(INNWillilmWINWWWIllHllHUll‘N

  

103
558

   

.THS_

711:: __
-
I Jul- ’ m» g
o: {L-
} E; 1?: .. "t“ '
S .t i... "jg “
t a.“ .0.“ ‘ _ .
O ' ’ .C'H
Ii c . '
Y ‘1”.
fi: .u'
i “ ‘ I
\ ’w'v
\
-—-- _ . ‘

This is to certify that the

dissertation entitled

EVIDENCE FOR THE UTILITY OF THE SCRIPT

CONCEPT AS A MODEL FOR CHILDREN'S
FRIENDSHIP FORMATION

presented by
William Michael Bukowski, Jr.

has been accepted towards fulfillment
of the requirements for

Ph-D- degreein mm

 

Date 7/27/83

‘

042771

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

 

MSU RETURNING MATERIAg:
Place in book drop to

Lian/mugs remove this checkout from

All-r.— your record. FINES will

 

 

be charged if book is
returned after the date
stamped below.

 

 

 

 

« . ~ .. _, is: r: ' 4r .2...

5 “ ‘ ‘i’
av“ } q 1, E .53.
i" ‘Ii. ‘3 }* 'E‘fifijl' {fl-J?

 

 

 

 

 

 

EVIDENCE FOR THE UTILITY OF THE SCRIPT
CONCEPT AS A MODEL FOR CHILDREN'S
FRIENDSHIP FORMATION

By

William Michael Bukowski, Jr.

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Psychology

1984

ABSTRACT

EVIDENCE FOR THE UTILITY OF THE SCRIPT
CONCEPT AS A MODEL FOR CHILDREN'S
FRIENDSHIP FORMATION

By

William Michael Bukowski, Jr.

TWO studies were conducted to test the hypothesis.that the script
(Schank and Abelson, l977) concept would be a useful model for
understanding of friendship formation among children. In the first
study, the analysis of written descriptions of friendship formation
provided by fifth grade boys and girls (n = ll7) revealed that (a)
only propinquity and mutual activity appear to be standard components
(i.e., REQUIREMENTS) of friendship formation for children of this age;
(b) that information seeking is more likely to precede mutual activity
than follow it whereas information giving is more likely to follow
mutual activity that precede it; and (c) that popular children relative
to their less popular peers are likely to include more components in
their descriptions. These results provide partial support for the script
concept although the latter finding could be attributed to better
verbal skills on the part of popular children.

In the second study, 40 previously unacquainted same-sex pairs of
fifth-graders were observed for 25 minutes in a laboratory playroom as
they interacted with each other for the first time. TWO coding schemes
were used to categorize their behavior during this session according

to the mutuality of their behavior and the content of their

conversation. The results indicate that information exchange was
most likely at the beginning of the interaction, that self-disclosure
increased slightly during the session, that girls were more likely
than boys to exchange information about school, family, and peers,
hand that boys tend to exchange information through the use of questions
and answers whereas girls are more likely to engage in self
disclosure.

These results are discussed in light of the proposed value of
"scripts" in social behavior and they are described as being
consistent with other studies of acquaintanceship formation among

children.

To Melissa, for her love,
support and friendship

and

To my parents, for their encouragement.
affection and foresight

11°

ACKNOWLEDGEMENTS

When New York City had returned to financial wealth in the late
l970's, Mayor Edward Koch claimed that he was responsible for the
city's improved fiscal condition. When asked by a reporter whether he
actually believed that he was solely responsible for this change, Koch
replied that "someone has got to take the credit and it might as well
be me." I feel much the same way about the project that is described
in this book: I did it. I might as well admit I did. Nevertheless,

I recognize that several persons have given me a great deal of assistance
on this and other projects I have conducted during graduate school.
First, several undergraduates have been a big help, particularly Jim
Heidrich, Wendy Kunce, Tom Shazer, Sharon Miller, Gwen Studnicki, and

Bob Freeland. I appreciate their patience and persistence. Second,

I would like to express my appreciation to students and administration
of East Lansing and Holt, Michigan, especially Tom Davis of the Hope
Middle School. Third, several graduate students have helped me greatly
and have prevented me from doing and thinking many stupid things. I
would like to thank Julie Stanley, Judy Meister, Tom Packard, Dan Stults,
Roger Buldain, Dewayne Moore, Debbie Hotch, Esther Dienstag, Mike
Fitzgerald, Jim Anderson, and many others. Fourth, I would like to

thank Ellen Strommen, Wade Horn, and Lucy Rau Ferguson for serving

on my committee, Gordon Wood for his support and his tolerance of my
computer expenses, Phil Marcus for his faith in me, and Annette Barclay

and Suzy Pavick for their service way above and far beyond the call of

duty; and to John McKinney for taking time out from his summer
sojourn to read this work. ‘

In addition, I would like to thank several people who have been
a source of help and encouragement for me in both personal and
professional ways during the six years I spent at MSU: Steve
Gitterman; MaryAnn Reinhart; Doug Carlson; Marianne Boltz; Bob Noll;
Art Pomerantz; Fran Bardoel Fujarczuk; Bob Varney; Lauren Harris;
Kristin Novak Arbaugh; and Jack Condon. I would like to particularly
acknowledge my appreciation to Andy Newcomb for his unending
encouragement and advice. He has made me a better psychologist and
person.

Finally, special thanks to three special people: Fred Rogosch for
his friendship; Laura Jean Steinberger Otwell for her unending

affection; and Melissa, fer four years of love and companionship.

iv

TABLE OF CONTENTS

Page

LIST OF TABLES ......................... vi

INTRODUCTION .......................... 1

Theoretical Foundations .................. 4

Story Grammars and Event Schemas ............ 6

Event Schemes and Social Understanding ......... 9

Script Theory ..................... lO

Metatheoretical Foundations ................ 13

Relevant Empirical Investigations ............. l4

Social Skill Training Programs ............. 15

Scripts and Social Behavior .............. 17

Studies of Peer Interaction ................ l9
Children's Descriptions of Strategies for

Friendship ..................... 21

Summary and Hypotheses .................. 23

STUDY ONE ........................... 25

Method...........; .............. 25

Subjects ........................ 25

Procedures ....................... 25

The Coding System ................... 27

Results .......................... 28

Response Frequencies .................. 28

Order of Use ...................... 30

Order of Presentation for Pairs of Items. ....... 32

Gender Differences ................... 33

Sociometric Differences ................ 35

Discussion ....................... 36

STUDY TWO ........................... 40

Method .......................... 40

Subjects ........................ 40

Procedures ....................... 4O

Coding of Behaviors .................. 42

Discussion ........................ 51

DISCUSSION ........................... 53

How Might Scripts be Valuable ............... 54

Do People Actually Use Scripts in Friendships ....... 58

REFERENCES ........................... 66

LIST OF TABLES

Table Page
1. Response Frequencies by Category ............. 29
2. Frequency of Each Category by Order ........... 3l
3. Frequencies of Each Code Summed Across Blocks ...... 44
4. Mean Frequencies of Statement Types for Each Time

Block .......................... 48
5. Mean Frequencies of Content Categories by Gender ..... 49

6. Mean Frequencies of Content Categories by Statement
Types .......................... 50

vi

Evidence for the utility of the script concept

as a model for children's friendship formation

INTRODUCTION

Friendships are a universal phenomenon: they are found in all
cultures, are enjoyed and experienced by nearly everyone, and serve as an
important focus of many persons' lives. For children, the ability to
engage in peer relations is regarded as an important developmental
milestone and recent research indicates that the quality of a child's
experience in the peer group is an excellent predictor of adjustment
during adolescence and adulthood. Recently, increasing numbers of social
scientists have taken steps toward understanding the processes underlying
children's frienships by investigating a variety of factors associated
with children's peer experiences. (see Hartup, in press; Asher 5
Gottman, l9Bl). These investigations have included behavioral studies
(e.g., Newcomb 5 Brady, l982: Dodge, l980: Puttalez 8 Gottman, l98l;
Gottman, 1983), social cognitive investigations (Bigelow, l975, l977:
Rubin. l973: Selman, l980; Berndt, l983:) and longitudinal investigations
of social effectiveness (Coie, Dodge 8 Brakke, l983: Bukowski 5 Newcomb,
in press). These investigations have provided information about
fundamental issues relating to the stability and consistency of
children's peer relations, the behavioral components of friendships, and

the structural qualities of children's friendship conceptions.

l

In spite of the recent increase of empirical interest in
children's peer relations, Hartup's (l970) observation that the basic
topography of peer interaction needs further exploration is probably
still true (c.f., Puttallez and Gottman, l98l). It is likely that this
persistent deficiency is due to a theoretical limitation that is
intertwined with an empirical shortcoming. Typically, investigators have
used a static conceptualization of relationships and consequently have
have chosen to study relationships as constant phenomena instead of
considering them as ongoing experiences that take place over time. Due
to this perspective, investigators have typically focused on frequencies
of behavior summed across time and they have paid little attention to
process oriented issues. Of course, this Previous research has been very
useful and necessary in that it has led to the identification of several
parameters that affect the quality of children's relations. Moreover it
has provided a large amount of important information concerning person
perception, interpersonal expectations, the important dimensions
distinguishing popular and less popular children and children's cognitive .
bases for their friendship relations. At this time, however, it appears
that a more process oriented approach is needed to unravel the sequential

aspects of children's friendship relations.

in the investigation reported here, children's impressions of
the process of friendship formation were studied by asking children to

explain how they become friends with other children. In addition. the

behavioral aspects of this process were examined by observing children
during their initial social encounter. These impressions and behaviors
were considered with respect to a theoretical model known as script
theory (Schank 8 Abelson,1977). According to this model, people organize
their knowledge of an event in a sequential manner that is arranged
according to temporal or causal relations among specific actions or
subevents. Moreover, according to this model, persons' impressions of an
event ought to fit the behavioral properties of the events themselves.

An additional proposal of this theory concerns the goal directedness of
behavior and cognitive schemas. That is, both behavior and persons'

impressions of behavior are largely determined by a specific goal.

in this paper, theoretical and empirical evidence is presented
in support of using the script model as a means of understanding the
process of friendship formation among children. The results of two
studies are described and discussed in terms of the support they provide
for the script concept. Study one was conducted to address the question
of whether of whether the script model could be used to describe
children's written explanations of how they would become friends with
another child. Study two was a behavioral investigation conducted in a
laboratory playroom. Pairs children were observed as they interacted
interacted with each other for the first time. The results of these two
studies are discussed in terms of their relevance to the script concept
and with respect to the complementarity between the written descriptions

and the behavioral data.

