Legal vs.

Quantified Definitions of Standards of Proof

Author(s): Dorothy K. Kagehiro and W. Clark Stanton
Source: Law and Human Behavior , Jun., 1985, Vol. 9, No. 2 (Jun., 1985), pp. 159-178
Published by: Springer on behalf of American Psychology-Law Society (AP-LS)

Stable URL: http://www.jstor.com/stable/1393751

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Springer and are collaborating with JSTOR to digitize, preserve and extend access to Law and
Human Behavior

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
Law and Human Behavior, Vol. 9, No. 2, 1985

Legal vs. Quantified Definitions of

Standards of Proof*

Dorothy K. Kagehirot and W. Clark Stantont

Three laboratory experiments were conducted to compare legal (unquantified) definitions of three
standards of proof ("preponderance of the evidence," "clear and convincing evidence," and "beyond
a reasonable doubt") with quantified definitions, in which the standards of proof were expressed i
probability terms (51%, 71%, and 91% probability of truth). In the first experiment, the quantifie
definitions had their intended effect-verdicts favoring the plaintiffs decreased as the standard of
proof became stricter-but the legal definitions did not have their intended effect. These results we
replicated in the second experiment; in addition, jury instructions that combined the two types of
definitions (legal and quantified) did not have their intended effect. Results of the third experimen
suggest that some legal definitions may be able to communicate their intended difficulty level whe
they appear in a comparative context.

INTRODUCTION

Candor suggests that, to a degree, efforts to analyze what lay ju

concerning the differences among these three (standards of proo
of a judge's instructions on the law may well be largely an ac
there are no directly relevant empirical studies. . . . We probabl
more than that the difference between a preponderance of the e
beyond a reasonable doubt probably is better understood than e
relation to the intermediate standard of clear and convincing ev
less, even if the particular standard-of-proof catchwords do not
great difference in a particular case, adopting a "standard of pr

* This research was supported by NIMH Training Grant No. 86-228-311-04, award
of Nebraska-Lincoln Law-Psychology Program. We wish to thank Dr. Ga
comments on an earlier draft.
t Behavioral Science Division, University of Wisconsin-Parkside.
t Musick, Peeler, and Garrett, Los Angeles.
159

0147-7307/85/0600-0159$04.50/0 ? 1985 Plenum Publishing Corporation

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
160 KAGEHIRO AND STANTON

an empty semantic exercise. ...." In cases involving individual right

criminal or civil, "[t]he standard of proof [at a minimum] reflects the v
places on individual liberty." (Addington v. Texas, 1979, pp. 424, 425)
omitted).
In our legal system, parties initiating lawsuits carry what is called the burden
of proof (Cleary, 1972); they must prove whatever facts are necessary to establish
their legal claims. Thus, in civil suits plaintiffs must show that the facts favor
their version of the case in order to obtain favorable verdicts. Similarly, in criminal
trials the prosecution must present evidence showing that the defendants com-
mitted certain acts, in order to obtain convictions. If the trier of fact-the judge
or the jury-is not satisfied with the plaintiffs' evidence, then the plaintiffs must
lose (while "plaintiff" typically refers only to the party initiating a civil suit, we
shall use it in this article to indicate the initiating party of any action, civil or
criminal).
Factual determinations in court must sometimes be based on very limited
evidence. The trier of fact has no generally available procedure by which to
suspend judgment pending additional evidence, however desirable or valuable
such evidence would be. At the end of the trial, the trier of fact must render a
final and definitive decision regarding the facts in dispute, based on the evidence
presented at trial (a jury may deadlock, but this would require another trial). Two
important and related points follow from this necessity to reach a verdict. First,
the factfinder can never be absolutely certain of the facts in dispute. As Justice
Harlan noted in In re Winship (1970), "All the factfinder can acquire is a belief
of what probably happened" (p. 370). The second point, which follows from the
first, is that the trier of fact will sometimes make erroneous decisions. Such errors
fall into two categories. First, plaintiffs may prevail when the true facts actually
warrant a determination for defendants (false positives). Second, defendants may
prevail when the true facts actually favor plaintiffs (false negatives).
This much is the same for any case, civil or criminal, that is tried in court.
What differs from case to case is the degree to which the trier of fact must be
satisfied that the necessary facts have been established. This degree of satisfaction
is called the standard of proof and takes three basic forms: (a) "preponderance
of the evidence," the standard used in most civil cases; (b) "beyond a reasonable
doubt," the standard used in criminal trials; and (c) "clear and convincing evi-
dence," an intermediate standard. It is the purpose of the standards of proof to
instruct the trier of fact as to the expected level of confidence in the factual
determination and to allocate between plaintiffs and defendants the risks of er-
roneous decision making (In re Winship, 1970).
The lowest standard of proof, "preponderance of the evidence," asks
whether the existence of a fact is more probable than its nonexistence. Thus, the
standard requires a plaintiff to establish the probability of the disputed facts by
something over 50% (McCauliff, 1982; Simon, 1969). The legal system assumes
that this standard divides the risk of erroneous decisions in roughly equal fashion
between plaintiffs and defendants (Addington v. Texas, 1979; Kaplan, 1968). This
assumed equal allocation of error is used exclusively in civil cases, since society's

