (1993), and Bohm and Vogel (2004) collected their data from students in Alabama.

Wright et al. (1995) collected their data from students in North Carolina. Data for Sandys (1995)

study were obtained from Indiana students. Clarke et al.’s (2000-2001) data came from an

unidentified university in the Midwest. Lambert and Clarke’s (2001) data were from Michigan.

Cochran and Chamlin (2005) and Cochran et al. (2006) obtained their data in Florida.

Sample Size and Composition

Sample sizes varied greatly. On the low end, Vidmar and Dittenhoffer (1981) had 21

subjects in an experimental group and 18 subjects in a comparison group. In this study 67

percent were female and only 33 percent were male. No other demographic data were given

other than the subjects were English Canadian. It does not appear that Vidmar and Dittenhoffer’s

sample was representative of the Canadian population. Sandys (1995) had 23 experimental group

subjects and no comparison group. Seventy-four percent of Sandys’ (1995) sample was male and

91 percent of her sample was white, hardly representative of the general population of the

United States. As mentioned previously, these were the only two studies that found a majority of

subjects opposed to the death penalty after becoming more informed about it and, thus, the only

studies that strongly supported Marshall’s second hypothesis.

On the high end, Lambert and Clarke (2001) had 488 experimental group subjects and

242 comparison group subjects for a total of 730 subjects. Ellsworth and Ross (1983) report 500

experimental group subjects but no comparison group. Table 2 provides sample sizes for all of

the studies discussed in this thesis.

Table 2

Sample Sizes of Previous Tests of the Marshall Hypotheses

STUDY E C TOTAL

Sarat and Vidmar (1976) 181 45 181

Lord, Ross and Lepper (1979) 48 No 48

Vidmar and Dittenhoffer (1981) 21 18 39

Ellsworth and Ross (1983) 500 No 500

Bohm (1989) 50 No 50

Bohm (1990) 59 54 109

Bohm et al. (1990) 44 27 71

Bohm et al. (1991) 190 82 272

Bohm and Vogel (1991) 70 35 105

Bohm et al. (1993) 106 No 106

Bohm and Vogel (2004) 69 No 69

Wright, Bohm and Jamieson (1995) 38 68 106

Sandys (1995) 23 No 23

Clarke, Lambert and Whitt (2000-2001) 110 dropped and 107 retained 103 324*

Lambert and Clarke (2001) 488 242 730

Cochran and Chamlin (2005) 70 No 70

Cochran, Sanders and Chamlin (2006) 365 No 365

* = The 110 subjects dropped had received experimental stimuli not germane to the study

discussed.

As noted in the previous section, all but two of the studies used student subjects.

Although student subjects are rarely representative of the general population, they were

employed in these studies because it was one of the only ways to expose subjects to the

experimental stimulus (death penalty knowledge) over an extended period of time. Brevity of

exposure to the experimental stimulus was a weakness of earlier studies (e.g., Sarat and Vidmar,

1976).

Comparison Groups

Only nine of the seventeen studies employed comparison groups. Thus, in those studies

without comparison groups it is not possible to determine if any changes in death penalty

opinions are a function of the experimental stimulus or some other factor or factors. Table 2

shows the studies that utilized comparison groups and those that did not.

Experimental Stimuli

The first few tests of the Marshall hypotheses used questionable experimental stimuli.

For example, after conducting an initial survey, Sarat and Vidmar (1976) administered their

experimental stimulus. This was in the form of two different 1,500 word essays. One of the

essays consisted of statistical studies, personal experiences, deterrence psychology, and

recidivism data for homicide offenders. They dubbed this the utilitarian essay. The second essay-

the humanitarian essay--dealt with the physical and psychological aspects of executions. The

sample was then split into four groups. These groups received either essays one and two, essay

one only, or essay two only. The fourth group—the control group—received an essay on

unrelated legal matters with no death penalty ramifications. After reading these essays the

subjects were requested to complete a section of the original survey which examined the

“application and effects” of the death penalty. Sarat and Vidmar warned of their study’s

limitations, stating that due to the limited manner of informing the subjects and the lack of time

given for reflection and discussion, their findings should be interpreted cautiously.

In Lord et al. (1979) subjects read two cards containing synopses of death penalty

deterrence study results: one containing pro-deterrent information and the other containing anti-

deterrent information. The subjects were then asked if they had changed their belief in the

deterrent effect of capital punishment. Subjects were then given methodological analyses of the

studies and asked whether or not they found the studies convincing. Then the subjects were

asked about their capital punishment views concerning its deterrent effect. A problem with the

experimental stimulus was that it was narrowly focused only on the death penalty’s deterrent

effect and did not address all of the other information a person would need to know to be

considered informed about the death penalty. Another problem, common to these early studies,

was that there was little time for subjects to reflect upon the information they had been given.

In the Vidmar and Dittenhoffer (1981) study, the stimulus was 8 papers and a

3,500 word essay about the death penalty, with the option of reading two books by respected

death penalty scholars. This was followed by a discussion period, which took place two weeks

after the stimulus. The discussion groups consisted of three to four individuals each and were

designed to improve the validity of the Sarat and Vidmar study of 1976 in which there was no

discussion period. However, despite this improvement, Vidmar and Dittenhoffer’s study did have

methodological problems with the experimental stimulus. First, there was no guarantee that

subjects read the required material. Second, no attempt was made to measure the amount of

information subjects understood and internalized. Third, with regard to the discussion groups,

there were no checks made to see if subjects were even discussing the death penalty and their

readings. Fourth, a one hour discussion may not have been long enough to achieve the desired

goal.

Ellsworth and Ross (1983) did not use an experimental stimulus. They only gathered

information about their subjects’ death penalty opinions at one point in time.

In a design similar to the earlier studies, Clarke et al. (2000-2001) and Lambert and

Clarke (2001) used three essays as the experimental stimulus. The first essay was a comparison

essay which dealt with the philosophical nature of punishment but did not discuss the death

penalty itself. The second essay discussed empirical evidence about the death penalty and its

deterrent effect. It showed that studies did not support a deterrent effect for capital punishment.

The third essay focused on the “frequency and probability of sentencing innocent people to

death” (p.328). These essays were distributed randomly to students who were enrolled in

criminal justice and general education classes. It does not appear that either study used a death

penalty class as a stimulus.