Theoretical Foundations

J. D. Salinger's (l9k5) novel IE2 Catcher in the 313 begins
with the following sentence: "If you really want to hear about it, the
first thing you'll probably want to know about is where I was born, and
what my lousy childhood was like and how my parents were occupied and all
before they had me." (p. 3). With this sentence the book's narrator,
Holden Caulfield, a troubled and cynical adolescent, introduces himself
to the reader and begins to tell his recollections of his childhood and
his experiences at prep school. Holden proceeds to tell the reader quite
a lot about himself and he proves to be a rather skilled story teller.

In this first sentence, however, Holden appears to be doing something
more than just telling the reader about himself; indeed, Holden starts by
saying something about the reader. He states that first the reader might
want to know something about where he was born, and then something about
his childhood and so on. Along with introducing himself to the reader,
Holden reveals his impression that readers expect to acquire certain
kinds of information and to acquire it in a certain order. In addition;
Holden appears to indicate that he considers order of presentation to be .
an important factor affecting the reader's ability to process
information. Eventually, Holden decides not to follow his initial plan,
but his original impression may have been a fairly close approximation of
what people expect from a story. Indeed, recent research has indicated
that people do have expectations of how information will be presented in
a story and that these expectations may serve as guides to both story

comprehension and the understanding of social behavior.

.\ -.

Investigations of the comprehension and structure of social.
processes fall beneath the topical heading of social cognition. Flavell
(l978) defined social cognition as "any sort of cognition that takes
human psychological and social phenomenon as its object. It therefore
includes our conceptions ("naive theories"), knowledge, inferences. and
observations concerning our own and other persons' feelings, perceptions,
motives, intentions, thoughts, personality traits, social interactions.
moral and other norms, and numerous other contents of our social world"
(p. A3). More recently, Nelson (l98l) provided a more succinct
definition when she wrote "social cognition is the process of
representing knowledge about people and their relationships" (p. 97).
The importance of studying social cognition and social processes was
indicated many years ago when Epictetus said "Hen are disturbed not by

things but by the views they take of them."

In the investigation reported here, a particular social
cognitive process was studied within the context of children's friendship
relations. Children's descriptions of the process of friendship
formation were collected and analyzed in order to determine the utility
of a model known as script theory as a general framework to describe
children's understanding of this social process. These descriptions were
then compared to actual behavior observed when two children interacted
with each other for the first time. The major thrust of this study is an

attempt to uncover not only the usefulness of script theory but to assess

the association between children's scripts and social behavior.
Theoretical foundations in support of this approach to the study of
children's understanding of social interaction were taken from three
sources: 1. literature related to story grammars and event schemes and
their relevance for the understanding of social behavior; 2. the notion
of a "script" and its role in the regulation and monitoring of behavior;
and 3. metatheoretical perspectives indicated in the work of both

psychologists and social philosophers.

Story Grammars and Event Schemas

 

Cognitive psychologists. especially psycholinguists, have
proposed a theoretical structure to account for children's construction
and comprehension of stories. These structures, referred to as story
grammars, have been defined as the feature of a sample of discourse that
gives the sample coherence (Brewer S Lichtenstein, l980). A number of
investigators have provided structural descriptions of story grammars and
have tried to identify the role of these grammars in story construction,
comprehension and memory. Handler and Johnson (l977) have represented
the underlying structure (i.e., the grammar) of simple stories with a
tree diagram that indicates the constituents of a story and the relations
among these constituents. Nodes in their conceptual tree structure are
representative of either a STATE or an EVENT. They define STATEs as
information about either internal (i.e., having to do with the person) or
external (i.e., having to do with the world) conditions and they define

EVENTs as "any occurrence or happening” (p. Il5). Children's nursery

rhymes are typically arranged according to a story grammar. Take for
instance "Jack and Jill":
Jack and Jill went up the hill 59 fetch a

crown and Jill came tumbling after him.

In this story the first sentence would be designated as the
STATE and the second would be regarded as the EVENT. In their
structural analysis of simple stories they note that typically STATEs
precede EVENTs. They argue that this order (i.e., STATEs before
EVENTs) facilitates the comprehension of the EVENTs and the story in
general. They propose that by initially describing something about
the character or setting it is easier for readers to understand the
subsequent events as well as to recognize the logic underlying the
sequences within or between the events. That is. they claim that
STATE information tells the reader about the actors' goals, needs,
desires, and plans and that this information is crucial for the

recognition of the logic that holds the action together.

Stein and Glenn (l979)and Stein and Goldman (l981) have
proposed a similar structural model. They have postulated that
stories consist of "two higher order syntactic categories: SETTING and
EPISODE" (lSBl, p. 30]). They note that the SETTING conveys
information about the "protagonist and the physical and cultural

context of the story" (p. 301) whereas the EPISODE consists of action

change. conflict or resolution. Like Handler and Johnson (l977). they
regard the SETTING as superordinate to the EPISODE and they point to
the importance of the contextual information conveyed in the SETTING

for the interpretation and comprehension of the EPISODE.

A number of studies have demonstrated the importance of story
grammars for comprehension and memory: Information that is arranged in
a hierarchical fashion is recalled easier than information that is
presented in a nonhierarchical fashion (Thorndyke, 1977) and is more
likely to be included in a summary of the story (Rumelhart, l977);
information presented in correct temporal order is better retained
than information presented in an incorrect sequence (Handler, l978:
Stein 8 Nezworski. l978). Brewer and Lichtenstein (1980) have noted
that “overall these results have been taken to support the position
that the structural relations represented in story grammars are used
to understand and remember stories." Nevertheless, these same authors
(Lichtenstein 8 Drawer, 1980) have also argued that this form of
organization is not restricted to the comprehension of stories
presented in prose. They have proposed and demonstrated that people
use a schema much like that of a story grammar in their organization
of information presented on video tape. In a series of five studies
(reported in Lichtenstein 5 Brewer, l980) they focused on the question
of whether story grammars might be simply a particular kind of
structure they call "event schemes." They define an event schema as a
set of behaviors or processes organized according to a plan of action

or designed according to accomplish a goal. The results of their

investigations indicate that: l. people can generally agree upon the
structure of an event; 2. that goal direction is an important feature
for memory and encoding; 3. that temporal order is a key parameter
for comprehension: A. that the temporal or plan oriented structures
of imformation are the most salient aspects of event information.
Most importantly, their studies, when considered in conjunction with
investigations of story grammars, indicate that hierarchical
structures may play a large role in persons' organization of
information regardless of its form of presentation (e.g., prose or
video). In fact, Lichtenstein 5 Brewer (I980) go so far as to say
that story comprehension or story construction has little to do with
literary skill or knowledge of linguistic conventions but has
everything to do with "knowledge of the structure of naturally
occurring sequences of behavioral events" ( p. AAZ). A similar

conclusion was reached by Baggett (l979).
Event Schemas egg Social Understanding

It appears that the event schema concept possesses some
validity as a model for understanding persons' comprehension and
memory of infomation obtained from a variety of media. Is there
reason to believe, however, that people use schemes in their
understanding and organization of actual behavior and interpersonal
interaction? This question was recently addressed by Stein and
Goldman (l981) and they argue that the answer is decidedly

affirmative. They note that in the comprehension of stories one is

IO

called upon to perform many of the same tasks required during actual
social interaction: making inferences about characters or other
people: recognizing individuals' goals or intentions; comprehending or
assuming other persons' perspectives; and developing expectations of
persons' subsequent behavior. Stated briefly, Stein and Goldman point
to the need for people to engage in social problem solving in their
comprehension of stories and in their day-to-day social behavior.
Drawing on a number of theoretical perspectives (DeCharms, I968;
Piaget, I968; Heider, l958), Stein and Goldman (I98l) state that the
ability to understand the behavior of another person depends on the
ability of the "observer to understand the (person's) plans and
actions" (p. 300). In particular, they stress the importance of a
person's ability to make inferences about others and to develop
expectations about others. Their assertion, than, is that schemes are
valuable for the comprehension of stories as well as for the
understanding of social behavior. Their conclusion is an extension of
Lichtenstein and Brewer (I980) who emphasized the pervasiveness of the
schema notion indicating its relevance for a broad array of events.
regardless of the manner in which information about the event is

presented or obtained.

Script Theory

So far, the discussion of the schema process has been focused
almost exclusively on the process of comprehension. Another

conceptualization of the schema notion, however, includes an emphasis

ll

on the role of schemes in the regulation and monitoring of behavior.
This conceptualization, referred to as script theory, has been
advanced by both social and cognitive psychologists as well as by
people from the field of artificial intelligence (Schank 5 Abelson,
1977; Nelson, l98l; Searle, l981, 1982; Tomkins, I980). As Abelson
(l98l) has noted, scripts "embody most of the conceptual issues raised
by other types of schemes, yet, they are simple and well structured
enough to permit more focused analysis and experimentation" (p. 7l5).
Indeed, scripts are a variation of the schema concept: they specify a
sequence that reflects the causal or temporal relations among the
actions (Schank 8 Abelson, l977). Schank and Abelson also argue that
people use scripts for both comprehension and in the regulation of
their behavior. Moreover, they have hypothesized that scripts are
”activated" in certain contexts, such as going to a restaurant or a
birthday party, making a campfire or a cake or taking an afternoon nap
in school. That is, they are goal dependent and they are used when

they are called upon by particular demand, needs or situations.

Nelson (1981) has pointed to the value of the script concept
as a useful framework from which to study children's conceptions of
relationships and the routine events that they engage in. She has
noted that the value of scripts lies in the dynamic structure of their
composition and in the flexibility that coexists with their very
concrete nature. This flexibility and structure is afforded by the
inclusion of both obligatory actions (REQUIREMENTS) and optional

actions (SLOTS). An example of a script for a particular context will

12

illustrate the concrete nature of a script as well as indicate the
distinction between REQUIREMENTS and SLOTS. Consider a "going to a

restaurant" script:

First you go up to the door and go in (REQUIREMENT)
Next you wait for the hostess to bring you to a table
(SLOT)

and then you can sit down. (REQUIREMENT)

Then you look at the menu (REQUIREMENT)

and order your meal. (REQUIREMENT)

Next you will have an appetizer (SLOT)

and wait for the meal. (REQUIREMENT)

When the meal arrives, you will eat it and talk to
your companions. (REQUIREMENTS)

When this part is finished you will order desert
(SLOT)

and after eating it (SLOT)

you pay your bill and leave. (REQUIREMENT)

This simple script is representative of a young child's (age
7) description of what happens when people go to a restaurant (Nelson,
l98l). Because of the inclusion of both REQUIREMENTS and SLOTS,
scripts are sufficiently rigid so as to facilitate accurate and
reliable expectations of events or experiences while having adequate

flexibility to permit variations among people or settings (Nelson,

13

I981). The importance of scripts for actual behavior is that they are
proposed to "govern a body of inferences" about what events will occur
and the order of their occurrence" (Abelson, l981). Schenk and
Abelson (1977) claim that the value of scripts lies in the
opportunities they offer for prediction. Because of the
opportunities, for better and more reliable prediction, they posit
that people will engage in more efficient and productive behavioral

selection and evaluation.