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 161

interest in the outcomes of such cases is regarded as minimal (S

Kramer, 1982).
In contrast, society's interest in the outcomes of criminal cases is
as quite high, and the trier of fact is asked to be satisfied "beyond a
doubt," the highest standard of proof. The disutility of convictng an
person is viewed as being much greater than the disutility of freeing
(In re Winship, 1970; Kaplan, 1968); hence, the probability of error is in
weighted in the defendant's favor (Speiser v. Randall, 1955). The s
been estimated by judges to be about 90% (McCauliff, 1982; Simon, 19
States v. Fatico, 1978).
The standard of "clear and convincing evidence" is intended to
intermediate position between the preponderance and reasonable d
dards. It is intended to afford defendants greater protection against
decisions than under the preponderance standard and has been applied
civil cases as denaturalization (Nishikawa v. Dulles, 1958), deportation
v. Immigration and Naturalization Service, 1966), civil commitment (A
v. Texas, 1979), and termination of parental rights (Santosky v. Kram
Mean probability estimates by judges of the degree of certainty requ
clear and convincing standard have ranged from 67% to 75% (McCauli
United States v. Fatico, 1978). Despite the clearly intended intermedia
of this standard, recent discussion of the standard's application h
uncertainty about standards of proof in general (see introductory qu
Addington v. Texas, 1979).
From the views expressed by judges in cases and in surveys, two as
would seem to be implicit in the concept of varying standards of pro
judges assume that jurors' probability estimates of the standards do n
from their own (Simon, 1969). Second, despite Chief Justice Burg
admission in Addington (1979) that there is no empirical evidence to in
jurors make reliable distinctions among the three standards, the fact
courts continue to maintain and to apply the different standards ind
assumption that jurors do make their decisions according to these dis
(Addington v. Texas, 1979; In re Winship, 1970; Santosky v. Kramer, 1
present paper attempts to test these two legal assumptions. Previous
ical research has tended to focus on the reasonable doubt standar
press; Kassin & Wrightsman, 1979; Kerr, Atkin, Stasser, Meek, Ho
1976; Nagel, Lamm, & Neef, 1981; Simon, 1970; Simon & Mahan, 1
one previous study has investigated the preponderance standard
Mahan, 1971), and no research has been conducted on the clear and co
standard [see, however, L. S. E. Jury Project's (1973) investigation of
intermediate standard of "sure and certain"]. Therefore, Experiment
ducted (a) to test the legal assumption that the higher standards of pro
greater protection to defendants by comparing the three standards wit
study; and (b) to determine whether jury instructions defining the diff
dards of proof in probability terms (quantified definitions) impro
ability to apply the intended standard, as compared with legal (un
definitions.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
162 KAGEHIRO AND STANTON

EXPERIMENT 1

Method

The subjects were 252 students from various undergraduate psychology

courses. The study was a 2 x 3 complete factorial, between-subjects design with
an additional control condition. Two levels of quantification of standard of proof
(legal or quantified) were crossed with three levels of standard of proof (prepon-
derance, clear and convincing, or reasonable doubt).
The subjects were instructed to assume the role of jurors in a trial and were
randomly assigned to experimental conditions. Experimental materials were ad-
ministered to the subjects in half-hour group sessions. Subjects engaged in no
deliberative discussion before responding to the dependent measures. Since the
study was concerned with the effects of jury instructions rather than with evi-
dentiary issues, a trial summary was utilized to maximize the salience of the
instructions concerning the standard of proof. The subjects were presented with
a summary of a civil trial based on Sargent v. Massachusetts Accident Co. (1940).
The judge's instructions on standard of proof were investigated in connection
with a civil case rather than with a criminal case for the following reasons: (a)
two of the three standards of proof (preponderance and clear and convincing) are
used to reach verdicts only in civil cases; (b) the reasonable doubt standard is
used to reach verdicts in selected civil cases, such as attempts to terminate Amer-
indian parental rights (Indian Child Welfare Act of 1978) and some civil commit-
ment cases (e.g., Estate of Roulet, 1979); and (c) it was felt that problems of
plausibility and confusion might arise if subjects were asked to apply any standard
other than reasonable doubt to a criminal case.
After reading the trial summary, subjects in the experimental conditions read
judge's instructions on the standard of proof to be used in reaching a verdict;
subjects in the control condition received no judge's instructions. The basic
format of the judge's instructions was taken from the California pattern jury
instructions (Richards, 1977):

(1) In this action, the plaintiffs have the burden of establishing [standard of proof]
all of the facts necessary to prove their case.
(2) [Definition of standard of proof.]
(3) If the evidence presented does not establish the plaintiff's case [standard of
proof], then your verdict must be for the defendant.
(4) In determining whether the plaintiffs' case has been proved [standard of proof],
you should consider all of the evidence regardless of who produced it.

In selecting the particular legal definitions to be tested in this study, we faced

the difficulty that such instructions may vary substantially from case to case.
Some states publish official pattern jury instructions which indicate the approved
verbal formulations for many instructions which a judge might give (e.g., Colo-
rado Supreme Court Committee on Civil Jury Instructions, 1980; Kansas District
Judges' Association Committee on Pattern Jury Instructions, 1977); however,
even these formulations vary from state to state. In an attempt to test widely
used legal definitions of standards of proof (i.e., from a large, populous state),

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 163

we selected the preponderance instruction (Richards, 1977) and the

doubt instruction (Committee on Standard Jury Instructions, Criminal,
the California pattern instructions. Since California does not have a pa
and convincing instruction, we selected the formulation given by th
preme Court in State v. Addington (1979), a subsequent decision in the
litigation (see Addington v. Texas, 1979). The probability estimates
the three quantified definitions were consistent with those given
(McCauliff, 1982; Simon, 1969; United States v. Fatico, 1978):
Legal Definitions. By a preponderance of the evidence is meant such evide
as, when weighed with that opposed to it, has more convincing force and the gr
probability of truth.
Clear and convincing evidence is defined as that measure or degree of proof w
will produce in the mind of the juror a firm belief or conviction as to the truth
allegations sought to be established.
Reasonable doubt is defined as follows: It is not a mere possible doubt; bec
everything related to human affairs, and depending on moral evidence, is open to
possible or imaginary doubt. It is that state of the case which, after the entire comp
and consideration of all the evidence, leaves the minds of the jurors in that cond
that they cannot say they feel an abiding conviction, to a moral certainty, of the
of the plaintiffs' case.
Quantified Definitions. [Standard of proof] is defined as evidence, favo
the plaintiffs, having at least a 51%/71%/91% probability of truth.

After reading the trial summary and the judge's instructions, sub
asked to respond to the dependent measures. The subjects were asked
to manipulation checks; to render a verdict, either in favor of the plain
favor of the defendant; to indicate to what extent the evidence favore
tiffs, on a scale of 0% to 100%; and to indicate their certainty tha
rendered a correct verdict on a scale of 0% (not at all certain) to 100%
certain). Experimental (but not control) subjects indicated their difficu
derstanding the judge's instructions, their confidence that they had
judge's instructions, and the plaintiffs' difficulty in meeting the stand
on scales of 0% to 100%, with endpoints labeled appropriately.

