‘: (’1‘ _...,AI..~~ «n. . - -- , , ~

‘-

 

THE EFFECTS OF SOCIAL CLASS;

MORAL ORIENTATION. AND SEVERITY'OF'
PUNISHMENT 0N CHILDREN-’3 MORAL 7

RESPONSES TO TRANSGRESSION

Thesis for the Degree ‘o‘f-Ph. D, ,

M‘CH‘GA“ STATE'UN'VERSW
DAVID. JOHN DePALM’A: ,.
- 19721   1*15%?st“rs;-€fé§fif§eiaj

 

   
 

._ 1W!gall/m/LI/lj/m/g/IW/l L ”13W???”

This is to certify that the

thesis entitled

THE E‘FFECTS 0F SOCIAL cums;

mega; oretENTA-TION, AND sevtRlTV

o? PumstmemT 0N LHlLDRELN‘S

MOKRL (LE—spamsEs TO TfiRNSG£E$$iW
presented by

DNA!) JoHN DE PALMA

has been accepted towards fulfillment
of the requirements for

PHD degree in PSVCHOkOGV

 

 

 
 
 
 
 

/// ‘If4{{(j)

Major professor

Dat6264bc’g /;7Z

0.7839

 

 

LIBRARY '

  
 
   

    
  

 
 

E, airtime av "
"ME & SUNS'
WK BINDERY INC.

LIBRARY BINDERS
"RINGPORY. MICHIGAI
t ‘ 5-:

      

ABSTRACT

THE EFFECTS OF SOCIAL CLASS, MORAL ORIENTATION,
AND SEVERITY OF PUNISHMENT ON CHILDREN'S
MORAL RESPONSES TO TRANSGRESSION

BY

David John DePalma

Recent studies (and earlier surveys) have attempted
to investigate the origins of moral responses. However,
because of the limited generalizability of the data
obtained, and various theoretical and/or methodological
problems, little understanding of the mechanism of moral
response has resulted.

Aronfreed (1961, 1963) recognized this difficulty,
and examined the variety of moral responses and their social
antecedents. His results indicated that two responses,
self-criticism and reparation, were differentially related
to the social class, sex, and parental discipline of the
child.

However, recent research by Hoffman (1970) has
important implications for the classification of parental
discipline. It would seem from Hoffman's research that
Aronfreed's categorization of disciplinary techniques

might be further extended. Aronfreed classified discipline

David John DePalma

as either "induction," or "sensitization." Hoffman found
that the moral orientation of both children and parents
could be coded as humanistic-flexible, conventional-rigid
(both internal), or external. The child's orientation was
related to his parents' moral orientation and disciplinary
techniques. Using different judgment categories, McKinney
(1971) and Olejnik (1971) also found that the orientation
of parents (prescriptive or proscriptive) was related to
the orientation of their children. The latter study found
this relationship existed in four-year-olds and their
parents. Thus, the orientation of the parents can have a
profound influence on their children, and at a very early
age. This substantiates Hoffman's contention that the
child-rearing pattern is directly related to the child's
moral development.

Aronfreed (1963) had subjects push for ten trials
through a formation of toy soldiers to remove a nurse doll
from the "battlefield." A modified version of this
technique was used in the present study. The subjects
were 128 eight-year-old boys classified according to social
class (Hollingshead, 1957), and moral orientation (Hoffman,
1970). The subjects were then randomly assigned to one of
four conditions (Baseline, Group I, Group II, or Group III).
In the Baseline condition, the subjects punished themselves
over two sets (Phases I and II) of ten trials. In Group I,

the subjects were punished (Phase I) at a level higher

David John DePalma

than the highest baseline average. In Group II, the
subjects received punishment equivalent to the baseline
average. In Group III, the subjects received punishment
lower than the lowest baseline average. In Phase II these
three groups ( as well as the baseline) punished themselves.
On trial eleven the subjects were tested for the presence
of self-critical and reparative responses (Aronfreed,

1963).

The response measures recorded were: the number
of self-critical responses over the ten trials, the presence
or absence of self-critical responses on trial eleven, the
presence or absence of reparative responses on trial
eleven, and the number of Tootsie Rools removed (self and
experimenter punishment) over the ten trials, in Phase I
and Phase II.

The analyses revealed a significant main effect of
punishment treatment, and a significant interaction between
moral orientation and punishment treatment. The interpre-
tation of the data suggested that subjects of high social
class and flexible moral orientation may have a more
highly "internalized conscience" and do not "need" the
overt stimuli (punishment and self-criticism) to control
their behavior. The present data when considered with
research by Luria (1961) indicated a developmental sequence

in the internalization of moral sanctions.

David John DePalma

The directions of the response data supported
Aronfreed's contention that self-criticism and reparation
are two different responses, and are not indicative of
some "unitary" structure such as conscience. The strong
influence of experimenter punishment, and the patterning
of self-critical and reparative responses suggested the
importance of imitation and past socialization experiences
in the child's responses to transgression.

Some suggestions for additional studies were
offered, and it was hoped that these ideas (and the present

data) would stimulate further investigations in the area.

THE EFFECTS OF SOCIAL CLASS, MORAL ORIENTATION,
AND SEVERITY OF PUNISHMENT ON CHILDREN'S

MORAL RESPONSES TO TRANSGRESSION

BY

David John DePalma

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1972

To Char

ACKNOWLEDGMENTS

Many people have facilitated the completion of
this project by the cooperation and assistance they so
generously offered.

My dissertation committee: Dr. John McKinney,
Dr. Elaine Donelson, Dr. Lucy Ferguson, and Dr. Gordon
Wood should be thanked for their judicious advice,
particularly in the planning stages of the research, and
for their support throughout the term of the project. I
would like to thank Dr. Elaine Donelson and Dr. Lucy
Ferguson for their help with the theoretical aspects of
the study. Dr. Gordon Wood also gave unselfishly of his
time, and provided invaluable counsel on methodology,
design, and analysis. I especially wish to express my
gratitude to Dr. John McKinney whose criticisms, encourage-
ment, and friendship have made this task both an enjoyable
and rewarding experience.

I am grateful to Dr. Joseph Bechard, Superintendent
of Okemos Public Schools, for permitting us to involve the
four elementary schools in the study. The principals of

these schools were: Dr. Marcia Boznango, Lucille Corkrey,

iii

Jeff Richburg, and Lynn Cherryholmes. These principals,
their staff, and the parents of the children involved
deserve appreciation for their patience and cooperation
during the entire program. The students who served as
subjects should also be thanked for their time and efforts.
I am particularly indebted to Dave Boumgarden and
Don Unsworth who were conscientious and hard-working
research assistants, and without whose involvement the
scheduled termination of this project would not have been.

realized.

iv

TABLE OF CONTENTS
Page
LIST OF TABLES . . . . . . . . . . . . . vi
LIST OF FIGURES. . . . . . . . . . . . . viii
Chapter
I. INTRODUCTION . . . . . . . . . . . 1
II. METHOD . . . . . . . . . . . . . 13
Subjects . . . . . . . . . . . . 19
Procedure and Design . . . . . . . . 22
III. RESULTS . . . . . . . . . . . . . 31
IV. DISCUSSION . . . . . . . . . . . . 39
REFERENCES . . . . . . . . . . . . . . 48
APPENDICES
Appendix
A. Moral Orientation Items. . . . . . . . 52
B. Experimental Design . . . . . . . . . 54

C. Summary Tables of Analyses of Variance. . . 55

Table

1.

LIST OF TABLES

Means and Standard Deviations of the Number
of Self-Critical Responses for Each
Punishment Treatment by Moral Orientation
by Social Class Cell for Phase I and
Phase II . . . . . . . . . . .

Means and Standard Deviations of Self-
Critical Responses on Trial 11 for Each
Punishment Treatment by Moral Orientation
by Social Class Cell for Phase I and
Phase II . . . . . . . . . . .

Means and Standard Deviations of Reparative
Responses on Trial 11 for Each Punishment
Treatment by Moral Orientation by Social
Class Cell for Phase I and Phase II . .

Means and Standard Deviations of the Number
of Tootsie Rolls Removed for Each
Punishment Treatment by Moral Orientation
by Social Class Cell for Phase I and
Phase II . . . . . . . . . . .

Summary of Analyses of Variance of the
Number of Self-Critical Responses for
Punishment Treatment by Moral Orientation
by Social Class for Phase I and Phase II

Summary of Analyses of Variance of Self-
Critical Responses on Trial 11 for
Punishment Treatment by Moral Orientation
by Social Class for Phase I and Phase II

vi

Page

33

34

35

36

55

56

Table Page

C-3. Summary of Analyses of Variance of Reparative
Responses on Trial 11 for Punishment
Treatment by Moral Orientation by Social
Class for Phase I and Phase II . . . . . 57

C-4. Summary of Analysis of Variance of the
Number of Tootsie Rolls Removed for
Punishment Treatment by Moral Orientation
by Social Class for Phase II. . . . . . 58

vii

LIST OF FIGURES

Figure

1. Two Dimensions of Value Development:
Reinforcement (Punishment and Reward) and
Behavioral Orientation (Wrong-Doing or
Right-Doing) . . . . . . . . . .

viii

Page

INTRODUCTION

Although the development of moral behavior has
interested psychologists for decades, it has only been
recently that the origins of moral responses have been
examined in the laboratory. These studies (and earlier
surveys) attempted to isolate the factors which influence
the acquisition and development of these responses.
However, because of the limited generalizability of the
data obtained, and various theoretical and/or methodological
problems, little understanding of the mechanism of moral
response has resulted.