The experimental stimulus in the remaining studies was a semester-long death penalty

class because, as Bohm (1990) has stated, “Brevity of exposure to the experimental stimulus is a

weakness of previous research” (p.287). The classes consisted of lectures, guest speakers, videos,

and assigned readings and were comprehensive examinations of capital punishment in the United

States. Although a semester-long class on the death penalty is a far better method for delivering

information than reading a few essays or a couple of books, a death penalty class is not the only

source of information about the death penalty. Witnessing an execution and talking to co-

victims, for example, would be other ways of becoming more knowledgeable about the death

penalty.

A problem with using a classroom experience for informing people about the death

penalty is the influence of the instructor—the “experimenter effect.” A charismatic or obnoxious

instructor, for example, could have an extraordinary influence on students beyond the class

material. In the Bohm and Cochran studies, the class instructor was forthcoming about his

abolitionist viewpoint but played advocatus diaboli to whatever position a student took. Both

researchers made it clear that students’ death penalty opinions would have no bearing on their

final grades. Other researchers either did not reveal their death penalty positions or, if they did,

did not report them in their studies. Bohm (1989) has suggested that having a proponent of the

death penalty teaching the class may produce different results.

Measures of Death Penalty Opinion

Many studies have found that the wording of questions and response categories used to

ascertain subjects’ death penalty opinions can make a great deal of difference to their response.

Williams, Longmire and Gulick (1988), for example, found that as questions become less

abstract and more concrete the level of reported support significantly decreases. They also found

that the public seemed to be surer about their general support for the death penalty than under

what specific circumstances to use it.

Sarat and Vidmar (1976) and Lord et al. (1979) tested opinion before and after the

experimental stimuli, as did a majority of the other studies. Sarat and Vidmar (1976) used a

seven-point Likert scale (“very strongly agree,” “strongly agree,” “somewhat agree,”

“uncertain,” “somewhat disagree,” “strongly disagree,” and “very strongly disagree”) to measure

death penalty opinion, but they did not report the question that was asked of their subjects. Also,

Lord et al. (1979) did not report how they nominally or operationally defined death penalty

opinion. The subjects, after administration of the experimental stimulus, were asked to simply

note their change in death penalty opinion on a sixteen-point scale.

Ellsworth and Ross’s (1983) survey measured death penalty opinion in a much more

thorough way than many of the other studies. The researchers asked one general death penalty

attitude question, “Do you believe in capital punishment or are you opposed to it?” (p. 126).

Response categories were “believed in,” “opposed,” and “undecided.” Questions were also asked

regarding mandatory or discretionary death penalty sentences for a range of scenarios and,

finally, questions were asked about the subjects’ predicted behavior if they were ever to be on a

capital jury. This question specifically dealt with the level of evidence they would require in

order to convict. In this study, the questionnaire was administered only once.

Bohm (1989), Bohm (1990), Bohm et al. (1990), Bohm et al. (1991), Bohm and Vogel

(1991), Wright et al. (1995), Bohm et al. (1993), and Bohm and Vogel (2004) used a seven-point

Likert scale on their surveys to gauge opinion. In Bohm (1989) and Bohm (1990), the survey was

administered after each class. In this way Bohm could ascertain the nature and degree of change

generated by each specific stimulus (class topic), allowing him to further discern what stimuli

had what effect. In these studies, the researcher asked the abstract question about death penalty

support for some people convicted of first-degree murder. Bohm and Vogel (1991) asked four

questions regarding death penalty opinion, measured using a seven-point Likert scale. However,

they only reported the results of one abstract question: “Which of the following statements best

describes your position toward the death penalty for some people convicted of first-degree

murder?” (p. 72).

Bohm et al. (1990) adopted and reported a more robust method of measuring death

penalty opinion in their sample. Their goal was to explain the reasons for death penalty support

and opposition. They asked their subjects’ death penalty opinion with respect to general

deterrence, retribution, incapacitation, administrative considerations, religious reasons and

support for law enforcement. All the responses were recorded and reported on a five-point Likert

scale (“strongly agree,” “agree,” “uncertain,” “disagree,” and “strongly disagree”).

Bohm et al. (1991) also used a more solid approach. The researchers were aware of the

potential problems Williams et al. (1988) had highlighted with death penalty opinion

measurement. Bohm et al. (1991), Bohm et al. (1993), and Bohm and Vogel (2004) used four

questions to measure death penalty opinion in the abstract and concrete. These questions were:

1. Which of the following best describes your position toward the death penalty for all

persons convicted of first degree murder?

2. Which of the following best describes your position toward the death penalty for some

persons convicted of first degree murder?

3. If you served on a jury where the defendant, if found guilty, would automatically be

sentenced to death, could you convict the defendant?

4. If asked to do it, could you pull the lever that would result in the death of an individual

convicted of first-degree murder? (Bohm et al. 1991, p. 368)

Responses were recorded on a seven-point Likert scale.

Wright et al. (1995) used the same four abstract and concrete questions that Bohm et al.

used in 1991. In addition, they asked two questions regarding alternatives to capital punishment,

again paying heed to Williams et al. (1988). The two additional questions were:

1. Would you support an alternative to the death penalty where a convicted first-

degree murderer is sentenced to life in prison, with no possibility of parole ever

(LWOP)?

2. Would you support an alternative to the death penalty where a convicted first-

degree murderer is sentenced to life in prison, with no possibility of parole ever

and have him/her work in a prison industry where his/her earnings would go to

the victim’s family?

The subjects answered using a five-point Likert scale.

Sandys (1995) measured opinion over the course of the semester. She asked the same

question before every class, “What is your attitude toward the death penalty for some people

convicted of first degree murder?” (p. 43, fn. 3). This is clearly an abstract question (with no

other question types) and therefore vulnerable to the dangers highlighted by Williams et al.

(1988). The subjects answered using a seven-point Likert scale. Sandys (1995) then followed up

to gauge opinion a year later using the same question.

Clarke et al. (2000-2001) and Lambert and Clarke (2001) used a five-point Likert scale to

measure death penalty opinion. However, they do not report the specific death penalty questions

they asked.

Cochran and Chamlin (2005) and Cochran et al. (2006) used two death penalty opinion

questions. They asked their subjects about the death penalty for “all people convicted of first-

degree murder” (p. 576-77), and if they would choose “life in prison, without any possibility of

parole” (p. 577) as an alternative to the death penalty. As noted previously, a problem with these

questions is that both are abstract. They used a seven-point Likert scale to measure responses.