Results

A total of 252 students participated in Experiment 1, but results from 54 of

these students were omitted from data analyses because of errors on one or more
of the three manipulation checks. The distribution of these discarded subjects
across experimental conditions is: legal/preponderance, n = 14; legal/clear an
convincing, n = 12; legal/reasonable doubt, n = 6; quantified/preponderance, n
= 7; quantified/clear and convincing, n = 7; quantified/reasonable doubt, n
8. Thus, reported results are based on a sample of 198 subjects: 29 subjects i
each of the experimental conditions (1 subject in the legal/reasonable doubt con
dition was excluded from data analyses because of missing data) and 24 subjects
in the control condition.
A 2 (Quantification of Standard) x 3 (Standard of Proof) MANOVA was
performed, with verdict, certainty in verdict, extent to which the evidence favored

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
164 KAGEHIRO AND STANTON

the plaintiffs, plaintiffs' difficulty in meeting the applicable standard, diff

understanding the jury instructions, and confidence in obeying the jur
tions as the dependent variables. Overall, these experimental subjects r
43% verdicts favoring the plaintiffs; expressed certainty in their ver
75); regarded 57% of the evidence as favoring the plaintiffs; regarded th
as having difficulty in meeting the applicable standard of proof (M = 7
regard the jury instructions as difficult to understand (M = 22); and we
confident that they had obeyed the jury instructions (M = 88). The co
jects rendered 46% verdicts favoring the plaintiffs. Significant multivar
were obtained for quantification of standard, approximate F (6, 16
< .013, and for standard of proof, approximate F (12, 326) = 2.46, p <
multivariate interaction of quantification and standard of proof was als
cant, approximate F (12, 326) = 3.40, p < .001. Significant univaria
the effect of quantification of standard indicated that subjects percei
evidence to favor the plaintiffs when they received quantified defini
63%) than when they received legal definitions (M = 52%, F (1, 167
< .009. Subjects also found the quantified definitions to be easier to un
(M = 16) than the legal definitions (M = 27), F (1, 167) = 8.32, p < .004.
Significant univariate tests of the interaction of quantification and standard were
obtained for verdict and plaintiffs' difficulty in meeting the applicable standard.
These results and the significant multivariate effect of standard of proof were
investigated further in MANOVAs performed separately for the legal definitions
and for the quantified definitions.
For the legal definitions, the multivariate effect of standard of proof was not
significant, approximate F (12, 158) < 1, indicating that the legal definitions of
the standards of proof had no effect on the dependent variables (see Table 1);
there were no significant univariate tests. The multivariate effect of standard of
proof was significant for the quantified definitions, approximate F (12, 160) =
5.50, p < .001, with significant univariate tests for verdict and plaintiffs' difficulty
in meeting the applicable standard. As would be expected if the different stan-
dards of proof operated as intended by the law, verdicts favoring the plaintiffs

Table 1. Experiment 1: Mean Scores by

Quantification of Standard and Standard of Proof
Plaintiff Plaintiffs'
Quantification/standard verdicts difficulty

Legal (Unquantified)
Preponderance 31% 76
Clear & convincing 38% 75
Reasonable doubt 43% 70
Quantified
Preponderance 66%a 69a
Clear & convincing 52%a 58a
Reasonable doubt 31%b 84b

Note. Means with different superscripts differ sign

based on contrasts for significant univariate tests of
standard of proof.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 165

decreased as the standard became stricter, F (2, 84) = 3.67, p < .030
indicated that the proportion of plaintiff verdicts at the reasonable dou
differed significantly from the proportion of plaintiff verdicts at the
standards, t (86) = 2.49, p < .015, but that the proportion of plaintiff
the two lower standards did not differ significantly from each other, t
N. S. Contrary to expectations, subjects regarded plaintiffs' difficulty
the applicable standard of proof to be lowest at the clear and convincin
F (2, 84) = 8.10, p < .001. Contrasts indicated that plaintiffs' diffic
reasonable doubt standard was perceived to be significantly greater than
difficulty at the two lower standards, t (86) = - 3.66, p < .001, but that
perceived difficulty at the two lower standards did not differ significa
= 1.68, p < .097 (see Table 1).

Discussion

In this first experiment, the quantified definitions of standards of proo

found to affect subjects' verdicts as intended by the law; the proportion of
favoring the plaintiffs decreased as the standard of proof became stricte
California and Addington (1979) definitions, however, had no significant ef
subjects' verdicts. Neither type of definition was found to affect subject
tainty in their verdicts or their certainty that they had obeyed the jury i
tions.

EXPERIMENT 2

Experiment 2 was conducted (a) to replicate the major findings

1, but utilizing two stimulus civil cases, instead of only one, and
of legal definitions; and (b) to determine whether combining the
definitions (legal and quantified) in the same jury instructions w
subjects' understanding of the instructions and would affect the
intended by the law. The legal profession has been reluctant to u
definitions of standards of proof; we are aware of no jurisdictio
instructions are expressed in quantified form. In light of the resu
experiment, we thought that combining both legal and quantifie
the same instruction-i.e., supplementing the legal definition wit
expression, rather than replacing the legal definition with a quant
might be a reform more acceptable to the legal profession, if its
could be demonstrated.

Method

The subjects were 220 undergraduate psychology students who participated

for partial fulfillment of a course requirement. The study was a 3 x 3 x 2
complete factorial, between-subjects design with an additional control condition.
Three levels of quantification of standard of proof (legal, quantified, or combined)

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
166 KAGEHIRO AND STANTON

were crossed with three levels of standard of proof (preponderan

convincing, or reasonable doubt) and two levels of order of presen
stimulus trial summaries (insurance claim case first or automobil
first).
The subjects were asked to assume the role of jurors in a trial and were
randomly assigned to experimental conditions. Experimental materials were ad-
ministered to the subjects in half-hour group sessions. The subjects were pre-
sented with a one-page summary of a civil case, judge's instructions on the ap-
plicable standard of proof, and the accompanying dependent measures for the
case (subjects in the control conditions received no judge's instructions). This
procedure was repeated with the second civil case, with the order of case pre-
sentation counterbalanced across all conditions. The subjects were asked to apply
the same standard of proof in both cases; they engaged in no deliberative dis-
cussion before responding to the dependent measures.
The subjects were presented with summaries of an insurance claim case
based on Sargent (1940) and of an automobile accident case based on Smith v.
Rapid Transit (1945). In an attempt to test widely used legal definitions of stan-
dards of proof, we selected federal pattern instructions for preponderance (Devitt
& Blackmar, 1977) and for reasonable doubt (Federal Judicial Center, 1982). The
clear and convincing definition-for which the federal system does not have a
pattern instruction-was again from State v. Addington (1979). The legal defi-
nitions of each standard are presented below:

To "establish by a preponderance of the evidence" means to prove that something

is more likely so than not so. In other words, a preponderance of the evidence in the
case means such evidence as, when considered and compared with that opposed to it,
has more convincing force, and produces in your minds belief that what is sought to be
proved is more likely true than not true.
Clear and convincing evidence is defined as that measure or degree of proof which
will produce in the mind of the juror a firm belief or conviction as to the truth of the
allegation sought to be established.
Proof beyond a reasonable doubt is proof that leaves you firmly convinced of the
plaintiff/s' [specified] case. There are very few things in this world that we know with
absolute certainty, and the law does not require proof that overcomes every possible
doubt.