Metatheoretical Foundations

In short, scripts are purported to facilitate planning as
well as comprehension. Clearly, the script model is predicated
heavily on the notion that individuals play an important role in the
planning and execution of their actions. This conceptualization of
the role of the individual is consistent with broad theoretical
descriptions of development. For example, drawing upon the TOTE
(Test-Operate-Test-Evaiuate) model proposed by Miller, Galanter and
Pribram (I960), Forbes and Lubin (l979) have pointed to the utility of
the script model as a means of conceptualizing how people plan their
behavior according to their needs and environment and how they monitor
the effectiveness of their actions and determine the need fogffurther
action. They note that in both the cybernetic TOTE model and in
Schenk and Abelson's description of how people use scripts there is a

strong appreciation of the intentional or planful character of human

14

action. The recognition of this goal oriented aspect of human
behavior has been previously emphasized in the work of both theoreticl
psychologists (Werner, l96l; Piaget, I952; Bruner, 197l) and social
philosophers (Peters, l958: Schwayder, I965; Harre', l97h). These
writers have stressed the importance of regarding individuals as
agents who play a central role in the conduct of their behavioral
interaction. This notion of the individual as an active organizer and
constructor of the environment and of behavior is consistent with the
organismic approach to human development. Jonas Langer (l969), in a
discussion of the organismic model has cited the importance of
considering the individual's behavior and development as part of a

self directed process.

RELEVANT EMPIRICAL INVESTIGATIONS

If scripts possess some value for understanding social
interaction, then empirical investigations of social behavior should
yield evidence that this form of organization facilitates social
relations. A number of empirical studies can be cited as evidence in
support of this argument. Support will be garnered from three
sources: 1. -studies of social skill training programs; 2. studies of
friendship formation among children; and 3. social cognitive and
behavioral studies of children, especially children representing
different levels of social effectiveness. In some of these

investigations the results have already been interpreted in light of a

15

script framework, whereas in other studies, the results will be

reinterpreted to indicate the salience of the script model.

Social Skill Training Programs

The results of at least two examinations of social skill
training programs can be interpreted to indicate that scripts
facilitate social behavior. In a well known study conducted by
O'Connor (I968), children who interacted infrequently with their peers
were shown a film that contained eleven modeling episodes designed to
increase the children's level of interaction with their peers. The
film was accompanied by a narration that was focused on the filmed
characters' intentions and the positive consequences of their
behavior. It also included a description of the characters' ongoing
behavior. O'Connor reported that the subjects who saw this film and
heard the narration become more interpersonally active than children
who saw a film about children and animals or those who saw nothing.
Asher and Renshaw (l98l) reported that other investigators who have
used this film have reported "significant increases in the children's
rates of interaction immediately following viewing the film and the
data with regard to maintainence are generally positive" (p. 282).
They noted that this success rate compares favorably with the results
of studies that used films without a narration or used a narration
that did not include instructional content regarding acts and
consequences like that of the O'Connor film., Stein and Goldman

expressed their impression of the reason for the success of this film

16

when they wrote ”the verbal narration makes explicit the goals,
thoughts and affective relations of the characters...and the causal
connections between these states" (p. 306). It would be difficult to
provide a better description of what a script is supposed to do than
what they said about the function of this film's narration. It is
conceivable then that the narration of the O'Connor film served to

provide a script for the children watching the film.

In another skill training program, Ladd (l98l) selected 36
third grade children with low peer acceptance and placed each child in
one of three experimental conditions: I. a treatment condition; 2.
an attention control condition; and 3. a no treatment control group.
The subjects in the treatment condition were encouraged to
”self-evaluate the social skill performance and peer responses "
(p.172). Ladd reported that children in this condition were taught to
monitor the effects of their behavior and to regulate their behavior
according to their expectations of peer responses. For example, the
children were encouraged to think of reasons to explain an undesirable -
outcome in an actual play session and to adjust their behavior
accordingly. Ladd reported that children who underwent this type of
treatment become more popular than children in either of the other two
conditions in spite of the similarity amomg these groups before the
study began. Although this study was not designed as a study of the
role of scripts in social interaction, its results can be interpreted
as support for the notion that script knowledge facilitates social

interaction. That is, it is conceivable that the subjects' increased

l7

recognition of the association between their own behavior and feelings
and the behavior and feelings of others led to an improvement in their

social behavior.

Scrigts ggg Social Behavior

 

David Forbes and David Lubin have directly addressed the
question of whether children who differ in their scripts will behave
differently with their peers. In a number of unpublished studies
(Forbes 8 Lubin, l980,l981) conducted in their laboratory at Harvard
University, they have observed young (preschool) children interacting
with their peers and conducted interviews with these same children in
an attempt to assess the children's script knowledge. The interview
consisted of a group of incomplete stories that the child was asked to
complete. Forbes and Lubin (l98l) were especialy concerned with two
aspects of their scripts: unity and flexibility. Unity was the extent
to which the child described the individual parts of the script as
forming an integrated process. Flexibility was the extent to which
the child's script allowed for individual differences between other
children and situations. Children whose scripts were both unified and
flexible were designated as high in script knowledge whereas those
whose descriptions were rigid, diffuse or both were designated as low
in script knowledge. Forbes and Lubin (1981) reported that children

who were designated as low in script knowledge were more likely to

l8

spend time alone than children who were either high or average in
script knowledge. They argued that children who lack the ability to
understand the continuity in other people's behavior while still
maintaining some flexibility will experience difficulty planning their
own behavior and becoming engaged with others. Forbes and Lubin
described these low script children as "out of touch" with their peers
and they noted that their parents reported that at home or in their
neighborhoods these children usually played with younger children.
Conversely, they reported that the children who were high in script
knowledge were the ones who were most actively involved with their
peers and, according to parental report, these children had a
preference for older playmates. They interpreted this preference for
older playmates as indicating that children with well developed
scripts are capable of functioning in social environments that are
more sophitcated than that of their own age group On the other hand,
it appears that children who have little script knowledge have trouble
functioning in the social environment appropriate for their age level.
They claim that this differential level of functionning may be due to
the accurate and reliable inferences that children with high script
knowledge can make whereas children without much script knowledge are
able to function in situations where they are called upon to make few

inferences.

19

Studies 9: Peer Interaction

Behavioral studies of children's interactions with their
peers also provide results that indicate the usefulness of the script
model. In one investigation, Putallez and Gottman (1981) examined the
way individual children became involved in the ongoing interaction of
two other children. They observed 60 second and third grade children
in a laboratory setting. Children were designated as either popular
or unpopular according to the number of times they were nominated by
their classmates as a ”friend." The children were assigned to triads
that consisted of either a pair of popular or unpopular children and
an additional child who was either popular or unpopular. This design
provided a fully crossed 2 X 2 factorial (i.e., two levels of dyad
popularity and two level of popularity for the third child). The
children in the dyad were instructed to play a game together in the
laboratory facility. After they began the game the experimenter
explained the rules to the third child prior to bringing this child
into the room. The dyad played for ten uninterrupted minutes before
the third child entered. When the third child (i.e., the entry child)
entered the laboratory playroom no specific direction were given.
Putallez and Gottman were interested in determining how the entry
child would enter the ongoing interaction and they wanted to discover
whether the unpopular child would try to enter the ongoing group in a
different manner than the popular children. They reported that
overall it appeared that children who successfully entered the ongoing

behavior were those who were able to determine the frame of reference

20 ,

common to the group members (i.e., their activities and goals) and
then to establish himself or herself as sharing in this activity.
That is, children who entered the group succesfully were able to
assess the nature and direction of the ongoing behavior and become
actively involved in it. In particular, they noted that unpopular
children, more so than popular children, disagreed with the two
children in the dyed, tried to change the direction of the dyadic
behavior or simply tried to draw attention to themselves. Overall,
these children (i.e., the unpopular children) were less successful at
becoming engaged in the ongoing behavior. In their discussion of
these findings, Putallez and Gottman noted that recognizing and
understanding the frame of reference of a particular pattern of
interaction is probably an important determinant of a child's success

at becoming part of that interaction.

In a more recent investigation, Gottman (1983) studied the
interaction of three- to nine- year old children as they became
friends. He was specifically interested in examining the sequential
aspects of their interaction. He followed l8 unacquainted dyads of
children playing in their homes for three sessions. His objective was
to "pinpoint a set of proceses that account for variation in
unacquainted children's progress toward friendship" (1983, p. 1). He
reported that among pairs of children who "hit it off" (i.e., became
friends) their interaction typically began with information exchange,
followed by the establishment of common ground activities and then

self-disclosure. It appears, then, that there is a specific pattern

21

in the interaction of children as they become friends. Perhaps most
important for the purposes of the current investigation, the sequence
of behavior reported by Gottman appears to fit the schema structures
reported earlier. That is, information search appears to precede
mutual activity which, in turn, appears to precede self disclosure
just as STATES (or SETTINGS) precede EVENTS (or EPISODES). Moreover,
this process appears to include at least three standard components
that could be construed to be what Schenk and Abelson would refer to
as REQUIREMENTS. The order of activities that he observed is also
consistent with the logic of a story grammar. For example, it is
conceivable that information search may be an important prerequisite
for mutual activity in order that the people engaged in the activity
are able to develop inferences or expectations of the other individual

and thus insure a more agreeable interaction.

Children's Descrigtions g; Stategies Lg; Friendshig

 

Studies of children's verbal descriptions of how to make
friends also provide evidence of a sequential structure to the process
of friendship formation. These studies also point to the differences
between the friendship strategies of popular and unpopular children.
For example, Gottman, Gonso and Rasmussen (1975) studied the
relationship between popularity and knowledge of how to make friends
among 198 third and fourth grade children. They assessed each child's
popularity by administering a free-choice sociometric questionnaire.

Knowledge of friendship was measured through individual interviews

with each child. The children were asked to pretend that they wanted
to become a friend of a child they had just met. The experimenter
asked the children to explain how they would go about doing this. The
items included in their answers were coded according to a scoring
system and a different number of points was given for different types
of answers. Specifically, the subjects were given points for
referring to particular behaviors in the following manner: I.
greeting- one point; 2. asking information- two points; 3. extending
inclusion- three points; and h. giving information- four points.
Gottman et al. (1975) reported that this scheme was sufficient for
the classification of the children's responses and, most importantly,
children refered to these behaviors in the order in which they are
listed here. They also reported that popular children received higher
scores than unpopular children. In their discussion of these findings
they claimed that popular children were more knowledgeable about
making friends than unpopular children. Because the scoring system
was dependent on both content and the sequential nature of the
responses it is conceivable that these results could be interpreted as
an indication that popular chderen have different scripts for
friendship formation than their less papoular peers. Taken together
with the results of Gottman's recent behavioral study the results of
the Gottman et. al. study indicate that there may be a certain
sequence to the behavioral subcomponents of friendship formation.
Moreover, it appears that the sequence of these actions is consistent

with what would be expected in light of the concept of event schemes.

