An assessment of the ecological validity of interpersonal cognitive problem-solving in elementary school children

MISSING IMAGE

Material Information

Title:
An assessment of the ecological validity of interpersonal cognitive problem-solving in elementary school children
Physical Description:
v, 86 leaves : ; 29cm.
Language:
English
Creator:
Kieffer, Michael L., 1957-
Publication Date:

Subjects

Subjects / Keywords:
Child Development   ( mesh )
Interpersonal Relations   ( mesh )
Personality Development   ( mesh )
Problem Solving -- Child   ( mesh )
Clinical and Health Psychology thesis Ph.D   ( mesh )
Dissertations, Academic -- Clinical and Health Psychology -- UF   ( mesh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1985.
Bibliography:
Bibliography: leaves 75-79.
Statement of Responsibility:
by Michael L. Kieffer.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000535819
oclc - 16865463
notis - ACV8828
System ID:
AA00009101:00001


This item is only available as the following downloads:


Full Text













AN ASSESSMENT OF THE ECOLOGICAL VALIDITY OF
INTERPERSONAL COGNITIVE PROBLEM-SOLVING
IN ELEMENTARY SCHOOL CHILDREN







BY


MICHAEL L. KIEFFER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE
UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY



UNIVERSITY OF FLORIDA


1985













ACKNOWLEDGMENTS


Acknowledging those people who have contributed to my graduate

school success is an easy and enjoyable task. First of all, many

thanks go to Larry Siegel, whose consistent support, guidance, and

expressed confidence in me enhanced my confidence in myself and

enabled me to be more productive, independent, and skillful in my

work. Second, thanks go to Jim Johnson, whose periodic and timely

advice and support helped me through difficult transition periods.

Third, I would like to thank the other members of my committee--

Rudy Vuchinich, Drew Bradlyn, and Janet Larsen--for their helpfulness

and their support. Finally, I thank, Fi, my wife, for her never-

ending support, patience, and, above all, her love, which helped see

me through the difficult times and made more enjoyable the good

times.














TABLE OF CONTENTS


ACKNOWLEDGMENTS. .

ABSTRACT .

INTRODUCTION .

ICPS and Sociome
Sociometric Find
Literature. .
Sociometric Find
Behavioral Asses
Problems in Soci
Potential Mediat
of ICPS Skill
Summary .

METHOD .

Subjects. .
Measures. .
Procedure .

RESULTS. .

Demographic Vari
Sociometric Grou
Sociometric Grou
Prediction of So
Translation of I

DISCUSSION .

REFERENCES .

APPENDIX .

BIOGRAPHICAL SKETCH.


tric Status . .
ings in the Social Competence

[ings in the Social Skills Literature.
sments of ICPS. . .
ometric Classification. .
ors of the Behavioral Manifestation





. .ables . . .
p Differences. . . .








p Differences by. . ...


ciometric . . ...

CPS Skill into Sociometric Status .


PAGE

ii

iv

1

4

7
9
13
25

29
30

32

32
32
43

47

47
49
53
56
59

62

75

81

86













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


AN ASSESSMENT OF THE ECOLOGICAL VALIDITY OF
INTERPERSONAL COGNITIVE PROBLEM-SOLVING
IN ELEMENTARY SCHOOL CHILDREN

By

Michael L. Kieffer

August 1985

Chairman: Lawrence J. Siegel
Major Department: Clinical Psychology


Assessments of the external validity of interpersonal cognitive

problem-solving (ICPS) have been few and inadequate. The present study

utilized sociometric and teacher-related behavioral criteria in inves-

tigating the external validity of the means-ends conceptualization of

ICPS in fourth, fifth, and sixth grade children.. The relative predic-

tive validity of ICPS for children's sociometric status was evaluated

against the predictive validity of both physical attractiveness and

academic achievement. In addition, self-report measures of anxiety,

depression, and perceived competence were utilized. It was found

that ICPS was not related to sociometric group status but was minimally

related to teacher-rated social competence. Physical attractiveness

was the best predictor of children's sociometric status. Evidence was

found which supports an interpretation that affective/motivational

factors interfered with children's translation of ICPS knowledge into







behavioral performance. Future ICPS validational studies are

necessary. These studies should investigate all ICPS conceptualiza-

tions in terms of both sociometric and behavioral validity criteria.

In addition, these validational studies should investigate the

viability of alternative ICPS conceptualizations, including chil-

dren's goals as well as strategies and children's responses to

unsolvable problems. Finally, the obtained significant relationship

between physical attractiveness and sociometric status suggests that

perhaps physical attractiveness should be incorporated as a treatment

component in social skills training studies.













INTRODUCTION


A number of studies have demonstrated a relationship between child-

hood social maladjustment and both concurrent and later impairments in

interpersonal functioning. Sociometric data suggest that children who

are rejected or not accepted by peers are more likely to interact

negatively with peers (Gottman, 1977; Hartup, Glazer, & Charlesworth,

1967), drop out of school (Ullman, 1957), become juvenile delinquents

(Roff, Sells, & Golden, 1972), present at community mental health

centers (Cowen, Pederson, Babigian, Izzo, & Trost, 1973), and receive

bad conduct discharges from the armed forces (Roff, 1961). Largely as

a result of the concurrent and predictive validity of these sociometric

data, there has recently been an increase in the number of attempts to

determine the salient interpersonal skills which contribute to the

social and emotional well-being of children (Hops & Greenwood, 1980).

These social skills assessment and treatment studies are predicated on

the assumption that there exists an interpersonal "skill deficit"

(Ladd & Mize, 1983). Deficits in both knowledge and performance of

skills have been implicated (Asher & Renshaw, 1981; Ladd & Mize, 1983;

Renshaw & Asher, 1982).

The present study deals with skill knowledge deficits--speci.fically,

deficits in interpersonal cognitive problem-solving (ICPS) skill.