The quantified definitions were the same as those used in Experiment 1. The
combined versions consisted of the legal definition and then the quantified defi-
nition, introduced by the transitional phrase, "In other words, .."
The definition of the standard of proof was embedded within the judge's
instructions from the federal pattern instructions (Devitt & Blackmar, 1977):

(1) The burden is on the plaintiff/s [specified] in a civil action, such as this, to prove
every essential element of his/her/their claim [standard of proof]. If the proof should fail
to establish any essential element of the plaintiff/s' [specified] claim [standard of proof]
in the case, the jury should find for the defendant [specified].
(2) [Definition of standard of proof.]
(3) In determining whether any fact in issue has been proved [standard of proof] in
the case, the jury may, unless otherwise instructed, consider all the evidence, regardless
of who may have produced it.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 167

After reading each trial summary and accompanying judge's instr

subjects were asked to respond to the dependent measures. Subjects r
to five multiple-choice questions designed to test their understanding
instructions and then rendered verdicts. Next, all subjects indicated t
tent the evidence favored the plaintiff/s and how important it was t
correct verdict. Finally, the experimental (but not the control) subject
their confidence that they had obeyed the jury instructions in reachin
dicts; their difficulty in understanding the jury instructions; and plaintiff
culty in proving the case by the applicable standard of proof. Subjects
to these last five dependent measures on scales of 0% to 100%, with e
labeled appropriately.

Results

Manipulation Checks and Perceived Importance. Overall, the control sub

jects gave 2.0 correct responses to the 5 manipulation checks; the experimental
subjects gave 4.2 correct responses. About three-fourths of the control subject
were already aware that plaintiffs carried the burden of proof in trials; jury in
structions increased subjects' awareness to 94%, overall. One-half of the contro
subjects were already aware that the jury instructions were rules, rather than
suggestions, governing jurors' decision making process; jury instructions i
creased their awareness to 76% (with the one exception of the combined prepo
derance definition). However, these jury instructions' most important role was
to inform subjects about the standard of proof which must be met by plaintiff
and about the defendants' lack of a burden of proof or of a standard of proof t
be met. Only one-third or fewer of the control subjects were aware of these leg
concepts; jury instructions raised subjects' awareness of plaintiffs' standard of
proof to 90%, of the defendants' lack of a burden of proof to 76%, and of the
defendants' lack of a standard of proof to 80%. In general, jury instructions d
not affect the perceived importance of reaching verdicts, which remained uni
formly high (overall M = 82). Thus, fears that different standards of proof m
communicate different levels of importance of reaching a verdict to jurors wer
not supported. The one exception was the quantified reasonable doubt defin
tion-perceived importance was greater for this definition (see Table 2).
Main Dependent Variables. A 3 (Standard of Proof) x 3 (Quantification o
Standard) x 2 (Order of Case Presentation) mixed MANOVA was performe
with case as the within-subjects factor and with verdict, favorability of evidenc
to plaintiffs, plaintiffs' difficulty in meeting the applicable standard of proof, con-
fidence in obeying the jury instructions, and difficulty in understanding the jur
instructions as the dependent variables. Overall, these experimental subjects (t
control conditions were omitted from this analysis) rendered 46% verdicts fa-
voring the plaintiffs; regarded 60% of the evidence as favoring the plaintiffs;
regarded the plaintiffs as having moderately high difficulty in meeting the appl
cable standard of proof (M = 69); were highly confident that they had obeyed
the jury instructions (M = 84); and did not regard the jury instructions to be ver
difficult to understand (M = 25).

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
168 KAGEHIRO AND STANTON

Table 2. Experiment 2: Mean Scores on Manipulation Checks and Perc

Importance by Standard of Proof and Quantification of Standard

Percentage of correct responses

on manipulation checks
Perceived

Standard/quantification A B C D E importance

Preponderance standard
Control 50a 73a 32a 16a 34a 80
Legal (unquantified) 75b 93b 66b 82b 86b 82
Quantified 84b 95b 66b 91b 84b 84
Combined 57a 9 b 77b 82b 80b 75
Clear & convincing
Control 50a 73a 32a 16a 34a 80
Legal (unquantified) 91b 98b 84b 93b 82b 84
Quantified 66C 93b 77b 93b 73b 84
Combined 73c 98b 73b 86b 70b 77
Reasonable doubt standard
Control 50a 73a 32a 16a 34a 80a
Legal (unquantified) 73b 93b 75b 91b 77b 80a
Quantified 84b 98b 86b 98b 86b 92b
Combined 82b 91b 80b 93b 82b 83a

Note. Means with different superscripts differ

iate tests of effect of quantification of standard

Significant multivariate effects we

imate F (10, 354) = 7.04, p < .001, f
F (10, 354) = 1.88, p < .047, and fo
.002. Significant multivariate interac
proximate F (5, 176) = 8.32, p < .00
proximate F (10, 354) = 2.15, p <.0
uninterpretable and, since they invo
terest, were not investigated further
dard of proof was investigated furth
level of quantification of standard in
the legal definitions, the quantified d
For the legal definitions, the multi
longer significant, approximate F
multivariate effect of standard was s
proximate F (10, 120) = 4.88, p < .
verdict, favorability of evidence for pla
the standard of proof, F (2, 63) value
from .044 to .006. As would be expec
erated as intended by the law, verdi
standard became stricter, evidence p
as the standard became stricter, and
of proof was perceived to increase as
the significant univariate tests indica
significantly from the clear and con

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 169

Table 3. Experiment 2: Mean Scores by Quantification of Standard and

Standard of Proof

Evidence Confidence
Verdicts for favoring Plaintiffs' obeyed
Quantification/standard plaintiffs plaintiffs difficulty instructions
Legal (unquantified) definitions
Preponderance 59% 59% 56 84
Clear & convincing 43% 54% 70 86
Reasonable doubt 41% 58% 72 87
Quantified definitions
Preponderance 64%a 53%" 61a 81
Clear & convincing 48%b 64%b 69b 87
Reasonable doubt 36%b 73%b 78b 89
Combined definitions
Preponderance 48% 52%a 70 77
Clear & convincing 34% 59%b 69 81
Reasonable doubt 39% 37%b 73 88

Note. Means with different superscripts

iate tests of effect of standard of proof

these measures, t (65) values r

.023 to .001, but that the two h
each other (see Table 3). For th
standard of proof was signific
a significant univariate test of
6.36, p < .003. Contrasts indic
significantly from the two hig
the two higher standards did n
- 1.94, p < .056. However, an ex
the combined definitions did not have their intended effect.