Measures of Death Penalty Knowledge

In the aforementioned studies, death penalty knowledge was measured in one of three

ways. Most studies used a pretest-posttest design, with identical questionnaires administered

before and after exposure to an experimental stimulus. A section of each questionnaire contained

a knowledge index consisting of various numbers of factual items about the death penalty. A few

studies simply assumed that subjects were more knowledgeable about the death penalty

following completion of a death penalty class than they were prior to it. Finally, the two follow-

up studies asked subjects to self report how much death penalty knowledge they had gained

along a continuum, ranging from zero (“know absolutely nothing”) to nine (“extremely

knowledgeable”).

Studies that employed a knowledge index with a pretest-posttest design were conducted

by Sarat and Vidmar (1976), Vidmar and Dittenhoffer (1981), Bohm et al. (1990), Bohm et al.

(1991), Wright et al. (1995), Cochran and Chamlin (2005), and Cochran et al. (2006). Sarat and

Vidmar (1976, p. 175) used six knowledge items:

1. Are there any people currently awaiting execution in the United States?

2. How many people were executed in the five years prior to the Furman decision?

3. Poor people who commit murder are more likely to be sentenced to death than rich

people who commit a similar crime.

4. The punishment of death has typically been imposed in only a small fraction of the cases

where it is an authorized punishment.

5. Most scientific studies of the effects of the death penalty show that it is an effective

deterrent to crime, do not show that it is an effective deterrent to crime, don’t know.

6. Studies have shown that the rate of murder usually drops in the weeks following a well

publicized execution.

Vidmar and Dittenhoffer (1981) used ten knowledge items, but did not provide them in

their published article (they were “available from the authors, upon request”).

Bohm et al. (1990) and Bohm et al. (1991) used fourteen items to measure death penalty

knowledge:

1. The death penalty has been abolished by a majority of Western European nations.

2. Over the years states which had the death penalty have shown lower murder rates than

neighboring states which did not have the death penalty.

3. Studies have not found that abolishing the death penalty has any significant effect on the

murder rate in a state.

4. Studies have shown that the rate of murder usually drops in the weeks following a well

publicized execution. (Sarat and Vidmar Q. 6)

5. The average prison term served by someone sentenced to life imprisonment is less than

ten years.

6. Poor people who commit murder are more likely to be sentenced to death than rich

people who commit a similar crime. (Sarat and Vidmar Q. 3)

7. After the Supreme Court struck down the death penalty in 1972, the murder rate in the

U.S. showed a sharp upturn.

8. On the average, the death penalty costs the taxpayer less than life imprisonment.

9. Currently, there are over one thousand people awaiting execution.

10. The punishment of death has typically been imposed in only a small fraction of the cases

where it is an authorized punishment. (Sarat and Vidmar Q. 4)

11. The majority of the executions in the United States take place in the South.

12. For the same crime men are more likely to be executed than women.

13. Currently, the leadership of organized religion in the United States (whether Catholic,

Jewish or Protestant) has abandoned its traditional support of (or silence on) the death

penalty and instead favors its complete abolition.

14. The majority of Americans currently favor the death penalty.

Wright et al. (1995) employed seventeen knowledge items. Twelve of the knowledge

items came from Bohm et al. (1990) and Bohm et al. (1991), and five of the items were new. The

twelve items from Bohm et al. were numbers 1, 2, 4-12, and 14, and the new items were:

1. Most of the states in the U.S. do not have capital punishment laws.

2. A black person is more likely to receive the death penalty than a white person for the

same crime.

3. Lethal injection is the most commonly used method of execution today.

4. No innocent people have been executed in error in the U.S. this century.

5. The killers of black people are just as likely to receive the death penalty as the killers of

white people.

Cochran and Chamlin (2005) and Cochran et al. (2006) included four death penalty

“truth” statements and seven death penalty “myth” statements as a knowledge measure. Only one

of the “truth” statements was original: “There is strong reason to believe that similar offenders

convicted of murder often receive dissimilar sentences; that is, some are sentenced to death while

others are sentenced to an alternative less than death.” Two of the other “truth” statements were

the same as in Wright et al. (1995)—numbers nine and 17—and the other one was the same as

Sarat and Vidmar’s (1976) number three. Three of the “myth” statements were original to the

study and four were the same items used in previous studies (numbers five and six in Sarat and

Vidmar, 1976 and numbers two and eight in Bohm et al., 1990, 1991). The three original

statements were:

1. The death penalty is more effective than life imprisonment without possibility of

parole (LWOP) in protecting society from the future actions of those who have

already committed capital crimes.

2. Only legally relevant criteria distinguish murderers sentenced to death from those

sentenced to a punishment less than death.

3. There is no evidence to support the claim that innocent persons have ever been

sentenced to death and executed in error. (Cochran et al., 2006, p. 211-212)

Ellsworth and Ross (1983) provided subjects with nine knowledge items on a pretest

only. Questions one through seven and question nine are the same as in Bohm et al. (1990,

1991). Question 8 is the only different one: “In several cases people executed for murder in the

U.S. were later proven innocent.”

Studies that only assumed subjects were more knowledgeable at the end of a death

penalty class included Lord et al. (1979), Bohm (1989), Bohm (1990), Bohm and Vogel (1991),

Sandys (1995), Clarke et al. (2000-2001), and Lambert and Clarke (2001). Bohm et al. (1993)

and Bohm and Vogel (2004) asked subjects to self-report their level of death penalty knowledge

in their follow-up studies.

Studies that assumed subjects were more knowledgeable at the end of a death penalty

class are vulnerable to the criticism that students may not have actually learned much in the

class, let alone retained any information that they did learn. Also, research shows that many

subjects may have assimilated information they learned in class biasedly. That is, they may have

learned the information that supported their initial death penalty position and ignored the

information that contradicted it. Another potential problem is that the class instructor may have

delivered factually inaccurate information during the course of the semester. A problem with

studies that relied on self-reported knowledge is that subjects may have incorrectly assessed their

level of death penalty knowledge. Even studies that provided knowledge indices can be criticized

for assuming the correct answers for all of the questions are noncontroversial. For example, in

Wright et al. (1995) one of the knowledge items was: “No innocent people have been executed in

error in the U.S. this [20

] century.” Although many authorities consider this statement false, as

did Wright et al. (1995), there is no incontrovertible evidence that the statement is false.

