lilillllllllllllllllllllllllillllllll

3 1293 00881 0925

This is to certify that the

thesis entitled

T/te thots of: Drswefany , OISLoA-firm‘fl‘on I and
terrific/e Elaborwrc‘on on A't'frrude (Junie

presented by

Chien—PIV Fa WI! La i

has been accepted towards fulfillment
of the requirements for

M0512? degree in O - O

 

Major p ofessor

Date f/Br/S‘rl

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

__ , 7, ..fi—#

 

 

 

LIBRARY
Michigan State
4 Untverslty

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

v a

 

 

 

 

 

 

   
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative ActioNEquel Opportunity Institution

cmmpme-pn

 

 

 

 

 

THE EFFECTS OF DISCREPANCY, DISCONFIRMATION, AND COGNITIVE

ELABORATION ON ATTITUDE CHANGE

BY

Chien-Ping Faith Lai

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Sociology

1991

L.

ABSTRACT
THE EFFECTS OF DISCREPANCY, DISCONFIRHATION, AND COGNITIVE
ELABORATION ON ATTITUDE CHANGE

BY

Chien-Ping Faith Lai

The effects of discrepancy and disconfirmation on
attitude change are reviewed. The role of cognitive
elaboration is discussed that source focus which is one kind
of cognitive elaboration should induce the effects of
disconfirmation on comparative elaboration and issue focus
which is the other kind of cognitive elaboration would lead
to the effect of comparative evaluation on final position.

The results support the two process - two opinion model
that discrepancy has a strong direct effect on final
position, whereas disconfirmation has a direct effect on
comparative evaluation. It is concluded that the more
subjects think about the source and engage in the issue, the
greater are the effects of disconfirmation on comparative
evaluation and comparative evaluation on final position,
respectively. Also, some problems of this experiment are

discussed.

To my parents

iii

Acknowledgements

I wish to express appreciation to my advisor, Prof.
Stan Kaplowitz, for his patience, enthusiastic support, and
detailed instructions. Also, I am grateful to Prof. Thomas
Conner and Cliff Broman for the valuable feedback they have
provided.

By the same token, I am indebted to a number of good
friends. Some have taught me new techniques; some provided
service to print questionnaires; further, some have always
been there when I need help. Among them are Sping Wang,
Jumbo Chang, Shyuee Chen, Jashinta Costa, Show-Hui Hsu,
Jackie Larivee, Grace Lin, Bangnan Liu, Peter Liu, Hsin-Ming
Liu, Hui-Jane Lu, Diana Pan, Rosa Teng, Elisabeth
Waeckerlin, and szy-Meng Yang.

Thank all of them and others whose names are not listed
here for their friendships which enriched my life at

Spartan, Michigan State University.

iv

TABLE OF CONTENTS

Page

 

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . Vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . Viii
Theoretical Review . . . . . . . . . . . . . . . . . . 1
Hypotheses . . . . . . . . . . . . . . . . . . . . . . 8

Method . . . . . . . . . . . . . . . . . . . . . . . . 10
Overview . . . . . . . . . . . . . . . . . . . . . 10
The Experimental Design . . . . . . . . . . . . . 10
Independent Variables . . . . . . . . . . . . . . 12
Major Dependent Variables . . . . . . . . . . . . 16
Data Collection . . . . . . . . . . . . . . . . . 16

Results . . . . . . . . . . . . . . . . . . . . . . . . 17
Data Transformations . . . . . . . . . . . . . . . 17
Manipulation Checks . . . . . . . . . . . . . . . 20

Discrepancy . . . . . . . . . . . . . . . . . 20
Disconfirmation . . . . . . . . . . . . . . . 21
Cognitive Elaboration . . . . . . . . . . . . 22
Effects on Attitude Change . . . . . . . . . . . . 27
Tests of Hypotheses . . . . . . . . . . . . . . . 29

Discussion and Conclusion . . . . . . . . . . . . . . . 33
REFERENCES . . . . . . . . . . . . . . . . . . . . . . 35

APPENDIX . . . . . . . . . . . . . . . . . . . . . . . 39

LIST OF TABLES

 

Page

Table 1 - Discrepancy and Disconfirmation Levels for
the Experimental Design. . . . . . . . . . . . . . 13
Table 2 - The Definition of Four Forms. . . . . . . . . 14
Table 3 - Skewness for Dependent Variables. . . . . . . 19
Table 4 - Homoscedasticity for Dependent Variables. . . 20

Table 5 - The Means of Subjects' Effort Spent Thinking about
Convict X Case (EFFORT) by Cognitive Elaboration
Condition. . . . . . . . . . . . . . . . . . . . . 23

Table 6 - The Means of Subjects' Time Spent on the Case of
Convict X (TIMSPD) by Cognitive Elaboration Condition23

Table 7 - The Means of Subjects' Total Thoughts on the Whole
Questionnaire (TOTTH) by Cognitive Elaboration
Condition. . . . . . . . . . . . . . . . . . . 24

Table 8 - The Means of Subjects' Source Thoughts by
Cognitive Elaboration Condition. . . . . . . . . . 25

Table 9 - The Means of Subjects' Issue Thoughts by Cognitive
Elaboration Condition. . . . . . . . . . . . . . 26

Table 10 - The Means of Subjects' Actual Time Spent on
Finishing the Whole Questionnaire (TOTATIM) by
Cognitive Elaboration Condition. . . . . . . . . . 27

Table 11 - Geometric Mean (95% Confidence Interval) of
Subjects' Final Position for Convict X (P1) by
Discrepancy and Disconfirmation Level. . . . . . . 28

Table 12 - Adjusted Geometric Mean (95% Confidence Interval)
of Comparative Evaluation of Convict X's by Discrepancy
and Disconfirmation Level. . . . . . . . . . . . . 29

Table 13 - The Effect of Disconfirmation on Comparative
Evaluation 0 O O O O O O O O O O O O O O O O O O 0 30

vi

Table 14 - The Effect of Comparative Evaluation on Final
Position. . . . . . . . . . . . . . . . . . . . . 31

Table 15 - The Effect of Final Position on Comparative
Evaluation . . . . . . . . . . . . . . . . . . . . 32

vii

LIST OF FIGURES

 

Figure 1 -

viii

Theoretical Review

Festinger and Aronson (1960) suggested that greater the
discrepancy between the recipient own opinion and the
opinion advocated by the communicator, the greater the
dissonance, and hence the greater the opinion change. Based
on cognitive dissonance theory, which was proposed by Leon
Festinger in 1957, a person experiences dissonance when s/he
finds the communicator's opinion is discrepant from what
s/he has in mind. But, how do people get rid of or reduce
the dissonance?

Aronson, Turner, and Carlsmith (1963) suggested two
modes: the recipient may reduce dissonance by changing
his/her own opinion or by derogating the communicator. Their
experiment tried to find out: Under what conditions, will
one change one's opinion and/or derogate the communicator
when an individual is exposed at various discrepant
circumstances? The key point is the communicator's
credibility. Aronson et a1. (1963) assumed that (1) when the
communicator is highly credible, the larger the degree of
opinion change advocated, the greater the dissonance, and
the greater the opinion change. (2) As the communicator is

presented as mildly credible, derogation will replace

2
opinion change as a method to reduce dissonance.
Accordingly, there would be little or no opinion change.

Their results showed that there is a significant
relationship between communication discrepancy and opinion
change. Under the condition of the highly credible
communicator, the greater the discrepancy between the
communicator's opinion and the opinion of the recipient, the
greater opinion change. That is, opinion change increases
with degree of discrepancy in the high credibility level.
However, in the mildly credible condition, it produces
greater opinion change with increasing discrepancy only up
to a point.