Discussion

In summary, the most important functional role of these jury instructi

these subjects was to inform them of the applicable standard of proof and
defendants' lack of a burden of proof to be met and lack of a standard of
to be met-legal concepts they were unlikely to know without such instru
The perceived importance with which subjects regarded their decision ma
these cases was not affected by the different standards of proof; thus, a
a lower standard of proof in a particular case does not lower the importa
rendering a verdict in the view of these subjects. The major findings of
ment 1 were replicated. In terms of their intended effect on verdicts, th
tified definitions worked; the federal and Addington (1979) legal definition
here did not. We had hoped that the combined definitions would be as ef
as the quantified definitions. Disappointingly, this did not prove to be th
Combining the legal and quantified definitions of standards of proof in th
jury instructions did not remedy the ineffectiveness of the legal definitio
haps because the legal definitions were presented first.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
170 KAGEHIRO AND STANTON

EXPERIMENT 3

Can legal (unquantified) definitions be presented to jurors in su

to communicate their intended difficulty levels as standards of proof fo
Experiment 3 investigated whether the presentation of legal definit
parative context-i.e., the presentation of all three standards of
improve subjects' understanding of the definitions and of the difficu
were intended to communicate. Another purpose of Experiment 3 w
subjects' understanding of several different sets of legal definitions

Method

The subjects, 96 students from various undergraduate psychology courses,

were group-administered the experimental questionnaire in their classes (one sub-
ject returned an imcomplete questionnaire). Subjects randomly received one of
four sets of legal definitions of standards of proof: (a) the complete set of Colorado
pattern definitions (Colorado Supreme Court Committee, 1980; 1983 Supp.); and
(b) the complete set of New Mexico pattern definitions ([New Mexico] UJI Civil
Committee, 1980; [New Mexico] UJI Criminal Committee, 1982). Jurisdictions
do not always agree upon pattern definitions of each standard of proof. Therefore,
two additional sets of definitions were created: (c) the California pattern defini-
tions of preponderance (Richards, 1977) and reasonable doubt (Committee on
Standard Jury Instructions, Criminal, 1979) were combined with the State v
Addington (1979) definition of clear and convincing; and (d) the federal pattern
definitions of preponderance (Devitt & Blackmar, 1977) and reasonable doubt
(Federal Judicial Center, 1982) were combined with the Kansas pattern definition
of clear and convincing (Kansas District Judges' Association Committee, 1977).
The first page of the questionnaire consisted of a brief description of the
plaintiff's burden of proof and the purpose of the standard of proof. The back-
ground description was followed by four definitions of the standard of proof,
identified only as Standards A, B, C, and D. These definitions consisted of one
set of the legal definitions and one of the quantified definitions (51%, 71%, or
91% probability of truth). Each of the quantified definitions was represented an
equal number of times across all four sets of legal definitions and duplicated one
of the legal definitions of each standard of proof an equal number of times. All
24 possible order presentations of the four definitions were used, and each defi-
nition was associated with each letter designation an equal number of times within
each set of four definitions. For illustrative purposes, one order presentation,
consisting of the New Mexico set of definitions and one quantified definition, is
presented (see Appendix for the complete set of definitions):
Standard A is proof based upon reason and common sense-the kind of doubt that
would make a reasonable person hesitate to act in the graver and more important affairs
of life. [Reasonable doubt standard]
Standard B is proof, favoring the plaintiff, having at least a 71% probability of truth.
[Clear and convincing standard]
Standard C is proof that something is more likely true than not true. [Preponderance
standard]

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 171

Standard D is proof which, when weighed against the evidence in opposition

you with an abiding conviction that the evidence is true. [Clear and convincing

The subjects were then asked to rate each of the four definitio
to how difficult it would be for a plaintiff to win his/her case at ea
proof on three different scales, for convergent validity purposes: (a)
where 1 = least difficult and 4 = most difficult (in case of "tie
standards, subjects were permitted to assign the same rank to stan
perceived as equal in difficulty); (b) a 9-point interval scale, where
difficult and 9 = extremely difficult; and (c) a percentile scale, wh
at all difficult and 100% = extremely difficult. The order in which
sponded to these dependent measures was randomized.

Results

Quantified Definitions. A 3 (Standard of Proof) x 4 (Set of Legal Definitions

in which the quantified definition was embedded) MANOVA was performed, with
the rank-order, interval scale, and percentile scale measures of perceived com-
parative difficulty as the dependent variables. Only the multivariate effect of
standard of proof was significant, approximate F (6, 162) = 6.81, p < .001, with
significant univariate tests for all three dependent variables, F (2, 82) value
ranged from 4.06 to 25.98, p values ranged from .021 to .001. Contrasts indicated
that, for all three dependent variables, the clear and convincing and reasonable
doubt standards differed significantly from the preponderance standard, t (93)
values ranged from -2.75 to -6.63, p values ranged from .007 to .001, and that
the clear and convincing standard differed significantly from the reasonable doub
standard on the interval scale measure, t (93) = -1.97, p < .052, and on the
percentile scale measure, t (93) = - 2.72, p < .008, but not on the rank-orde
measure, t (93) < 1. Thus, for the most part, each quantified definition wa
perceived to be distinctively different in its difficulty as a standard of proof to b
met by a plaintiff (see Table 4).
Legal Definitions. A mixed MANOVA was performed, involving the sam
three dependent variables and with set of legal definitions as the between-subject
factor and standard of proof as the within-subjects factor. Both the multivariate
effect of set, approximate F (9, 270) = 2.25, p < .019, and the multivariate effec
of standard of proof, approximate F (6, 358) = 3.62, p < .002, were significant.
The multivariate interaction of set and standard was also significant, approximat
F (18, 540) = 2.43, p < .001.
Standard of proof was then examined separately for each set of legal defi-
nitions. The multivariate effect of standard remained significant for the Colorado
definitions, approximate F (6, 90) = 2.82, p < .015; for the New Mexico defi-
nitions, approximate F (6, 90) = 2.88, p < .013; and for the federal/Kansas
definitions, approximate F (6, 86) = 3.28, p < .006; but not for the California/
Addington definitions, approximate F (6, 86) < 1. These results indicate that the
subjects were unable to distinguish among the California/Addington definitions
of standard of proof. Contrasts were performed for the significant univariate tests
of the interval scale and percentile scale measures for the other three sets of