Measures of Retribution

Only nine of the studies measured retribution, and only five of the studies actually tested

Marshall’s third hypothesis. All of the studies that measured retribution used one or more items

and asked subjects to respond on Likert-type scales. The studies that tested Marshall’s third

hypothesis were Sarat and Vidmar (1976), Bohm et al. (1991), Bohm and Vogel (2004), Cochran

and Chamlin (2005), and Cochran et al. (2006). Sarat and Vidmar (1976, p.181, fn. 54) used

three retribution items:

1. It is only right that people who hurt others should be hurt themselves.

2. It is simple justice that criminals should be punished for their crimes.

3. The only proper justification for punishing criminals is the punishment has a deterrent

or retributive purpose.

Bohm et al. (1991, p. 382) used a series of eight items to measure retribution:

1. If a murderer is not executed for the crime, the friends and family of the victim are

likely to take it upon themselves to seek revenge.

2. The very worst of the Nazi war criminals should have been executed for their “crimes

against humanity.”

3. Those who take a life should forfeit their own in return.

4. Killing is all right if the right people do it and think they have a good reason for doing

it.

5. Society has a right to get revenge when a very serious crime like murder has been

committed.

6. Sometimes I feel a sense of personal outrage when a convicted murderer was

sentenced to a penalty less than death.

7. There are some murderers whose death would give me a sense of personal

satisfaction.

8. An execution would make me sad, regardless of the crime the individual committed.

(Reverse scored)

Cochran and Chamlin (2005) and Cochran et al. (2006) both used two retribution

statements taken from Bohm et al. (1991):

1. Those who take a life should forfeit their own in return.

2. Society has a right to get revenge when a serious crime like murder has been

committed.

Studies by Lord et al. (1979), Sandys (1995), Wright et al. (1995), Clarke et al. (2000-

2001), Bohm and Vogel (1991), and Lambert and Clarke (2001) did not measure retribution and,

therefore, could not test Marshall’s third hypothesis.

A problem with the retribution measures is that the concept of retribution is ambiguous

(Bohm, 2003). Philosopher John Cottingham (1979) contends there at least nine different

definitions of retribution. As noted previously, Justice Marshall equated retribution with

vengeance, but vengeance is only one type of retribution. Another, for example, is “just deserts.”

Therefore, a problem is whether the measures of retribution employed in the studies cited above

truly captured Justice Marshall’s conception of retribution. A related problem is that none of the

retribution items actually contain the word retribution; consequently, it must be inferred that

subjects interpreted the items as retributive.

This thesis is another test of Marshall’s hypotheses. It differs from the previous studies in

several ways. First, it is one of the only studies that was conducted in a post-9/11 environment.

Second, it was conducted at a time in United States history when lethal injection, the preferred

method of execution in the United States, was coming under both greater scrutiny and attack

from many areas of American society. Various state legislatures and executive branches are

presently acting in response to this nationwide concern. This is shown by the fact that two states

(Illinois and New Jersey) have official moratoria in place, five other states (Arkansas, Delaware,

Maryland, Nebraska, and North Carolina) have ceased executions to either allow further

investigation into the administration of lethal injection executions or to allow the courts to

consider Eighth Amendment claims, and California has been enjoined from carrying out lethal

injection executions by order of a federal court. No test of the Marshall hypotheses has been

carried out at a time of such flux in capital punishment.

In addition, this thesis, unlike previous studies, uses 19 items to measure death penalty

opinion. Three of these items ask about abstract support for the death penalty and 16 items ask

about support for the death penalty in a number of specific circumstances (e.g., the murder of a

child, the murder of a law enforcement officer, rape in the absence of murder). Two of the items

inquire about support for life without the possibility of parole (LWOP) and life without the

possibility of parole plus restitution (LWOP +). Research shows that these two measures greatly

attenuate death penalty support (see, for example, Gallup, 2006). This thesis also uses a larger

sample than Lord et al. (1979), Vidmar and Dittenhoffer (1981), Sandys (1995), and many of the

Bohm experiments and employs a comparison group unlike Lord et al. (1979), Ellsworth and

Ross (1983), Bohm (1989), Sandys (1995), Cochran and Chamlin (2005), and Cochran et al.

(2006). At the very least, this study, which was conducted under different circumstances than

previous studies, is another replication of tests of the Marshall hypotheses. As such, it is another

check on the reliability and validity of previous tests and the veracity of Marshall’s claims. It

also provides another check on the generalizability of Marshall’s contentions.

CHAPTER THREE: METHODOLOGY

Experimental Design

This study used a pretest-posttest comparison group design. Subjects of the study were

undergraduates from a large public university in Florida. One hundred forty-six subjects took the

pretest (82 in the experimental group and 64 in the comparison group), and 175 subjects

completed the posttest (95 in the experimental group and 80 in the comparison group). The

disparity in the number of subjects who took the pretest and posttest is a function of students

adding the courses after the pretests were administered. The experimental group was comprised

of students enrolled in a semester-long death penalty class, and the comparison group consisted

of students from another criminal justice class. Each course was taught during the fall semester

of 2006. Subjects in the comparison group had not taken, or were not currently taking, the death

penalty class. Most of the subjects were criminal justice majors or minors, and many of them had

been exposed to some material on the death penalty in previous classes. None of the subjects was

randomly selected or assigned. As noted in the critique of previous studies (e.g., Sarat and

Vidmar, 1976), student subjects were used to overcome the problem of other studies that only

briefly exposed subjects to the experimental stimulus. Participation in the study was optional,

and responses were provided anonymously.

Table 3 shows demographic characteristics of experimental and comparison groups on

the pretest and posttest. On the pretest, 53 percent of experimental group subjects were female,

and 47 percent were male; 10 percent were black, 12 percent were Hispanic, 71 percent were

white, and 7 percent were Asian or other. On the posttest, 45 percent of experimental group

subjects were female, and 52 percent were male; 6 percent were black, 10 percent Hispanic, 71

percent were white, and 13 percent were Asian or other. On the pretest, 45 percent of comparison

group subjects were female, and 55 percent were male; 14 percent were black, 16 percent were

Hispanic, 67 percent were white, and 3 percent were Asian or other. On the posttest, 45 percent

of comparison group subjects were female and 54 percent were male; 13 percent were black, 13

percent were Hispanic, 71 percent were white, and 3 percent were Asian or other. These data

show that the experimental and comparison groups were relatively equivalent with regard to

gender and race or ethnicity.