The experiment designed by Bochner and Insko (1966)
extended the results of Aronson et a1. (1963). It measured
three dependent variables: opinion regarding sleep,
disparagement of the communication, and disparagement of the
source. It manipulated three independent variables:
discrepancy (nine levels of hours of sleep advocated),
source credibility (high or medium), and order of dependent-
variable measurement (opinion-disparagement or
disparagement-opinion).

The findings confirmed, in one hand, the study of
Aronson et a1. (1963) that there was significant
curvilinearity between opinion change and discrepancy for
the medium credibility source. The expectation that the high

credibility source would show a curvilinear trend, however,

3
was not supported by the data. In the high credibility
condition, opinion change was found to be linearly related
to discrepancy between the communicator and recipients.
Also, the data showed that the medium and high credibility
sources do not differ in opinion change at the moderate
discrepancy levels since theoretically the high credibility
source should have been equal to or superior to the medium
credibility source.

Therefore, if the source is highly credible, the
relationship between discrepancy and opinion change reveals
linearity. If the source is moderately credible, the
relationship tends to be curvilinear. The above studies
treated the communicator's (source) credibility as the
important factor to affect recipients' attitude change when
they experience discrepancy. But, what if recipients found
that what they expected from the source disconfirm the
source's actual position?

Eagly, Wood, and Chaiken (1978) studied the effect of
disconfirmation on attitude change. They suggested that
recipients would be influenced by two types of communicator
biases during the process of message persuasiveness:
knowledge bias and reporting bias. Knowledge bias refers to
a recipient's belief that a communicator's knowledge about
external reality is not truthful because of personal
attributes. Reporting bias refers to the belief that a

communicator's willingness to bring an accurate message of

4
external reality is compromised since s/he suffers
situational pressures.

The results indicated that the communicator gained in
perceived sincerity by taking the expected position in the
knowledge-bias conditions since the source's disposition
corresponded to his behavior. In contrast, in the reporting-
bias conditions, the communicator was regarded as insincere
when he advocated a position that confirmed subjects'
expectancies. However, the source was judged more biased
when he confirmed rather than disconfirmed expectations. No
matter whether the source disconfirmed a knowledge-bias or
reporting-bias expectancy (or both), he was significantly
more persuasive than when expectancies were disconfirmed.

Wood and Eagly (1981) also studied persuasive messages
through attribution analysis. They manipulated subjects'
initial position, expectancy confirmation, and discrepancy
of the source advocated position as three independent
variables. As for dependent variables, they measured causal
attributions, communicator bias, message comprehension, and
opinion change.

The findings of this study asserted that both perceived
communicator bias and message comprehension were
determinants of opinion change. When the communicator's
position disconfirmed subjects' expectancies based on the
communicator's background, he was perceived as unbiased and

more persuasive since recipients believed that he was

5
influenced by factual evidence. This result confirms the
earlier study of Eagly et al. (1978) that message
disconfirmation leads to greater opinion change.

From the above studies, it is obvious that two factors
influencing opinion change are message discrepancy and
message disconfirmation. Studies have defined message
discrepancy as the difference between the position advocated
by the source and the initial position of the recipient.
Message disconfirmation has been defined by Kaplowitz and
Fink (1990) as the difference between the position advocated
by the source and the position which the recipient expected
from the source.1

The effects of discrepancy and disconfirmation on
opinion change seem, however, to have been confounded by
many studies. Do both of them have distinct effects on
opinion change from each other? Kaplowitz et al.'s (1990)
study has further clarified the relationship between
discrepancy and disconfirmation.

First of all, they specify that discrepancy studies
typically use positional measures, but disconfirmation

studies do not. Also, results of Wood and Eagly (1981) and

Nemecek (1985)'s findings, suggested that discrepancy and

 

1Discrepancy and disconfirmation have been defined as
follows: Dp = PA- P0
Df=PA-PE
where Up is discrepancy; Df is disconfirmation; P’ is the
position advocated by the source; P’ is the recipient's
initial position, and.Pg is the p051tion expected from the
source.

6
disconfirmation affect different aspects of the receiver's
opinion regarding the attitude object, via different
processes.

Thus, they tested a model, the two process - two
opinion model (see Figure 1), which assumes that discrepancy
directly affects the receiver's final position and
disconfirmation directly affect the receiver's comparative
evaluation of the current issue relative to others in some
class. Kaplowitz et al. (1990) created the questionnaire
booklet which described the sentence guideline for armed
robbery, Judge Walter's background and his past judicial
experience, the case of Convict X, and, then, asked
subjects' opinions about criminal issues.

The results of the two experiments successfully
disentangle the effects of discrepancy and disconfirmation
on attitude change and support the two process-two opinion
model that discrepancy has a strong direct effect on final
position, whereas disconfirmation has a direct effect on
comparative evaluation.

In these two process, what is the role of cognitive
elaboration? Kaplowitz et al. (1990) concluded that
subjects responded to varying levels of discrepancy by
simply taking a weighted average of their initial position
and the position advocated, without deeply engaging in
cognitive elaboration. Elaboration, however, is required for

the indirect effect of disconfirmation on final position.

Figurel-

Initial

Position

 

Discrepancy

Disconfirmation

/

/

Final
Position

 

 

Comparative
Evaluation

8

Therefore, Kaplowitz et al. (1990) suggested that the
effects of disconfirmation need a greater degree of
cognitive elaboration and future research should manipulate
and measure elaboration levels to see the results under
various conditions of cognitive elaboration.

The current study has taken the above suggestion in
consideration to do further research. The amount of thinking
(elaboration) is manipulated as the third independent

variable in addition to discrepancy and disconfirmation.

Hypotheses

According to Kaplowitz et al.'s study (1990), the two
process model implies that focus of attention can moderate
the effect of disconfirmation on comparative evaluation.
They suggested that the greater when subjects were asked to
focus on the position expected from the source, there was
the greater the effect of disconfirmation on comparative
evaluation than when subjects were asked to focus on the
issue. Based on this, we hypothesize -

H1) The more subjects think about the source, the
greater should be the effect of disconfirmation on subjects'
comparative evaluation of the current case, that is, the
source focus should induce the effect of disconfirmation on

comparative evaluation.

9

Moreover, the results of Kaplowitz et al's (1990) prior
experiments showed that for comparative evaluation and
position to affect each other, cognitive elaboration was
required. In the first experiment, the issue focus group
indicated the larger number of thoughts than the source
focus group. Similarly, the second experiment showed that
subjects, who had received a large amount of knowledge about
the source's previous behavior, had more thoughts (see
Kaplowitz et al. 1990).

These two groups, but not the others, showed
significant effects of comparative evaluation on final
position. Accordingly, there seem to be a larger effects of
comparative evaluation on subjects' final position when
recipients think more about the issue under considerations,
i.e. when they cognitively elaborate on message contents. If
this is the case, then, we have hypothesis 2 -

H2) The greater the amount of cognitive elaboration, on
the issue, the greater will be the effect of comparative
evaluation on final position.

Hypothesis 1 emphasizes source focus, which is thought
to increase the effect of disconfirmation on comparative
evaluation. Hypothesis 2 stresses issue focus, which is
thought to enhance the effect of comparative evaluation on
final position.

If (1) and/or (2) above are correct, this would lead to

an indirect positive effect of disconfirmation on position.

Method

Overview

This study employed a 2 (source focus encouraged vs.
not) x 2 (issue focus encouraged vs. not) x 3 (discrepancy
levels: small, moderate, or extreme) x 3 (disconfirmation
levels: small, moderate, or extreme) factorial design.
Subjects were told that we were doing a study of public
attitudes regarding the criminal justice system, and were

given a questionnaire booklet to complete.

The Experimental Design

This study continued to use the questionnaire booklet
similar to one designed by Kaplowitz et a1. (1990), which
concerned public attitudes towards the criminal justice
system. The topic of criminal sentencing was chosen because
it can express the message and subjects' position on a
numerical scale with ratio measurement. Armed robbery was
selected from 16 different crimes based on a pilot study in
which the recommended years imprisonment for armed robbery
had the lowest dispersion relative to its central tendency
(Kaplowitz et al. 1990).