 

23

Summary egg Hyggtheses

Evidence from theory and research related to event schemes,

the script concept and children's social relations has been presented.

In light of this evidence, the current investigation was designed to

address the following hypotheses:

1.

It was predicted that children's descriptions of how they
become friends with other children would fit a script model.
Specifically, it is expected that:

i. There will be standard as well as optional actions
included in children's descriptions of this process.

2. The presentation of actions will be order dependent; that
is, some actions will be more likely to appear in one
part of the description than another.

3. In these descriptions, information seeking will be more
likely to precede mutual activity than follow it whereas
information giving will be more likely to follow mutual
activity than precede it.

It was further predicted that popular children would have
better organized scripts than their less popular peers. The
better organization of these scripts will be manifested in
the following ways:

1. Popular children will include more actions in their
scripts;

2. popular children will cite a broader range of activities
in their scripts;

3. Popular children will be more likely than others to place
information seeking before action and to place
information giving after action.

It was predicted that the behavior of children during their
initial interaction would resemble a script format. That is,
there will be a standard order of actions which would

24

resemble the written descriptions of friendship formation.
A. It was also predicted that this script-like organization to

the behavioral data would be most pronounced when the
children were specifically given a goal of "making friends."

This investigation consists of two studies. Study One is an
analysis of of children's written descriptions of how they would
become friends with another boy or girl. In Study Two, children were
observed in a laboratory playroom as they interacted with each other
for the first time. Fifth-grade children served as subjects in each
of these studies. Study One is described first followed by a
presentation of the results and a brief discussion. Study 2 is then
presented in the same manner followed by a discussion of the complete

investigation.

STUDY ONE

Method

Subiects

Fifth-grade boys (n-6l) and girls (n-56) from five elementary
schools in a suburban midwestern community were subjects in this study.
The mean age of the subjects was 11 years, 2 months. Over 753 of the
children in the classrooms chosen for this study were given permission by

their parents to participate.

Procedures

The subjects were met in their classrooms at a time that was
convenient for the investigator and the classroom teacher. At these
meetings the investigator was assisted by a graduate and an undergraduate
student. Each of the participants was asked to complete a sociometric
questionnaire and to answer a question about friendship formation. The
sociometric questionnaire consisted of two sections. First, the subjects
were asked to choose from among the same-sex members of their school
grade the three persons whom they thought of as their best friends.
Second, they were asked to choose from among this same group of children
the three persons whom they would not want to play with. These answers

were used in the construction of four sociometric variables. First, the

25

26

number of times that an individual child was chosen as a best friend
served as the child's acceptance score. Second, the number of times that
a child was chosen as an undesireable playmate was used as a rejection
score. These two scores were standardized (i.e., transformed into
z-scores) according to the mean and standard deviation of the scores from
the same-sex members of their school grade. These standardized scores
were used in the construction of the third and fourth scores. The third
score was created by simply subtracting a child's rejection score from
the acceptance score. The difference was then standardized by gender and
school. This standardized score was referred to as preference and served
as a measure of a child's relative likeableness. The fourth score was
the sum of acceptance and rejection. The sum was then standardized by
gender and school. This standardized score was referred to impact and it

provided a measure of a child's visibility within the peer group.

Children's descriptions of friendship formation were collected
by presenting the children with a situation and a question and then
asking them to write a response to the question. The situation and
question were as follows: "Let's pretend there is a girl named Linda who
is the same age as yourself. She just moved into a house in your
neighborhood. How would you try to become her friend? (Write as much as
you can to answer this question).l It took the subjects approximately 25

minutes to complete these questions.

lThis was the question for girls. The question for boys was exactly the

same except the name "Tom" was used instead of "Linda."

27

122 Coding System

The children's responses were coded into eight mutually
exclusive and exhaustive categories by two undergraduate assistants who
were blind to the purpose of the study. The eight categories included in
this scheme were: introductions/propinquity; information search;
information: more introductions; ego reinforcement: play/activity
involvement; friendship bid: other/miscellaneous. The coders were
instructed how to use the coding scheme and were given 15 children's
responses to code as part of their training. For this training different
reponses were used with each of the two coders.. After they completed the
first ten answers, the investigator compared their classifications to his
own. The percent of agreement was computed and the disagreements were
noted. These disagreements were discussed with the coders and they were
given additional instruction with the coding scheme. With one coder, the
initial percent agreement was over 908; with the second coder the percent
agreement was only 782. After the completion of the remaining five
training responses the percent agreement was computed again. At this
time both coders agreed with the investigator on more than 903 of the
items. Eighty-five answers were assigned to each coder. Twenty-five
answers were assigned to both coders in order that interecoder
reliability could be assessed. Inasmuch as theses are‘nominel data, the
proper procedure for computing reliability is Cohen's Kappa (Cohen,
1960). This procedure provides a reliability estimate based upon both

percent agreement and the amount of agreement that is expected by chance.

28

The investigator considered the Kappa procedure appropriate and adequate
in this situation because of the nominal nature of the coding scheme. In
as much as the purpose of assessing relaibility was to determine the
extent to which the two coders classified the statements in the same
manner the Kappa computation is appropriate procedure because it is
focused on the coding of particular statements. Had the investigator
been interested in the aggreementas to the number of times that a code
was used (i.e., on agreement of frequency of occurrence only) then
alternative strategies would have been entertained (e.g., Cronbach, l970;
Cronbach, Gleser, Nanda, 8 Rajaratnam, l972). For the present data the
percent agreement expected by chance is .l6 and the Kappa value is .78,

considerably higher than the .6 minimum recommended by Hartman (1977);

Results

Resggnse Freguencies

In the first analysis of these data the number of times that
items from particular categories were used in the subjects responses
served as an index of the items role as either a "REQUIREMENT" or a
"SLOT" in the subjects' scripts (Schank 8 Abelson, l977). Lichtenstein
and Brewer (1980) have previously used 753 as a criterion for the
designation of items as either standard components or options in an event
schema. If an item was used by at least 752 of the subjects they
regarded it as a standard component whereas those that occured less

frequently were regarded as optional. Only two of the categories used

29

here (introduction/propinquity and mutual activity/play) were included in

the responses of at least 753 of the subjects. The total number of times
\.

that responses from each of the categories was used and the number of ,

subjects who used each is presented in Table I.

Table l

Response Frequencies By Category

targets
Introduction/Propinquity 88 107
Other Introductions 28 32
Information Search 32 52
Information Giving 35 38
Ego Reinforcement AS 52
Mutual Activity 9A 137
Friendship Bid/Statement 25 . 27

gThis figure represents the number of subjects who used the category at
least once.

gThis figure represents the total frequency of use for the category

30

Order f Use

 

Next, the order in which the different categories were used was
investigated. The number of times that each type of response was used as
the first, second, third, fourth or fifth items in the subjects responses
was computed. The results are shown in Table 2 in three forms: 1. the
absolute frequency of this category for each of the five positions; 2. a
percent based on the number of times the category is used at each time
relative to the number of items used overall in each position; and 3.
the number of times each category is used relative to the number of times
it is used overall among the first responses for each subjects.
Inspection of this table will reveal the following: First,
introduction/propinquity frequency drops markedly from the first response
to the fifth. Second, mutual activity is used rarely as the first item
but is used very often throughout the rest of the response, particularly
as the second response. Three, information giving and the more
introduction categories are used more so at the end of the responses than
at the beginning. Four, information searching increases in frequency and
then decreases. _Five, ego reinforcement and friendship statements are
each used with roughly the same frequency in the beginning, middle and
ends of the subjects' responses. Chi2_ analyses were conducted to
determine whether some of the response categories were more likely to be
found in one order position than another. Separate analyses were
conducted for each response category. Four of the chi; computations
provided significant results (i.e., p< .05, with four degrees of

freedom). That is, of the seven categories, four were more likely to be

Category
Intro/Propinquity

Other Introductions
Information Search
Information Giving
Ego Reinforcement
Mutual Activity
Friend Bid/State
Other
Note.

1,
72/61/67
01/01/03
06/05/12
04/03/11
18/15/34
12/10/09
03/25/11
01/01/08

31
Table 2

2_
26/22/24
06/05/18
13/11/25
06/05/16
09/08/17
48/42/35
04/04/15

02/02/16

Frequency of Each Category by Order

Qgggg.
2.
04/04/04
09/09/28
20/20/38

10/10/26

14/14/26

35/35/25
08/08/30
03/03/24

9.
03/04/03
11/14/34
08/10/16
11/14/29
06/08/11
26/33/18
09/11/33
04/05/33

5_
02/08/02
05/11/15
05/11/10
07/15/18
05/11/10
16/35/11
03/06/11
02/04/16

The first figure in each column represents the number of times that

category appeared in that order; the second figure is the percent of the

total number of responses in that order comprised by the particular

category: the third figure is the percent of the total for the

particular category comprised by the number at that particular order.

32

found in one position than another. Specifically,
intoductions/propinquity, more introductions, information search, and
play appear to depend partially on order as a determinant of their use.
The chi; values for these categories were 167.2, 11.4, 31.3 and 30.5,

respectively, all with a probability of less than .02.

In the next analysis the order in which specific pairs of items
were used was examined. Two pairs were studied: 1. Information search
and mutual activity; and 2. self disclosure and mutual activity. Only
those subjects who used both items in these pairs could be used in this
analysis. For the first analysis, there were 26 subjects who mentioned
both mutual activity and information giving. In light of the results
from Gottman's (1983) behavioral investigation, it was expected that
information search would be mentioned before mutual activity. The chi;
procedure was used to determine whether information search was more
likely to occur before mutual activity than after it. As described in
Runyon and Haber (1971), in a situation where there is only one degree of
freedom the chi square distribution is a poor approximation of the normal
distribution. In order that the chi square distribution would be a
closer approximation of the normal distribution .5 was subtracted from
the absolute value of the difference between the observed and expected
frequency for each cell. This procedure is described in Runyerd and
Haber (1971). Of the 26 subjects considered in this computation, 18

mentioned information search before mutual activity and eight mentioned

33 -

it after mutual activity. The chi;, for these data was 3.115, p< .075.

This finding was regarded as marginally significant.