Table 3

Demographic Characteristics of Experimental and Comparison Groups on Pretest and Posttest

(in percentages)

SEX RACE

Pretest Male Female White Black Hispanic Asian or Other

Exp 47 53 71 10 12 7

Comp 55 45 67 14 16 3

Posttest

Exp 53 47 71 6 10 13

Comp 54 46 71 13 13 3

Both classes met twice a week for one and a quarter hours for an entire semester (16

weeks). The death penalty class text was Deathquest II: An Introduction to the Theory and

Practice of Capital Punishment in the United States (2003), which was written by the instructor

and is known as a balanced assessment of many aspects of capital punishment (Blecker, 2007).

The course consisted of lectures by the instructor, videos, and discussion. Discussion was

encouraged at all times by the instructor. The instructor was forthright about his anti-death

penalty opinions but presented the information in a manner that was as unbiased as possible. The

instructor played “devil’s advocate” when students asked questions, making sure that both sides

of the death penalty debate were represented. During the semester many death penalty-related

topics were covered including its history, Supreme Court death penalty cases, public opinion,

deterrence, incapacitation, religious arguments, retribution, arbitrariness, and discrimination (see

Appendix A). This class was the experimental stimulus and, as such, suffers from all of the

problems with this type of stimulus discussed in the critique of previous studies. However, it did

allow subjects to thoroughly immerse themselves in death penalty knowledge, to discuss the

subject, and to reflect upon it over a protracted period of time.

Measurement Procedures

Subjects completed the same survey (see Appendix B) on the first and the final days of

the semester. The survey was designed to obtain three kinds of information. Section one

contained items that measured subjects’ death penalty opinions. Four of these items asked about

abstract support for the death penalty; however, only three of the items were used in the analysis,

and 16 items asked about support for the death penalty in a number of specific circumstances

(e.g., the murder of a child, the murder of a law enforcement officer, rape in the absence of

murder). Two of the abstract items—support of the death penalty for all and for some people

convicted of first-degree murder—have been used in previous research. The “death penalty for

all” item was dropped from the analysis because upon further reflection it was decided that the

item was no longer relevant since the Court’s decisions in Woodson v. North Carolina (428 U.S.

280, 1976) and Roberts v. Louisiana (428 U.S. 325, 1976), which made mandatory death penalty

statutes unconstitutional. The other two abstract items dealt with alternatives to the death

penalty: 1) support for life in prison without the possibility of parole (LWOP), and 2) support for

life in prison without the possibility of parole plus forcing the inmate to work in prison industry

and give all his/her earnings to the victims’ family or the community (LWOP+). Subjects

responded on a five-point Likert scale (“strongly agree” to “strongly disagree”).

The first section also contained five items used to measure retribution:

1. I believe the government has the right to kill in certain circumstances.

2. The death penalty is “purposeless vengeance” (reverse scored).

3. I believe the main purpose of the death penalty is retribution/revenge.

4. Sometimes I have a sense of personal outrage when a convicted murderer is

sentenced to a penalty less than death.

5. When I hear about an execution it makes me sad, regardless of the crime the

individual committed (reverse scored).

All but the second item had been used in previous studies to create a retribution index

(see Sarat and Vidmar, 1976; Bohm et al., 1991; Bohm et al., 1993; and Wright et al., 1995).

Cronbach’s alpha coefficient of reliability for the retribution index is .77.

Section two of the survey contained 21 knowledge items with three response categories:

“true,” “false,” and “don’t know.” The “don’t know” category was scored as an incorrect answer

and was used to discourage subjects from simply guessing and potentially distorting the results.

The 21 knowledge items were combined to form an index. Cronbach’s alpha coefficient of

reliability for the knowledge index is .83. Finally, the third section asked for demographic

details about the subjects (See Appendix B)

CHAPTER FOUR: RESULTS

The results of the tests for each of Marshall’s hypotheses are presented below.

Hypothesis I: The American people are largely ignorant about the administration of capital

punishment

Results of the study support Marshall’s first hypothesis. On the pretest, the experimental

group answered 39 percent of the knowledge items correctly, and the comparison group

answered 33 percent of the knowledge items correctly. Thus, neither group could be considered

well informed about the death penalty. Looked at somewhat differently, the experimental group

answered a mean of approximately 8 of 21 knowledge items correctly, while the comparison

group answered a mean of about 7 of the 21 knowledge items correctly. A two-tailed

independent t-test showed the difference in knowledge items answered correctly by the

experimental and comparison groups was statistically significant at p = .009. These results are

displayed in Table 4. However, because the experimental group, on average, answered only one

more knowledge item correctly than the comparison group, the difference between the two

groups is not considered substantively significant. Either way, neither group could be considered

knowledgeable about the death penalty.

Table 4

Comparison of Pretest Percentages and Means of Experimental and Comparison Groups on the

Knowledge Index

Group

N % Mean SD Std. Error Of Mean

Comparison 64 32.5 6.8281 3.22960 .40370

Experimental 82 38.9 8.1707 2.91782 .32222

t df Sig. (2-tailed) Mean Difference

-2.632 144 .009 -1.34261

Hypothesis II: An informed American public would not support the death penalty

The first issue addressed in this section is whether the experimental stimulus was

successful in producing informed subjects. The second issue examined was whether increased

death penalty knowledge changed subjects’ death penalty opinions (Marshall’s second

hypothesis). After exposure to the experimental stimulus, the experimental group increased the

percentage of knowledge items answered correctly about 20 percentage points, from 38.9 percent

on the pretest to 58.8 percent on the posttest. The comparison group, on the other hand, increased

the percentage of knowledge items answered correctly only six percentage points from 32.5

percent on the pretest to 38.5 percent on the posttest. Looked at in terms of knowledge mean

score differences, the experimental group increased its mean score of knowledge items answered

correctly from about eight items on the pretest to approximately 12 items on the posttest. This 50

percent increase in death penalty knowledge for the experimental group was statistically

significant at p < .000. As for the comparison group, the mean knowledge score increased from

about seven items answered correctly to approximately eight items answered correctly. This 14

percent increase in death penalty knowledge for the comparison group was statistically

significant at p = .014, but the approximately one point difference in means is not considered

substantively significant. These data suggest that the experimental group was significantly more

informed than the comparison group (p< .000) after the experimental group was exposed to the

experimental stimulus. Yet, as noted previously, the experimental group cannot be considered

well informed even after exposure to the experimental stimulus.