First of all, the questionnaire begins with the

introduction of criminal justice system and its functions

10

11
and the purpose of this study. Subjects then are given some
information about the Michigan sentencing guideline for the
crime of armed robbery, which is ten years imprisonment.
Based on the sentencing policies, subjects give their
initial views of the appropriate sentence for the crime of
armed robbery, which is treated as the initial position
(Po).

Next, subjects were told a story about Judge Walters,
who has served as a judge for several years. During his
years as a judge, Judge Walters had sentenced many
defendants, often relying on his own points of view, instead
of simply relying on the sentencing guidelines. Subjects
were informed that Judge Walters had imposed fifteen prison
sentences for armed robbery and were informed of the
average, the smallest, and the greatest of these sentences
in prison. After reading the above descriptions, subjects of
the source focus group (H & S forms) were asked to write
down their viewpoints about Judge Walters' degree of
severity or leniency.

Then, a particular defendant called Convict X, was
introduced to subjects. They read the statement Judge
Walters made as he sentenced Convict X. Convict X's sentence
was varied in different conditions. Then, subjects indicated
their final position regarding Convict X's proper sentence

and also answered other questions about it.

12

Independent Variables

There are three independent variables in this study -
message discrepancy, message disconfirmation, and cognitive
elaboration. In order to create low, moderate, and extreme
message discrepancy and message disconfirmation, we varied
the position advocated by the source (PR) to be 10, 22.5,
and 50 years, and the position expected from the source (Pd
was manipulated by varying the previous sentences of the
source for armed robbery. These were 2, 4.5, 10, 22.5, and
50 years, respectively. The individual subject's initial
position (P5) should be the sentencing guideline - 10 years.
Hence, a position advocated of ten years leads to a zero
level of message discrepancy.2 The nine (3x3) combinations,
in which discrepancy and disconfirmation are orthogonal, are

presented in Table 1.

 

2We used the logarithms to express three levels and
kinds of discrepancy and disconfirmation instead of the raw
numbers. According to Kaplowitz et al. (1990), the constant
ratio should be made to increase discrepancy and
disconfirmation in steps which will be equally far apart to
subjects. Thus, the logarithms is used. Since the lowest and
highest number is 10 and 50, respectively, they can be
expressed like 10, 10X, and 10X2.

Since, 10x2 = 50, x = 2.24
Therefore, Dp = logzz,‘ (PA/P0) = 1n (PA/Po)/ In (2.24)
Similarly, Df = logzz,’ (PA/PE) = in (P /PE)/ ln (2.24)

From the above steps, three levels 0 d1screpancy and
disconfirmation (0, 1, 2) are created.

13
Table 1 -

Discrepancy and Disconfirmation Levels for the Experimental
Design

 

     
   
 

I Disconfirmation level

 

 

 

 

 

Discrepancy level ————————_
I O 1 2

0 10/10 10/4.5 10/2
(8, 12.5) (3.6, 5.6) (1.6, 2.5)

1 22.5/22.5 22.5/50 22.5/4.5>
(18, 28) (8, 12.5) (3.6, 4.5)

2 50/50 50/22.5 50/10
(40, 62.5) (18, 28) (8, 12.5)

 

 

 

 

 

 

 

Note: The number of the right of each slash is the average of
previous sentences (P’) which manipulates expected position.
The number of the left of each slash is the sentence for
Convict X (P3). The numbers in parentheses are the lowest and
highest of previous sentences for armed robbery. The initial
position.(Pb) of the sentence for armed robbery is assumed to
be 10 years.

The third independent variable, which is the main
different feature from Kaplowitz et al.'s study (1990), is
the amount and kind of thinking (cognitive elaboration)
encouraged. We designed four forms to manipulate the levels
of cognitive elaboration. Four forms (H, S, I, L) are
distinguished from each other in terms of specific key
instructions at the end of the introduction to the form.
Simultaneously, two different kinds of focuses (on source
vs. on issue) were encouraged within four forms. For the

definitions of these forms and what the abbreviations, H, S,

I, and L stand for, see Table 2.

14

 

 

 

 

 

 

 

 

 

Table 2 -
Issue Source Focus
Focus
Yes No
Yes H I
No S L
Note:
H - High: Cognitive elaboration encouraged on both the

source and the issue

L'"HUJ
I

Cognitive elaboration encouraged on the source only
Cognitive elaboration encouraged on the issue only
Low: Cognitive elaboration encouraged neither on the

source nor on the issue

Two of the forms (H & 8 forms), asked subjects to think

about the

source, i.e. the judge. The H & I forms asked

subjects to pay attention to the crime, the sentence, and

the reasoning behind the sentence. As for the low level of

elaboration (L form), we ask subjects to give their first

impressions without carefully thinking about other things.

The key instructions of H and L forms are stated,

respectively, as follows:

For H form -

"The Commission is interested in the reasoned
opinions of a broad spectrum of the community. It
is especially interested in your thoughts about
how the characteristics of the judge, as well as
the crime and the reasoning behind the sentence,
can affect the sentence. ........ Be sure to
focus primarily on thegiudqe. the crime. the

sentence, and the reasoning behind the sentence

15
and try not to let other things distract you. This
is not a speed test."

For L form -

"Sometimes people form their opinions after a
great deal of thought. At other times, they must
act on relatively fast first impressions. While
first impressions are sometimes different from
judgments formed from extensive thought, the
evidence suggests that they are sometimes more
valid. We suspect that this may even be in true
issues of criminal sentencing. Hence, we would
like you to answer this questionnaire with your

first impressions."

In the S form, the instructions put stress on the judge
and ask for subjects' thoughts about "how the
characteristics of the judge can affect the sentence". In
the I form, we ask subjects think about "the crime, the
sentence, and the reasoning behind the sentence".

In both conditions in which source focus is encouraged
(S & H), subjects were also be asked their view of Judge
Walters' severity, his future sentences for armed robbery,
and to indicate their confidence in their own prediction.

These questions, however, are omitted in the I and L forms.

 

16
Major Dependent Variables

Subjects' attitudes were measured in two ways: the
subjects' final position was their view of the sentence
Convict X should receive after reading Judge Walters'
sentencing speech for Convict X. The comparative evaluation
of Convict X, asked how bad he was. It used a magnitude
scaling question (see Lodge 1981 for discussion of this
technique) which rated 100 as bad as the average armed
robber. The higher the numerical value, the worse is
subjects' comparative evaluation of Convict X. The scale has
no maximum value. We also assume that thinking about the
object should be reflected in thoughts listed. Subjects were
then asked to list all thoughts about the sentence of
Convict X. This thought listing was also used to measure
different kinds of elaboration.

In addition, we assumed that the more subjects think
about the object, the more effort and time they should spend
on it. We measured two different time, one is the time
subjects say they spent on thinking about the case of
Convict X; the other is the actual time they spent finishing

the whole questionnaire.

Data Collection
Three hundred forty-eight male and female
undergraduates in several communication courses at Michigan

State University participated as subjects who received extra

17
credit for their participation. We have 36 different forms,
which were randomly administered to subjects in order to get
as near an equal number as possible in each form.

Before subjects started to fill out the questionnaire,
we explained some of the opinion scales to assist them with
the magnitude scaling questions. We set the number 100 as a
moderate value and subjects were free to choose any number
they like. In other words, there was no upper limit. To
enhance motivation, subjects were told that their opinions
would be important to judges and be reported to the US
Commission on Judicial Reform. They were encouraged to
answer every question as well as possible.

So that we could measure the amount of time spent on
the questionnaire, subjects were asked to write down their
time of starting on the cover sheet and their time of
completion on the final page. For this purpose, we helped
them by putting the time on the blackboard and updating it
every minute.