Thirty subjects mentioned both information giving and mutual
activity. Chi; was computed to determine whether information giving was
more likely to follow mutual activity than precede it. Twenty-one
subjects mentioned mutual activity before information giving whereas only
nine mentoned informaton giving before mutual activity. The chi; value

for these data was b.033, p< .05.

Gender Differences

Four sets of analyses were conducted to study the differences
between the responses of boys and girls. First, the number of actions
mentioned by each subject was used as the dependent measure in an
analysis of variance with gender as the sole independent measure. The
results indicated significant differences between boys and girls, i
(1,115)-lh.306, p< .001, with the girls ( fl -A.4) citing a greater number
of actions than the boys ( M -3.5). Next,the differential use of
particular response categories was examined by computing the number of
times that each subjects used each individual category. These totals
were subjected to a multivariate analysis of variance. In this analysis,
there were seven dependent variables (i.e., all the categories but
"other") and only one independent variable (i.e., gender). The results
of the MANOVA computation revealed that there were differences between

the responses of the boys and girls, F (7,109)-2.57, p<.02. Univariate

34

analyses of variance were then conducted to identify the sources of these
differences. Significant differences were found for two of the
categories (ego reinforcement, L (l,115)-4.26, p< .05, and mutual play, L
(1,115)-6.l3, p<.01) and a marginally significant difference was found
for a third category (information giving, 5 (1,115)-2.4, p< .10). For
each of these categories girls made more responses than boys (for ego
reinforcement Mg - .58 for girls and .34 for boys; for mutual activity,
fig I1.416 for girls and 1.0 for boys; for information giving M3 -.48 for

girls and .29 for boys).

The second set of analyses was conducted to determine whether
boys and girls used the same number of action types in their answers.
The number of categories referred to by each subject was computed and the
sum was used as the dependent measure in analysis of variance. The
results indicated that boys refered to fewer categories than girls, E

(1,115)-9.557, p< .002, M3 -3.1 and 2.6 for girls and boys, respectively.

The focus of the third and fourth sets of analyses was directed
toward the differences in the order of responses given for boys and
girls. In terms of the order in which the catgories were used boys and
girls were nearly identical to one another. In regard to the placement
of information search, mutual activity and information disclosure there

were also no differences between boys and girls.

35

Sociometric Differences

The final group of analyses was conducted to assess the
differences between children representing different levels of sociometric
status. Pearson product moment correlations were computed between the
four sociometric scores and the number of responses that the subjects
gave and the number of categories they cited. Acceptance, rejection, and
preference were significantly related (p<.05) to the number of items that
the subjects included in their responses (r- .17, -.15, .18,
respectively) and to the number of categories that they cited (r-.18,
-.l7, .20, respectively). Impact was not significantly related to any of
the response variables. In regard to individual categories, these same
scores (i.e., acceptance, rejection and preference) were significantly
related to the subjects' use of information giving (r-.2, -.2, .22,
respectively) and ego reinforcement (r-.lh, -.l7, .17, respectively).

Again, there were‘po significant relationships for impact.

The final analysis was conducted to determine whether popular
children were more likely than other children to place information search
before mutual play than after it and to place information giving after
mutual play than before it. Subjects were assigned to sociometric groups
according to a procedure described by Newcomb and Bukowski (in press).
The .10 level of probability was used for the group assignments. Of the
26 children who mentioned mutual play and information search, only 1 was

from the group known as "populars" (i.e., children who have many friends

36

and are rarely nominated as an undesireable playmate), two were in the
rejected group, and three were from the isolate group. The remaining
children were from the average group. Because such a small number of
children were from the non-average groups five of the eight cells in the
A X 2 table (i.e., order by group) had only 1 or no persons in it.
Because of this low representation in some of the cells an analysis of

the data was abandoned.

This same problem was encountered when the data related to the
order of mutual play and information giving were considered. Again,
analyses of the data were abandoned because of the low representation in
several of the cells. Nevertheless, inspection of the raw data idicated
that subjects in the the four different groups presented theses items in

approximately the same order.

We

These results confirm five of the six hypotheses proposed for
this study. The sixth hypothesis could not be adequately addressed
because of the unavailability of a sufficient number of subjects.
Specifically, it appears that: 1. some actions are more likely to be
found at one stage of the subjects' descriptions than at another part; 2.
the order in which informaton giving, information searching and mutual
play are present relative to one another is not entirely dependent on
chance; 3. there are at least two standard components

(introductions/propinquity and mutual activity) to children's

37

descriptions of friendship formation; A. popular children are likely to
include more actions in their descriptions than their less popular peers;
and 5. popular children tend to cite a greater number of category types,

too.

The final two findings suggest that popular children have
scripts that are more extensive and more sophisticated than those of
their peers. It is conceivable that this greater sophistication may
facilitate and enhance the quality of popular children's social relations
and allow them to interact in agreeable manners with their peers. This
interpretation bears a resemblence to Roistacher's (1973) micro-economnic
model of children's friendship relations. He argued that popular
children attained their social success by simply recognizing and taking
advantage of the social opportunities available to them. Conversely,
these results may be indicative of a social cognitive deficit among less
popular children. Their relatively less sophisticated scripts may make
it difficult for them develop expectations of others and consequently
limit their social interaction. This notion of a social cognitive
deficit among children of low popularity status has been previously
discussed by Asher and Renshaw (1981) and Ladd and Mize (1983). In each
of these articles, the authors specifically suggest that recognizing the
direction in the behavior of other individuals may be especially
problematic for unpopular children. There is also empirical support for
this argument. Putallez, (in press) has recently reported that children
who have difficulty entering into the ongoing behavior of a dyad of

children of their own age also appear to have trouble making inferences

38

about the other children's behavior. She also reported that it is these
same children who eventually have trouble making friends in school and
they appear to be less popular in school than children who exhibit
superior skills atmaking social inferences. Renshaw and Asher (in press)
have also recently described the results of a study designed to examine
children's goals and strategies in social interaction. They concluded
that less popular children had different goals in social contexts than
their more popular peers and these children also recommended the use of
differented strategies in social situation. Renshaw (in press)
speculated that the difference in strategies may be a function of the of

the differences in their goals.

The findings pertaining to the overall content and organization
of the children's descriptions are consistent with the results of the
Gottman studies (Gottman, 1975, 1983: Putallez and Gottman, 1981) and
with what would be expected based on the concept of event schemes. That
is, SETTING information such as introductions, establishment of
propinquity and information search appears at the beginning of
descriptions, typically before play. Moreover, the direction of the
actions seems to go from superficial (e.g., introductions and infomation
search) to more personal activities (e.g., mutual activity and

information giving).

Overall, these results are taken as support for the script

model. Nevertheless, it is important to note that there is a

considerable amount of variance in these data that has not been explained

by the analyses reported here.

39

A second study was undertaken to further explore the
organization of the acquaintanceship process. A major difference between
the first and second studies was the unit-of analysis. In the first
study it was the written responses of individual children. In the second
study, it was the behavior of dyads of children. Of particular interest
in the second study were three issues: 1. the different kinds of
behavior that are observed at different times during two children's
initial encounter; 2. the manner in which children obtained information
about each other: and 3. whether the specification of a particular goal

would affect the pattern and nature of children's behavior.

STUDY TWO

Method

Subiects

The participants in this study were A0 boys and A0 girls from
six fifth grade classrooms in six schools. These children were randomly
chosen from among 179 children who had received parental permission to
take part in the study. Fifty-nine of these subjects had also
participated in Study 1. The sociometric scores of the subjects who were
chosen to participate were compared with the scores of the the other
potential subjects in order to determine whether they constituted an
unbiased representation of the entire sample. An analysis of variance
revealed that the participants did not differ from the non-participants
on any of the four sociometric values described earlier. The subjects'
parents were contacted on the telephone in order to schedule a convenient

time for them to bring their child to the laboratory.
Procedures

Children were randomly paired with a same-sex child from a
different school. This provided 20 pairs of boys and 20 pairs of girls.
Initially parents had been asked whether their children knew the child

whom they would be paired with in order to ensure that this meeting would

40

41

be the children's first encounter. When the subjects arrived at the
laboratory they were greeted by the project staff (i.e., the investigator
and some undergraduate assistants) and escorted to separate waiting
rooms. Caution was used to prevent the subjects from meeting or
interacting before the session began. The subjects were brought
individually to the playroom where they were greeted by the investigator.
He asked them to enter the playroom and gave them one of two sets of
instructions. One half of the subjects (10 pairs of boys and ten pairs
of girls) were told "What we would like you to do is to play in this room
for 30 minutes. When 30 minutes are over we will come back and then we
will ask you some question." The other subjects (i.e., the other 10
pairs of boys and the other 10 pairs of girls) were told "What we would
like you to do is to play in this room for 30 minutes. During this time
we really want you to become friends with each other. When 30 minutes
are over we will come back and then we will ask you some questions. In
the mean time, we really want you to become friends." The difference in
the instructions was intended as an experimental manipulation whose
purpose was to alter the subjects' goal during the interaction. In this
study the experimental manipulation and gender constituted two factors (

each with two levels) providing a fully crossed 2 X 2 design.

The playroom was approximately 3m x 5m and equiped with two
tables, four chairs, a number of toys and games, puzzles, drawing
materials and a small basketball hoop that was mounted above the
entranceway. One-way mirrors were embedded In the wall surrounding the

room. During the 30 minute session, a Sony Betamax video camera placed

42

behind one of the mirrors was used to record the subjects' interaction.
A microphone that was placed in the room was used to record the audio

portion of their behavior.

At the end of the thirty minutes, the investigator returned to
the playroom and the children were escorted to separate interview rooms.
Here they were asked if they were aware that their behavior was being
recorded and if they would allow the investigator to keep the recording.
Also, as part of a check on the experimental manipulation, the subjects
were asked to repeat the instructions they were given at the beginning of
the session. All of the children gave the investigator their permission
to use the recording and all the subjects but 2 correctly recalled the

instructions.
Coding 21 Behaviors

Two coding schemes were developed for this study. The first
scheme was designed as a "macro" scheme that was focused on broad
categories of behavior whereas the second scheme was designed to capture

the specific components of the subjects' infomation exchange.

gggigg 353593 In the first coding system, there were two
categories (conversation and action) included in each coding
classification. There were eight levels of the action category (mutual
competitive play, mutual cooperative play, solitary play, exploration,

one child explore/one child play, no activity, one child plays/one child

43

does nothing, one child explores/one child does nothing. Conversation
and nonconversation were the only levels for the conversation categories.
Information about what the dyad was doing and whether or not they were
talking was included in each code. Two undergraduate coders individually
watched the video tape and coded the children's behavior at eight second
intervals. The coders recorded their classifications in the memory of a
microprocessor designed for event recording. The event recorder also
informed the coder when each eight second epoch had ended. Twenty seven
tapes were assigned to each coder. The tapes that were assigned to both
coders were used to determine iter-rater reliability. Again, because of
the nominal character of the data Cohen's kappa was used as the
reliability measure. For these data, the percent agreement for the two
coders was .86 and the percent of chance agreement was .26.