Conceptualizations of moral development which use
a unitary structure of acquisition seem to be partly
responsible for this difficulty in understanding moral
responses. The assumption of a prerequisite unity in the
" . . . forms and sources of moral behavior may obscure
some important differences between specific responses and
their distinct antecedents" (Aronfreed, 1963a). Further
support for a more flexible interpretation of the moral

development process comes from Hoffman (1970a):

Knowing a person's moral orientation alone then does
not enable us to predict with confidence how he will
behave. Whether he resists or submits to temptation
'f—— . . .

is a function of a complex balance of forces involv1ng
his achievement needs and specific goal strivings,
guilt and other aspects of his moral structure, his
system of ego controls and various aspects of the
immediate situation (p. 345).

The socialization of moral behavior therefore plays
an important role in the understanding of the acquisition
of moral responses. Moral behavior can be motivated by
internal as well as by external forces. Internal standards
are learned through processes called introjection and
identification.1 Allinsmith (1954) interprets the child's
compliance with parental prohibitions and his adoption of
these ideals as the child's wish to master the environment.
After the child has "taken in“ these injunctions, he acts
as though he had within him a force (the "voice of
conscience") which threatens him as he previously perceived
himself threatened in reality. Thus, a person in a moral
(transgression) situation, depending upon the degree of
his conformity to the internal demands, punishes himself
even when there is no possibility of external censure.

Allinsmith also found that the individual's
severity with himself is related to types of parental
discipline and socioeconomic status. Burton, Maccoby, and

Allinsmith (1961) supported this notion and showed that

restrictive parents have children who more thoroughly

‘

1For a further discussion of these and similar
Processes see Hoffman (1970a).

internalize parental values. Aronfreed's (1961) expla-
nation is that parents who focus on withdrawal of affection
in their discipline may be those who are also most oriented
toward inducing an internal governor of conduct in their
children. Parents who sensitize their children to the
punitive external consequences of transgression induce an
external control in their children. Disciplinary techniques
then can have a profound influence on the moral behavior
of the child. Sears, Maccoby, and Levin (1957) observed
that parental withdrawal of love makes the child take on
characteristics of his parent. That is, after transgressing,
the child will criticize, evaluate, punish and control his
behavior as his parent had formerly done. The child
responds to power assertion, however, by trying to avoid
the parent. This and other studies led Hoffman (1970a) to
conclude, " . . . the direction of causality is from the
child-rearing pattern obtained, to the child's moral
development."

More recently, Bandura (1969) described the social
learning of moral judgments as proceeding from a variety
of modeled responses and transmitted on the basis of a
modeling process. This modeling is such a potent factor
that Bandura found that children would imitate models who
were acting in ways which were contrary to the children's
moral beliefs. This confirmed Aronfreed's (1964) belief

that the child's responses to social punishment reflect

lU

fl)

1"

.(Z‘)

explicit standards or values which have been adopted from
a model. Dworkin (1968) also found that modeling was
important for the learning of moral behavior. However, he
argues that in order to promote long-lasting changes,
rather than the overt adoption of an imitative response
(parroting), training must focus on cognitive mediating
responses. This might be accomplished by communicating
the salient cognitive dimensions of the moral concept
being taught. Dworkin believes that socioeconomic status
may be relevant for such learning.

Support for class differences in moral response to
transgression was obtained in research by Aronfreed (1961).
This author found that middle-class children were less
dependent upon external reinforcement than working-class
children. A sex difference existed within these data. In
both socioeconomic classes boys relied less upon an external
definition of moral consequences than the combined classes
of girls. These differences were not attributable to the
variable of intelligence as Aronfreed measured it. Such
relationships between moral responses and social positions
indicated to Aronfreed (1961) that:

. . . different moral orientations do not emerge
sequentially with advancing age or experience, as has
been argued in some interpretations of moral develop-
ment, but are rather the stable end-results of
different patterns of social reinforcement (p. 239).

The relation of social class to differences in

moral behavior has been obtained in many studies. An

1111'

OF

.
t e

o.‘

of

hil

rel

19'

investigation by Hess and Shipman (1965) provided evidence
that socioeconomic status differences are related to the
early experiences and the socialization of "cognitive
styles" of children. Using the Siegel Sorting Task, these
authors found a significant difference between upper- and
lower-class children in the descriptive part-whole category.
This difference was influenced by the disciplinary method
of the mother. The lower-class child (compared to the
upper-class child) was characterized as one who relates

to authority rather than rationale; who although often
compliant is not reflective in his behavior; and for whom
the consequences of an act are largely considered in terms
of immediate punishment and reward, rather than in terms
of future effects and long-range goals.

Although psychologists have focused on the pro-
hibitive and punitive components of the parent-child
relationship in moral behavior, recent research (McKinney,
1971; Olejnik, 1971) has examined both positive and negative
values. Instead of the psychoanalytic or social learning
theory conceptualizations, McKinney (1971) postulates a
four-fold model for the development of behavioral values.
The two dimensions on which this model is based are the
reinforcement dimension (reward or punishment), and the
behavioral orientation dimension (right-doing or wrong-

doing) [see Figure 1].

 

 

Reinforcement
Punishment Reward
Proscriptive Doing wrong Avoiding wrong
Behavioral
Orientation

Prescriptive Neglecting to Doing good
do good

 

 

 

Figure 1. Two dimensions of value development: rein-
forcement (punishment and reward) and behavioral
orientation (wrong-doing or right-doing).

This assumption of a four-fold process prompted
McKinney to examine his hypothesis empirically. In this
research, McKinney was not only trying to determine the
existence of these dimensions, he was also attempting to
relate these dimensions in a systematic way to the
individual's perception of his parents' reward and
punishment behaviors. The results of this research
enabled McKinney to conclude:

l. . . . subjects consider the positive effects of
their behavior to be the result of the right
things they do, while they consider the negative
effects to be the result both of the wrong things
they do and the right things they fail to do. In
psychoanalytic terms, it would appear that the two
aspects of the super-ego, i.e., the conscience and
the ego-ideal, are developed in very different
ways.

2. The data confirmed the existence of a dimension of
value development concerned with a prescriptive vs.
prescriptive orientation.

3. The results supported the hypothesis that " . . .
a prescriptive orientation is related to one's
perception of his parents as being more rewarding,
while a proscriptive orientation is related to
one's perception of his parents as being more
punitive."

 

More recently, Olejnik (1971) extended McKinney's
(1971) findings to altruistic behavior in four-year olds.
Olejnik observed that the child's altruism (giving candy
to a hypothetical "needy" child) was related to his
parents' value orientation. That is, children of pre-
scriptive parents were significantly more altruistic (gave
more candy) than were children of proscriptive parents.
Olejnik's data not only show that the orientation of
parents can have a profound influence on their children,
but also that this influence can be effective at a very
early age. These studies (McKinney, 1971; Olejnik, 1971)
provide further support for Hoffman's (1970a) contention
that parental discipline plays a very important role in
the moral development of the child.

Additional relevant data concerning the child's
moral behavior, particularly his reactions to transgression,
appear in the research of Aronfreed. In his 1961 study,
Aronfreed found that children's internalized responses to
transgression assume a great variety of forms. Responses
were attributed to different patterns of social rein-
forcement, since they were predictably related to the
socioeconomic status and the sex roles of the children,
and to a lesser degree to maternal disciplinary practices.
However, recent unpublished research by Hoffman (in Hoffman,
1970a) seems to justify criticism of Aronfreed's discipline

Classification.

I").

(§

rf

‘3:

r0“

(T

f?

Aronfreed recognized the importance of internal and
external orientation, and maternal discipline in moral
behavior. From a questionnaire interview, Aronfreed
determined the predominant disciplinary techniques of the
mothers. These techniques were labeled either "induction,"
or "sensitization." Hoffman classified discipline as
power assertion, love withdrawal, or induction. Thus,
there may actually be Egg varieties of Aronfreed's
"induction" (love withdrawal and induction). Hoffman also
discovered that these parental discipline types were
differentially related to moral orientation.

By examining the moral standards or principles
which children gave in response to several moral judgment
items, Hoffman was able to identify two internalized but
different value orientations. Children who " . . .
emphasized the consequences of behavior for others or
stressed certain interpersonal moral values such as mutual
trust were coded as humanistic." Subjects who responded to
the items in terms of the " . . . violation of an insti-

"2 A second

tutional norm were coded as conventional.
criterion which was used to evaluate the subjects' responses
was the flexibility with which standards were applied.
Categorization was again based on the subjects' responses

to moral judgment items.

2See Hoffman (1970a, pp. 336-343) for more detailed
discussion of these orientations.

Responses which justified the acts because of extenu—
ating circumstances were coded as flexible; responses
which criticized them for being against religious
precepts or against the law without regard to the
extenuating circumstances or simply stated that they
should never be done were coded as rigid (p. 336).