Researchers in this area, particularly Shure and Spivak (1972, 1978)

and Spivak and Shure (1974), whose seminal work is exemplary of the







ICPS field, suggest that the ability to cognitively solve interpersonal

problems media-tes social adjustment. As postulated by these authors,

ICPS ability encompasses several skills, including (1) sensitivity to

the existence of a problem, (2) an ability to formulate alternative

solutions to a problem, (3) an ability to devise the means to solve a

problem, (4) an ability to consider the consequences of proposed solu-

tions, and (5) an awareness of the causes of one's behaviors, feelings,

and motivations as well as those of others. The ICPS skill is typically

assessed via verbal self-report instruments which measure the number of

conceptually different responses per category which children provide

to several interpersonal problems. Although Shure and Spivak suggest,

on the basis of their treatment research, that the cognitive ability to

solve interpersonal problems may be causally related to better social

adjustment, there are a number of problems with their research which

undermine the validity of this conclusion. Conger and Keane (1981) note

that these problems include (1) the fact that Shure and Spivak's

strongest data are derived from samples of black, inner-city, low

socioeconomic status (SES) preschoolers, (2) ratings of posttreatment

behavioral adjustment were made by nonblind teachers, (3) it is diffi-

cult to determine which trained skills accounted for their results, and

(4) the external validity of the behavioral adjustment measure has not

been established in terms of either independently rated behavior or

sociometric status. In addition, Kendall and Fischler (1984) note that

Shure and Spivak's validational efforts have involved comparisons

between groups of children broadly defined as disturbed and nondis-

turbed on the basis of scores on a single teacher rating measure, the

Devereaux Child Behavior Rating Scale (Spivak & Spotts, 1966). As a







consequence of these methodological problems, it is not clear that

Shure and Spivak's (1972) conception of ICPS is tapping a critical

cognitive dimension of social skill. Perhaps this is the reason that

the results of ECPS training programs appear to be equivocal at this

point in time (Rickel, Eshelman, & Loigman, 1983). In general, it

appears that treatment efforts are premature and of limited practical

utility (Urbain & Kendall, 1980) given that ICPS has at best a tenuous

empirical tie to sociometric status (which has demonstrated predictive

validity in terms of later interpersonal maladjustment) and to observ-

able behavior (which is objectively verifiable and appears to be the

data level upon which sociometric rating decisions are based). In

addition, the assessment methodology in these treatment studies is

inadequate given that the lack of assessment of pre- and/or post-

treatment correlations of ICPS with sociometric data or independently

rated behavior makes difficult the interpretation of the studies which

have found that significant improvement in ICPS skill has not been

accompanied by changes (i.e., improvements) in behavior or sociometric

status. In such a case, any of a number of interpretations could be

made: (1) lack of a functional relationship between the variables,

(2) insensitivity of the measures (Urbain & Kendall, 1980), (3) lack of

correspondence of ICPS training content with the content of the external

validity measures, (4) intractibility of the subjects' behavior problems

regardless of improvements in ICPS skill (Urbain & Kendall, 1980),

(5) pretreatment low level of significant behavior problems upon which

treatment might significantly impact (Urbain & Kendall, 1980), (6) lag

of positive changes in external validity measures behind ICPS gains,

(7) failure of generalization of ICPS gains to "real-life" situations







because of lack of ability (i.e., performance deficit) or lack of

incentive (Sharp, 1981), and (8) failure of generalization of ICPS to

the natural environment due to resistant group opinion regarding the

individual (which suggests that intervention may be necessary at the

level of the peer group)(Strain & Fox, 1981; Urbain & Kendall, 1980).


ICPS and Sociometric Status


Three studies have investigated the relationship between ICPS and

sociometric status. Richard and Dodge (1982), utilizing nonreferred,

second through fifth grade boys of unspecified SES and race (but

apparently predominantly white, middle class), found that more poorly

adjusted children [i.e., those children classified as aggressive or

withdrawn via a combination of positive and negative peer nominations,

peer-rated behavior description scores (e.g., isolated, aggressive, or

cooperative), and teacher ratings (favorability of social relationships,

frequency of aggressive and prosocial behaviors emitted)] offered fewer

solutions to interpersonal problems and a significantly lower percen-

tage of effective solutions after the initial solution than did well-

adjusted children. Their results suggest that ICPS skills may be

important mediators of interpersonal adjustment in children who are

demographically similar to those employed in their study. Unfortunately,

the relationship between ICPS and sociometric status cannot be ac-

curately assessed in this study because sociometric status (i.e.,

positive and negative peer nominations) was combined with peer-rated

behavior description scores and teacher ratings in defining interper-

sonal adjustment.





5

In an ICPS training study, Weissberg, Gesten, Rapkin, Cowen,

Davidson, Flores de Apodaca, and McKim (1981), utilizing nonreferred

samples of black, low SES, urban and white middle SES, suburban third

grade children and controlling for age, sex, race and SES differences

between treatment children and no-treatment controls, found that ICPS

training was effective in terms of problem identification, alternative

solution thinking, and consequential thinking on the Rochester Means-

Ends Problem-Solving Test (a modification of Shure and Spivak's, 1972,

Means-Ends Problem-Solving Procedure--MEPS), but produced no changes on

a sociometric rating scale in either treatment group. In another train-

ing study, Allen, Chinsky, Larcen, Lochman, and Selinger (1976) also

failed to find a relationship between training-improved ICPS skill

(alternative solution thinking and total score on a problem-solving

measure analogous to Shure and Spivak's MEPS) and a sociometric rating

of friendship in a sample of predominantly white, middle-class, third

and fourth grade children in a normal school setting.

The results of the latter two training studies are difficult to

integrate with Richard and Dodge's (1982) assessment study given the

variation in definitions of social adjustment and in sample characteris-

tics. In addition, the fact that the treatment studies failed to find

a significant sociometric improvement of ICPS-trained children over

untrained children cannot be clearly interpreted for the reasons pre-

viously enumerated regarding the difficulty of interpreting such nega-

tive results. Given that no data have been collected to support the

existence of a relationship between ICPS and sociometric status, it is

clearly indicated that future research should investigate to what extent







sociometric status and ICPS are related--and should do so independently

of efforts to train ICPS skill.

In summary, Richard and Dodge (1982) found ICPS differences between

groups of children classified in terms of interpersonal adjustment on

the basis of combined peer and teacher ratings. It is unclear whether

group differences would have been obtained had children been classified

solely on the basis of sociometric ratings. The failure of Allen et al.

(1976) and Weissberg et al. (1981) to demonstrate a functional rela-

tionship between training-improved ICPS and sociometric status provides

little information on the relationship between ICPS and sociometric

status given that there are a number of interpretations for this failure.

Thus, it is important at this point in the development of the ICPS

literature to investigate to what extent (1) there is a correlational

relationship between ICPS and sociometric status, and/or (2) groups of

children which differ in sociometric status also vary in terms of ICPS

skill. In addition, and perhaps most importantly, the assumed mech-

anism by which ICPS affects sociometric status needs to be more

thoroughly investigated. This mechanism, which appears to underlie any

theory in which a particular cognitive skill is hypothesized to affect

sociometric status, proposes that children's ICPS ability is expressed

in behavior which is perceived by peers, forming the data base for

their sociometric ratings. The present study, given its correlational

nature, was limited to speculation in terms of causal attributions

regarding the processes within the mechanism. However, the magnitude

of the correlations between ICPS scores, independently rated behavior,

and positive and negative peer nominations did allow general, noncausal







formulations regarding the processes by which ICPS skill does or does

not affect positive and negative peer nominations.