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
172 KAGEHIRO AND STANTON

definitions. For the federal/Kansas definitions, the clear and convinci

sonable doubt standards differed significantly from the preponderanc
F (1, 22) = 15.51, p < .001 for the interval scale measure and F (1, 22)
p < .001 for the percentile scale measure, and the clear and convincin
differed significantly from the reasonable doubt standard, F (1, 22)
.019 for the interval scale measure and F (1, 22) = 7.71, p < .011 for th
scale measure. Thus, each of these definitions was perceived as d
different in difficulty as standards of proof. For the New Mexico def
clear and convincing and reasonable doubt standards differed signifi
the preponderance standard, F (1, 23) = 16.05, p < .001 for the int
measure and F (1, 23) = 18.24, p < .001 for the percentile scale me
ever, the clear and convincing standard did not differ from the reaso
standard, F (1, 23) < 1 for the interval scale measure and F (1, 23) = 1
for the percentile scale measure. Thus, for these definitions, subject
to distinguish the preponderance standard from the two higher stan
could not distinguish between the clear and convincing and reaso
standards. Finally, for the Colorado definitions, the clear and con
reasonable doubt standards did not differ significantly from the pre
standard on either the interval scale or percentile scale measures,
1.54 and < 1, respectively; the clear and convincing standard was sign
different from the reasonable doubt standard on the interval scale me
123) = 16.81, p < .001, but not on the percentile scale measure,
3.95, p < .059. An examination of the mean ratings for these definiti
4 indicates considerable confusion among subjects concerning their c
difficulty. If anything, subjects regarded the preponderance and clea
vincing standards, as defined, to indicate greater difficulty than the
doubt standard, as defined (see Table 4).
In summary, of the sets of legal definitions tested in this study, the q
and federal/Kansas sets performed best, with subjects able to disting
and correctly among the three standards of proof. For the New Mex
tions, subjects were able to distinguish the lesser difficulty of the pre
standard, but were unable to distinguish between the two higher sta
"clear and convincing evidence" and "beyond a reasonable doubt
were unable to discern any differences among the California definitio
Colorado definitions appeared to communicate erroneous difficulty le
the reasonable doubt standard perceived as the least difficult of the
dards. It should be noted that the first experiment, which reported
the legal definitions on verdicts, utilized the California/Addington set.
study replicates the ineffectiveness of those definitions. The present
suggests-but does not demonstrate-that the federal set of definition
in Experiment 2, might affect verdicts as intended by the law, if the
sented in comparative context.

Discussion

The legal definitions tested in Experiments 1 and 2 did not have their inte

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 173

Table 4. Experiment 3: Mean Ratings of Comparative

Difficulty of Definition by Set and Standard of Proof
Rank Interval Percentile
Set/standard of proof order scale scale

Quantified
Preponderance 2. a 3.5a 40a
Clear & convincing 2.6b 5.2b 626
Reasonable doubt 2.8b 6. 75
Federal & Kansas
Preponderance 2.1 4.2a 48a
Clear & convincing 2.5 5.6b 64b
Reasonable doubt 2.9 6.9c 81c
New Mexico
Preponderance 2.0 3.5a 39a
Clear & convincing 2.7 4.9b 52b
Reasonable doubt 2.6 5.0b 58b
California & Addington
Preponderance 2.4 4.5 53
Clear & convincing 2.3 5.1 62
Reasonable doubt 2.6 5.1 61
Colorado
Preponderance 2.8 5.8a 55a
Clear & convincing 2.5 6.3a 60a
Reasonable doubt 2.5 3.8b 40b

Note. Means with different superscripts differ si

on contrasts for significant univariate tests of eff
proof.

effect. Experiment 3, however, suggests that all legal definitions are not created
equal; these definitional formulations varied considerably in their ability to com-
municate to subjects their intended difficulty level as standards of proof for plain-
tiffs. Our results suggest that legal definitions may have their intended effect on
verdicts if they are presented in some form of comparative context. For example,
jury instructions might indicate the applicable standard's ordinal position vis-a-
vis the other possible standards of proof (e.g., "There are three possible standards
of proof. In this particular case, we wish you to apply , which is the
lowest/intermediate/highest standard of proof"), in addition to defining the stan-
dard. Research is being planned to test this approach. The quantified definitions
achieved their intended effect while deriving no contextual benefit from the legal
definitions in which they were embedded. It would appear that quantified defi-
nitions provide their own implicit context through their positions on a scale with
known endpoints. Thus, when presented alone, they seem to be better able to
communicate their intended difficulty level.