On the basis of these criteria, Hoffman was able
to identify two internalized groups: the conventional-
rigid and the humanistic-flexible. Hoffman also selected
a third group with an externally focused moral orientation.
This group was lower on IQ, but by excluding some subjects
a very close match between the three groups was achieved.3

Hoffman then investigated the relationship of these
moral orientations to certain personality characteristics
and parental discipline.4 As Hoffman summarized:

. . . the humanistic-flexible and conventional-rigid
groups appear to be two variants of an internalized
conscience which differ not only in the manifest
content but also the hierarchial arrangement and
motivational basis of their moral standards. Thus,

in making moral judgments about other people's
violations, the humanistic-flexible subjects tend to
stress the consequences for others and are more likely
to take extenuating circumstances into account (p. 339).

Such people are high on guilt and other indices of inter-

nalization, but they experience guilt " . . . primarily as
a consequence of their behavior for others." The
conventional-rigid subjects, however, are " . . . more

likely to give a religious or legal basis for their moral

 

3This is consistent with other data which show a
positive relation between IQ and internal orientation.
Also see Hoffman (1970a).

4See Hoffman (1970a, pp. 336-343) for further
discussion.

10

judgments and to ignore extenuating circumstances." 'Their
guilt stems more from the awareness of "unacceptable
impulses" than from the amount of harm done others. As
Hoffman concludes: " . . . although the two groups were
selected on the basis of conscious moral judgments, they
appear to reflect moral syndromes which differ beyond the
conscious and rational aspects of moral orientations"

(P. 339).

Regarding parental discipline, Hoffman observed
that both internal groups obtained higher scores on
induction and affection, and lower scores on power
assertion, than the external group. However, some
remarkable differences between the two internal groups
emerged. Parents of the conventional-rigid group reported
using love withdrawal (or some variant) as the predominant
discipline technique. Parents in the humanistic-flexible
group used a more varied and discriminating disciplinary
method. This method ranged from power assertion to
permissiveness, depending on the situation, and utilized
more induction and suggestions of reparation (where
possible).5 From the McKinney (1971) and Olejnik (1971)
studies previously mentioned, one might expect that the
moral standards of the children would be similar to those

of their parents. Hoffman's research supports this

 

5Similar results were recently obtained by
Baumrind (1971).

11

expectation, since parents of humanistic-flexible children
fit the humanistic-flexible pattern. Similarly,
conventional-rigid children had conventional-rigid
parents. Whereas the humanistic-flexible group tends to
look at behavior in its larger context, the conventional-
rigid parents are " . . . more likely to ignore contextual
factors, to compare standards of good or bad behavior, and
act accordingly."

The problem with Aronfreed's (1961) classification
then is that it is based on structure, rather than
function. Aronfreed classified maternal discipline
techniques as either "sensitization" if the mothers used
physical punishment (e.g., power assertion), or as
"induction" if the mothers used other kinds of punishment
(e.g., love withdrawal). However, it is clear from
Hoffman's data that a structural classification is not as
powerful a discriminator as a functional classification.
Only with a classification similar to Hoffman's (which

emphasized the way a disciplinary method functions for the

 

parent and child) can one make the important discrimination

between humanistic-flexible and conventional-rigid values.
If Aronfreed (1961) had classified maternal

discipline techniques in this way, it is quite possible

that he would have obtained different relationships, or

at least interpreted his data differently. Although

Aronfreed has shown that moral responses to transgression

12

can be examined empirically, his classification weakens
his conclusions.

Thus, the present study was designed to analyze
the relationship of control over punishment, parental
discipline, socioeconomic status, and moral orientation
(parent and child) to intensity of punishment and self-
critical and reparative responses in children. Modifi-
cations of Aronfreed's (1963) design were made so that the

impact of experimental socialization could be assessed

more accurately.

METHOD

Aronfreed (1961, 1963) investigated the relation-
ship of self-criticism and reparation to socioeconomic
status, sex, maternal discipline, control over punishment,
and cognitive structure of instructions. The present
study focused on the relationship of specific subject
variables and experimental paradigms to severity of
punishment and self-critical and reparative responses in
children.

The subjects came from middle- and working-class
families. Subjects from each class were classified as
either internal or external moral orientation. The
internal orientation subjects were further classified as
either humanistic-flexible, or conventional-rigid (see
Hoffman, 1970b). The children were then randomly assigned
to one of four groups: the baseline (or control) group,
or one of three experimental groups.

In the baseline group, treatment involved a
modified version of Aronfreed's (1963a) technique (see

Appendix A). The subjects in this group punished

13

14

themselves. In the experimental groups, control over
punishment changed. In Phase I, the experimenter punished
the subjects in the experimental groups. The level of
punishment was different for each group. In Group I the
subjects received a punishment intensity greater than the
average level of punishment used by baseline group subjects.

Group II subjects received punishment equivalent to the

 

baseline average. And Group III subjects received a
punishment less than the baseline level. In Phase II,
the same subjects in each of the experimental groups
punished themselves.6 This design allowed for comparisons
within and between groups on the variables mentioned
earlier.

In 1961, Aronfreed found that middle-class children
(boys and girls) gave more self-critical responses than
working-class children. Later, Aronfreed (1963a) dis-
covered that self-critical responses are more likely to
appear when the socialization paradigm provides explicit
standards of evaluation than when the cognitive structure
of the instructions is minimal. He also observed that
self-critical responses were not related to the degree
of control over punishment given the child. In the present
investigation, self-critical responses were maximized by

providing the subject with only the high cognitive structure

 

6The baseline group subjects punished themselves
during both phases of the experiment.

15

instructions (see Aronfreed 1963a). In view of Aronfreed's
(1961, 1963a, 1963b) data on socioeconomic status, and
Hoffman's (1970) research on moral orientation, it was
expected that for all groups throughout the experiment,

self-critical responses would be:

1. greater for middle-class than working-class

subjects;

2. greater for internal orientation than external

orientation subjects;

3. greater for middle-class internal than working-

class internal orientation subjects;

4. greater for middle-class external than working-

class external orientation subjects.

Regarding reparative responses, Aronfreed (1963a)
found that the frequency of these responses was positively
correlated with control over punishment, rather than with
the structure of the instructions. In an earlier investi-
gation, Aronfreed (1961) observed that reparative responses
were not directly related to the social class or the sex of
the subject. He did find, however, that reparation occurs
more often in subjects from "induction" than from "sensi-
tization" disciplinary backgrounds. Hoffman (1970b) also
reports a relationship between discipline and reparative
responses. These data indicated that there might be a

complex relationship between social class, discipline, and

16

reparative responses. It was possible, for example, that
working-class subjects would give as many reparative
responses as middle-class subjects. Reparative responses,
unlike self-critical responses, are usually expressed
overtly. Therefore, the working-class subjects might use
reparation as a reaction to fear of punishment. The middle-
class children, however, might use reparation for other
reasons (e.g., high degree of guilt). It was hoped that
analyses of the interaction between reparation, level of
punishment, and the subject variables might clarify the
ambiguities in this relationship.

Although reparation might not be related to social
class, it might be related to moral orientation. That is,
it was postulated that internal orientation subjects of
both socioeconomic classes would give more reparative
responses than the external orientation subjects of both
classes. Reparative responses would be lower when the
experimenter punished the subjects (low control), than when
they punished themselves (high control) (Aronfreed, 1963a).

Confirmation of these hypothesized relationships
would also support Aronfreed's contention that self-
criticism and reparation cannot be thought of as alternative
or equivalent responses to transgression deriving from a
single pattern of socialization.

The amount of punishment used by the subjects

should be negatively related to socioeconomic status and

17

moral orientation. One reason for this relationship may
be that middle-class and internal subjects have higher
guilt scores (see Hoffman, 1970b) and are more internally
controlled. It was expected that they would "need" less
punishment to control their behavior than working-class
and external subjects. Hoffman (1970b) reports that
humanistic-flexible parents do not feel it necessary to
punish a child for accidental damage. Conventional-rigid
parents, however, believe one should punish accidental

damage. Therefore, it was hypothesized that:

l. middle-class children would punish themselves less

than working-class children punish themselves;

2. internal orientation subjects would punish

themselves less than external orientation subjects;

3. middle-class internal orientation subjects would
punish themselves less than working-class internal

orientation subjects punish themselves;

4. middle-class external orientation subjects would
punish themselves less than working-class external

orientation subjects punish themselves.

These results would also show a strong negative relation-
ship between self—criticism and intensity of punishment.
Subjects who do not "need" punishment (humanistic-flexible,
and to a lesser degree, conventional-rigid), would use

more self-critical responses. Subjects who employ a high

18

level of punishment would have a lower number of self—
critical responses. That is, as the number of self-
critical responses increased, the level of punishment
would decrease.