After excluding some subjects who have previously heard
or participated or did not believe in the cover story, the

final valid sample is reduced from 348 to 331.

Results
Data Transformations
By examining the skewness and the distribution of the

residuals of raw data in this study, we found that the

l8

assumption of normality and homoscedasticity appeared to be
violated. Therefore, we transformed the raw data to meet the
assumptions. Since the distribution of residuals of the data
has a tail in the positive direction, it's helpful to take
logs of the dependent variables. Zero is the lowest value
among the magnitude scale variables and we cannot use a log
transformation without adding a constant value. Thus, we
tried some different additive constants and selected the
most appropriate one which had the minimal skewness.3 The
data appeared to less skewed and more homoscedastic after

taking the logarithm (See Table 3 & 4).

 

3For the variable of EFFORT, one outlier was found and
we recoded the extreme values greater than 1500 to 1500
before transforming.

19

 

 

 

 

 

 

 

 

 

 

 

Table 3 -
Skewness for Dependent Variables

Dependent Skew before Skew after

Variables Transformation Transformation Value of K
INIT POS 2.514 -.213 0
FIN POS 2.158 .178 0
SURPRISE 3.627 -.635 15
PSYDISC 2.302 -.772 15
UNEXPECT 2.877 -.636 15
EFFORT 6.483 .480 15
BADPRE 5.419 -.396 25
CONBAD 2.975 .571 .001
TIMSPD 1.095 -.l86 l
TOTATIM 3.745 .691 l

 

 

 

 

 

 

Note: The logarithmically transformed data used the formula X*
= Ln(X+K) where X is the raw ‘variable and X' is the
transformed. The adjusted geometric mean is found by first
taking the anti-logarithm of the mean of X*, then, subtracting
the constant K. Here, the value K of each variable is listed
on the last column as the above. For the abbreviations, INIT
POS and FIN POS are subjects' initial and final positions,
respectively. SURPRISE - "how surprising did you find the
Judge's sentence on Convict X"; PSYDISC - "how different was
the Judge's sentence on Convict X from the prior sentence";
UNEXPECT - "how unexpected did you find the Judge's sentence
on Convict X"; EFFORT - "how much effort did you spend
thinking about this case"; BADPRE - "how bad are the previous
armed robbers"; CONBAD - "how bad is Convict X"; TIMSPD - "the
time spent thinking about the case of Convict X"; TOTATIM -
"the time spent completing the whole questionnaire".

20

 

 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 4 -
Homoscedasticity for Dependent Variables
Dependent Cochran's C Cochran's C Sample
Variables before after Size
Transformation Transformation
INIT POS .2148 .1343 328
FIN POS .3387 .1994 330
BADPRE .5009 .1742 331
CONBAD .2906 .1420 331
SURPRISE .2900 .1609 331
PSYDISC .2592 .1917 330
UNEXPECT .4123 .1746 331
EFFORT .5082 .1916 331
TIMSPD .2024 .1307 330
TOTATIM .3160 .2000 331
Note: The value of Cochran's C (max variance/sum of

variances), which was used to test for homoscedasticity, was
computed by nine group (three levels of discrepancy * three
levels of disconfirmation) in ANOVA.

Manipulation Checks

Discrepancy - Our assumption that a position advocated of

ten years has a discrepancy of zero requires that subjects

initial position (P5) be that ten years is the proper

sentence for armed robbery. In other words, I3 should have a

mean of 10 years and low variance. The geometric mean of

subjects' initial attitude was 10.87 years. There were 37.2%

of subjects choosing ten years and 89.1% of subjects in the

range from five to fifteen years.

21
Also, we checked the correlation between PSYDISC (how
psychological discrepant Judge Walters' sentence was) and
the actual positional discrepancy, r=.429, F(1, 328)=73.911,

p=.000. Thus, subjects were aware of the discrepancy.

Disconfirmation - It was manipulated by varying both the
position Judge Walters was expected to pass and the actual
position he decided to sentence Convict X. After reading
Judge Walters' average sentences in the previous armed
robbery (PREVAVE), subjects of source focus group (H & S
forms) were asked about expected average sentence from Judge
Walters' in the future. We found that PREVAVE predicted
subjects' expectation of Judge Walters' average sentence in
the future (FUTAVEl) with r=.843, F(1, 112)=272.887, p=.000.
It showed that subjects expected Judge Walters to behave the
same way as he had behaved before.

For the manipulation of disconfirmation, both of
magnitude scaling questions - SURPRISE and UNEXPECT (how
surprising and how unexpected Convict X's sentence was) were
significantly affected by Df. For Df and SURPRISE, analysis
of variance showed that, r=.503, F(1, 329)=111.662, p=.000:
for UNEXPECT, r=.495, F(1, 329)=106.660, p=.000. The results
mean that subjects responded to the two questions according

to the different levels of disconfirmation.

22

Cognitive Elaboration - There were two independent cognitive
focus variables, source focus and issue focus, which were
presented by four forms. Six variables were measured as
manipulation check: subjects' effort spent thinking about
the case of Convict X (EFFORT); their report time on this
case (TIMSPD): their total number thoughts while completing
this questionnaire (TOTTH) - source thoughts (SOTH) and
issue thoughts (ISTH); and their time finishing the whole
questionnaire (TOTATIM). If manipulation of elaboration is
successful, we would expect to see either the source or the
issue focus groups (H, S, & I forms) spend significantly
more effort, and more time, and generate more thoughts than
the group of L form. We would also expect more source
thoughts in the source focus conditions and more issue
thoughts in the issue focus conditions. In other words, the
main or interaction effects of the two elaboration
encouragement on the above six dependent variables would be
anticipated.

For EFFORT (see Table 5), we found that there was no
significant effort of source focus (F=.168, p=.683) or issue
focus (F=.116, p=.734) since subjects in L form contributed

more effort (M=4.97) than those in H form (M=4.93).

23

Table 5 -

The Means of Subjects' Effort Spent Thinking about Convict x
Case (EFFORT) by Cognitive Elaboration Condition

 

 

 

 

 

 

 

Issue Source Focus
Focus
Yes No Row means
Yes 4.93 (H) 5.01 (I) 4.97
No 5.00 (S) 4.97 (L) 4.99
Column means 4.97 4.99 4.98

 

 

 

 

Note: The variable of EFFORT has been transformed prior to
doing ANOVA. There are between 73 and 83 subjects per cell,
for a total N=308.

As for TIMSPD, though subjects of both elaboration
conditions did spend more time on the current case on the
average (Table 6), it failed to show significant effects of

either of the two independent variables. For source focus,

F=1.286, p=.258, and for issue focus, F=1.378, p=.241.

Table 6 -

The Means of Subjects' Time Spent on the Case of Convict x
(TIMSPD) by Cognitive Elaboration Condition

 

 

 

 

 

 

 

 

Issue Source Focus
Focus
Yes No Row means
Yes 2.23 (H) 2.13 (I) 2.18
No 2.08 (S) 2.09 (L) 2.08
Column means 2.15 2.11 2.13

 

 

 

Note: The unit of time is ndnutes. TIMSPD was transformed
before doing ANOVA. There are between 80 and 85 subjects per
cell, for a total N=330.

24
As another check on the focus manipulation, subjects
were asked to list all thoughts in their minds. The results
indicated that subjects in either the source or issue focus
groups generated more thoughts than those in L form (Table
7), however, the slight difference of the means of total
thoughts was not statistically significant since the values
of F for source focus and issue focus are .028 (p=.867) and

.235 (p=.628), respectively.

Table 7 -

The Means of Subjects' Total Thoughts on the Whole
Questionnaire (TOTTH) by Cognitive Elaboration Condition

 

 

 

 

 

 

 

Issue Source Focus
Focus
Yes No Row means
Yes 2.68 (H) 2.79 (I) 2.74
No 2.66 (S) 2.64 (L) 2.65
Column means 2.67 2.71 2.69

 

 

 

 

Note: There are about 73 to 83 subjects per cell, and the
total N=308.