Sociometric Findings in the Social Competence Literature


Although not explicitly addressed by researchers in the ICPS

literature, the growing volume of research regarding children's social

competence addresses issues similar to those with which the ICPS

literature is concerned and also appears to assume a similar underlying

mechanism between children's cognitive competence and its effect on

their sociometric status. Social competence researchers have primarily

investigated the manner in which children develop a friendship and, to

a lesser extent, the manner in which they maintain relationships with

and solve conflicts with peers. Within this literature, at least two

studies have investigated the relationship between children's socio-

metric status and their ability to provide solutions to interpersonal

situations involving uncertainty or potential conflict.

Renshaw and Asher (1983) obtained significant sociometric findings

while investigating the interpersonal strategies of third through sixth

grade children who had been classified into high- and low-status groups

on the basis of positive or negative z scores, respectively, derived

from peer play ratings. These authors studied children's interpersonal

strategies in response to four sets of hypothetical social situations:

initial contact, entry, friendship, and conflict. Although no signifi-

cant sociometric effect was found for the four situations combined, in

the conflict situation, high-status children suggested significantly

more positive-accommodating and rule-oriented strategies than did low-

status children. High-status children's responses were generally







less aversive and left open the possibility of continuing a friendly

interaction.

Asher and Renshaw (1981) investigated high- and low-status kinder-

garten children's responses to nine hypothetical interpersonal situa-

tions, which included initiating social relationships, maintaining a

conversation, and management of conflict. The high-status group was

composed of children who received a large number of positive peer nomina-

tions and a high average play rating. Low-status children were those

who received few positive nominations and a low play rating. The

authors note that there was considerable overlap (70%) between the three

most common strategies given by both popular and unpopular children to

each of nine situations. However, two findings highlighted the dif-

ferences between these groups: (1) "strategies that were suggested

exclusively by unpopular children were often inappropriately negative,

whereas this was never the case among the popular children" (p. 279),

and (2) in terms of percentage usage, a) on the items concerning manage-

ment of conflict, unpopular children were much more likely than popular

children to suggest aggressive strategies (12% of unpopular children's

responses versus 2% of popular children's responses), and b) on the

items concerning initiation of social relationships or maintenance of

ongoing relationships, unpopular children were much more likely than

popular children to suggest vague ideas or to appeal to authority

figures. As Asher and Renshaw (1981) note, "the unpopular children's

responses indicate a lack of sophistication in handling the everyday

social tasks of childhood" (p. 279).

In these two studies, primary consideration in terms of group

differences was given to qualitative differences as opposed to







quantitative differences (e.g., number of solutions), which are the

primary focus of the ICPS literature. The results suggest that, rela-

tive to unpopular children, popular children may have a more quali-

tatively sophisticated knowledge of appropriate social behavior.

Popular children are apparently more adept at handling hypothetical

conflict situations in a positive, relationship-enhancing fashion. The

latter findings appear to substantiate the claims by ICPS researchers

that the ability to cognitively solve interpersonal problems is related

to better social adjustment--in this case, better sociometric status.


Sociometric Findings in the Social Skills Literature


An additional potential source of clarification regarding what

might be expected from future research on the relationship between ICPS

and sociometric status may be contained in social skills training

studies which have employed a cognitive component in the training pro-

gram and have assessed posttreatment group differences in sociometric

status. Although the specific content of the skill knowledge (i.e.,

cognitive) component employed in each of these social skills training

studies, unlike the ICPS studies, does not contain information regarding

how to deal with peer conflict situations, it is clear that both sets

of research have an interpersonal, cognitive focus and view inter-

personal, cognitive training as an essential treatment component in

the modification of children's social adjustment (i.e., their socio-

metric status). That is, both sets of research assume a similar

underlying process: cognitive ability is expressed in terms of

behavior which is perceived by peers and constitutes the data on which

they base their sociometric choices. Instead of simply assuming the







operation of these processes, however, these social skills training

studies are designed in such a fashion as to test the effect of manipu-

lation of children's social cognitions on their sociometric status. The

purpose of the presentation of these studies is to address the ques-

tion regarding the validity of the hypothesis that cognitive instruc-

tion regarding interpersonal behavior (which is what ICPS training

studies purport to do, but have been generally unsuccessful in doing)

can result in bettered social adjustment. A brief summary of social

skills training studies employing cognitive components may provide

additional data regarding (1) to what extent ICPS skill is related to

peer acceptance, and (2) whether intervention studies which employ an

ICPS component should be expected to produce changes in children's

levels of peer acceptance. Social skill training methods which employ

cognitive components include modeling, instruction, coaching, and

verbal rehearsal.

Gottman (1977) found that 32 Head Start children (defined as

socially withdrawn on the basis of low rate of interaction) who viewed

a modeling film developed by O'Conner (1969) did not differ from with-

drawn control children in terms of sociometric status at 7 weeks

following treatment.

Gottman, Gonso, and Schuler (1976) assigned two socially withdrawn

third grade females (defined as withdrawn on the basis of a "best

friend" sociometric instrument) to a treatment condition consisting of

a combination of exposure to a modeling film which included self-coping

statements and strategies in the narrative (similar to the O'Conner,

1969, film), coaching by an adult on making friends, and training on a

perspective-taking task. Two other withdrawn females were assigned







to a control condition which consisted of participation in games and

discussion. Results showed significant improvement on a play rating

sociometric measure for both trained children as well as improvement on

a work rating measure over a 9-week follow-up.

Oden and Asher (1977) trained third and fourth grade children who

had been defined as low peer-accepted via peer play and work ratings.

Children received six 1-hour sessions of instruction, rehearsal, and

review of skill concepts (i.e., participation, cooperation, communica-

tion, and validation/support) in addition to being encouraged to

generalize the usage of these skills with peers to the classroom

setting. Results showed significant sociometric gains on the peer play

ratings measure which were maintained at a 1-year follow-up. However,

gains on the peer work rating measure and the friendship nomination

measure were not significant. In addition, the significant increase by

trained children on the play rating measure is difficult to interpret

given that experimenter attention was uncontrolled (i.e., no attention-

placebo group was used).

Gresham and Nagle (1980) compared a coaching procedure similar to

that of Oden and Asher (1977) to a filmed modeling condition and to a

combination coaching-modeling condition. Training was conducted during

six 20-minute sessions over a 3-week period. Results showed that

middle-class, third and fourth grade children who were low on socio-

metric work and play ratings improved in one of five sociometric status

measures (i.e., play rating, but not work rating, work nomination or

work, play, or best friend choices) relative to control group children.

However, as with the Oden and Asher (1977) study, it may be that







children's sociometric gains were the result of experimenter attention,

given that attention appears to have been inadequately controlled.