It was expected that the basic relationships
described above would exist within all four groups
throughout the experiment (i.e., during both Phases I and
II). If one were to compare the data between groups, the
hypotheses become more complex. But basically the
relationships should operate on the principle (Aronfreed,
1963b) that:

The child in adopting a socializing agent's critical
evaluative responses reduces the anxiety attached to
transgression by reproducing punitive stimulus aspects
of the agent which originally came to serve as cues
signifying the termination of the anxiety that
accompanied their anticipation (p. 292).
Thus, the subjects in the experimental groups should respond
differently to the model (the experimenter in Phase I).
These responses to different punishment levels Should be
consistent with the subjects' previous socializing experi-
ences. That is, middle-class and internal subjects should
respond to increased level of punishment by increasing
self-critical responses, while working-class and external
subjects should respond by increasing the punishment level.
In Group III, the subjects received lower-than-baseline
punishment. The subjects in this group should "compensate"

for this lower level by responding again in a way more

consistent with their "pre-experimental disposition"

l9

(Aronfreed, 1963b). The middle-class and internal subjects
should increase self-critical responses; the working-class
and external subjects should increase punishment level.
Thus, the three intensities of experimenter punishment
would have differential effects on the dependent variables
(self-criticism, reparation, and self-punishment) as a
function of the social class and moral orientation of the
child. It was hoped that by analyzing this data, it would
be possible to determine the importance of "pre-experimental
dispositions” (naturalistic socialization factors) in the
experimental socialization setting. These findings may
provide valuable (meaningful) relationships between
socialization effects and moral responses to transgression,
and perhaps more generally, for the entire class of

responses in moral behavior.

Subjects

The subjects were sampled from a population of
232 second (110) and third (122) grade boys from the four
public elementary schools in the Okemos school system.
Only 21 of these boys were not allowed (by their parents)
to participate in the study. The remaining subjects were
then classified by social class, moral orientation, and
age.

The subject's social class was determined by a

combination of the father's (head of the household)

20

occupation and educational background according to
Hollingshead (1957).

The subjects were interviewed individually to
discover their moral orientation (humanistic-flexible,
conventional-rigid, or external). Four moral judgment

items adapted from Hoffman (1970b) were used in these

sessions (see Appendix A). Coding of these responses was

based on the procedure used by Hoffman (1970b; personal

communication):

- Two criteria were used in coding the responses. One
was the moral standards or principles given in support
of the judgments. Thus item 1 tapped the extent to
which the subject believes that human needs are the
underlying basis for complying with social norms:
responses which stressed the harm done others by the
stealing were coded as humanistic; responses phrased
in terms of the violation of an institutional norm,
such as the law or the ten commandments, were coded as
conventional. Responses to item 2 which were coded as
conventional stated that breaking into the store was
worse because stealing is against the law or may
involve property destruction, or that the two acts are
equally bad because they both violate religious
precepts or involve stealing the same amount of money.
Responses coded as humanistic generally stated that
the "borrower" was worse because of the violation of a
trust and personal deception involved, the possible
lessening of the benefactor's faith in people, or the
ultimate loss to others who really need the money.
Responses in which the stealing was judged worse
because the loss would be greater to the storekeeper,
who was often perceived as less affluent than the
benefactor, were also coded as humanistic.

The second coding criterion was the flexibility with
which standards were applied, particularly the extent
to which extenuating circumstances were taken into
account. Items 3 and 4 were used to tap flexibility.
Responses which justified the acts because of the
extenuating circumstances were coded as flexible;
responses which criticized them for being against
religious precepts or against the law without regard
to the extenuating circumstances, or simply stated

21

that they should never be done, were coded as rigid.

In coding, the reasons given were more important than
the direction of the choices, and perfunctory responses
were not coded at all (pp. 95-96).

As a result of this coding, a sample of 182 subjects
remained; the others being discarded because of response
ambiguity. Twenty-two of the 182 subjects were then
dropped because of conflict between their responses to
items 1 and 2, and their responses to items 3 and 4 (i.e.,
humanistic responses to 1 and/or 2; rigid responses to
3 and/or 4). Additional subjects (23 classified as social
class II, 3 from class IV, and 2 from class V) were omitted
in order to increase the difference between the high and
low social class groups. This procedure permitted a
sample size of 132 subjects. However, two of these
subjects moved during the course of the experiment; and
two more were randomly dropped in order to equalize the
number of subjects in each cell (eight).

Thus, the final sample consisted of 128 boys
averaging 8.4 years of age (range, 7.2-9.8; standard
deviation, .67). This age was chosen since it appears to
be part of a "transition period" in child development
(e.g., from egocentric to non-egocentric, or from moral
realism to moral relativism--Piaget, 1960; Gutkin, 1970;
Einhorn, 1971; and others). Half of these subjects were
classified as high social class (mean Hollingshead value,

11.56; standard deviation, 1.39), and half as low social

class (mean, 38.42; standard deviation, 10.18). Further

22

categorization by moral orientation provided for equal

representation of the flexible and rigid internal orien-
tation subjects in each social class. Unfortunately, no
clear examples of the external orientation were observed

in the interview sessions.

Procedure and Design
The basic procedure used by Aronfreed (1963a,
pp. 440-446) was followed in this study with these

modifications:

1. Aronfreed (1963a) conjectured that the terms
"careless" and "rough" might have created a greater
perceived intensity of punishment, and thereby affected
the frequency of self-critical and reparative responses.
However, Aronfreed controlled the intensity of punishment
so that the experimenter or the subject could only remove
one, two, or three Tootsie Rolls (TR's). Therefore, the
relationship of "careless" and "rough" to perceived
intensity of punishment is confounded. It is also possible
that similar intensities of punishment are perceived
differently. That is, a loss of two TR's might be

extremely_punishing for one subject, but only mildly

 

punishing for another. The latter subject might utter
more self-critical responses to compensate for his lower
(perceived) intensity of punishment. Or, he might not

respond at all.

23

Therefore, in the present study, Aronfreed's
statement telling the child that he might lose one, two,
or three TR's, either by experimenter punishment, or by

self-punishment, was omitted.

2. Once the TR's were shown to the subject at the
beginning of the experiment, they were removed from his
sight. The TR's were put in a box with an opening in it so
the subject might remove the TR's to punish himself. The
subject was just aware that there were many TR's in the
box, but was not allowed to count them. In this way, the
subject did not know how many TR's he had at any given
time in the experiment; and the removal of TR's retained
its "punishing" character. It is possible under Aronfreed's
method that the subject began to see his pile of TR's
diminish, and subsequently took fewer TR's away from
himself (since he could keep the remaining TR's). This
decrement in punishment number might affect the perceived
severity of the punishment and the number of self-critical
responses (see 1 above).

After the subjects were classified according to
social class and moral orientation, they were randomly
assigned to one of four groups. In the baseline condition,
the subjects were tested as were the self-punishing subjects
in Aronfreed's (1963) study, with the modifications
mentioned above. The number of TR's each subject took

from himself, the number of self-critical responses he

24

made over the ten trials, and the presence or absence of
self-critical and reparative responses on trial eleven
(the test trial) were recorded for each subject.

In the experimental groups, the basic design was
the same. However, these groups were tested on the
experimenter punishment condition in Phase I. There were

three levels of experimenter punishment:

1. In Group I, the experimenter punished the subjects
at a level greater than the highest average

punishment intensity of the baseline subjects.

2. In Group II, the experimenter punished the subjects
at a level corresponding to the average punishment

intensity of the baseline group.

3. In Group III, the experimenter punished the
subjects at a level lower than the lowest average

punishment intensity of the baseline group.

The number of self-critical responses and the presence or
absence of self-critical and reparative responses ontrial
eleven were recorded for each subject.

Following the ten trials and test trial in Phase I,
the subjects (baseline and experimental) punished themselves
(Phase II) for another set of ten trials of the task.

Again, the number of TR's removed, the number of self-
critical responses, and the presence or absence of self-
critical and reparative responses on trial eleven were

recorded (see Appendix B for the design).

25

The subjects were equally divided by social class
and moral orientation, and randomly assigned to one of the
four conditions and to one of two male experimenters.7
The experimenters were not aware of the subject's social
position or moral orientation.

Each subject was individually taken by the experi-
menter from his classroom to the experimental room, where
he was asked to sit in front of one end of a rectangular
piece of plywood resting upon a small table. The board
was roughly 2x3.5 feet in size. Twenty-four small toy
soldiers were thickly clustered in a triangular formation
at the other end of the board. Behind the soldiers, at
the edge of the board, stood a rubber and wood doll about
four inches in height. Two dolls were used in this study--
a doctor doll and a nurse doll--in order to provide the
subjects with a different doll from Phase I to Phase II.
This was done to increase the credibility of the doll's
breaking "by accident" on trial eleven. The order of the
doll presentation was controlled for each experimenter and

across all subjects.

 

7Adult males were the models (experimenters) here
since Bandura and Kupers (1964) found that adults were more
often imitated than peers. They also observed that subjects
used the same self-critical responses as the model.
Aronfreed (1963) observed a very low frequency of identical
verbal imitation. Also, Bandura and MacDonald (1963) found
that social reinforcement could change children's moral
judgments.

26

A large padded cardboard box rested on the floor
just beyond the far edge of the board and behind the doll.
The box was below the height of the board's surface, so
that the interior was not visible to the subject when he
was seated. The experimenter sat to the right side of
the board toward the far end. Next to the experimenter's
left hand was a box containing small Tootsie Rolls. This
box was hidden from the subject by a small, vertical
wooden barrier. The subject could reach around this
barrier for the TR's, but he was not permitted to look
around it at the box. On the subject's left was a small
box in which the subject (and experimenter) placed the TR's
lost.

After the experimenter and the subject were seated,
the experimenter used one of two procedures. In one
procedure, the experimenter evaluated and punished the
subject; in the other, the subject evaluated and punished
himself. In either procedure, the experimenter read the
following instructions:

I have something here for you to do. The Army uses
it to pick people for a special kind of work. To do
this work, you have to be very careful and gentle.
Being careful and gentle is the most important thing.