Among total thoughts, we distinguished two kinds of
thinking - source thoughts and issue thoughts. There were
significant effects of source focus and issue focus on these
two dependent variables. For the manipulation of source
thoughts (SOTH), Table 8 showed that subjects of source
focus group engaged in more significantly source thinking,
F(1, 327)=6.102, p=.014. We also saw that an issue focus

actually reduced the number of source thoughts,

25
F(1, 327)=31.464, p=.000. In addition, there was a
significant source focus by issue focus interactions on

SOTH, F(1, 330)=11.920, p=.001.

Table 8 -

The Means of Subjects' Source Thoughts (80TH) by Cognitive
Elaboration Condition

 

 

 

 

 

Issue Source Focus
Focus
Yes No Row means
Yes .20 (H) .29 (I) .25
No .93 (S) .46 (L) .69
Column means .57 .38 .47

 

 

 

 

 

 

Note: There are about 80 to 85 subjects per cell, and total
N=331.

For the issue thoughts (ISTH), significant effects of
both source focus and issue focus groups were found, F(1,
330)=6.036, p=.015, and F(1, 330)=12.035, p=.001,
respectively. Table 9 indicates the result. While an issue
focus increased the number of issue thoughts, it decreased

the number of source thoughts.

26

Table 9 -

The Means of Subjects' Issue Thoughts (ISTH) by Cognitive
Elaboration Condition

 

 

 

 

 

Issue Source Focus
Focus
Yes No Row means
Yes 2.13 (H) 2.31 (I) 2.22
No 1.35 (S) 1.96 (L) 1.66
Column means 1.74 2.13 1.94

 

 

 

 

 

 

Note: The number of subjects is from 80 to 85 per cell, and
total N=331.

Finally, for the time of completion of this
questionnaire, there was a highly significant effect of
source focus variable (H & 8 forms) on TOTATIM, r=.242,
F(1, 330)=20.464, p=.000. It indicated that subjects in the
source focus conditions spent more time on the whole
questionnaire. There were also two-way interaction effects
of source and issue focus, F(1, 330)=6.322, p=.012. (see

Table 10).

27
Table 10 —

The Means of Subjects' Actual Time Spent on Finishing the

 

 

 

 

 

Whole Questionnaire (TOTATIM) by Cognitive Elaboration
Condition
Issue Source Focus
Focus
Yes No Row means
Yes 3.13 (H) 3.07 (I) 3.10
No 3.21 (S) 3.01 (L) 3.11
Column means 3.17 3.04 3.11

 

 

 

 

 

 

Note: The unit of time is minutes. TOTATIM was transformed.
The number of subjects is from 80 to 85 per cell, and total
N=330.

In sum, it appeared that both SOTH and ISTH variables
were manipulated successfully for source focus and issue

focus, and TOTATIM indicated success only for source focus

manipulation.

Effects on Attitude Change

Four way analysis of covariance was performed to
examine the effects of independent variables on attitude
change, which were measured by subjects' final position (PH
and comparative evaluations of Convict X. The independent
variables include discrepancy, disconfirmation, source
focus, issue focus, and subjects' initial position (P3) as
the covariate.

For subjects' final position (Pg), we found highly
significant effects of the covariate P0,IF(1, 327)=145.08,

p=.000, and Discrepancy, F(2, 327)=69.254, p=.000. It showed

28
that there was a strong linear relationship between
discrepancy and attitude change (final position), but
disconfirmation had a non-significant relationship with
position, consisted with Kaplowitz et al.'s two opinion-two
process model. See Table 11 for the geometric mean and 95%
confidence interval of subjects' recommended sentences for

Convict X, by discrepancy and disconfirmation level.

Table 11 -

Geometric Mean (95% Confidence Interval) of Subjects' Final Position for
Convict X (P1) by Discrepancy and Disconfirmation Level

 

 

 

 

 

Discrepancy Disconfirmation level
level
0 1 2 Row
means
0 11.02 9.97 10.87 10.56
(9.44,12.22) (8.5,11.69) (9.72,12.16)
l 15.56 13.63 14.06 14.47
(13.59,17.83) (12.3,15.1) (12.23,16.17)
2 21.37 20.41 20.68 20.81
(17.43,26.23) (l7.32,24.05) (17.15,24.95)
Column 15.67 14.76 15.32 15.27
means

 

 

 

 

 

 

Note: The geometric mean is the anti-logarithm of the arithmetic mean of
the logarithmically transformed data. The limits of the confidence
interval for the geometric mean are obtained by taking anti—logarithms of
the limits of the confidence interval for the transformed data. There are
between 32 and 42 subjects per cell, and the total sample is 330.

For the comparative evaluation of the badness of
Convict X (CONBAD), however, only the independent variable

of Df showed a highly significant effect, F(2, 327)=16.240,

p=.000. The means for low,

disconfirmation are;

moderate,

119.4,

29

and extreme

166.9, and 174.8, respectively.

It indicated that subjects' comparative evaluation of the

badness of Convict X increased with the rising of

disconfirmation levels (see Table 12).

Table 12 -

Adjusted Geometric
Evaluation of Convict X's Badness by Discrepancy and Disconfirmation.Level

Means

(95% Confidence

Interval)

of Comparative

 

 

 

 

 

 

 

 

 

 

 

 

 

Discrepancy Disconfirmation level I
level
0 l
0 125.3 159.2 145 142.1
(105,149.5) (136.6,185.7) (117,180)
1 123.8 175.4 140.8 144.9
(107.4,142.6) (141.7,216.8) (118.6,167)
2 107.3 166.7 236.3 171.9
(90 3,127.7) (135.8,204.4) (192 5,289.7)
Column means 119.4 166.9 174.8 149

 

Note: For a definition of the adjusted geometric mean, see Table 3 (Here,
K-.001). There are between 32 and 42 subjects per cell, for a total N-330.
100 is a moderately bad level, and the higher the value, the worse Convict

X is viewed.

Tests of Hypotheses

As for the effects of cognitive elaboration, subjects

using the H & S forms should think more about the source

than those using the other forms. This should lead to a

greater effect of disconfirmation on comparative evaluation

of the Convict X case. The result clearly confirms

 

30
the greater should be the effect of disconfirmation on
subjects' comparative evaluation of the current case (see

Table 13).

Table 13 -

The Effect of Disconfirmation on Comparative Evaluation

 

 

 

 

Issue Source Focus
Focus
Yes No
Yes 8=.315** (H) p=.155 (I)
No 6=.375*** (S) B=.156 (L)

 

 

 

 

 

Note: Simple regression was run with comparative evaluation as
the dependent variable and disconfirmation as the predictor.
Beta (6) is the standardized coefficient. ** p<.01, and ***
p<.001.

On the other hand, we predicted that a greater issue
focus should increase the greater effect of comparative
evaluation on final position. This second hypothesis was
confirmed by subjects of I form (the issue focus group), who
showed a significant effect of comparative evaluation on
final position (see Table 14). That is, the greater the
amount of cognitive elaboration subjects engage in the

issue, the greater is the effect of comparative evaluation

on final position.

31

Table 14 -

The Effect of Comparative Evaluation on Final Position

 

 

 

 

 

 

 

Issue Source Focus
Focus
Yes No
Yes 6=.097 (H) 6:.166" (I)
No 6:.080 (S) 8:.070 (L)

 

 

Note: Multiple regression was run with final position as the
dependent variable and discrepancy, initial position, and
comparative evaluation as predictors. * p<.05.