CONCLUSION

In these experiments, we tested the legal profession's hypothe

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
174 KAGEHIRO AND STANTON

v. Texas, 1979; In re Winship, 1970) that the three standards o

protection to a defendant on a sliding scale-that the preponderan
able doubt standards constitute the endpoints on the scale, w
providing the least protection and the latter providing the most
that the clear and convincing standard occupies an intermediat
first two experiments demonstrated that legal, quantified, and co
tions can be understood by subjects-in the sense that subjects
rote memory of the standard of proof label (i.e., "prepondera
dence," "clear and convincing evidence," and "beyond a reaso
However, these experiments also demonstrated that only the quan
tions were understood by subjects in the sense that the subjec
apply the definitions' intended difficulty level to their decision ma
tified definitions had their intended effect; the proportion of verdicts
plaintiffs decreased as the standard of proof became stricter. How
periments were unable to determine the precise intermediate posit
and convincing standard. In Experiment 1, the proportion of plai
indicated that the effect of the clear and convincing standard wa
to the effect of the preponderance standard than of the reasonable
in Experiment 2, the effect of the clear and convincing standard w
to that of the reasonable doubt standard. These particular legal de
not have their intended effect. The third experiment suggested-bu
onstrate-that the ineffectiveness of the legal definitions may be
a comparative context; jury instructions may have to communicate
standard's ordinal position vis-a-vis the other standards of proof
understanding and utilization by jurors.
Having demonstrated the effectiveness of quantified definition
of proof in the laboratory, we now consider the issue of giving q
nitions to jurors. The legal profession appears to be adverse to th
standards of proof in probability terms. As noted, we know of no
which quantified instructions are given. This apparent aversion to
of legal terms and concepts may be based, at least in part, on mis
its effects. Tribe (1970; 1971) has articulated some of these concer
tions to quantification appear to be based on two assumptions: (a)
creates intentional and permissible risk of convicting an innocent d
(b) specifying the required certainty of a juror for a verdict favoring t
will reduce it to a level below that elicited by a legal definition of
(Tribe, 1971). Focusing on criminal trials and the reasonable doubt s
(1971) has argued that the imposition of criminal sanctions "i
recognized and quantitatively measured doubt" (p. 1373) is legal
even immoral. While Tribe is willing to admit that, out of a large g
some criminal defendants are convicted erroneously, he opposes t
of quantifiable doubt in any individual trial. He regards quantific
doubts as irremediably reducing the subjective certainty that he b
must possess when they reach verdicts against the defendants. Th
parently equates quantifiable doubt with reasonable doubt, and if
he does not believe that a juror can vote to convict.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 175

Tribe's (1970; 1971) objections are not supported by our results. Fi

subjects did not report absolute certainty in their verdicts, nor were
differences found in subjects' certainty levels under the legal definitio
the quantified definitions. This suggests that, contrary to Tribe's ass
jurors are aware of some measurable degree of uncertainty, even whe
against a defendant, and that quantification has no effect on their aw
Second, quantification of standard or proof merely makes explicit
implicit in the original legal definition. If there is an assumption of
risk in the guilt determination process expressed in a quantified defi
because the assumption of risk is present in the original legal definitio
1968)-for example, the legal standard of proof is "beyond a reasonabl
not "beyond any doubt." Simply because the risk is stated less precise
legal definition does not indicate its absence.
Third, Tribe (1971) appears to assume that legal definitions of stan
proof afford greater protection to defendants than quantified definiti
standards. However, without knowing how legal vs. quantified standard
affect their verdicts, whether jurors describe themselves as being
possible" or as being "certain to a .91 probability level" can only b
as differences in verbal expression; it cannot be assumed that "sure a
reflects a higher certainty level than ".91 certainty" without examinin
dicts rendered. Our results indicate that these quantified definitions o
dards of proof have the effect on verdicts intended by the law; the legal de
tested in these studies did not.
Fourth, the probability values in these experiments were presented as the
minimum certainty levels, or thresholds, below which a juror should not go and
still render a verdict for the plaintiffs. The selected probability values do not
preclude certainty levels for jurors which are higher than these minimums. More-
over, the selection of a particular probability value for a standard of proof for
research purposes is not intended nor should it be interpreted as advocacy of that
same value for trial purposes. The probability values were selected to demon-
strate that subjects may have a better understanding of a standard of proof when
it is quantified and that the standard of proof is more likely to have its intended
effect on the rendering of verdicts when it is quantified. The legal profession can
thereafter set the probability level anywhere it wishes-at .91, .95, or .99999.
Fifth, our results need not be interpreted to mean that unquantified defini-
tions of standards of proof should never be used or that quantified definitions
must be used. The results of our third experiment suggest (but do not demon-
strate) that providing a comparative context for legal definitions may produce
their intended effect on verdicts. At the very least, our results suggest that the
legal definitions tested in these experiments should be reexamined (cf. Elwork,
Sales, & Alfini, 1982). The quantified definitions could function as yardsticks
against which the effectiveness of legal definitions is measured.
The preponderance and clear and convincing standards would seem to pro-
vide their own separate challenges to the argument against quantification. One
cannot object that quantification of such instructions would impair jurors' cer-
tainty, since uncertainty is presumed by these instructions. The issues under these

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
176 KAGEHIRO AND STANTON

standards are to determine what degree of doubt is acceptable an

quantification is helpful in making this determination. Our results i
it may be.
One last comment concerning the findings should be made. As with any
laboratory simulation study, criticism is sometimes expressed concerning the lim-
ited generalizability, and hence the verisimilitude, of the findings (Diamond,
1979). While we recognize the need for further testing of the effectiveness of the
quantified definitions under field conditions, it is more difficult to fault our find-
ings concerning the ineffectiveness of these legal definitions on such grounds. If
anything, the laboratory conditions under which the definitions were tested (i.e.,
short written trial summaries, immediately followed by the standard of proof
instruction, immediately followed by collection of the dependent measures) would
seem to provide an opportunity for the definitions to have their maximum in-
tended effect. Actual courtroom conditions (i.e., the presence of additional
sources of influence, such as greater amounts of evidence, longer trial duration,
other required jury instructions) would seem to provide a more rigorous testing
situation in which to demonstrate any effect due to the standard of proof instruc-
tions.
In summary, our concern has been to determine what role the standards of
proof play in juror decision making, specifically to determine whether the appli-
cable standard in any given case achieves the law's professed goal of appropri-
ately distributing the risk of erroneous decision making. If definitions in current
use fail to achieve this goal, then it seems important to determine how the goal
might be reached. We cannot agree with Tribe (1971) that such errors are irrele-
vant because they are not intended. Nor can we agree with Chief Justice Burger
that the actual effect of the standard of proof in any particular trial is of less
concern than the symbolic obeisance rendered to the general principle (see intro-
ductory quote from Addington v. Texas, 1979). If the redistribution of the risk of
erroneous decision making is as essential to our system of justice as Tribe and
the Supreme Court assert, then justice demands that a relationship between the
standard of proof and that risk be demonstrated.

APPENDIX: LEGAL DEFINITIONS OF STANDARDS OF PROOF

FROM EXPERIMENT 3

California and Addington Definitions

. . . proof as, when weighed with that opposed to it, has more convincing f
the greater probability of truth.
. .. proof which will produce in the mind of the juror a firm belief or con
to the truth of the allegations sought to be established.
. . . proof that does not involve a mere possible doubt; because everything
to human affairs, and depending on moral evidence, is open to some possible
inary doubt. It is that state of the case which, after the entire comparison a
eration of all the evidence, leaves the minds of the jurors in that condition t
cannot say they feel an abiding conviction, to a moral certainty, of the trut
plaintiff's case.