Here's the way we do it. Back here there is a
nurse (doctor) of a special type (point to doll). She
(he) is important in case anyone gets sick, and all the
soldiers love her (him) because she (he) is so good to
everyone. Whenever there's any danger she (he) has to
leave the field by going back into this box (lift up
box). Here's how she (he) goes into the box. You push
her (him) off the board with this pusher; you must keep

it down on the board when you push (demonstrate--give
to the subject).

27

Now the soldiers guard the nurse (doctor) very
carefully. They stand all around her (him) like this
(point to soldiers). Of course, to push her (him) off,
you will have to push through the soldiers, and so you
will knock some of them over, even though you're trying
to be careful and gentle. The idea is to see how many
of the soldiers you will knock down. If you knock
down just a few, that's good. If you knock down a
lot of them, that's not so good. When you use the
pusher try to be as careful and gentle as you can.

The experimenter then followed the baseline or

experimental condition procedure.

Baseline Condition
Phase I

The experimenter showed the subject the TR's in
the box. He then read:

”When we're all through, you may keep however many
you have left. Each time after you push off the nurse
(doctor), you look at the number of soldiers you
knocked down, and you decide how careless and rough
you've been. Then you decide how many Tootsie Rolls
you should take from your pile. You take the number of
Tootsie Rolls you think is right, and put them in the
box here (point to small cigar box). Those are the
Tootsie Rolls you lose. The amount you take depends on
how careless and rough you think you've been."

The experimenter then began trial one:

"Remember to keep the pusher down on the board.
Push off the nurse (doctor), and knock down as few
soldiers as you can. And be as careful and gentle as
you can. All right, go ahead."

After the doll was pushed over, the experimenter
said:

"All right. You knocked down some soldiers, so
you decide how careless and rough you've been. Take as
many Tootsie Rolls as you think is right, and put them
into the box."

The experimenter then reset the doll and soldiers,
recorded the subject's responses and the number of

28

TR's removed. When the subject had finished putting
the TR's into the box, the experimenter said casually:

"Good."
Then the experimenter began the next trial(s):
"Okay, go ahead."

The eleventh trial of all conditions was planned as
an "unexpected" disruption due to the breaking of the
doll. Both dolls were arranged so that the experimenter
could easily and unobtrusively "break" them while he
picked the doll up from the box. Thus, after the
subject pushed the doll off the board on trial eleven,
the experimenter held up the "broken" doll and (looking
at the doll, not the child) said:

"Auh oh--it's broken. And we don't have another
nurse (doctor) here to use for this."

The experimenter then indirectly asked:
(1) "I wonder why it broke."

If the subject gave any response to this stimulus
that was relevant to the cause of the doll's breaking,
whether or not it was self-critical, the experimenter
went on to the third stimulus (given below). If the
response was not clearly relevant to the doll's
breaking or if the subject gave no response, the
experimenter said to the child:

(2) "Why do you think it broke?"

Then regardless of the responses to the first two
stimuli, stimulus three was:

(3) "Well, now that it's broken, I wonder what we
should do."

This stimulus was intended to indirectly elicit
reparative responses. If the subject offered any
response that was relevant to the implied question,
whether or not it was reparative, the experimenter
terminated the eleventh trial procedure. The experi-
menter then read the Phase II instructions, or
terminated the experiment if Phase II was completed
(see below). If the subject gave no response, or one
that was not relevant to the question, the experimenter
then presented the final, stronger stimulus four:

29

(4) "What do you think we should do now?"

Regardless of the subject's response to the last
question, the eleventh trial procedure was terminated.
The experimenter than read the Phase II instructions,
or ended the experiment if Phase II was completed (see
below). .

After the Baseline condition subjects had been run,
the number of TR's to be removed by the experimenter
in Groups I, II, and III was determined. The average
number of TR's removed by the Baseline subjects was
5 TR's, and the range was from 1.5 to 1236 TR's over
the ten trials for both Phases I and II. Thus, the
punishment intensity was 13 for Group I, 5 for Group II,
and 1 for Group III. The experimenter punished the
subjects over the ten trials according to the following
scheme:

Group I: 13, 12, 14, 12, l4, 13, 14, 13, 12, 13

Group II: 5, 4, 6, 4, 6, 5, 6, 5, 4, 5

Group III: 1, 0, 2, 0, 2, l, 2, l, 0, l
The three experimental conditions were employed by the
experimenters in random order to control for sequence
effects and experimenter bias (e.g., fatigue, or
practice).

The following procedure for the three experimental
conditions was very similar to the Baseline condition
instructions. The experimenter showed the subject the
TR's in the box. He then read the following:

"When we're through you may keep however many you
have left. Each time after you push off the nurse
(doctor), I look at the number of soldiers you knocked
down, and I decide how careless and rough you've been.
Then I decide how many Tootsie Rolls I should take from
your pile. I take the number of Tootsie Rolls I think
is right, and put them in the box here. Those are the
Tootsie Rolls you lose. The amount I take depends on
how careless and rough I think you've been."

The experimenter then began trial one:

"Remember to keep the pusher down on the board.
Push off the nurse (doctor), and knock down as few

 

8Since the mean punishment intensity was 5 TR's,
Aronfreed's (1961) procedure of punishing at levels of l,
2, or 3 TR's can be questioned. Thus, it is quite possible
that his "punishment" confounded the moral response a
behavior he observed.

30

soldiers as you can. And be as careful and gentle as
you can. All right, go ahead."

After the doll was pushed over, the experimenter
said:

"All right. You knocked down some soldiers, so I
decide how careless and rough you've been. I'll have
to take--let's see--(one, two, . . .) Tootsie Rolls."

The experimenter then reset the doll and soldiers,
and recorded the subject's responses. The experimenter
began the next trial(s):

"Okay, go ahead."

The eleventh trial was similar to the procedure
previously described (see above).

For Phase II, the experimental as well as the
baseline condition subjects were instructed as follows.

Phase II

The experimenter told the subject to leave the room
for a minute, put the doll away, and brought out the
other doll and reset the soldiers. Then the subject
was called in. The experimenter told him:

"Well, we have a different doll now."

After this statement the experimenter repeated the
instructions from Baseline condition, Phase I (see
above).

In terminating the experiment (for all subjects),
the experimenter showed the subject that he had fixed
both dolls. He asked the subject not to talk with the
other children (or anyone else) about the experiment.
The experimenter looked into the box briefly, without
the subject's seeing, and gave the subject six Tootsie
Rolls in a plastic bag. The procedure ended with the
experimenter thanking the subject for participating in
the experiment.

RESULTS

The procedure for quantifying the data from the
task varied as a function of the responSe measure in
consideration. The self-critical and reparative responses
were identified by criteria identical to those used by
Aronfreed (1963a): "A response was classified as self-
critical if the child, in accounting for the doll's
breaking, referred to his behavior in pushing it--for
example, any response indicating that he had not pushed
the nurse (or doctor) 'the right way,‘ had pushed it too
hard . . ." This criterion was used to identify self-
critical responses during the ten trials (before the doll
broke) as well (Aronfreed did not include this in his
research). The self-critical responses before the doll
broke were typically very similar to those which occurred
after the doll broke. This made identification of self-
critical responses straightforward.

Responses were classified as reparative when they
indicated the child's perception that the effects of

transgression could be corrected or ameliorated through
his own resources for constructive action. These

31

32

responses invariably took the form of suggestions for

repairing the doll or continuing the procedure, even

without the doll, in some alternative way (p. 441).
Of course, no reparative responses were made during the
ten trials because the doll did not "break" until trial
eleven.

The self-critical and reparative responses could
occur to either stimulus l or 2, or stimulus 3 or 4
respectively, on trial eleven. Thus, a coding procedure
was devised to indicate the presence or absence of these
responses to the various stimuli. If the subject responded
to the indirect eliciting stimulus (item 1 for self-
critical, item 3 for reparative responses) with an
appropriate response, he was given a score of 3. An
appropriate response to the second, more direct stimulus
(item 2 for self-critical, item 4 for reparative responses)
was given a score of 2. An irrelevant response, or no
response at all, was assigned a score of 1.

The Tootsie Rolls removed were recorded by the
experimenter for each of the ten trials, and an average
was computed. Thus, the response measures recorded for
each subject were: the number of self-critical responses
over the ten trials, the self-critical responses on trial
eleven, the reparative responses on trial eleven, and the
mean number of Tootsie Rolls removed over the ten trials.