If disconfirmation (Df) significantly affects
comparative evaluation (Eval), and comparative evaluation
also has a significant effect on final position (IQ), then,
it should lead to the relationship that disconfirmation
indirectly affects final position. However, none of the four
forms indicated such indirect positive effect of
disconfirmation on final position.“

According to cognitive consistency theories, subjects'
comparative evaluation of Convict X's badness and their
final position regarding the appropriate sentence should be

related to each other. Based on Kaplowitz et al's finding

(1990), we would expect that comparative evaluation should

 

‘The diagram is like as follows:
Df 4 Eval
prep,
Eval » P1

Though H & S forms showed strong effects of Df on Eval, they
did not have any significant effect of Eval on.P’. Thus, the
indirect positive effect of Df on PH is always small.

32
have a greater effect on position than vice versa.
Hypothesis 2 has predicted that subjects' comparative
evaluation affect their final position. But, conversely,
their final position can also predict comparative evaluation
(see Table 15). This result is not expected because the
finding of Kaplowitz et al. (1990) seems to be an one way
effect. This study, however, shows two-way effects instead
of one way. It seems that position has more impact on
comparative evaluation than the reverse, at least in our
study. However, these ordinary least squares (OLS)
regressions are somewhat questionable because the causal

model has two-way causation.

Table 15 -

The Effect of Final Position on Comparative Evaluation

 

 

 

 

 

 

 

 

Issue Source Focus
Focus
Yes No
Yes 6=.220* (H) ,8=.239* (I)
No 6=.083 (S) 6=.262* (L)

 

Note: Multiple regression was run with comparative evaluation
as the dependent variable and. disconfirmation and final
position as predictors. * p<.05.

Discussion and Conclusion

Consistent with Kaplowitz et al.'s study (1990), the
results of this study confirmed the assumptions of the two
process model that discrepancy had a strong and direct
effect on final position and disconfirmation had a
significantly direct effect on comparative evaluation of the
current case.

The results also provided strong support for hypothesis
1 that the more subjects focus about the source, the greater
the effect of disconfirmation on their comparative
evaluation. Hypothesis 2 was confirmed partly by one of the
issue focus groups (I form), which showed a significant
effect of comparative evaluation on final position.
Originally, we would expect that the source or issue focus
conditions should encourage subjects' to spend more effort
and time, and to produce more thoughts. Since the
manipulation of elaboration by these variables did not work
very successfully, it influenced the experimental result.

We suspect that some errors in the experimental forms
make such difference. Within the four forms, the first
paragraph on page 44 (see Appendix), which introduces Judge

Walter's sentence for Convict X, should not be the same in

33

34
the four focus groups. This error may have had some effect
on the result.

In addition, through the manipulation of source focus
and issue focus, we asked subjects think about the source,
but it seemed that may not be enough because one kind of
thinking (source) induced the other kind of thinking
(issue), which weakened the effects of comparative
evaluation on final position.

One important question about cognitive elaboration
remains unaddressed by the present study. It concerns the
limitations on increased the issue focus facilitating
message processing. Theoretically, characteristics of the
issue content should have a greater impact under high
cognitive elaboration conditions. If the issue, however, is
not so important and is non-self—relevant to subjects, there
would be some limitations to enhance subjects' issue focus
even in the high cognitive elaboration condition (Petty and
Cacioppo, 1979). It may provide one of the reasons that
subjects of the issue focus group of this study did not show
strong effects of their comparative evaluation on final
position since they were not highly involved in the issue of

criminal justice system.

LIST OF REFERENCES

REFERENCES

Aronson, E., Turner, J.A., and Carlsmith, J.M.
1963 "Communicator Credibility and Communication Discrepancy
as Determinants of Opinion Change". Journal of Abnormal

and Social Psychology 67: 31—36.

Bochner, S., and Insko, C.A.
1966 "Communicator Discrepancy, Source Credibility, and
Opinion Change". Journal of Personality and Social

Psychology 4: 614-621.

Chaiken, Shelly
1980 "Heuristic versus Systematic Information Processing and
the Use of Source versus Message Cues in Persuasion".

Journal of Personality and Social Psychology 39: 752-

766.

Chaiken, S., and Stangor, C.

1987 "Attitudes and Attitude Change". Annual Review

PSYChOlOgy 38: 575-630.

35

36
Chaiken, Shelly

1987 "The Heuristic Model of Persuasion". In M.P. Zanna,

J.M. Olson, and C.P. Herman (Eds.), Social Influences:

The Ontario Symposium, Vol. 5, Pp. 3-39. Hillsdale, NJ:

Erlbaum.

Eagly, A.H., Wood, W., and Chaiken, S.

1978 "Causal Inferences about Communicators and Their Effect

on Opinion Change". Journal of Personality and Social

Psychology 36: 424-435.

Festinger, L., and Aronson, E.

1960 "The Arousal and Reduction of Dissonance in Social

Contexts". In D. Cartwright & A. Zander (Eds.), Group

Dynamics: Research and Theory. (2nd ed.), Pp. 214-231.

Evanston, IL: Row, Peterson.

Hovland, C.I., Harvey, O.J., and Sherif, M.
1957 "Assimilation and Contrast Effects in Reactions to

Communication and Attitude Change". Journal of Abnormal

Social Psychology 55: 244-252.

Johnson, B.T., and Eagly, A.H.
1989 "Effects of Involvement on Persuasion: A Meta-

Analysis". Psychological Bulletin 106: 290-314.

37
Kaplowitz, S.A., Fink, L., Mulcrone, J., Atkin, D., and
Dabil, S.
1990 "Disentangling the Effects of Discrepancy and
Disconfirming Information". Unpublished paper, Michigan

State University, East Lansing, MI.

Lodge, M.
1981 Magnitude Scaling: Quantitative Measurement of

Opinions. Beverly Hills, CA: Sage.

Nemecek, P.J.
1985 "The Effects of Disconfirmation and Discrepancy on
Attitude Change". Unpublished master's thesis, Michigan

State University, East Lansing, MI.

Petty, R.E., and Cacioppo, J.T.

1979 "Issue Involvement Can Increase or Decrease Persuasion
by Enhancing Message - Relevant Cognitive Responses".
Journal of Personality and Social Psycholoqy 37: 1915-

1926.

Petty, R.E., Cacioppo, J.T., and Goldman, R.
1981 "Personal Involvement as a Determinant of Argument-
Based Persuasion". Journal of Personality and Social

Psychology 41: 847-855.

38
Petty, R.E., and Cacioppo, J.T.
1981 Attitudes and Persuasion: Classic and Contemporary

Approaches. Dubuque, IA: Wm.C. Brown.

Petty, R.E., and Cacioppo, J.T.
1986 Communication and Persuasion: Central and Peripheral

Routes to Attitude Change. New York: Springer-Verlag.

Wood, W., and Eagly, A.H.

1981 "Stages in the Analysis of Persuasive Messages: The
Role of Causal Attributions and Message Comprehension".
Journal of Personality and Social Psycholoqy 40: 246-

259.

APPENDIX

CARD 1
SID

Study of Public Attitudes
Towards the Criminal Justice System

Michigan State University

Form 22H

1990

Today's date

 

Course Number

 

Section Number

 

Class Time

 

Time Started Questionnaire

 

40
Study of Public Attitudes
Towards the Criminal Justice System

Crime has long been a major concern of our society, but
different people have different views as to the prOper
treatment of criminals. Some people feel that our legal
system treats criminals too leniently to be an effective
deterrent. Others, however, feel that time in prison makes
criminals more likely to commit crimes in the future and that
overly long terms require the state to spend too much of its
money on prisons.

If the law is to be supported, people must believe that the
criminal justice system serves the needs of society. While not
everyone need agree with every judicial decision, it is
necessary that the public understand the way in which the
criminal justice system functions and the kinds of factors
which can enter into sentencing decisions.