In a modification of the Oden and Asher (1977) study, Ladd (1981)

instructed low peer-accepted (defined via the peer play and work ratings

used by Oden and Asher, 1977), middle-class third grade children in

three relatively specific skill concepts (asking questions, leading

peers, and offering support), gave guided rehearsal and feedback, and

encouraged children in self-directed rehearsal, self-evaluation of skill

performance, and performance-enhancing attributions for success and

failure. Training consisted of eight 45- to 50-minute sessions con-

ducted on alternating school days. Compared to attention-placebo con-

trols, trained children showed significant gains on two of the three

behavioral skills (asking questions and leading peers) and in socio-

metric status which were maintained at a 4-week follow-up.

La Greca and Santogrossi (1980) found that low peer-accepted third,

fourth, and fifth grade children who viewed videotapes (in groups which

met for 2 hours per week for 4 weeks) followed by a review and dis-

cussion, received guided rehearsal and videotaped feedback, and were

encouraged to use the trained skills (i.e., smiling/laughing, joining

conversations, extending invitations, greeting others, conversational

skills, sharing and cooperation, complimenting, and physical appearance/

grooming) with classmates at school, had greater social knowledge (as

measured by their responses to interpersonal problem-solving tasks),

but made no gains in sociometric status compared to nontrained children.

Ladd and Mize (1983) suggest that La Greca and Santogrossi's (1980)

failure to find sociometric gains--as compared to the positive socio-

metric findings of Oden and Asher (1977), Gresham and Nagle (1980), and







Ladd (1981)--may be due to the age composition of the sample. Specifi-

cally, La Greca and Santogrossi (1980) utilized a substantial number

of fifth grade children while the other authors employed third and/or

fourth grade children.

All but one of these social skills training studies support the

hypothesis that cognitive instruction regarding interpersonal behavior

can increase third and fourth grade children's peer-rated social

adjustment (i.e., sociometric status). Thus, support is provided for

the validity of the proposed mechanism underlying researchers' focus on

ICPS skill in children: cognitive ability is expressed in terms of

behavior which is perceived by peers and forms the basis for their

sociometric choices. However, the extent to which the sociometric

changes are the result of improvements in skill knowledge is uncertain

given that each of these studies except the La Greca and Santogrossi

(1980) study (which found that improvements in skill knowledge were not

accompanied by sociometric improvements) did not assess posttreatment

group differences in skill knowledge.


Behavioral Assessments of ICPS


The relationship between independently rated (i.e., directly

observed or teacher-rated) behavior and ICPS has been investigated in

relatively few studies. A demonstration that such a relationship

exists and is significant would appear to be essential, given that,

if ICPS skill affects children's sociometric status, it is children's

behavioral manifestation of their ICPS skill on which peers base their

sociometric choices. Butler and Meichenbaum (1981) note that "the lack

of attention to behavioral validation is a serious omission in the







social problem-solving literature, in view of the emphasis in ICPS

theory on the role ofsocial problem-solving cognition as a mediator of

social and behavioral adjustment" (p. 215). Only one of these studies

(Sharp, 1981) which has investigated the relationship between ICPS skill

and behavior has included an assessment of pre- and posttreatment

correlations between these variables as opposed to an assessment of

within-group pre-post change scores and posttreatment group differences.

Allen et al. (1976), utilizing nonreferred samples of third and

fourth grade children of predominantly white, middle-class origin,

found that children trained in ICPS generated significantly more

alternative solutions and obtained a significantly higher total score

on a problem-solving measure similar to Shure and Spivak's (1972)

Means-Ends Problem-Solving Procedure than did no-treatment controls.

Trained children also tended to show greater improvement on the number

of steps generated for each solution but not on the number of obstacles

cited to solutions. These ICPS gains were not accompanied by teacher

rated behavioral adjustment improvements on an abbreviated version of

the Walker Behavior Problem Identification Checklist. On the Struc-

tured Real-Life Problem Situation (SRLPS), which involved the child's

independently rated behavioral problem-solving skills in a child-adult

interaction, the authors found, as Butler and Meichenbaum (1981) note,

"that although significantly more problem-solving trained children

generated two solutions on a verbal test, training and control groups

did not differ significantly in the substantial number of children in

both groups who failed to generate even one solution to the actual

problem" (p. 216). Allen et al. (1976) note that the low response rate

of subjects might not necessarily have stemmed from inadequate








problem-solving skill but rather could have resulted from children's

possible perception of the problem as one primarily for the adult

experimenter to solve. The results on the SRLPS are difficult to

interpret given that group differences, although significant, were not

marked and that problem-solving performance was assessed in a child-

adult dyad as opposed to a peer dyad. The latter limitation is a result

of the fact that although a child's peer relations are predictive of

later social maladjustment, the behaviors of interest in this study

were assessed in the context of a child-adult interaction (i.e., an

authority situation) rather than a peer interaction. This distinction

appears to be an important one in terms of both the previously stated

reason and on the basis of Elias, Larcen, Zlotlow, and Chinsky's (cited

in Butler & Meichenbaum, 1981) findings that in hypothetical problem-

solving situations involving an authority figure (as opposed to a peer),

children's means, means in response to obstacles, plans, and unique

means were significantly fewer in number, problem-solving strategies

were qualitatively different, and their expectations regarding outcomes

were also more negative. Given the demonstration of Elias et al. of a

disparity between children's cognitions in problem-solving situations

involving adults as opposed to peers--and the potential extension of

these differences to overt behavior, it appears that the findings of

Allen et al. (1976) regarding children's problem-solving behavior in a

child-adult interaction might not be representative of those obtained

in a peer interaction. Another limitation of this study is that the

social problem-solving treatment condition, which involved a consider-

able amount of adult attention, was not evaluated against an attention-

placebo group. As such, attentional factors may have constituted







an active treatment component--in addition to or instead of the ICPS

training.

McClure, Chinsky, and Larcen (1978) found that nonreferred rural

and suburban third and fourth grade children trained in one of three

ICPS programs (video modeling tapes, video modeling tapes plus dis-

cussion exercises, or video modeling tapes plus role play exercises) as

a whole generated more solutions and a greater number of effective

solutions than did no-treatment control children. Behavioral changes

were assessed two months posttreatment in a "Friendship Club interaction"

in which individual groups of six children were observed on (1) two

explicitly stated social problem-solving tasks which demanded the

cooperation of all members and in which children were rewarded for

good performance and (2) three tasks embedded in the interaction set-

ting: (a) five chairs for six subjects, (b) five officer cards for six

subjects, and (c) the process of distributing officer titles. This

observational assessment revealed that, relative to controls, in three

of the five situations, at least one of the three trained groups gener-

ated a greater number of alternative solutions, and, in two of the

five situations, at least one of the groups generated a greater number

of specific steps for each solution. These relatively weak positive

findings were accompanied by no significant group differences on an

additional behavioral problem-solving task involving a child-adult

interaction. The validity of the authors' conclusion regarding the

effectiveness of ICPS training in improving behavioral problem-solving

performance is reduced by the fact that the training groups were evalu-

ated against a no-treatment control group instead of an attention-placebo

group.