Tables 1-4 represent the means and standard

deviations of the response measures for each cell (n:8) in

33

TABLE 1

Means and Standard Deviations of the Number of
Self-Critical Responses for Each Punishment
Treatment by Moral Orientation by Social
Class Cell for Phase I and Phase II

 

 

 

 

High Low
Flexible Rigid Flexible Rigid
Phase I i
Baseline i 2.00 2.13 1.38 2.34 1.96
2.69 2.34 1.65 3.20
Group I i .38 .75 2.13 2.75 1.50
.48 .83 2.62 2.38
Group II 2 1.13 1.38 1.63 2.50 1.66
2.32 1.58 3.24 2.65
Group III 2 1.38 1.50 2.75 1.00 1.66
1.58 1.41 2.63 1.50
i 1.22 1.44 1.47 2.15
Phase II
Baseline i 1.00 .63 .63 2.13 1.10
1.50 1.11 .99 3.06
Group I i .25 .88 .88 2.25 1.07
.66 1.69 1.05 1.64
Group II 2 1.00 .50 1.38 .63 .88
2.65 1.32 2.91 .86
Group III 2 1.00 1.00 2.75 1.00 1.44
1.22 1.50 3.11 1.50
i .81 .75 1.41 1.50

 

34

TABLE 2

Means and Standard Deviations of Self-Critical
Responses on Trial 11 for Each Punishment
Treatment by Moral Orientation by Social

Class Cell for Phase I and Phase II

 

 

 

 

High Low

Flexible Rigid Flexible Rigid

Phase I i

Baseline i 1.38 1.13 1.38 1.38 1.32
.48 .33 .70 .48

Group I i 1.00 1.38 1.63 1.38 1.35
.00 .48 .48 .48

Group 11 i 1.25 1.25 1.25 1.25 1.25
.43 .43 .43 .43

Group 111 i 1.38 1.38 1.13 1.25 1.29
.48 .48 .33 .43
i 1.25 1.29 1.35 1.32

Phase II

Baseline i 1.38 1.38 1.25 1.50 1.38
.48 .70 .43 .87

Group I i 1.25 1.25 1.63 1.38 1.38
.43 .43 .70 .70

Group 11 i 1.50 1.25 1.25 1.38 1.35
.71 .66 .43 .48

Group III R 1.63 1.38 1.25 1.25 1.38
.86 .70 .43 .43
i 1.44 1.32 1.45 1.38

 

35

TABLE 3

Means and Standard Deviations of Reparative
Responses on Trial 11 for Each Punishment
Treatment by Moral Orientation by Social

Class Cell for Phase I and Phase II

 

 

 

 

High Low
Flexible Rigid Flexible Rigid
Phase I 2
Baseline i 2.13 2.00 2.00 1.63 1.94
.78 .71 .50 .70
Group I i 2.13 1.63 1.75 1.63 1.79
.78 .48 .66 .48
Group II R 2.00 1.88 1.75 1.88 1.88
.00 .33 .66 .60
Group III 2 2.25 2.13 2.25 1.75 2.10
.66 .60 .66 .66
i 2.13 1.91 1.94 1.72
Phase II
Baseline i 2.38 1.88 1.75 2.00 2.00
.70 .60 .83 .87
Group I i 2.13 1.63 1.75 2.00 1.88
.78 .48 .66 .71
Group II 2 2.75 2.13 1.88 2.00 2.19
.43 .60 .60 .71
Group III E 1.88 2.00 2.25 1.50 1.91
.60 .87 .66 .50
i 2.29 1.91 1.91 1.88

 

36

TABLE 4

Means and Standard Deviations of the Number of
Tootsie Rolls Removed for Each Punishment
Treatment by Moral Orientation by Social

Class Cell for Phase I and Phase II

 

High Low

 

 

Flexible Rigid Flexible Rigid

 

Phase I

Baseline i 5.09 4.96 5.80 4.90
.95 3.37 2.95 2.17

Group I 2 13.00 13.00 13.00 13.00

Group II 2 5.00 5.00 5.00 5.00

Group III R 1.00 1.00 1.00 1.00

Note: The baseline subjects punished themselves in Phase I,
while the experimenter punished the subjects in the
experimental groups as indicated.

Phase II 2

Baseline i 5.03 3.81 4.26 4.25 4.34
.93 2.52 2.01 1.41

Group I 2 12.28 13.68 11.18 14.21 12.84
2.62 2.13 2.41 1.34

Group II 2 4.96 4.61 4.19 5.24 4.75
.74 .47 .95 1.10

Group III 2 1.78 1.46 1.86 1.79 1.72
.49 .37 1.88 1.42
i 6.01 5.89 5.37 6.37

Note: In Phase II, all subjects punished themselves.

 

37

Phase I and Phase II. It should be remembered that the
same subjects were used in Phases I and II.

The data represented by these tables were analyzed
by seven 4 by 2 by 2 (punishment treatment by moral
orientation by social class) analyses of variance. The
summary tables of these analyses are represented in
Tables C-l to C-4 (see Appendix C). An analysis was not
performed on the Phase I Tootsie-Rolls-removed data (see
Table 4), since the experimenter punished the subjects in
the experimental groups during this phase.

The analyses of the data for three of the four
response measures failed to yield any significant main or
interaction effects for either the Phase I or Phase II
data (see Tables C-l to C-3).

However, the analysis of the number of Tootsie
Rolls removed in Phase II provided significant main and
interaction effects (see Table C-4). Punishment treatment
provided a strong main effect (f = 252.31, df = 3/112,

p < .001). Inspection of the data revealed a relationship
between the intensity of punishment delivered by the
experimenter to the subject and the intensity of punishment
the subject provided for himself. That is, in terms of

the numbers of TR's removed in each punishment treatment
from Phase I to Phase II: Baseline, 166.0 to 138.8;

Group I, 416.0 to 410.7; Group II, 160.0 to 152.0; and

Group III, 32.0 to 55.1. In Phase II then, the subjects

38

in the three experimental groups clearly imitated the
experimenter punishment they had received in Phase I.
Only the Group III subjects punished themselves more than
the experimenter had.

The interaction between punishment treatment and
moral orientation was also found to be significant
(F = 4.26, df = 3/112, p < .01). Examination of the data
indicated the main source of variance in the interaction
to be the difference between the flexible and rigid moral
orientation subjects at Group I. The rigid subjects
(223.1) punished themselves more than the flexible subjects
(187.6) did in this punishment treatment. None of the

other main or interaction effects was significant for the

data.

DISCUSSION

The present research attempted to investigate the
relationship of social class, moral orientation, and
severity of punishment to children's moral responses to
transgression. The hypotheses concerning this relationship
were not confirmed statistically. However, consideration
of the significant main effect of punishment treatment,
and the interaction between moral orientation and punish-
ment treatment is important to the interpretation of the
mechanism of moral response.

The punishment treatment the subject received (in
Phase I) had a strong influence on the way he punished
himself. In Phase I, the subjects were punished, Group I >
Group II > Group III. The subjects imitated this punish-
ment in Phase II when they punished themselves, Group I > ~
Group II > Group III. However, Group III subjects punished
themselves more than they had been punished by the experi-
menter. These subjects may have been "compensating" for
the lower-than-baseline punishment they received in

Phase I. That is, they perceived the Group III punishment

39

40

as being too weak (insufficient) to control their behavior,
so they punished themsleves more severely. It might be
postulated that some minimal amount of punishment is
necessary to "involve" the subject in the task. Aronfreed
(1963b) hypothesized a similar "prerequisite amount" with
regard to nurturance.

The punishment compensation may have influenced
the interaction between moral orientation and punishment
treatment as well. For Group I (particularly) and
Group II, the rigid moral orientation subjects punished
themselves more than the flexible subjects. The Group III
difference was in the opposite direction--f1exible subjects
punished themselves slightly more than rigid subjects.

A possible interpretation of this finding might be that
the flexible subjects have a more highly "internalized
conscience" than the rigid subjects. The Group III
flexible subjects felt more "guilty" about not having a
sufficient level of punishment for the task, so they
punished themselves slightly more than the rigid subjects
did. In Groups I and II, the necessary "minimal level of
punishment" existed, and was sufficient for the flexible
subjects. This minimal level was not, however, sufficient
for the rigid orientation subjects to control their
behavior, so they punished themselves even more.

This internalization difference is supported in

the Baseline condition. The flexible subjects there

41

punished themselves more (but equivalent to Group II) than
the rigid subjects. Although the Group I (and II) differ-
ence was in the opposite direction, it might be hypothesized
that the task instructions were related to the way the
subjects punished themselves. That is, the reasoning or
"induction" of the instructions was characteristic of the
discipline flexible children typically received from their
parents (see Hoffman, 1970). Therefore, when these
subjects were provided with "cognitive standards"
(instructions) for evaluating their behavior, they felt
"involved" (guilty) regarding their performance on the
task. These same standards, however, were not sufficient
for the less "internalized" rigid subjects. The high
intensity punishment in Group I was more similar to the
rigid children's parental discipline. The rigid subjects
responded to this additional standard (punishment) for
evaluation by punishing themselves more than the flexible
subjects did (in Groups I and II). Thus, the rigid
subjects may be more dependent on concrete guidelines to
control their behavior. Although the other data were non-
significant, they provide trends consistent with this
postulated differences in internalization.

If the response data for Group I (especially) are
reviewed, it can be observed that the rigid moral orien-
tation and low social class subjects gave more self-

critical responses over the ten trials than the flexible

42

and high social class subjects. In the "Tootsie-Rolls-
removed" data, the rigid group punished themselves more
than the flexible subjects in the low social class,

9 These non-significant

relative to the high social class.
findings (together with the significant) suggest that
self-critical responses (on the ten trials) and punishment
(TR's removed) may serve the same function for the low
social class and rigid moral orientation subjects. That
is, these subjects may use punishment and self-criticism
in concert to control their behavior. The high social
class and flexible moral orientation subjects used fewer
of these responses (punishment and self-criticism) because
they have a more highly "internalized conscience." Thus,
the high social class and flexible subjects do not "need"
these overt stimuli (cues) to control their behavior ~
patterns.