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
STANDARDS OF PROOF 177

Federal and Kansas Definitions

.. proof that sometimes is more likely so than not so. In other words, it means
such evidence as, when considered and compared with that opposed to it, has more
convincing force, and produces in your minds belief that what is sought to be proved is
more likely true than not true.
. . . proof which is "clear" in the sense that it is certain, plain to the understanding,
unambiguous, and "convincing" in the sense that it is so reasonable and persuasive as
to cause you to believe it.
. . . proof that leaves you firmly convinced of the plaintiff's case. There are very
few things in this world that we know with absolute certainty, and the law does not
require proof that overcomes every possible doubt.

Colorado Definitions

. . . proof which is most convincing and satisfying in the controversy between the
parties, regardless of which party may have produced such evidence.
... proof which is stronger than a "preponderance of the evidence" and which is
highly probable and free from serious or substantial doubt.
... proof based upon reason and common sense which arises from a fair and
rational consideration of all the evidence, or the lack of evidence, in the case. It is doubt
which is not vague, speculative or imaginary doubt, but such a doubt as would cause
reasonable persons to hesitate to act in matters of importance to themselves.

New Mexico Definitions

. . . proof that something is more likely true than not true.

. . . proof which, when weighed against the evidence in opposition, leaves you with
an abiding conviction that the evidence is true.
... proof based upon reason and common sense-the kind of doubt that would
make a reasonable person hesitate to act in the graver and more important affairs of
life.

REFERENCES

Addington v. Texas, 441 U.S. 418 (1979).

Cleary, E. W. (Ed.) McCormick's handbook of the law of evidence. St. Paul, M
pp. 784-802.
Colorado Supreme Court Committee on Civil Jury Instructions (1980). Colorado
Civil [CJI-Civil 2d], ?? 3.1, 3.3. San Francisco: Bancroft-Whitney.
Colorado Supreme Court Committee on Civil Jury Instructions (1983 Supp.). Col
tions 2d: Civil [CJI-Civil 2d], ? 3.2. San Francisco: Bancroft-Whitney.
Committee on Standard Jury Instructions-Criminal of the Supreme Court of L
California (1977). California jury instructions-Criminal [CALJIC], ? 2.90. S
West.

Dane, F C. (in press). In search of reasonable doubt: A systematic examination of selected quanti-
fication approaches. Law and Human Behavior.
Devitt, E. J., & Blackmar, C. B. (Eds.) Federal jury practice and instructions, Vol. 2, ? 71.14. St.
Paul, Minnesota: West.
Diamond, S. S. (Ed.) Simulation research and the law [Special issue]. Law and Human Behavior, 3,
(1/2).

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms
178 KAGEHIRO AND STANTON

Elwork, A., Sales, B. D., & Alfini, J. J. (1982). Making jury instructions understandab
ville, Virginia: Michie.
Estate of Roulet, 23 Cal. 3d 219 (1979).
Federal Judicial Center Committee to Study Criminal Jury Instructions (1982). Pattern
instructions. St. Paul, Minnesota: West, p. 28.
Indian Child Welfare Act of 1978, 23 U.S.C. ? 1912(f) (1976, Supp. III).
In re Winship, 397 U.S. 358 (1970).
Kansas District Judges' Association Committee on Pattern Jury Instructions (1977). Pa
tions for Kansas-Civil 2d [PIK-Civil 2d], ? 2.11. San Francisco: Bancroft-Whitne
Kaplan, J. (1968). Decision theory and the factfinding process. Stanford Law Review, 2
Kassin, S. M., & Wrightsman, L. S. (1979). On the requirements of proof: The tim
instruction and mock juror verdicts. Journal of Personality and Social Psycholo
1887.

Kerr, N. L., Atkin, R. S., Stasser, G., Meek, D., Holt, R. W., & Davis, J. H. (1976).
a reasonable doubt: Effects of concept definition and assigned decision rule on the
mock jurors. Journal of Personality and Social Psychology, 34, 282-294.
L. S. E. Jury Project (1973). Juries and the rules of evidence. Criminal Law Review,
McCauliff, C. M. A. (1982). Burdens of proof: Degrees of belief, quanta of evidence, or
guarantees? Vanderbilt Law Review, 35, 1293-1335.
Nagel, S., Lamm, D., & Neef, M. (1981). Decision theory and juror decision making.
(Ed.), Perspectives in law and psychology: The trial process (Vol. 2). New York: P
pp. 353-386.
[New Mexico] UJI Civil Committee (1980). Uniform jury instructions-Civil [UJI-Civ
10.17. Charlottesville, Virginia: Michie.
[New Mexico] UJI Criminal Committee (1982). Uniform jury instructions-Criminal [UJ
? 40.60. Charlottesville, Virginia: Michie.
Nishikawa v. Dulles, 356 U.S. 129 (1958).
Richards, P. H. (Ed.) (1977). California jury instructions civil: Book of approved jur
[BAJI], ? 2.60. St. Paul, Minnesota: West.
Sargent v. Massachusetts Accident Company, 29 N.E. 2d 825 (Mass., 1940).
Santosky v. Kramer, 455 U.S. 745 (1982).
Simon, R. J. (1970). "Beyond a reasonable doubt"-An experimental attempt at qu
Journal of Applied Behavioral Science, 6, 203-209.
Simon, R. J. (1969). Judges' translations of burdens of proof into statements of proba
Lawyers Guide, 103-114.
Simon, R. J., & Mahan, L. (1971). Quantifying burdens of proof: A view from the ben
and the classroom. Law and Society Review, 5, 319-330.
Smith v. Rapid Transit, 58 N.E. 2d 754 (1945).
Speiser v. Randall, 357 U.S. 513 (1958).
State v. Addington, 588 S.W. 2d 569 (Tex. Sup. Ct. 1979).
Tribe, L. H. (1970). An ounce of detention: Preventive justice in the world of John Mitc
Law Review, 56, 371-407.
Tribe, L. H. (1971). Trial by mathematics: Precision and ritual in the legal process. Harvard Law
Review, 84, 1329-1393.
United States v. Fatico, 458 F Supp. 388 (E.D.N.Y. 1978).
Woodby v. Immigration and Naturalization Service, 385 U.S. 276 (1966).

This content downloaded from

201.234.181.53 on Thu, 08 Dec 2022 23:04:44 UTC
All use subject to https://about.jstor.org/terms