Weissberg et al. (1981) found that nonreferred samples of black,

low SES, urban preschoolers and white, middle SES, suburban third grade

children who showed training-related improvements in ICPS skill (1) re-

ceived higher posttreatment ratings by teachers on the Health Resources

Inventory (Gesten, 1976) and the Devereaux Elementary School Behavior

Rating Scale (Spivak & Spotts,1966), and (2) tried more solutions and

persisted longer (the latter was a marginally significant finding,

p <.08) than matched controls in behavioral problem-solving performance

with an adult confederate. These authors suggest that "cognitive

problem-solving skills acquired through training generalize to relevant

behavioral situations removed in time, place, and person" (p. 259). A

limitation of behavioral results of Weissberg et al. (1981) is that,

as in Allen et al. (1976), children's behavioral problem-solving per-

formance was assessed in the context of a child-adult interaction

instead of a peer interaction. In addition, it is difficult to identify

the active treatment component in this study given that an attention-

placebo group was not utilized.

Ridley and Vaughn (1982) found that, relative to a control group,

middle-class preschoolers trained in an expanded version of Spivak and

Shure's (1974) training program, including empathy training, generated

a greater number of peer problem solutions on the Preschool Inter-

personal Problem-Solving Test (PIPS) (Shure, Spivak, & Jaeger, 1971)

at both posttest and at a 3-month follow-up. These ICPS gains were

accompanied by a greater number of solutions generated in a real-life

problem situation involving a peer confederate and a greater percentage

of relevant (as opposed to irrelevant) solutions on this structured

behavioral test of social problem-solving. In light of the







predominantly negative findings between ICPS and behavioral problem-

solving obtained by other authors, a possible explanation for the

positive relationship obtained by Ridley and Vaughn (1982) is (1) their

assessment of each child's behavior in an actual problem-solving situa-

tion involving a peer (as opposed to assessment of each child's general

classroom behavior or assessment of each child's problem-solving

behavior in a child-adult interaction) and (2) the high correspondence

between the content of the verbal problem-solving measure and the

content of the behavioral assessment situation. A question is raised

regarding the active ingredient of Ridley and Vaughn's (1982) training

program given that their control group, which was apparently designed

as an attention-placebo group, did not receive attention from the

experimenter (as the training group did) but rather received attention

from the regular classroom teacher, whose attention may have held less

stimulus value than the attention of the experimenter.

Rickel et al. (1983) found, utilizing black, low-income preschool

children who had been classified as adjusted or maladjusted (i.e.,

inhibited or impulsive) on the basis of teacher ratings on the Hahnemann

Preschool Behavior Rating Scale, that trained subjects made significant

gains in terms of alternative solutions and consequential thinking

but made no concomitant gains on independent observers' ratings of

classroom behavior on the Sharp Behavior Identification Checklist or

blind teacher ratings on the Hahnemann scale. The authors note that

the lack of behavioral effects may have been the result of the absence

of a consistent, strong relationship between ICPS and behavior at the

preschool level or the result of an inadequate or insufficient training

program. In addition, the authors' expectation that ICPS training would







have an effect on children's general classroom behavior as opposed to

effects on their behavior in social problem-solving situations may be

unrealistic. It also appears that trained children's lack of behavioral

gains is difficult to interpret for the same reasons enumerated regard-

ing the difficulty in interpreting the predominantly negative findings

concerning sociometric status and ICPS. Finally, the fact that all four

groups of Rickel et al. (1983) (i.e., adjusted and maladjusted children

in both the training and attention-placebo conditions) increased on

alternative solution and consequential thinking (which the authors

suggest may be due to normal maturational processes and general educa-

tional stimulation) also suggests that the component which was active

in increasing both training and attention-placebo subjects' problem-

solving skill was an attentional component, thus reiterating the impor-

tance of attention-placebo groups in treatment outcome research.

In a methodologically more rigorous study investigating the rela-

tionship between social problem-solving skill and directly observed

behavior, Sharp (1981) attempted to replicate Shure and Spivak's (1974)

results via utilizing the latter authors' ICPS training program as well

as a subject sample designed to be as similar to their sample as pos-

sible (i.e., black, low-income preschoolers). A unique and important

aspect of this study is that both pre- and posttreatment assessment of

the relationship between ICPS and behavior were conducted. Sharp's

(1981) findings strongly suggest that (1) ICPS skill has no direct

relationship to either teacher rated behavior on the Hahnemann scale

or behavior ratings on the Sharp Behavior Identification Checklist

and (2) ICPS training, which increased children's ability to generate

alternative solutions (although it did not increase their ability to







conceptualize the potential consequences of an act) did not improve

preschoolers' independently observed behavioral competency. In view of

the obtained negative findings regarding the relationship between ICPS

and behavior, Sharp (1981) notes that Shure and Spivak's (1974) hypothe-

sis that positive behavioral adjustment in preschoolers is mediated by

the ability to cognitively generate a large number of alternative

solutions to interpersonal problems must be questioned.

In summary, in terms of the findings regarding the relationship

between teacher ratings of behavioral adjustment and ICPS skill, three

of the four studies which employed teacher ratings (excepting Weissberg

et al., 1981) found that training-improved ICPS skill was not reflected

in improved behavioral adjustment scores. The positive findings of

Weissberg et al. (1981) may be a result of the fact that they used two

teacher rating measures which have reliable and valid factor structures

and which may, thus, be more sensitive and psychometrically sound

instruments than those used by the other authors. For instance, Sharp

(1981) and Rickel et al. (1983) found no relationship between ICPS skill

and the Hahnemann scale, a psychometrically poor measure which con-

sists of only seven items (which load on one of three factors) and has

very little data to support its reliability and validity.

In terms of the overall findings regarding directly observed

behavior, although each of the studies investigating the relationship

between ICPS and behavior have demonstrated some form of improvement in

ICPS skill, only two studies (McClure et al., 1978; Ridley & Vaughn,

1982) have assessed behavior in the context of a problem-solving situa-

tion with a peer. This type of assessment appears to be much more

relevant than both problem-solving behavior in a child-adult dyad and







general (i.e., nonproblem-solving) classroom behavior. Both of these

studies found behavioral differences between treatment and control

children. However, McClure et al. (1978) failed to employ an attention-

placebo group, and Ridley and Vaughn (1982) used an inadequate attention-

placebo group.