Marshall’s second hypothesis was tested using one-tailed independent t-tests. Results

show that Marshall’s second hypothesis is only partially supported by the data. Increased death

penalty knowledge, albeit minimally increased knowledge, did not reduce death penalty support

for some people convicted of first-degree murder (see Table 5) or for any of the 16 concrete

items (see Table 7). Alternatives to the death penalty, however, fared better. Following exposure

to the experimental stimulus, support for LWOP and LWOP + increased in the experimental

group. As shown in Table 8, the mean score on the LWOP measure in the experimental group

increased from about 3.5 on the pretest to approximately 3.9 on the posttest. The difference was

statistically significant at p = .007. On the LWOP+ measure, the mean score increased from

about 3.7 on the pretest to approximately 4.3 on the posttest. This difference was statistically

significant at p < .000 (see Table 9).

Table 5

Comparison of means: Death penalty support for some people convicted of first-degree murder

COMP EXP

Pre Post Pre Post

64 80 82 93

44.4 55.5 46.8 53.2

Mean

3.610 3.300 3.451 3.484

1.341 1.479 1.306 1.508

seM

.168 .165 .144 .156

1.300 -.152

142 173

Sig. (1-tail)

.100 .440

ΔMean

.309

-.033

Table 6

Experimental Group: Concrete Items

Pre Mean Post Mean t df Sig. (1-tail)

I support the death penalty for the

adult murderers of children.

4.171 4.258 -.480 173 .316

I support the death penalty for

murderers of law enforcement

officers.

3.817 3.979 -.866 173 .194

I support the death penalty for

murderers of correctional officers.

3.817 4.044 -1.227 172 .111

I support the death penalty for

prisoners who murder fellow

inmates.

3.210 3.280 -.385 172 .350

I support the death penalty for

serial killers.

4.451 4.413 .213 172 .416

I support the death penalty for

those who kill during a rape.

4.146 4.280 -.719 173 .237

I support the death penalty for

those who kill during a robbery.

3.732 3.742 -.054 173 .479

I support the death penalty for

those who kill during a

kidnapping for ransom.

3.963 4.141 -.976 171 .166

Pre Mean Post Mean t df Sig. (1-tail)

I support the death penalty for

rapists who do not kill their

victims.

2.890 2.742 .832 173 .204

I could pull the lever to initiate an

execution by lethal gas (gas

chamber).

3.110 2.989 .552 173 .291

I could pull the trigger in a firing

squad.

2.939 2.926 .062 174 .476

I could pull the gallows lever in a

hanging execution.

2.744 2.830 -.407 174 .342

I could flip the switch in an

execution by electrocution.

2.976 2.957 .085 173 .466

I could inject the chemicals in a

lethal injection execution.

2.902 2.946 -.199 173 .421

I could witness an execution. 3.561 3.763 -1.049 173 .148

If I were a member of a jury in a

death penalty case, I could

sentence a defendant to death if

the evidence supported it.

4.000 4.043 -.229 173 .410

Table 7

Comparison Group Concrete Items

Pre Mean Post Mean t df Sig. (1-tail)

I support the death penalty for the

adult murderers of children.

4.422 4.225 1.062 142 .145

I support the death penalty for

murderers of law enforcement

officers.

3.938 3.963 -.120 142 .453

I support the death penalty for

murderers of correctional officers.

3.828 3.900 -.341 142 .367

I support the death penalty for

prisoners who murder fellow

inmates.

3.422 3.375 .224 142 .412

I support the death penalty for

serial killers.

4.484 4.375 .606 142 .273

I support the death penalty for

those who kill during a rape.

4.266 4.238 .142 142 .444

I support the death penalty for

those who kill during a robbery.

3.953 3.763 .919 142 .180

I support the death penalty for

those who kill during a

kidnapping for ransom.

4.125 3.975 .763 142 .224

Pre Mean Post Mean t df Sig. (1-tail)

I support the death penalty for

rapists who do not kill their

victims.

3.206 2.975 .957 141 .170

I could pull the lever to initiate an

execution by lethal gas (gas

chamber).

2.906 2.813 .377 142 .353

I could pull the trigger in a firing

squad.

2.781 2.913 -.527 142 .300

I could pull the gallows lever in a

hanging execution.

2.703 2.788 -.340 142 .367

I could flip the switch in an

execution by electrocution.

2.828 2.886 -.228 141 .410

I could inject the chemicals in a

lethal injection execution.

2.750 2.738 .051 142 .480

I could witness an execution. 3.313 3.250 .264 142 .396

If I were a member of a jury in a

death penalty case, I could

sentence a defendant to death if

the evidence supported it.

4.032 3.838 .952 141 .172

Table 8

Comparison of means: LWOP as an alternative to death penalty

COMP EXP

Pre Post Pre Post

64 80 82 94

44.4 55.5 46.5 53.5

Mean

3.500 3.563 3.488 3.926

1.260 1.271 1.250 1.080

seM

.157 .142 .138 .111

-.294 -2.493

142 174

Sig. (1-tail)

.385 .007

ΔMean

.142

-.438

Table 9

Comparison of means: LWOP + as a death penalty alternative

COMP EXP

Pre Post Pre Post

64 80 82 91

44.4 55.5 47.4 52.6

Mean

4.110 3.888 3.712 4.297

.994 1.158 1.103 .810

seM

.124 .130 .122 .085

1.216 -3.949

142 171

Sig. (1-tail)

.113 .000

ΔMean

.222

-.577

Hypothesis III: If the primary reason people support capital punishment is based on the principle

of retribution, then increased knowledge about the subject would make little or no difference to

their death penalty opinions

As described in the methodology section, a retribution index was employed to measure

the level of retribution in the experimental and comparison group. To test Marshall’s third

hypothesis, experimental and comparison group subjects were divided into those who scored

“high” on the retribution measure and those who scored “low” on it. The response categories to

the retribution items ranged from “strongly agree” to “strongly disagree” on a five-point Likert

scale. These items were combined into an index, which provided individual scores from zero to

25. The mean and median scores of the four groups (comparison pretest, comparison posttest,

experimental pretest, and experimental posttest) were all approximately 17. Therefore, subjects

with retribution scores of 17 or less were considered “low” on the retribution scale, and those

with scores of above 17 were considered to be high on the retribution scale.