Consequently, kappa was .79.

For each dyad, frequencies for each code were computed for five
blocks of time. Each block of time included 32 eight second epochs. In
each block, then, there were five minutes and four seconds of time. The
frequencies for each coding category summed across the five blocks is
presented in Table 3. As illustrated in this table, the interaction of
these children was dominated by cooperative play and they appeared to
spend more time in conversation than they spent without talking. In
order to determine how how the interaction of the subjects changed across
the five time blocks and to assess the effects of the experimental

manipulation and those of gender, four analyses were conducted.

Table 3

Frequencies of Each Code Summed Across Blocks

 

With Without

Category Conversation Conversation 12531
Play-Compete 2.6 1.8 A.A
Play-Cooperate 67.6 6A.3 131.2
Solitary Play 12.A 10.2 22.6
Explore 6.A 2.6 8.9
One play/One explore 3.8 2.0 5.9
No Activity 6.1 1.3 7.A
One Play/One Nothing 5.07 1.9 7.0
One Explore/One Nothing 0.9 0.7 1.6

Note: Each unit represents one eight second epoch.

In the first analysis the dependent measure was the amount of
mutual behavior. Mutual behavior was the sum of the mutual play, mutual
explore, and no activity conversation codes. A 2 X 2 X 5 (gender by
condition by time block) analysis of variance revealed no significant
(i.e., p< .05) main effects and only one significant interaction (time by
gender, 5 (A,156)-A.25, p< .01). For the boys the mean number of epochs
coded as mutual behavior for times blocks 1,2,3,A, and 5 were 32.8, 3A.l,
32.2, 30.3, and 26.0, respectively, whereas for the girls the mean number

of epochs was 29.7, 29.1, 28.7, 28.6, and 30.7, respectively. Next, the

45

exact same type of analysis (i.e., a 2 X 2 X 5 analysis of variance) was
conducted with the amount of conversation as the dependent measure. This
analysis revealed no significant main effects nor interactions. In the
final analysis, a more specific category of behavior was used as the
dependent measure. A 2 X 2 X 5 analysis of variance (gender by condition
by time) was performed first with the amount of cooperative play as the
dependent variable. The results indicated a significant main effect for
time, L (A,156)-2.7A, p< .03,( M s- 25.8, 29.0, 27.8, 25.0, and 2A.0 for
time blocks 1,2,3,A and 5, respectively) and a significant interaction
between time and gender, 5 (A,156)-3.18, p< .01. In regard to the
interaction, there was an average of 27.3, 30.5, 28.7, 2A.0, and 21.2
epochs of cooperative play for boys in time blocks 1,2,3,A, and 5 and an
average of 23, 27.5, 29.A, 26.1, and 27.3 epochs of cooperative play in

these same blocks for girls.

Egglgg’ggggmg Processes of information exchange were staudied
with the second coding scheme. This coding system was based on two
aspects of the subjects' conversation. One aspect was the actual content
of the conversation and the other aspect had to do with the type of
statement that was used. There were nine content categories: interests
(what one likes to do), school, peers, family, self-objective (e.g.,
name, place of birth, current residence), self-ability (i.e., someone's
skills) self-personal (i.e., intimate information) and other. (Only the
first eight were used in subsequent data analyses.) There were three
types of statements: questions, answers to questions and volunteered

information.

A transcript was made of each dyad's conversation. Two
undergraduate assistants coded the statements included in the
transcripts. Twenty-five transcripts were assigned to each coder. The
transcripts that were assigned to both coders were used to determine the
extent of inter-rater reliability. Cohen's kappa was used as the
reliability measure. The percent agreement was .88 and the percent
agreement expected by chance was .20, resulting in a kappa of .85.
Finally, as with the first coding scheme, the dyadic interaction was

divided into five five minute blocks.

The eight content categories were not subjected to any further
data reduction for two reasons. First, the investigator had designed the
coding scheme so as to categorize statements into categories that were
both consistent with the hypotheses being considered and qualitatitively
distinct. That is, they were supposed to respresented conceptually
different types of behaviors. Second, inspection of the correlations
amomg the classification groups revealed only small correlations (18 of
21 correlations were between -.15 and .15: the remaining 3 were less than

.25) among the coding classifications.

Before any analyses were conducted, correlations were computed
between the question and answer codes for each content category within
each of the five time blocks. It was expected that because the question
and answer codes were to a great extent mutually inclusive, the

association between them ought to be very strong. Not suprisingly, then,

47

the observed correlations were rather high, (average r- .77 and the range
was from .65 to .92). Inasmuch as it appeared that these codes were
probably representing the same underlying process the investigator had to
decide whether to combine these codes into a single measure or to choose
one as a index of the question and answer method of information exchange.
In the analyses planned for these data, the investigatior wanted to
compare the amount and the time related use of the question and answer
format with the amount and time related use of the volunteer format. The
use of a combined score was rejected because the investigator feared that
it might provide an inflated measure of information exchange. The
investigator chose to use the answer codes only because they were was
more similar in nature to the volunteer codes than were the question

codes.

The data from this coding scheme were analyzed with a 7 X 5 X 2
X 2 X 2 (content category by time block by type of statement (i.e.,
answer or volunteered) by condition by gender) analysis of variance.
This type of analysis would reveal main effects for all the independednt
measures considered and.indicate whether there were any interactios among
them. The results indicated a significant main effect for time, E
(1,156)-3.01. 6<.01 ( g s fr time blocks 1 through 5 were 8.6, 7.3, 4.5,
A.l, and 5.1, respectively) and for content ( M s for the interests,
school, peer, family, self-objective, self ability, and self-personal
categories were A.2, 9.5, 5.3, 3.2, 5.3, 2.2, and l.A, respectively).
Significant two-way interactions were observed bewtween gender and type

of statement,( fi (l,39)-3.57, p< .05, with the boys using an average of

48

1A.7 answers and 10.7 volunteered statements and the girls using 16.6
answers and 20.0 volunteered statements), time block and type of
statement, ( fi (A,156)-2.53, p< .05), content and gender ( L
(6,23A)-2.28, p< .03), and type of statement by content ( E (6,23A)-5.6,
p< .001). The means for the final three interactions are presented in

Tables A, 5, and 6.

Table A
Mean Frequencies of Statement Types For Each Time Block

Tyge 2: Statement

 

1153 Block Answer Volunteered
I 11.1 3.1.
2 7.1 3.9
3 5.5 3.A
A 4.0 1.9

5 A.5 ' 2.9

49

Table 5

Mean Frequencies of Content Categories by Gender

.c_e_nd_e;

Content 52y; Gigi;
Interests A.8 3.6
School 6.6 12.38
Family .8 5-7
Peer . 3.9 6.6
Self-Objective 6.A A.2
Self-Ability 1.6 2.7

self-Personal 1.3 l.A

50

Table 6

Mean Frequencies of Content Categories by Statement Types

Statement Tyges

Content Answers Volunteered
Interests 2.A 1.8
School 5.5 A.0
Peer 5.5 A.O
Family 1.5 1.7
Self-Objective 3.3 2.0
Self-Ability .2 1.9

Self-Personal .A .9

51

Discussion

The results taken from both coding schemes indicate the
following: 1. that for boys mutual behavior appears to have increased
and then decreased during the course of the their interaction whereas for
girls the amount of mutual interaction appears to have remained
relatively constant throughout the thirty minute session; 2. that the
amount of information exchange is greatest at the beginning of the
interaction, apparently because at the beginning, relative to the middle
and end, the children are more likely to ask many questions whereas, on
the other hand, they tend to volunteer information at about the same rate
throughout the thirty minute interaction; 3. that the amount of
information exchange varies according to the type of information being
discussed (i.e., information amount school and the self was discussed
more than information about peers, family and interests); A. that the
amount of information exchanged within a particular category appears to
be mediated by gender, with girls being more likely than boys to exchange
information about school,family and peers; and 5. that boys are more
likely than girls to acquire information through questions and answers
than through volunteering, whereas girls are more likely to volunteer it
than acquire it through question and answers. These results are
consistent with the results of the first study in that they point to the
preponderance of play and mutual activity in children's social
interactions, to the use of information search at the beginning of

interactions whereas information giving, relative to information seeking,

52

is more likely to occur in the later stages of an interaction, and to the
greater amount of information giving among girls than boys. these
results, then, are supportive of hypothesis four stated earlier. The

fifth hypothesis, however, was not supported as there were no significant

differences attributable to the experimental manipulation.

DISCUSSION
\3

These results provide support for the concept of scripts in that
they generally support hypotheses that were based on script theory
(Schank 8 Abelson, 1977: Schenk, 1981: Nelson, 1981). In this sense,
these findings are consistent with the notion that this event scheme
concept can be used to describe the processes of friendship formation
among children. Moreover, the current data lend credence to the
reinterpretaton of previous studies according to a script model.
Generally speaking, then, it appears that there is a is a script for
children's initial social encounter with another child that can be
observed in their actual behavioral interaction with another peer as well

as in their descriptions of how people become friends.

Of course, anumber of other issues (e. g., the association
between a particular person's scripts and the person's actual social
behavior, or the degree to which people with more orgianized or flexible
script are more skillful at predicting or understanding the behavior of
others) need to be addressed before the utility of the script model will
be adequately understood. Nevertheless, when the data presented here are
considered in conjunction with other theory and empirical work from the
psychological literature, it is possible to address at least two issues
related to the role of scripts in friendship development. The first of
these issues is related to the role of scripts in social interaction,
whereas the second has to do with whether or not persons actually use

scripts in the development of friendships.

53

54

Mg! might scrigts Mg valuable?

The results of the studies conducted in this investigation
revealed that the early stages of an interaction are typically
characterized by information searching whereas later stages typically
consisted of mutual activities or information giving. This pattern has
been taken as evidence of a script format in that it indicates a
sequential organization of the components of this type of social
interaction. Along with deciding whether or not these data support the
script model, it is equally important to determine what value this
particular pattern of interaction has for the process of friendship
formation. That is, if the script concept is to be regarded as
meaningful then it should be possible to point to something other than
its statistical "fit" to a set of date as support for its legitimacy. In
this section, an attempt will be made to identify the ways that the
concept of a-script is consistent with other theories related to

friendship and social interaction.