In this study, you. will first. be asked about. your' own
experience with the criminal justice system and your
acquaintance with people who work in this system. You will
then be given some information about sentencing policies in
Michigan and asked your initial views of the proper sentence
for a crime. You will then be told of a real-life sentencing
decision by a judge and will be asked some questions about
this decision.

The answers you provide will help make judges more aware of
your concerns since your answers will be reported to the US
Commission. on. Judicial Reform. 'This body is currently
studying sentencing policies and practices in various states
and will make recommendations about them.

The Commission is interested in the reasoned opinions of a
broad spectrum of the community. It is especially interested
in your thoughts about how the characteristics of the judge,
as well as the crime and the reasoning behind the sentence,
can affect the sentence. 'Fherefore, please read all questions
carefully and give each of them the carefully thought out
answers they deserve. Be sure to focus primarily on Q1;
iudqe1,the crime. the sentence. and the reasoning behind the
sentence and try not to let other things distract you. This
is not a speed test.

We are aware that you do not have full information about the
case you will evaluate, but many decision in life must be made
without complete information. Please answer all questions on
the form.

 

 

41

Part I
Below you will be asked some questions about your experience
with the criminal justice system. Please answer them
honestly. Be assured that your responses are totally

anonymous and confidential.

 

1. Are you personally acquainted with anyone who has been a
judge in criminal trials?
Yes (1)
No (0)
2. Are you personally acquainted with anyone who has been an
attorney in criminal trials?
Yes (1)
No (0)
3. Are you personally acquainted with anyone who has been a
member of a parole board?
Yes (1)
No (0)
4. Have you, or anyone you know well, ever been a witness in
a criminal trial?
Yes (1)
No (0)
5. Have you or anyone you know well, ever been a defendant
in a criminal trial?
Yes (1)
No (0)

 

42
Part II

Sentencing Guidelines

The State of Michigan, along with many other states, has
issued a Sentencing Guidelines Manual. These Guidelines,
which are based on a consensus of legal experts, are to assist
judges and provide some degree of consistency in sentencing.
Below is a copy of the cover of the Sentencing Guidelines
Manual of the State of Michigan.

State of Michigan

 

The sentence you will be examining was for the crime of armed
robbery. The Michigan Sentencing Guideline for the crime of

-armed robbery is 10 years imprisonment. The sentence for this

crime is not only a consensus of legal experts, but has also
been found to be supported by a large majority of the public.

These Guidelines, however, are recommendations, not laws.
Because many people feel that a judge must be able to take
into account the special features of each case, the law
permits a judge to pass a sentence which is considerably
greater or considerably less than the Guideline.

To make sure that you have absorbed all of the information
above, please answer the following questions. REFER TO THE
INFORMATION ABOVE, IF YOU NEED TO REFRESH YOUR MEMORY.

43

On the views of which group of people are the Sentencing
Guidelines based?

 

 

What is the Sentencing Guideline for the crime of armed
robbery? years imprisonment

 

HOW'does the public feel about this particular Guideline?

 

 

Think of some reasons which might be used to justify the
selection of 10 years as the Guideline for the crime of
armed robbery.

 

 

 

How many years in prison do you think is an appropriate
sentence for the crime of armed robbery?

years
(Note: You need not choose a whole number of years but
do choose a number of years. Hence, please do not choose
"life imprisonment" or "death." 'The death penalty is not
allowed in Michigan.)

Are judges required to follow these Guidelines?

___ Yes (1)
__ N0 (0)

Why or why not?

 

 

 

44
Part III

As stated earlier, while the State of Michigan has Sentencing
Guidelines, the State still allows.Iudges to make up their own
minds in passing sentences. “Therefore, a sentence may deviate
considerably from the Guidelines, for a variety of reasons.

We will now give you some information about a particular
judge, whom we shall call Judge Walters. The following are
excerpts from a report on various Michigan Judges, which was
released one year ago.

Judge Walters is a judge in one of the larger
metropolitan areas in Michigan. He is in his fifties,
has gray hair, is married, and has grown children.

He has had many years of experience as a judge in
criminal cases. In imposing sentences, he sometimes
imposes the sentence recommended by the Sentencing
Guidelines. However, he places the greatest weight on
his own judgment.

In his many years of judicial experience, Judge Walters
has had to pass sentence on many defendants for a variety
of crimes. . . .

Of those defendants, fifteen were convicted of, and
sentenced for, the crime of armed robbery. Of these
fifteen.sentences.for'armed robbery, the average»sentence
he imposed was about 10 years in prison. The smallest of
these sentences was 8 years, while the greatest of them
was 12.5 years. . . .

(Criminal sentences are typically expressed as a certain
number of years plus a certain number of months. To
improve comprehension, however, this report has expressed
parts of a year in decimal format. For example, 6 months
is expressed as .5 years.)

45

Like all people, different judges have different points of
view. Some tend to be more severe with defendants who have
been found guilty, while others tend to be more lenient.

Please answer the following questions, before you turn to the
next page.

1. Keeping in mind what you have learned about Judge
Walters, how severe do you view him?

 

 

If Judge Walters continues to be a judge for many more years,
he may pass many more sentences for armed robbery. Suppose he
passes 100 more sentences for armed robbery. These 100
sentences may not all be the same. We would like to know,
however, what you think will be the average of the next 100
sentences he may pass for armed robbery.

2. If, in the future, Judge Walters passes 100 additional

sentences for armed robbery, what is your best guess as
to what the average of the additional sentences will be?

My guess is that the average of Judge Walters' future

 

sentences for armed robbery will be years
imprisonment
3. How sure do you feel about the guess you have just made?

If you feel completely certain it is correct, choose
100%. If you feel completely certain it is igcorrect,
choose 9%. If you feel it is equally likely to be right
or wrong. choose 50%. You may use any number between 0%
and 100%.

I feel % sure of my answer to question 2 above.

46
Part IV

This past year, after the release of the report which we
quoted, Judge Walters had to again sentence someone for armed
robbery. We will refer to this person as Convict X. We will
now present the statement Judge Walters made as he sentenced
Convict X. As you read it, keep in mind both what you know
about Judge Walters the reasons he gave for his decision.

By threat of force and violence, you gained access to
money which was not rightfully yours. You brandished a
lethal weapon and made quite clear that you would not
hesitate to use it if your crime were in any way
resisted. Since there was no resistance, you did not
fire your weapon, but the terror you instilled in all of
those present will be with them for a very long time.
Clearly, you played a major role in the planning and
execution of this crime. jFinally, your record shows that
this is not the first time that you have violated the
laws which create a civilized society. . . Therefore, I
sentence you to 50 years in the penitentiary.

Before preceding further, let us be sure that certain facts
are clear in your mind. To assure this please take a moment

to answer the following questions. YOU MAY TURN BACK TO
REFRESH YOUR MEMORY IF NECESSARY.

1. What was the Sentencing Guideline for the crime of armed
robbery?
years imprisonment

2. What was your best guess as to the average of the
sentences Judge Walters would pass in the future?

years imprisonment

3. What sentence did Judge Walters pass on Convict X?

years imprisonment

47

Remember it is important that you think about how the
characteristics of the judge and those of the criminal and
crime can influence the sentence. To help you with this
thinking, please answer the following questions.

1) Taking into account what you know about Judge Walters, and
about the case of Convict X, why do you think Judge Walters
imposed that particular sentence on Prisoner X?

 

 

 

 

2) Do you think this was an appropriate sentence? Why or why

not?

 

 

 

 

 

You will soon be asked to evaluate the case of Convict X. .As
you answer these questions, please be sure to think about

1) the Sentencing Guideline for the crime of Armed
Robbery

2) Judge Walters' sentence for Convict X

3) How strict or lenient Judge Walters generally is,

and the sentences he has given before
and

4) the reasons Judge Walters gave for his sentence.

 

48
Part V

Earlier in this study, you expressed an opinion about the
usual sentence for armed robbery but since then, you have
received information about this specific case and have had
additional time to think about this issue. Therefore, please
feel free to give a different answer than you gave before, if
you feel that is appropriate.