Luria (1961) observed that language and cognitive
processes are related in a developmental sequence. When
a child is very young, he can inhibit his behavior only
with overt verbal cues. As the child matures, the need for
overt cues diminishes, and the child eventually inhibits
his behavior with covert verbal cues. This developmental

sequence in the use of overt and covert cues to control

 

9That rigid children used more punishment than
flexible children is supported by Hoffman's (1970)
observation that rigid parents are more concerned about
punishing accidental damage by their children than are
flexible parents.

43

behavior may be influential in the interpretation of the
child's use of self-criticism and punishment.

It might be argued that the interaction effect has
nothing to do with internalization of sanctions, but was
obtained because the high social class and flexible
subjects were merely less involved in the task (felt less
concerned) than the low social class and rigid subjects.
Or, that the high social class and flexible subjects did
not respond from fear of punishment. The reparative
response data, however, do not indicate support for these
arguments. The flexible moral orientation and high social
class subjects consistently (albeit non-significantly) gave
more reparative responses than the rigid and low social
class subjects. The subjects used more reparative
responses when they punished themselves than when they
were punished by the experimenter (supported by Aronfreed,
1961). And, there was no relationship between punishment
treatment and reparative responses. The subjects who were
punished more than the baseline average (Group I) in
Phase I, did not use more reparative responses than the
other punishment treatments. If a subject gave a reparative
response, it appeared that he felt "involved" (concerned)
in the task. It seems reasonable to suggest that subjects
who gave more reparative responses were exhibiting a more
highly internalized moral orientation. These differences

in internalization between the two social classes, and

44

between the two moral orientations may not be equivalent.
That is, the two moral orientations which are both
"internal" types (Hoffman, 1970) might be closer in
internalization than the two social classes. One would
expect such a relationship from other research (e.g.,
Hoffman, 1970; Aronfreed, 1961). Thus, the high social
class and flexible moral orientation subjects may possess
to varying degrees a greater internalization of moral
sanctions than the low social class and rigid orientation
subjects.

The directions of the differences in the self-
critical and reparative response data imply that Aronfreed
(1961) may have been correct in claiming that these
responses are "distinct moral phenomena" and are not
"equivalent reflections of an underlying unitary phenomenon
such as 'conscience.'" The powerful influence of experi-
menter (model) punishment, and the patterning of self-
critical and reparative responses indicate the importance
of imitation and past socialization experiences in the
child's responses to transgression (Aronfreed, [see p. 18];

Bandura, 1969; Hoffman, 1970; and others).10

 

10It should be noted that these antecedent con-
ditions can also include very subtle cues from the child's
socialization agents. In the baseline condition, for
example, one experimenter had subjects who consistently
gave more self-critical responses over the ten trials than
the subjects of the other experimenter. Upon investigation
of this "phenomenon," it was discovered that the former
experimenter had maintained eye-to-eye contact while
giving the instructions, while the latter experimenter had

45

Imitation of socialization agents then is one of
the important influences in the child's internalization of
moral sanctions. In this study, adult models were used.
But Piaget (1969) emphasizes the role of peer interactions
in moral development. Thus, further research utilizing
the present methodology modified to include peers (and
peer interaction) would be useful in investigating Piaget's
theory of moral development.

Since moral orientation played a significant role
in this study, the criticism of Aronfreed's discipline

11 The moral orientation

classification seems justified.
categorizations Hoffman (1970) suggested appear to have
functional relevance to the patterning of children's moral
responses to transgression, and should prove invaluable for
future examinations of moral behavior.

Aronfreed (1961) believed that "different moral
orientations do not emerge sequentially with advancing age
or experience," while Hoffman (personal communication) has

observed that children generally become more humanistic-

flexible with age. The data obtained in the present study

 

not. This eye contact may have changed the subject's
perception of the task, and affected the responses given.
The importance of nonverbal cues to moral responses has
been recently examined by Womack (1971). Although self-
critical responses over the ten trials was the only
response measure which showed experimenter bias, the
experimenters were cautioned in order to minimize the
occurrence of this confounding factor in Groups I, II, and
III. The data for these three groups do not show experi—
menter bias.

11See pp. 8-12 of this paper.

46

when considered in terms of Luria's (1961) findings
indicate that the rigid moral orientation subjects are
less "mature" than the flexible subjects. This notion is
supported by Hoffman's observation, and implies that age
and experience are important in the development of moral

12 In the present study, it was found that

orientations.
only 37 of the 128 subjects felt that their performance
was related to the doll's breaking. The children may have
been too young to exhibit such self-critical responses (on
trial eleven) in greater numbers. Perhaps, with age more
subjects would assume responsibility for the "damage." It
is evident that further research (longitudinal and cross-
sectional) is necessary to investigate the role of age and
experience in moral development.

The social classes used in this investigation were
structurally quite different from one another. The social
classes were equally represented in the two internal
orientations, and no external moral orientation subjects

13 The social

were obtained in either social class.
classes did, however, imply differential trends in the
patterning of responses to transgression. Thus, the

social class categorization of Hollingshead (1957) may be

 

12See Piaget (1969) for additional discussion of
this topic.

13This equal representation may be indicative of a
"transition period" as hypothesized earlier (see p. 21,
this paper).

47

a functional, as well as structural, classification system
(as Hollingshead proposed). Additional research is
necessary with subjects from dissimilar locations, and
with external moral orientation subjects. Greater subject
differences in social class and moral orientation might
result in more significant findings than were obtained in
the present study. Such data might permit a broader
understanding of the mechanisms involved in moral responses
to transgression.

The data also demonstrated the significant influ-
ence of punishment in the child's responses to trans-
gression. The use of punishment in behavior control has
been examined quite frequently in past research in moral
behavior. In order to achieve a greater understanding of
moral behavior and internalization, it is imperative that
future research investigate the effectiveness of reward as
a sanctioning device.

In retrospect, the research procedure has proved
valuable as a modus operandi for the examination of the
relationship of social class, moral orientation, and
severity of punishment (and social antecedents in general)
to children's moral responses to transgression. The data
have been integrated in an attempt to furnish a more
cohesive and functional explanation of the social mechanisms
involved. Some suggestions for additional studies have
been offered, and hopefully these will provide impetus for

further investigations in these areas.

REFERENCES

REFERENCES

Allinsmith, W. The learning of moral standards. Disser-
tation Abstracts, 1954, 1441), 1095-1096.

 

Aronfreed, J. The nature, variety, and social patterning
of moral responses to transgression. Journal of
Abnormal and Social Psychology, 1961, 63, 223-240.

Aronfreed, J. The effects of experimental socialization
paradigms upon two moral responses to transgression.
Journal of Abnormal and Social Psychology, 1963(a),
66, 437-448.

Aronfreed, J. The origin of self-criticism. Psychological
Review, 1964, 11, 193-218.

 

Aronfreed, J., Cutick, R. A., and Fagen, S. A. Cognitive
structure, punishment, and nurturance in the
experimental induction of self-criticism. Child
Development, 1963(b), 34, 281-294.

Aronfreed, J., and Reber, A. Internalized behavioral
suppression and the timing of social punishment.
Journal of Personality and Social Psychology, 1965,
1’ 3.16.

Aronfreed, J. Conduct and conscience. New York: Academic
Press, 1968.

Bandura, A. Social learning of moral judgments. Journal
of Personality and Social Psychology, 1969, ll(3),
'275-279.

Bandura, A., and Kupers, C. Transmission of patterns of
self-reinforcement through modeling. Journal of
Abnormal and Social Psychology, 1964, 52, I-9.

Bandura, A., and MacDonald, F. Influence of social
reinforcement and the behavior of models in shaping
children's moral judgments. Journal of Abnormal
and Social Psychology, 1963, E2, 274-281.

48

49

Baumrind, D. Current patterns of parental authority.
Developmental Psychology Monograph, 4(l-Part 2),
1971, 1-103.

Biaggio, A. M. Relationships among behavioral, cognitive,
and affective aspects of chilren's conscience.
Dissertation Abstracts, 1968, 28(11-8), 4742.

 

Boehm, L. The development of conscience: A comparison of
American children of different mental and socio-
economic levels. Child Development, 1962, 33,
575-590.

 

Burton, R. V., Maccoby, E. E., and Allinsmith, W.
Antecedents of resistance to temptation in four-
year-old children. Child Development, 1961, 32,
689-710.

 

Caring, L. C. The relation of cognitive styles, sex and
intelligence to moral judgment in children.
Dissertation Abstracts International, 1971, 11

Douglas, V. I., and Shaffer, H. B. Developmental changes
in the attribution of blame. Canadian Psychologist,
1971, 12(2), 248-253.

 

Dworkin, E. The effects of imitations, reinforcement, and
cognitive information on the moral judgments of
children. Dissertation Abstracts, 29(1-8),
365-366.

Einhorn, J. A test of Piaget's theory of moral develop-
ment. Canadian Journal of Behavioral Science,

 

Ferguson, L. R. Personality develo ment. Belmont,
California: BrooRs/Cole PuEIisHing Company, 1970.

Grinder, R. E. Parental childrearing practices, and
resistance to temptation of sixth-grade children.
Child Development, 1962, 33, 803-820.

 

Gump, P. V., and Korinin, J. S. Milieu influences in
children's concepts of misconduct. Child Develop-
ment, 1961, 32, 711-720.