.In the past decade theory related to friendship and
interpersonal interaction has received a considerable amount of
attention. For example, Stephen Duck, a British social psychologist, has
conducted a number of empirical studies to investigate the role of
similarity in interpersonal attraction and friendship formation.

Overall, the results of his studies (Duck, 1982) demonstrate that

55

interpersonal similarity can be an important determinant of the formation
and maintainence of friendships. That is, people who are similar to one
another will be attracted to each other and, provided that they are
similar to one another on increasingly profound issues, their relation
will continue and/or become more intimate (Duck and Craig, 1978). Based
on these and other findings, Duck has argued that friendship development
will consist of a systematic gathering of information about a partner's
personality and that people will choose as their friends those persons
whom they consider themselves to be most similar to. The reason for the
importance attributed to similarity is the need for individuals to find
people who will reinforce their attitudes, habits, ways of thinking and
living and so on. In light of this arguement, it is not suprising that
information searching was especially apparent at the beginning of the
subjects' interaction. It may be that they were assessing the areas in
which they were similar and attempting to establish a common ground
activity. This search for common ground also may be be an important
prerequisite for the establishment of a mutual activity or a common play
activity. Indeed, Gottman (1983) noted that dyads of children who were
unable to establish a common ground were those who did not eventually

”hit it off."

The utility of initially obtaining information about another
person is also supported by recent evidence indicating that individuals
use person schemes in the processing of information about other people.
Cantor and Mischel (1977) have demonstrated that people appear to use

person schemes in their organization of information about a particular

56

individual and that the initial information that one receives about the
person is an important determinant for the particular scheme that is
chosen for this organization. For example, Cantor and Mischel (1977)
have reported that information about a person is processed differently
when the person was initially described as, say, an introvert relative to
when the person was originally described as an extrovert. Their
proposition can be interpreted as an indication that persons need
information at the beginning of an interaction in order that the
processing of subsequent information can be done according to

schema-based fashion.

Another use of initial information may be in the development of
expectations. Darley and Fazio (1980), have proposed that the "first
step in the interaction sequence is the perceiver's formation of an
expectancy regarding the behavior of (another) person" (p. 869). That
is, during the beginning of an interaction, persons are purported to
develop expectations about the specific behavior of another person or
about the person's intentions or dispositions that, in turn, are used to
predict the person's behaviors. Darley and Fazio (I980) cite two sources
of support for this notion. First, they draw upon the results of
actor/observer studies (Jones and Nisbett, 1971: Jones, 1976) indicating
that observers typically make judgements about a target person's traits
or motives based on their situational observation of the person even when
their is considerable evidence that there are environmental or temporary

factors affecting the person's behavior.

57 —

The second source of support that they cite is not related to
observations of the person's behavior but instead has to do with
information indicating to which class of individuals a person is assumed
to belong. As both Brigham (1971) and Hamilton (1979) have shown, a
person's knowledge that someone belongs to a particular racial or ethinc
group may lead to expectations of how the person will behave. In
addition, Snyder, Tanke and Berscheid (1977) have reported It is
important to recognize, however, that, in addition to racial and ethnic
stereotypes, expectations may be based on other types of information.
For example, Snyder, Tanke and Berscheid (1977) have reported that people
adjusted thier behavior in an interaction according to the information
they had been given about the person whom they are interacting with.
Specifically, college men who were told they were going to interact (on
the telephone) with a physically attractive women had different
expectations of this individual and subsequently interacted with her
differently than men who had been told they were interacting with a

physically unattractive person.

Just as there appears to be a number of reasons why an
information search should be found at the beginning of an interaction
there are also theoretical bases to explain the increased levels of
information giving and mutual play in the later stages of an interaction.
Principaily, Duck and Craig (1978) have proposed that individuals follow
a filtering process as their relationships develop. According to this

filtering model, relationships become increasingly intimate as they

58

develop provided that at each stage of the relationship the dyed members
are able to negotiate increasingly intimate demands. In addition,
Gottman (1983) has noted that pairs of children who were unable to become .
engaged in a mutual activity typically did not volunteer much information
to one another. Gottman interpreted this observation as indicating the
importance of dyad's to establish strong ties on increasingly intimate
levels before the relationship can become a friendship. Apparently, than
without the ties established during mutual play self-disclosure will be
unlikely. It appears, then, that there may an important reason for the
order of events that was observed in the scripts provided by the subjects
of this investigation. and to social interactions in general. This
order appears to be consistent with theory related to friendship

development and with social interaction in general.

22 232213 actually g3; scrigts lg friendshig?

Although the data presented here are consistent with the script
concept and with other theories pertaining to social interaction they
provide little information about whether people actually use scripts in
their behavior and they tell us very little about friendship as a special
form of relationship. In this section, the issues of whether persons
actually use scripts in a conscious manner and whether there is a special

script for friendship formation will be discussed.

59

The current data generally theory related to relationship
development but they are somewhat inconsistent with what one would expect
in light of the results of recent studies of children's friendship
expectations. Bigelow (1975, 1977). Furman and Bierman (1983) and others
(Youniss, 1979; Damon, 1979) have reported that children of the same age
as the subjects in this study, begin to describe friendship in terms of
constructs such as intimacy, loyalty and longevity. Nevertheless, in the
subjects' descriptions of how to make a friend these issues were rarely
mentioned. Play, the most frequently cited component, is actually what

one might expect of younger children.

A possible interpretation of this inconsistency suggests is that
the children simply don't have a distinct script for friendship
formation. That is, they simply do not think of the development of
friendships as different from other types of relations or, alternatively, .
that they do not have a particular script at all for making friends.
There are several reasons why this script may not exist. .First, in the
script concept it is implied that a pattern of behavior is goal oriented
or structured around a goal. It may be that a specific script for
friendship formation does not exist because friendship is not a product
of a particular process but instead is something that happens when a
certain set of conditions are met. That is, friendship may be more of a
"happening" than the result of a particular process. It is conceivable,
then, that the script concept may have greater utility for social

processes that are more goal specific, such as getting to know someone,

60

or establishing a work relationship and so on. For this reason the
script concept may be most appropriate for understanding how two persons
become acquainted. Of course, at an age when friendship may not be
distinct from a play relationship the script concept may have some value

as a way of understanding this type of relationship.

On the other hand. it may be that theprocess of friendship
formation is comprised of a series of subscripts. The data presented
here may shed light on the subscript that has to do with the initial
stage of an interaction. Perhaps after this "getting-to-know-you“ script
has been activated and completed, and provided that the interaction has
been satisfactory, persons may begin a subsequent script to further this
relationship. Perhaps this script has to do with the establishment of
more intimate forms of interaction (e. g. sharing secrets), the
continuation of self disclosure or the further identification of common
ground. Even later subscripts may serve to maintain the relationship,
add to its intimate nature or allow for conflict resolutions that will

strengthen instead of weaken the relationship.

It perhaps through this consideration of subscripts that the
value of Miller, Galanter and Pribram's cybernetic TOTE model is most
apparent. For example, as one script has been executed a person may
check to see if certain criteria have been reached (e.g., Do I feel
comfortable with this person?, Do we get along easily?, Do we share any
interests?, Do we have fun together?) and if so a subsequent script

could be activated. If the criteria were not achieved, then the script

61

could be repeated, the goal modidfied or even abandoned. In this
cybernetic fashion, a friendship may eventually develop, not out of a
single goal, but rather out of the achievement of a series of smaller

goals and scripts.

Just as the goal dependent character of scripts may be central
to the value of scripts scripts in the regulation of social behavior, it
may also explain why some person are experience greater social success
than others. If a child has a different set of goals for social
interaction, then it is likely that a different set of scripts or
strategies may be chosen. That is, different strategies will be used
because of the differences in goals. This issue poses an important
question for child-clinical psychologists who are concerned with
improving the peer experiences of of rejected and neglected children.
Simply coaching them on a certain set of skills or training them to
follow a certain script may not be productive if considerationis not
given to recognizing the child's goals or needs in social interaction.
Indeed, just as goals serve as the central focus of a script, the child's
own need and goal may be the starting point of understanding how they try
to make friends.

Along these lines, it is important to recognize that the
question of goals is central to the issue of how persons usexscripts.
Schenk and Abelson's (I977) enthusiasm for the script concept stems
mostly from the value it holds for the development of a kind of computer

program known as artificial intellegence. In the past they have

62

programmed computers to perform a number of human-like tasks such as
summarizing newspaper stories or thinking like, of all people, Barry
Goldwater. They found that the use of scripts as a means of telling a
computer about action and relationships was very productive. Based on
their success, these authors (Schank, in particular) have argued that the
script concept has enormous potential as a means of understanding the
human mind and human behavior. Clearly, teaching a machine to understand
newspaper stories is a pretty remarkable accomplishment. Nevertheless,
in their programs there is a important feature that distinguishes then
from human behavior. In programs the programmer can (and, in fact, is
required to) state goals very explicitly. In human behavior the goal may
be much less clear and in some instances some experiences may not be the
direct result of goal oriented behavior at all. Just as persons may not
have a specific goal to acquire a new cognitive structure, they may not
have specific goals to develop a particular friend either. Of course,
each of these may be influenced by some goal oriented actions but the

actual final experience may not be.

In regard to the question of whether persons actually
consciously use scripts, there is anecdotal evidence that in some
instances people may be unaware of how they go about some tasks. For
example, when the late Canadian pianist Glenn Gould, who is not only one
of the best known musicicians of the twentieth century and
unquestionnably among the greatest Bach interpretors of all time
(Rothstein, 1982), was asked to describe his method of creativity, he

replied:

63

Thats very difficult, and its one of those
centipedal questions- you know Schoenberg once
said that he would not willingly be asked by any
of his composition students exactly why
such-and-such a process served him well, because
the question made him feel like that centipede
who was asked in which order it moved its hundred
legs, and afterwards he could move no legs at
all- there's something impotent-making about that
question. I am rather afraid of it. (Gould,

1968) .
Nevertheless, in his work, Gould is know to have followed a number of
fairly rigid techniques to accomplis various and sundry tasks
associated with the production of recordings. For each of the
specific parts of making an album (the musical preparation, his
perfomance on the piano, the editting and mixing, and the final
preparation of the record for marketting) were conducted in an almost
ritualistic manner. But just as Gould did not have a specific plan
for making music, perhaps people do not have specific plan for
becoming friends. The lack of one large ultimate goal, however, may

be overcome by the successive achievement of smaller criteria.

An alternative explanation for the lack of findings
indicating how friendship differs from other types of social
interaction is the methodological constraints of both the studies
reported here. Perhaps the written answers were an inadequate means
of determining what was special about friendship in that they had to
do mostly with the initiation of an interaction. The solicitation of
answers having to do with the nature of relationships over a longer
period of time may have been more useful for identifying the unique

features of friendships. Furthermore, if there had been a question

64

about non-friendship relations it would have been possible to make
comparisons in order to discover how these two types of relations may
differ. These methodolgical changes may provide a more direct and

appropriate measure of children's scripts for friendship formation.