Thus, these studies failed to demonstrate that behavioral gains were

the result of training-induced improvement in ICPS skill as opposed to

experimenter attention. These studies would apparently have been more

elucidating regarding the relationship between ICPS and behavior had

pre- and posttreatment correlations of this relationship been conducted

--given that conclusions regarding the functional relationship between

these variables are confounded with experimenter attention.

Two studies (Allen et al., 1976; Weissberg et al., 1981) found that

ICPS improvements were accompanied by behavioral improvements in the

context of a child-adult interaction. However, the generalizability of

these findings to peer interaction situations, as previously suggested,

is questionable at best, given that the child's perception of his/her

role vis-a-vis an adult as opposed to a peer may influence his/her

behavior. This analysis is supported by the fact that McClure et al.

(1978) found that ICPS improvements were accompanied by behavioral

improvements in children's problem-solving interactions with peers,

but not in their problem-solving interactions with adults.

Two studies (Rickel et al., 1983; Sharp, 1981) utilizing attention-

placebo groups found no behavioral differences between trained children

and control children on a measure of general classroom behavior. This

finding may not be surprising given that it may be unrealistic to

expect ICPS training and demonstrated improvement on ICPS skill to







generalize beyond specific social problem-solving behavior to behavior

in nonproblem-solving situations (i.e., general classroom behavior).

The comparability of studies is generally very limited given the

variation in sample characteristics, assessment formats (Butler &

Meichenbaum, 1981), assessment measures, assessment context, and ade-

quacy of internal control as well as the previously described diffi-

culty in interpreting the negative results of the investigations which

failed to find behavioral gains as a result of ICPS treatment (disregard-

ing, for the moment, the adequacy of internal control of the studies).

In summary, although there is no conclusive evidence supporting a

relationship between ICPS skill and independently rated behavior, the

findings of Ridley and Vaughn (1982) and McClure et al. (1978)--although

limited by an inadequate and absent attention-placebo group, respec-

tively--suggest that specific training-induced improvements in ICPS

skill produce concomitant changes in children's behavioral problem-

solving skill with peers. Beyond these studies, however, little conclu-

sive data can be assembled given the previously described problems of

insufficient internal control as well as the failure to assess problem-

solving behavior in a peer interaction. The correlational relationship

between ICPS and peer problem-solving behavior as opposed to general

classroom behavior has apparently not been assessed in any population.

The closest approximation to this ideal is Sharp's (1981) pre- and

posttreatment assessment of the relationship between ICPS and general

classroom behavior. Finally, teacher ratings of behavioral adjustment

have, in three of four studies, failed to reflect significant ICPS

gains. As was suggested before, failures to demonstrate a functional

relationship between training-improved ICPS skill and behavioral







adjustment provides little readily interpretable information on this

question, given that there are a number of interpretations for these

failures.

The presence of a significant causal relationship between ICPS

skill and independently rated behavior has received only tentative

support in the studies reviewed. The hypothesis that ICPS is a sig-

nificant contributor to sociometric status, which is the focus of the

present study, however, demands that ICPS skill be reflected in behavior

if it is to be perceived by peers and thus influence their sociometric

ratings. Regarding the first portion of this postulated mechanism by

which ICPS skill affects sociometric status, given that previous studies

have not conclusively established that ICPS skill is translated into

actual behavior, the present study attempted to provide evidence, in

postulating a mechanism by which ICPS affects sociometric status, that

varying degrees of ICPS skill are reflected in corresponding degrees of

independently rated skill in terms of behavioral manifestation of ICPS

skill (i.e., behavioral adjustment). In terms of the second half of

this postulated mechanism regarding the potential effect of ICPS behavior

on sociometric status, given that little data have been generated

regarding the extent to which behavioral manifestations of ICPS are

related to sociometric status, the present study assessed the degree of

relationship between teacher-rated behavioral adjustment and sociometric

status. Thus, given (1) the sparse and relatively inconclusive data

regarding the relationship between (a) ICPS skill and behavioral

adjustment and (b) behavioral adjustment and sociometric status, and

(2) the fact that the plausibility of the proposed mechanism by which

ICPS and sociometric status are related depends upon the strength







of the relationships obtained between these two sets of variables, the

present study used a teacher rating measure of behavioral adjustment in

order to gauge (1) the extent to which (a) ICPS skill is reflected in

behavioral adjustment and (b) behavioral adjustment is related to

sociometric status, and (2) the plausibility of the hypothesis that

ICPS skill affects sociometric status. Given that the data which are

derived from this analysis of the process by which ICPS skill affects

sociometric status are strictly correlational in nature, the causal

nature of this relationship can only be speculated upon by way of

examining the strength of the component processes by which ICPS skill

and sociometric status are hypothesized to be related.

An additional method of assessing the potential importance of ICPS

skill to sociometric status is to assess the extent to which ICPS skill,

relative to other variables which have been previously found to be

related to sociometric status, predicts sociometric status. Hallinan

(1981) reports that variables which have been found to discriminate

between popular and unpopular children include physical attractiveness,

intelligence, academic achievement, athletic ability, and social class.

For instance, the results of Kleck, Richardson, and Ronald (1974) and

Langlois and Stephan (1977) suggest that children (ranging in age from

4 to 14 years across these studies) who were socially accepted by peers

received independent ratings of physical attractiveness by peers which

were significantly higher than those of low-accepted children. In

addition, Green, Forehand, Beck, and Vosk (1980) found, in a sample of

rural third-grade children, that academic achievement was significantly

positively correlated with positive nominations (r =.334) and negatively

correlated with negative nominations (r=-.205). The present study







assessed the relative extent to which ICPS, physical attractiveness, and

academic achievement predicted individual's sociometric status and

discriminated among groups differing in terms of sociometric status.


Problems in Sociometric Classification


The sociometric classification system designed by Coie, Dodge,

and Coppotelli (1982) and modified by Coie and Kupersmidt (1983) was

used in investigating the relationship between ICPS skill and sociometric

status. This system seeks to deal with the problem of behaviorally

heterogeneous groups which has plagued past sociometric classification

systems. It employs both positive and negative nominations and con-

ceptualizes sociometric status in terms of two dimensions: social

preference and social impact. Social preference, which is a variation

on traditional conceptions of sociometric status, consists of the

score obtained by subtracting negative nominations (e.g., "who do you

like the least?") from positive nominations (e.g., "who do you like the

most?"). Social impact consists of the score obtained by adding both

positive and negative nominations. The social impact score is designed

to assess the "impact" of a particular child on other children. From

these social preference and impact dimensions are derived five distinct

sociometric groups: (1) a popular group which consists of children

with a social preference standardized score greater than 1.0, a like

most standardized score greater than 0, and a like least standardized

score less than 0; (2) a rejected group which consists of children who

receive a social preference standardized score less than -1.0, a like

least standardized score greater than 0, and a like most standardized

score less than 0; (3) a neglected group which consists of children with







a social impact standardized score of less than -1.0 and an absolute

like most raw score of 0; this group differs from the rejected group in

that children in the latter group receive many nominations as liked

least, whereas neglected children do not; (4) a controversial group

which consists of children with a social impact standardized score of

greater than 1.0 and like most and like least standardized scores which

are each greater than 0; and (5) an average group which consists of

children with social preference and social impact standardized scores

which are both greater than -0.75 and less than 0.75. This definition

of the average group is more stringent than that of Coie and Kupersmidt

(1983) who specified limits of -1.0 and 1.0 for the social preference

and social impact dimensions. This classification system has been

validated in terms of peer-rated behavior (although the behavior ratings

were not independent of the nomination procedure) (Coie et al., 1982)

and directly observed behavior (Coie & Kupersmidt, 1983; Dodge, 1983).