Then, changes between pretest and posttest were examined for each of these groups on

the 19 death penalty opinion measures using one-tailed independent t-tests. Results show a

statistically significant change for the experimental group on only two items: LWOP and

LWOP+. However, contrary to Justice Marshall’s expectation, for LWOP and LWOP+ a

statistically significant change occurred for both “low” retributivists and “high” retributivists

(see Tables 10 and 11).

Table 10

Experimental Group “Low” Retributivists: Pretest v. Posttest

Condition Pre/Post N Mean SD Std. Error of Mean

Death Penalty Pre 38 3.0526 1.31411 .21318

Post 45 3.5333 1.53149 .22830

LWOP Pre 38 3.8158 1.13555 .18421

Post 46 4.2391 .89901 .13255

LWOP + Pre 38 3.8684 .99107 .16077

Post 43 4.4186 .79380 .12105

Condition t df Sig. (1-tailed) Mean Difference

Death Penalty -1.519 81 .070 -.48070

LWOP -1.907 82 .030 -.42334

LWOP + -2.771 79 .004 -.55018

Table 11

Experimental Group “High” Retributivists: Pretest v. Posttest

Condition Pre/Post N Mean SD Std. Error of Mean

Death Penalty Pre 44 3.7955 1.21195 .18271

Post 48 3.4375 1.50044 .21657

LWOP Pre 44 3.2045 1.28641 .19393

Post 48 3.6250 1.16006 .16744

LWOP + Pre 44 3.5909 1.18750 .17902

Post 48 4.1875 .81623 .11781

Condition t df Sig. (1-tailed) Mean Difference

Death Penalty 1.252 90 .107 .35795

LWOP -1.648 90 .050 -.42045

LWOP+ -2.828 80 .003 -.59659

Marshall did not expect “high” retributivists to change their death penalty opinions.

There were no other statistically significant changes for the experimental group on the other

death penalty opinion measures or for the comparison group on any of the other death penalty

opinion measures (see Tables 10 and 11).

CHAPTER FIVE: CONCLUSION

This thesis tested three hypotheses suggested by former Supreme Court Justice Thurgood

Marshall in his Furman opinion. Marshall’s first hypothesis –“The American people are largely

ignorant about the administration of capital punishment” –was supported by the results of this

thesis. On average, the experimental group initially answered only about 39 percent of the 21

knowledge items correctly, and the comparison group initially answered only about 32 percent of

the 21 knowledge items correctly. Although Marshall did not stipulate how much death penalty

knowledge a person must possess to be considered informed about the subject, he surely would

not have considered answering only 32 percent or 39 percent of knowledge items correctly as

being well informed. This conclusion is more compelling because many of the experimental and

comparison group subjects were criminal justice majors or minors and probably were exposed to

at least some death penalty information in other criminal justice classes. One wonders how

much knowledge would be found in a sample drawn from outside a criminal justice or, indeed, a

university environment. This is certainly a subject for future research.

Marshall’s second hypothesis –“An informed American public would not support the

death penalty” –was only partially supported by the data. The experimental (informed) group did

not change their death penalty support (to a statistically significant degree) for “some people

convicted of first-degree murder” or for the 16 concrete items. The only statistically significant

changes in the experimental group were on the two alternative items: LWOP and LWOP +.

Although Marshall never discussed these two alternatives in his Furman decision, this thesis

shows that informed subjects would support either LWOP or LWOP + as an alternative to the

death penalty more than less informed subjects.

The data support Marshall’s third hypothesis –“If the primary reason people support

capital punishment is based on the principle of retribution, then increased knowledge about the

subject would make little or no difference to their death penalty opinions.” After exposure to

death penalty knowledge, subjects who scored “high” on the retribution index did not change (to

a statistically significant degree) their death penalty opinion for “some people convicted of first-

degree murder” or for any of the concrete items. What Marshall did not anticipate, and a finding

of this thesis, was that subjects who scored “low” on the retribution index also did not change

their death penalty opinion for any of those items after exposure to the experimental stimulus.

Thus, at least in this thesis, a subject’s level of retribution does not appear to be the sole reason

that a subject might be impervious to death penalty knowledge for these measures of death

penalty opinion. This finding is unique in the literature; thus, further research of it is warranted.

The only statistically significant findings with regard to retribution were for the LWOP

and LWOP+ items. Experimental group subjects who scored both “high” and “low” on the

retribution index increased their support for both LWOP and LWOP+ to a statistically significant

degree. Again, although Marshall did not discuss LWOP or LWOP+ as alternatives to the death

penalty in his third hypothesis, the finding about subjects who scored “high” on the retribution

index does not support Marshall’s general contention about informed retributivists’ death penalty

opinions being obdurate. The second finding about subjects who scored “low” on the retribution

scale being supportive of LWOP and LWOP+ was expected. As noted above, these findings

should be researched further.

In sum, results of this thesis provide unqualified support for Marshall’s first hypothesis

and partial support for his second and third hypotheses. Thus, findings for hypothesis one are

consistent with previous studies, while the results for hypotheses two and three differ in some

ways from the findings of previous research. Five findings were unique to this thesis. First was

the finding that death penalty knowledge did not have a statistically significant effect on any of

the 16 concrete measures of death penalty opinion. Second was the finding that death penalty

knowledge produced a majority of subjects willing to support LWOP and LWOP+ as alternatives

to the death penalty. Third, although the finding that experimental group subjects who scored

“high” on the retribution scale did not change their death penalty opinion for “some people

convicted of first-degree murder” or for any of the concrete items was expected, the finding that

experimental group subjects who scored “low” on the retribution scale also were intransigent on

these measures was not expected. The fourth unique finding of this thesis was that experimental

group subjects who scored both “high” and “low” on the retribution scale increased their support

for both LWOP and LWOP+ as alternatives to the death penalty. Fifth, results of this thesis

showed that exposure to death penalty knowledge did not alter subjects’ scores on the retribution

index. In fact, an examination of retribution score means for both experimental and comparison

group subjects showed almost no change between pretest and posttest. In other words, results of

this thesis suggest that “once a retributivist always a retributivist.”