In regard to Study Two, methodological limits may be
the reason for the lack of success associated with the experimental
manipulation. Although the laboratory could conceivably be a place
where one could observe intimacy, it is difficult to determine how
persons were to display consructs such as loyalty or longevity. The
atypical setting in which the study took place (i.e., two early
adolescents cooped up in a little room, isolated from everyone else
and surrounded by mirrors, and then, to top it all off, a person whom
neither child knows tells half of them to become friends) may have
been too artificial to actually elicit what might be the behaviors of
friendship formation. Indeed, one of the subjects went so far as to
tell the investigator that he was crazy if he expected her to become

friends with someone in "that tiny little room.“

In light of these considerations it may be more
realistic to consider these results reported here as providing
information about how children initiate relationships instead of
regarding them as telling us something about friendship formation.
More generally, it may be that the script concept itself is more
appropriate for the initiation of interactions than it is for

explaining how relationships are maintained or further developed.

65

Moreover, it may be that they are most appropriate for behavioral
processes that are largely governed by social conventions. As Nelson
(1981) has previously noted, the most scripted events are probably
those that are straight out of etiquete books or those that are highly
ritualized such as a Japanese tea ceremony. Indeed, when this
investigator was a young adolescent his mother gave him an etiquette
book whose title was the epitome of a script. It was called §£ggg Mg;

Shake Hands m 59y "How 19 y_o_g_ do?" (Young 5 Buchwald, 1977).

REFERENCES

REFERENCES

Abelson, R.P. (1981). The psycholo ical status of the script concept.
The American Psychologist, 35(7 , 715-729.

Asher, S.R., & Gottman, J.M. (1981). The development of children's
friendships. New York: Cambridge UniversityPress.

Asher, S.R., & Renshaw, P.D. (1981). Children without friends:
Social knowledge and social skill training. In S.R. Asher &
J.M. Gottman (Eds.), The development of children's friendships.
New York: Cambridge university Press.

Baggett, P. (1979). Structurally equivalent stories in movie and text
and the effect of the medium on recall. Journal of Verbal Learning
and Verbal Behavior, 18, 333-336.

Bigelow, B.J. (1977). Children's friendship expectations: A cognitive
developmental study. Child Development, 48, 246-253.

Bigelow, B.J., & LaGaipa, J.J. (1975). Children's written descriptions
of friendship: A multdimensional analysis. Developmental Psychology,
11. 857-858.

Brewer, W., & Lichtenstein, E. (1980). Event schemas11§tory schemas and 2
story grammars. Technical Report No. 197:Center for the Study Off
Reading, University of Illinois, Champaign, IL.

Brigham, J.C. (1971). Ethnic stereotypes. Psychological Bulletin, 26,
15-38.

Bukowski, W.M., & Newcomb, A.F. (In press). The stability and
determinants of sociometric status and friendship choice: A
longitudinal study. Developmental Psychology.

Cantor, N., 8 Mischel, W. (1977). Traits as prototypes: Effects of
recognition memory. Journal of Personality and Social Psychology,
;§(1), 38-48.

Cohen, J. (1960). A coefficient of agreement for nominal scales.
Educational and Psychological Measurement, ;9, 37-46.

Darley, J.M., 8 Fazio, R.H. (1980). Expectancy confirmation processes
arising in the social interaction sequence. The American Psychologist,
35(10), 867-881.

DeCharms, R. (1968). Personal causation. New York: Academic Press.

66

67

Dodge, K.A. (1980). Social cognition and children's aggressive
behavior. Child Development, 51, 162-170.

Duck, S.W., 8 Craig, G. (1978). Personality similarity and the
development of friendship: A longitudinal study. British Journal
of Social Clinical Psychology, 11, 237-242.

Flavel, J.H. (1978). The development of knowledge about visual perception.
In C.B. Keasy (Ed.), The Nebraska Sym sium on Motivation, (V01. 25).
Lincoln, NB.: University of Nebraska ress.

Forbes, D., 8 Lubin, D. (1980). Reasoning and behavior in children's
friendly interactions. Paper presented at the meetings of the
AmericanTPsychological Association, New York.

Ferbes, D., 8 Lubin, D. (1981). The development of applied strategies
in children's social behavior. Paper presented at theIBiennial
Meeting of the Society for Research in Child Development, Boston.

Gottman, J.H. (1983). How children become friends. Mono ra h of the
Society of Research in Child Development, 48(3), Serial NE. 201.

Gottman, J.M., Gonso, J., 8 Rasmussen, B. (1975). Social interaction,
social competence and friendship in children. Child Development,
46, 709-718. '

Gould, G. (1968). Album cover comments from Glenn Gould: Concert
Dropout. New York: Columbia Records.

Hamilton, D.A. (1979). Cognitive-attributional analysis of psychology_
(Vol. 12). New York: Academic Press.

Harre, R. (1974). The conditions for a social psychology of childhood.
In M. Richards (Ed.), The inte ration of a child into a social
world. London: Cambridge UniversityPress.

Hartup, W.W. (1970). Peer interaction and social organization. In
P.H. Mussen (Ed.), Carmichael's manual of child psychology (3rd ed.,
Vol. 2). New York: Wiley.

Hartup, W.W. (In press). The peer system. In P.H. Mussen (Ed.-in-
chief) 8 E.M. Hetherington (Ed.), Carmichael's manual of child
psychology (4th ed., Vol. 4). New York: Wiley.

Heider, F. (1968). The psychology of interpersonal relations.
New York: Wiley.

Jones, E.E. (1976). How do people perceive the causes of behavior?
American Scientist, 55, 300-305.

Jones, E.E., 8 Nisbett: R.E. 71971). _The actor and the observer:
Diver ent perceptions of the causes of behavior. Morristown,
NJ.: General EearningPress.

68

Ladd, G. (1981). Effectiveness of a social learning model for enhancing
children's social interaction and peer acceptance. Child
Development, 5;, 171-178.

Ladd, G., 8 Mize, J. (1983). A cognitive-social learning model of
social skill training. Psychological Review, 59(2), 127-157.

Lichtenstein, E., 8 Brewer, W. (1980). Memory for goal-directed
events. Cognitive Psychology, 1;, 412-445.

Mandler, J.M. (1978). A code in the node: The use of a story schema
in retrieval. Discourse Processes, 1, 14-35.

 

Mandler, J.M., 8 Johnson, N.S. (l977). Remembrances of things parsed:
Story structure and recall. Cognitive Psychology, 5, 111-115.

Miller, G., Galanter, E., 8 Pribram, K. (1960). Plans and the
structure of behavior. New York: Henrv Holt.

 

Nelson, K. (1981). Social cognition in a script framework. In J.H.
Flavel, 8 L. Ross (Eds.), Social cognitive development. New York:
Cambridge University Press.

Newcomb, A.F., 8 Brady, J.E. (1982). Mutuality in boys' friendship
relations. Child Development, 55, 392-395.

Newcomb, A.F., 8 Bukowski, W.M. (In press). Social impact and social
preference as determinants of children's peer status.
Developmental Psychology.

O'Connor, R. (1972). Relative efficacy of modelling, shaping and the
combined procedure for the modification of social withdrawal.
Journal of Abnormal Psychology, 15, 327-334.

Peters, R.S. (I950). The concept of motivation. London: Routledge
and Kegan Paul.

Piaget, J. (1952). The origjns of intelligence in children. New York:
International University Press.

Piaget, J. (1968). The psychology of intelligence. Totowa, NJ.:
Littlefield, Adams.

Putallez, M. (In press). Predicting children's sociometric status from
their behavior. Child Development.

Putallez, M., 8 Gottman, J. (1981). An interactional model of children's
entry into peer groups. Child Development, 5;, 986-994.

Renshaw, P., 8 Asher, S.R. (In press). Children's goals and strategies
in social interaction. Merril-Palmer Quarterly.

69

'

Roistacher, P. (1974). A micro-economic model of sociometric choice.
Sociometry, 51, 219-238.

Rothstein, E. (1982, October 5). Glenn Gould, pianist, is dead:
Saw recordings as art form. The New York Times, 1, l7.

Rubin, K. (1973). Egocentrism: A unitary construct? Child
Development, 55, 102-110.

Rumelhart, D.E. (1977). Understanding and summarizing stories.
In D. LaBerge 8 S.J. Samuels (Eds.), Basic procg§sing in reading:
Perception and comprehension. Hillsdale, NJ.: Erlbaum.

Runyon, R.P., 8 Haber, A. (1971). Fundamentals of behavioral
statistics. Reading, MA.: Addison-Wesley.

Salinger, J.D. (1945). The catcher in the rye. Boston: Little, Brown.

 

Schenk, R.C., 8 Abelson, R.P. (1977). Scripts, plans, goals and
understanding, Hillsdale, NJ.: Erlbaum.

Schwayder, D.S. (1965). The stratification of behavior. London:
Routledge and Kegan Paul. ‘

Searle, J. (1980). Minds, brains, and programs. The Behavioral and
Brain Sciences, 5,

Searle, J. (1982, April 29). The myth of the computer. The New York
Review(of Books).

Selman, R. (1980). The growth of interpersonal understanding.
New York: Academic Press.

Snyder, M., Tanke, E.D., 8 Berscheid, E. (1977). Social perception
and interpersonal behavior: On the self fulfilling nature of social
stereotypes. The Journal of Personality and Social Psychology, 55,
656-666.

Stein, N.L., 8 Glenn, C.G. (1979). An analysis of story comprehension
in elementary school children. In R.O. Freedle (Ed.), New directions
in discourse processing (Vol. 2). Hillsdale, NJ.: Ablex.

Stein, N.L., 8 Goldman, S.R. (1981). Children's knowledge about social
situations. In J.H. Flavel 8 L. Ross (Eds.), Social cognitive
development. New York: Cambridge University Press.

Stein, N., 8 Nezworski, M.T. (1978). The effects of organization and
instructional set on story memory. Discourse Processes, 1, 177-193.
Tomkins, S. (1979). Script theory: Differential magnification effects.
In R.A. Dienstbier (Ed.), The Nebraska Symposium on Motivation
(Vol. 26). Lincoln, NB.: Univer§ity ofiNébrEEka Press.

70

Werner, H. (1961). Comparativecpsychology of mental development.
New York: Science Editions, Inc.

Young, M., 8 Buchwald, A. (1977). Stand up, shake hands and say
"How do you do?". Sacramento: Luce, Inc.

Youniss, J. (1979). The nature of social development. In H. McGurk
(Ed.), Social development. London: Methuen.