The development of this sociometric classification system by Coie

et al. (1982) stems from their conclusion that previous definitions of

peer acceptance and peer rejection have not been uniform--consisting of

rating scale measures and/or nomination measures of acceptance and/or

rejection. They note that "since these two dimensions--acceptance and

rejection--are only slightly negatively correlated, the kinds of social

status distinctions that can be drawn from sociometric data vary

greatly depending on whether acceptance and rejection scores are used

together to define types of status or whether acceptance alone is used

as the index of status" (p. 557). These authors suggest that low

scorers on a positive nomination measure may consist not only of rejected

children, but neglected children as well. This distinction has been







validated in terms of directly observed behavioral differences between

these groups (Coie & Kupersmidt, 1983; Dodge, 1983; Renshaw & Asher,

1983). Rejected children are a high visibility group in terms of their

higher rates of negative behavior while neglected children are a low

visibility group in that they tend to be overlooked/ignored by peers and

appear to be more often "tuned out" or off task when alone and higher

on a set of shy, anxious, fearful behaviors termed "hovering" (Gottman,

1977, p. 513). Dodge (1983) notes that "the profile of rejected boys

. is one of antisocial behavior and inept peer interaction, whereas

the profile of neglected boys is one of inept peer interaction in the

absence of antisocial behavior" (p. 1397). In addition, Coie et al.

(1982) found that in terms of peer ratings of behavior (i.e., peer-

perceived behavior), rejected children were high on "disrupts group,"

"fights," and "seeks help," while neglected children displayed the

opposite profile.

Another difficulty which Coie et al. (1982) note with previous

sociometric classification systems is that when only positive nomina-

tions are used, the popular group may consist of a mixture of popular

children and controversial children, who, like popular children, receive

a large number of positive nominations but who would also receive a

similar number of negative nominations if a negative nomination measure

was utilized. This distinction between popular and controversial

children is important in that Coie et al. (1982) found that contro-

versial children received high peer-rated behavior scores on "disrupts

group" and "fights" and were frequently nominated for "seeks help."

In contrast, popular children received high scores on "cooperates" and

low scores for the "disrupts group," "fights," and "seeks help"







descriptions. In addition, Dodge (1983) found in a study of the devel-

opment of sociometric status in second grade boys' peer groups over

time that, over the course of eight hour-long free-play sessions of

individual groups of eight previously unacquainted peers, boys who

became popular were found to be generally cooperative and unaggressive

in terms of directly observed behavior. On the other hand, boys who

became controversial were distinguished by their high frequency of

both prosocial and antisocial behaviors.

Further evidence of the heterogeneity of the children in peer-rated

unpopular groups is Vosk, Forehand, Parker, and Rickard's (1982) finding

that the standard deviations were greater for unpopular children than

popular children on a teacher estimate of depression, the Conduct Prob-

lems and Hyperactivity scales on the Conners Teacher Rating Scale,

directly observed negative interactions, and the Children's Depression

Inventory (a self-rating measure of depression). Regarding this greater

variability, Vosk et al. (1982) suggest that their failure to dis-

tinguish within the unpopular group among rejected and neglected

children (as a result of not using negative nominations) may be respon-

sible for this greater variability.

Although the Coie et al. (1982) classification system appears to

generate homogeneous groups and has been currently validated in terms

of peer-rated and directly observed behavior, Putallaz and Gottman

(1983) note two shortcomings of the system. First, there exist

no data regarding the predictive validity of the social impact dimen-

sions of the system--in contrast to the extensive data regarding the

predictive validity of the social preference dimension. In addition,

Putallaz and Gottman (1983) note that the classification system of Coie







et al. (1982) is of questionable practical value in terms of compre-

hensiveness given that only 57.3% of their total sample fit one of

the five sociometric types--which suggests (and Coie et al., 1982,

provided no evidence to the contrary) that these types do not represent

clusters of children, but rather the extreme cases along what may well

be a normal distribution. Despite these shortcomings, the classifica-

tion system of Coie et al. (1982) was used in the present study be-

cause it represents the most detailed system to date and appears to

generate homogeneous groups.


Potential Mediators of the Behavioral
Manifestation of ICPS Skill


In terms of the proposed mechanism by which ICPS affects socio-

metric status, the present study investigated potential affective and

motivational mediators of the translation of ICPS skill into behavior

(as rated by independent observers). In terms of motivational mediat-

ing factors, the present study addressed, on a correlational level, the

role of perceived competence in the translation of ICPS skill into

behavior. According to Harter (1979),

perceived competence is viewed as an important correlate
and mediator of the child's intrinsic motivation to be
effective, to engage in independent mastery attempts in
the anticipation of a competent outcome. It is postu-
lated that the more a child is intrinsically motivated,
the greater will be his or her sense of competence.
Conversely, children with an extrinsic motivational
orientation, who are highly dependent on external ap-
proval and feedback, will perceive themselves as less
competent. (p. 1)

In terms of potential effective and motivational mediators of

ICPS skill, the relationship between children's ICPS skill and their

levels of anxiety, depression, and perceived competence was assessed.







Regarding the five sociometric groups of Coie et al. (1982) it was

expected that if behavioral differences between groups were obtained,

but were not reflected in corresponding ICPS skill differences between

the groups, then perhaps high levels of affect (i.e., depression and/or

anxiety) were interfering with the translation of skill knowledge into

skill performance.


Summary


The results of the three studies which have investigated the rela-

tionship between ICPS skill and sociometric status are inconclusive.