As noted, three “abstract” and 16 “concrete” death penalty opinion measures were

employed in this thesis. Previous research has found that death penalty opinion question type and

response categories can affect results (see Ellsworth and Ross, 1983; Williams et al, 1988; Bohm

et al., 1991). Specifically, results have been found to vary depending on whether the death

opinion question is “abstract” or “concrete.” “Concrete” items used in this thesis included

whether a person, as a member of jury in a death penalty case, could sentence a defendant to

death if the evidence supported it; whether a person could witness an execution; and whether a

person could pull the lever to initiate an execution by lethal gas, pull the trigger in a firing squad

execution, pull the gallows lever in a hanging execution, flip the switch in an execution by

electrocution, or inject the chemicals in a lethal injection execution. Although these items are not

the same as actually serving on a jury in a capital case or participating in an execution, they are

more “concrete” than simply asking whether a person favors or opposes the death penalty for

some people convicted of first-degree murder. Still, research shows that what a person says and

what a person does are often very different (see Wicker, 1969).

Another methodological weakness involved the samples. As noted previously, some

subjects only took the pretest (e.g., those students who dropped the class before the posttest was

administered), and some students only took the posttest (e.g., those students who added the class

after the pretest was administered). A problem was that these students could not be identified

because questionnaires did not ask for names or any other identifying information. What is

known is that 13 more students in the experimental group and 16 more students in the

comparison group took the posttest. Thus, the pretest-posttest experimental and comparison

groups were not equivalent and caution should be exercised when interpreting changes in means.

Further research on this subject should employ true experimental designs with randomly

selected and assigned experimental and comparison groups. The groups should be large enough

and diverse enough to be able to more confidently generalize findings. If a classroom experience

is used as the experimental stimulus it should be varied from the type of experience used in

previous research. For example a class could be team-taught with a death penalty opponent and

proponent as instructors, or a class could be taught by a death penalty proponent to compare

findings from that class with classes taught by death penalty opponents Other experimental

stimuli besides a classroom experience also should be utilized in future research, such as

extensions to the rudimentary Sarat and Vidmar (1976), Lord et al. (1979), Vidmar and

Dittenhoffer (1981) “essay stimuli,” death-row visits, witnessing an execution, witnessing a

capital trial, serving on a capital jury, and monitoring pro- and anti-death penalty groups (see

Bohm, 1990).

Finally, no previous study has examined a possible “testing threat.” Future research

should use a Solomon Four Group design to remedy this problem.

This thesis does not fully support Justice Marshall’s contention that a knowledgeable

public would oppose the death penalty, if Marshall had in mind that knowledge would be

achieved by participation in a death penalty class such as the one used in this thesis. That does

not mean that death penalty knowledge cannot change death penalty opinions. In this thesis, it

did for LWOP and LWOP+. Rather, it only means that it likely does not change death penalty

opinions significantly under the other specific conditions of this thesis and similar studies.

Clearly, more research using different conditions is warranted.

APPENDIX A:

CCJ 4361-0001: DEATH PENALTY FALL, 2006

CCJ 4361-0001: DEATH PENALTY

Fall, 2006

INSTRUCTOR: Professor Robert M. Bohm

OFFICE: Room 362 HPA PHONE: 823-5944

EMAIL: [email protected]

OFFICE HOURS: TR, 12:20-1:20 and 2:50-4:20 or by appt.

CLASS MEETINGS: 4:30-5:45, TR, HPA 112

REQUIRED TEXT: Bohm, Robert M. (2003) Deathquest II: An Introduction to the Theory and

Practice of Capital Punishment in the United States, Second Edition. Cincinnati, OH: Anderson

Publishing Co.

COURSE OBJECTIVES:

1. To learn about the death penalty in the United States, especially its administration under

current statutes.

2. To provide an opportunity to discuss issues about the death penalty that the student finds

troublesome or about which the student seeks clarification.

3. To help the student develop critical thinking abilities.

4. To allow the student to develop an informed opinion about the death penalty.

COURSE DESCRIPTION AND REQUIREMENTS:

1. Each chapter will be introduced with a PowerPoint presentation, summarizing the main

points of the chapter. At the end of each presentation, students should be prepared to

respond to end of chapter discussion questions. Other relevant questions will also be

addressed. Discussion is encouraged.

2. Satisfactory performance on three examinations. Exams will be given during the class period

following the completion of each section as indicated on the course outline. Learning objectives

for each chapter /exam are attached to the syllabus and should be used for class and exam

preparation. Exams include multiple choice and true-false items and cover reading assignments,

presentations, videos, and anything discussed in class. Each exam covers only the material in a

particular section. Three raspberry-colored scantron sheets, one for each exam, are required. Use of

#2 pencils with good erasers is suggested. Grades are based on the percentage of items answered

correctly. A conventional scale (i.e., A = 90-100%, B = 80-89%, etc.) will be employed. No plus

or minus grades are given. Each exam counts as one-third of the final grade. Make-up exams are

given in only exceptional circumstances (e.g., commitment to a hospital), otherwise missed exams

are recorded as zeros.

Students may pick up their exam results from their professor during his office hours anytime

during the semester. Students who wish to examine their results with a copy of the exam booklet

may do during their professor’s office hours anytime during the semester. Alternatively, students

may access their exam grades anytime after they have been posted by using “myUCF Grades” on

the UCF website. If you need help accessing myUCF Grades, see the online tutorial at

https://myucfgrades.ucf.edu/help/.Exam grades are not provided by phone or email.

COURSE OUTLINE: READINGS:

Section 1

1. History of the Death Penalty in the U.S. Chap. 1

2. Capital Punishment and the Supreme Court Chap. 2

Exam #1

Section 2

3. Current Data None

4. The Federal Death Penalty and the Military Chap. 3

5. Methods of Execution Chap. 4

6. General Deterrence Chap. 5

7. Incapacitation and Costs Chap. 6

Exam#2

Section 3

8. Miscarriages of Justice Chap. 7

9. Arbitrariness and Discrimination Chap. 8

10.Retribution and Religion Chap. 9

11.Death Penalty Opinion Chap. 10

Exam #3 (Thurs., April 27 @ 4 p.m.)

All UCF students have the responsibility to be familiar with and to observe the requirements of

the Rules of Conduct described in The Golden Rule: A Handbook for Students, which may be

obtained from the Office of Student Rights and Responsibilities, Student Resource Center, Rm.

155, UCF Phone: 407-823-6960.