Gutkin, D. Unpublished M.A. thesis, Michigan State Uni-
versity, 1970.

Hardeman, A. M. Children's moral reasoning. Dissertation
Abstracts, 1967, 38(5-8), 2126.

 

50

Hess, R., and Shipman, V. Early experience and the
socialization of cognitive modes in children.
Child Development, 1965, 33, 869-886.

Hicks, D. J. Imitation and retention of film-mediated
aggressive peer and adult models. Journal of
Personality and Social Psychology, 1965, 2, 97-100.

Hoffman, M. Moral development. In Mussen, P. (ed.),
Carmichael's manual of child s cholo (Vol. II).
New York: J6hn Wiley and Sons, Inc., 1970(a),
pp. 261-359.

 

Hoffman, M. Conscience, personality, and socialization
techniques. Human Development, 1970(b), 13,
90-126.

Hollingshead, A. B. Two factor index of social position.
Yale University. Mimeographed paper, 1957.

Irwin, D. M., and Moore, S. G. The young child's under-
standing of social justice. Developmental
Psychology, 1971, 5(3), 406-410.

Keasey, C. B. Social participation as a factor in the
moral development of preadolescents. Developmental
Psychology, 1971, 5(2), 216-220.

LeFurgy, W., and Woloshin, G. Immediate and long-term
effects of experimentally induced social influence
in the modification of adilescents' moral judgments.
Journal of Personality and Social Psychology, 1969,

I " .

Lewis, M., Wall, A., and Aronfreed, J. Developmental
change in the relative values of social and non-
social reinforcement. Journal of Expgrimental
Psyghology, 1963, 63(2), I33-I372

Longstreth, L. Psychological development of the child.
New York: Rona Press Co., .

Lorimer, R. Changes in the development of moral judgments
in adolescence: The effect of a structured
exposition vs. a film and discussion. Canadian
Journal of Behavioral Science, 1971, 3(1), I-IU.

 

 

Luria, A. R. The role of speech in the regulationrof
normal and abnormal behavior. New Yofk: LIVeright,
I961.

51

Maccoby, E. E. Role taking in childhood and its conse-

MacRae,

quences for social learning. Child Development,

D. A test of Piaget's theories. Journal of
Abnormal and Social Psychology, 1954, 42, 14-19.

 

 

McKinney, J. P. The development of values--Prescriptive

or proscriptive? Human Development, 1971, 14,
71-80.

 

Olejnik, A. Unpublished M.A. thesis, Michigan State Uni-

Piaget,

Sears,

Selman,

versity, 1971.

J. The mgral_jud ment of the child. New York:
The Free Press, 69 (1932).

 

R. R., Maccoby, E. E., and Levin, H. Patterns of
child rearigg. Evanston, 111.: Row, Peterson,
1957.

 

 

R. The relation of role taking to the development
of moral judgment in children. Child Development,
1971, 42(1), 79-91.

 

Whiting, J. M., and Child, D. L. Child training and

Winer,

Womack,

 

personality: Aggross-cultural study. New Haven:
Yale University Press, 1953.

B. J. Statistical_principles in ex erimental
design. New York: McGraw-Hill Boo? Co., 1962.
M. O. Utilization of two nonverbal cues by
children giving moral judgment responses.

Dissertation Abstracts International, 1971, 31(8-A),
3893.

 

 

 

 

APPENDICES

APPENDIX A

MORAL ORIENTATION ITEMS

APPENDIX A

MORAL ORIENTATION ITEMS

Why shouldn't someone steal from a store?

Two young men, Al and Joe, were in trouble. They
were secretly leaving town in a hurry and needed
money. So Al broke into a store and stole $500.
Joe went to a man who helped people in town. Joe
told the man he was very sick and needed $500 to
pay for an operation. Really Joe wasn't sick at
all, and he wasn't going to pay the man back. Even
though he didn't know Joe well, the man loaned Joe
the money. 50 Al and Joe both left town, each
with $500. If you had to decide who did worse, Al
who broke into the store and stole $500, or Joe who
borrowed $500 but wasn't going to pay it back,
which one would you say did worse? Why do you
think he did worse?

Let's pretend that your good friend's mother is
dying and needs some medicine to save her life.
But only one man in town has this medicine. He
wants a lot of money for it too ($1,000). Your
friend's father went to everyone he knew to try to
borrow the money (for the medicine). He even went
to banks. But your friend's father can only get
$500. So he asks the man if he can get the medicine
now with the $500 and pay him the rest later. The
man says no--he wants all the money. So your
friend's father gets desperate and breaks into the
man's house and steals the medicine for his wife.
Do you think your friend's father was right or
wrong to do that? Why?

Jim is the best bowler of all his friends. He

usually bowls around 155 (score). One day some of
the boys are teasing his friend, Bobby, about how
poorly Bobby bowled (the low score he had) the day

52

53

before. They keep saying Bobby doesn't know how to
bowl and never did. Bobby isn't smiling at all.
Finally Bobby says, "I didn't bowl very well last
night, but once I bowled 145." The other boys
don't believe Bobby. They just laugh at him. Jim
hears them teasing his friend Bobby. Even though
Jim never saw Bobby bowl 145, Jim says, "It's

true what Bobby says. I was there when he bowled
145. I saw him myself." Do you think Jim was
right or wrong to say that? Why?

APPENDIX B

EXPERIMENTAL DESIGN

Baseline Condition
T§s punishithemiélves)

Flexible
Orientation
Rigid

High 535

APPENDIX B

EXPERIMENTAL DESIGN

PHASE I

Low SES

 

 

 

 

 

 

PHASE II

Same subjects

punish themselves

Group I g punishes g > highest baseline average

Flexible
Orientation
Rigid

Group II p punishes § ~ equivalent baseline average

Flexible
Orientation
Rigid

High SES

Low SES

 

 

 

 

 

 

High SES

Low SES

 

 

 

 

 

 

Group III E punishes g < lowest baseline average

Flexible
Orientation
Rigid

High SES

Low SES

 

 

 

 

 

 

54

Same subjects

punish themselves

Same subjects

punish themselves

Same subjects

punish themselves

APPENDIX C

SUMMARY TABLES OF ANALYSES OF VARIANCE

Summary of Analyses of Variance of the Number

55

TABLE C-1

of Self-Critical Respones for Punishment
Treatment by Moral Orientation by Social

Class for Phase I and Phase II

 

 

Source df MS F p
Phase I
A (punishment treatment) 3 1.24 .22 NS
B (moral orientation) 1 1.32 .24 NS
C (social class) 1 17.26 3.09 NS
AB 3 3.68 .66 NS
AC 3 5.99 1.07 NS
BC 1 .01 .002 NS
ABC 3 3.15 .56 NS
Within cell (error) 112 5.59
Total 127
Phase II
A 3 1.76 .45 NS
B l .01 .003 NS
C l 14.44 3.70 NS
AB 3 6.59 1.69 NS
AC 3 .91 .23 NS
BC 1 .20 .05 NS
ABC 3 4.74 1.22 NS
Within cell (error) 112 3.90
Total 127

 

Summary of Analyses of Variance of Self-Critical
Responses on Trial 11 for Punishment Treatment

56

TABLE C-2

by Moral Orientation by Social Class for
Phase I and Phase II

 

 

Source df MS F p
Phase I
A (punishment treatment) 3 .05 .22 NS
B (moral orientation) 1 .00 .00 NS
C (social class) 1 .13 .57 NS
AB 3 .06 .26 NS
AC 3 .35 1.52 NS
BC 1 .03 .13 NS
ABC 3 .30 1.30 NS
Within cell (error) 112 .23
Total 127
Phase II
A 3 .007 .02 NS
B l .07 .16 NS
C l .01 .02 NS
AB 3 .11 .26 NS
AC 3 .34 .79 NS
BC 1 .19 .44 NS
ABC 3 .16 .37 NS
Within cell (error) 112 .43
Total 127

 

57

TABLE C-3

Summary of Analyses of Variance of Reparative Responses

on Trial 11 for Punishment Treatment by Moral
Orientation by Social Class for

Phase I and Phase II

 

 

Source df MS F p
Phase I
A (punishment treatment) 3 .55 1.31 NS
B (moral orientation) 1 1.53 3.64 NS
C (social class) 1 1.13 2.69 NS
AB 3 .18 .43 NS
AC 3 .02 .05 NS
BC 1 .00 .00 NS
ABC 3 .27 .64 NS
Within cell (error) 112 .42
Total 127
Phase II
A 3 .63 1.21 NS
B 1 1.32 2.54 NS
C 1 1.32 2.54 NS
AB 3 .07 .13 NS
AC 3 .40 .77 NS
BC 1 .94 1.81 NS
ABC 3 1.32 2.54 NS
Within cell (error) 112 .52
Total 127

 

58

TABLE C- 4

Summary of Analysis of Variance of the Number
of Tootsie Rolls Removed for Punishment
Treatment by Moral Orientation by
Social Class for Phase II

 

 

Source df MS F p

A (punishment treatment) 3 739.26 252.31 <.001
B (moral orientation) 1 6.21 2.12 NS
C (social class) 1 .19 .07 NS
AB 3 12.48 4.26 <.01
AC 3 .35 .12 NS
BC 1 10.02 3.42 NS
ABC 3 .75 .26 NS
Within cell (error) 112 2.93

Total 127