We know that you do not have all of the information that Judge
Walters had when he imposed the sentence he didc Nonetheless,
please answer the following questions the best you can. ‘While
some of the questions may seem very similar to others, please
answer all of them.

1. To how many years in prison do you think Convict X should
have been sentenced?
(Note: You need not choose a whole number of years) but
do choose a number of years. Hence, please do not choose
"life imprisonment " or "death." The death penalty is
not allowed in Michigan.)

I think Convict X should have been sentenced to years
in prison.

We will next be asking you some questions using a kind of
scale with which you may not be familiar. While some scales
have 100 as a maximum value, the scales which follow treat 100
as a moderate value and have no upper limit.

2. Think back to the sentence you expected Judge Walters to
impose on Convict X, after you were introduced to Judge
Walters, but before you read his sentencing speech. How
surprising did you find the sentence he imposed on
Convict X?

If you found the sentence not at all surprising, write 0
(zero). Iijou found his sentence moderately surprising,
write 100. If you found his sentence twice as surprising
as moderately surprising, choose 200 (2 x 100). If you
found his sentence half as surprising as moderately
surprising, choose 50 (1/2 x 100). You may use any
number between 0 and 100 to indicate a level of surprise
which is less than moderate and any number greater than
100 for to indicate a level of surprise which is greater
than moderate. While 0 is the lowest number you may use,

there is no "highest number."

Judge Walters' sentence of Convict X was units
surprising to me.

PLEASE THINK ABOUT EACH QUESTION BEFORE YOU ANSWER IT. THINK
ABOUT THE JUDGE, THE SENTENCE AND THE REASONING BEHIND THE
SENTENCE.

49

3. Think back to the sentence you thought armed robbery
should receive, before you read Judge Walters' sentencing
speech.

How different was Judge Walters' sentence of Convict X from
your own prior view of a proper sentence for armed robbery?

Imagine that 100 is moderately different from your own
view. If you think the difference between your view and
Judge Walters' sentence is twice as much as a moderate
difference, rate it as a 200. If you think the
difference between your view and Judge Walters' view is
half as much as a moderate difference, rate it as a 50.
If Judge Walters' sentence is not at all different from
your view, rate it Q (zero). You may use any number you
wish, from zero on up.

Judge Walters' sentence was units different from the
sentence I initially would have imposed.

4. How expert a judge do you think Judge Walters is?
(Imagine that 100 is moderately expert and zero is not at
all expert.)

 

I think Judge Walters is units expert.

5. How fair' a judge do you think; Judge Walters' is?
(Imagine that 100 is moderately fair and zero is not at
all fair.)

I think Judge Walters is units fair.

6. How serious a crime is armed robbery? (Imagine that 100

is moderately serious and zero is not at all serious).

I think armed robbery is units serious.
7. Think back to the sentence you expected Judge Walters to

impose on Convict X, after you were introduced to Judge
Walters, but before you read his sentencing speech. How
unexpected did you find the sentence that Judge Walters
imposed on Convict X? (Imagine that 100 is moderately
unexpected and zero is not at all unexpected.)

I found the sentence Judge Walters imposed on Convict X
units unexpected.

PLEASE THINK ABOUT EACH QUESTION BEFORE YOU ANSWER IT. THINK
ABOUT THE JUDGE, THE SENTENCE AND THE REASONING BEHIND THE
SENTENCE.

50

8. How objective a judge do you think Judge Walters is?
(Imagine that 100 is moderately objective and zero is not
at all objective.)

I think Judge Walters is units objective.

 

9. How biased a judge do you think Judge Walters is?
(Imagine that 100 is moderately biased and zero is not at
all biased.)

I think Judge Walters is units biased.

 

10. How strong, or compelling, do you think.were the reasons
Judge Walters had for the sentence he imposed on Convict
X? (Imagine that 100 is moderately compelling and zero
is not at all compelling.)

I thought the case for Judge Walters' sentence was
units compelling.

We are now going to ask you how bad you think various
criminals are. To measure badness of criminals, let us now
use a new "yardstick".

As your new "yardstick", imagine that the average armed
robber is 100 units bad and zero is not at all bad. You
will then be comparing the badness of various criminals
with the badness of the average armed robber.

11. Think back those who Judge Walters sentenced for armed
robbery, prior to his sentence of Convict X. How bad are
they?

I think those‘previously sentenced for armed robbery, by.Judge
Walters, were units bad.
12. How bad is Convict X? (Again, imagine that the average

armed robber is 100 units bad.)

I think Convict X is units bad.

 

PLEASE THINK ABOUT EACH QUESTION BEFORE YOU ANSWER IT. THINK
ABOUT THE JUDGE, THE SENTENCE AND THE REASONING BEHIND THE
SENTENCE.

51

13. Imagine that -7 is strongly favoring, for Convict X, the
ten year sentence recommended by the Sentencing
Guidelines and +7 is strongly favoring Judge Walters'

actual sentence for Convict X. Further, imagine that
numbers between -7 and +7 represent views in between
those extremes. Where would you put your own view on

 

that scale? (Circle the number which best represents
your View.)

10 Actual
years Sentence

 

-7 [-6 [-5 I-4 l—3 |-2 [-1 I 0 [+1 1+2 |+3 |+4 |+5 1+6 1+7

PLEASE THINK ABOUT EACH QUESTION BEFORE YOU ANSWER IT. THINK
ABOUT THE JUDGE, THE SENTENCE AND THE REASONING BEHIND THE
SENTENCE.

52

 

Part VI
We would now like to ask some factual questions. Please
answer them, without looking’ up the information in the
questionnaire. Do not worry if your answers are incorrect.

The questionnaire is strictly anonymous and any wrong answers
will have no effect on you.

 

1. Prior to his sentence of Convict X, how many defendants
did Judge Walters sentence for armed robbery?
defendants
2. The average of those prior sentences was years
imprisonment.
3. The lowest of those prior sentences was years.
4. The highest of those prior sentences was years.

For the next few questions, place a check next to True, for
statements, which are TRUE, according to the speech made by
Judge Walters as he sentenced Convict X. Place a check next
to False, for statements, which are FALSE, or which were not
stated or implied in Judge Walters' sentencing speech.

5. In the armed robbery in which Convict X took part,
someone was shot.
True (1)
False (0)
6. Convict X played a major role in the crime.
True (1)
False (0)
7. This robbery was the first crime Convict.X had committed.
True (1)
False (0)
8. Hostages were taken in that armed robbery.
True (1)

False (0)

53
Part VII

We know that we have asked you many questions and some of them
may seem very similar to others. We want to assure you that
we do have reasons for asking them and to request your
patience and co-operation in answering them.

1. Please list all thoughts which were at all related to the
sentence received by Convict X. Please list each thought
on a separate line. Moreover, please categorize each
thought as either favoring severity (marked S), favoring
leniency (marked L), or favoring neither (marked N).

 

 

 

 

 

 

2. How much effort did you spend thinking about this case?
(Imagine that 100 is moderate effort and zero is no
effort.)

I spent units of effort on this case.
3. About hOW'mUCh time did you spend thinking about the case

of Convict X?

minutes and seconds

54

4. Have you previously heard of this study?

Yes (1)
No (0)

If "yes", please explain briefly what you heard.

 

 

 

5. Have you previously participated in this study?

___ Yes (1)
_ N0 (0)

If "yes", please explain briefly.

 

 

 

 

 

 

6. In your own words, please describe the purpose of this
study.

7. Please indicate your time of completion of this

questionnaire.

Time of Completion

 

We very much appreciate your participation.

 

MICHIGAN STATE UNIV. LIBRARIES
llHlllmllmlll"I"llllllllllllllllmllllllllllllllllllllll
31293008810925