Findings in the social competence literature suggest that popular

children, relative to unpopular children, tend to handle hypothetical

peer conflict situations in a more positive, relationship-enhancing

fashion. These findings support ICPS researchers' claims that the

ability to cognitively solve interpersonal problems is related to

better social (i.e., behavioral) adjustment. Social skills training

studies reviewed here suggest that increasing children's general

knowledge of appropriate interpersonal behavior can result in altered

interpersonal behavior and, in turn, in enduring improvements in socio-

metric status. Given that ICPS skill can be considered as one type of

knowledge of appropriate interpersonal behavior, it appears reasonable

to expect that ICPS skill level should be manifested in terms of

behavior and, subsequently, reflected in a corresponding level of

sociometric status. Despite the positive expectations which might be

derived from the findings in these related areas, the fact remains that

the validity of the ICPS construct is tenuous. It is clear that

investigations should focus on the external validity of the ICPS







construct via assessing the validity of the various conceptualizations

of the ICPS construct (i.e., means-ends thinking, alternative solutions,

consequential thinking, and indicated obstacles) and the specific

measures of these particular conceptualizations.

The present study investigated to what extent groups of children

differing in terms of sociometric status also vary in terms of ICPS

skill (i.e., means-ends thinking). The sociometric classification sys-

tem of Coie et al. (1982), which classifies children into relatively

homogeneous groups, was utilized. The relative ICPS skill of the five

sociometric groups was considered in terms of potential variables which

may mediate the translation of ICPS skill into actual problem-solving

behavior: self-rated perceived competence, anxiety, and depression.

In addition, the process by which ICPS skill affects sociometric status

(i.e., positive and negative nominations) was investigated via a

correlational assessment of the relationship between (1) ICPS skill

and teacher rated behavior and (2) teacher rated behavior and socio-

metric status. Finally, the present study investigated the relative

strength of the relationship between ICPS skill and sociometric status

as opposed to the relationships of physical attractiveness and academic

achievement with sociometric status.













METHOD


Subjects


Participants were 157 boys and 150 girls from fourth (N=102),

fifth (N=81), and sixth (N=124) grade classrooms in two Polk County

schools and one school each from Pinellas, Alachua, and Bradford

Counties in Florida. Social class distribution, as determined by the

Two Factor Index of Social Position (Hollingshead, 1957), was as

follows: Class I, 21.8%; Class II, 20.2%; Class III, 37.8%; Class IV,

17.1%; and Class V, 3.1%. The ethnic composition of the sample was

77.7% white non-Hispanic, 14.5% black, 4.1% white Hispanic, and 3.6%

other.


Measures


Problem-Solving Procedure (PSP)


The PSP, which was developed by Richard and Dodge (1982) and used

by these authors with second through fifth grade boys, is similar to

Shure and Spivack's (1972) Means-Ends Problem-Solving Procedure. The

PSP, which was individually administered, presents each subject with

six hypothetical stimulus situations involving peer conflicts resolu-

tion (three stories) or friendship initiation (three stories). For

each problem the subject was told the beginning of the story (in which

the problem is presented) and the end of the story (in which the problem







has been resolved) and was asked to supply the middle of the story

(i.e., the intervening events by which the target child solves the

problem). Standardized prompts were given to the subject after every

one or two responses. Professionally drawn cartoons of each situation

were used to aid children's understanding of and to help maintain their

interest in the task.

The verbatim responses of each subject were recorded in writing by

the experimenter or one assistant. Five dependent measures were

derived from the subjects' responses to each story. The first was the

number of different solutions which the subject provided. Credit was

given for each conceptually different response. Credit for one solu-

tion was given if the subject offered several responses which dif-

fered only in terms of peripheral elements. For instance, a child who

suggested "give him a toy" and "give him candy" to the same problem was

given credit for one solution.

The second, third, fourth, and fifth dependent measures were

qualitative in nature and were derived by coding each solution according

to content as "effective," "ineffective," or "aggressive." Categoriza-

tion of each solution was performed according to criteria provided by

Richard and Dodge (1982). The first of these two measures consisted of

the percentage of initial solutions which were rated as effective. The

second measure consisted of the percentage of effective solutions fol-

lowing the first solution. Finally, the third measure was the per-

centage of initial solutions which were rated as aggressive. The fifth

measure was the percentage of subsequent solutions rated as aggressive.

The experimenter scored each child's responses and was blind to

sociometric categorization. Coefficient alpha for the number of







solutions was .76, which is somewhat better than the .65 figure reported

by Richard and Dodge (1982). In terms of interscorer reliability,

these authors reported that scorers completely agreed on 115 of 120

stories from 20 randomly selected protocols for an overall interscorer

agreement of 97%. The magnitude of the coefficient alpha and inter-

scorer agreement coefficients suggest that the PSP is internally con-

sistent and can be reliably scored, respectively.

Richard and Dodge used additional PSP procedures in their inves-

tigation. However, because these procedures and the resulting

additional ICPS variables generated from these procedures failed to

discriminate poorly-adjusted from well-adjusted children, these pro-

cedures were omitted in the present investigation.


Peer Nomination Measures


The peer nomination measures employed were similar to those used

by Coie et al. (1982) with third, fifth, and eighth grade children. The

positive nomination measure asked children to "circle the names of

the three children that you like the most." The negative nomination

measure asked children to "circle the names of the three children that

you like the least. Remember, this doesn't mean that you don't like

these children. Everybody likes some kids more than others." For

each measure, a child's score consisted of the total number of times

his/her name was circled by classmates, divided by the number of

classmates. Nomination scores were standardized by classroom and

transformed into social preference and social impact scores for each

subject. Social preference and social impact scores were then

standardized by classroom so that subjects were selected equivalently







across classrooms for placement into one of five sociometric groups

within the Coie et al. (1982) system. The rules for group assignment

were as previously described.

Coie et al. (1982) found that the 12-week test-retest reliability

coefficients (Pearson product-moment correlation coefficients) of

these two peer nomination scores, as obtained in an individual adminis-

tration format, were similar across third, fifth, and eighth grade

samples, ranging from .46 to .88 with a median correlation of .65.

These authors note that the magnitude of the reliabilities was similar

to those obtained by Roff et al. (1972). Kane and Lawler (1978), in a

review of peer assessment techniques, reported that peer nominations

are generally reliable, valid across time, and relatively bias-free.

Evidence for the concurrent validity of positive and negative peer

nominations is supplied by Green, Forehand, Beck, and Vosk (1980) and

Vosk et al. (1982) in terms of teacher ratings of behavior. Green et

al. (1980) found that higher numbers of positive nominations predicted

significantly lower total scores on the Conners Teacher Rating Scale

(CTRS), while higher numbers of negative nominations predicted signifi-

cantly higher scores on this scale. Vosk et al. (1982) found that

children classified as unpopular on the basis of a low number of

positive nominations and a lower peer rating score were rated by

teachers on the CTRS as more unpopular, depressed, and maladjusted.

Evidence for the concurrent validity of positive and negative peer

nominations in terms of directly observed behavior is supplied by

Hartup et al. (1967) and Charlesworth and Hartup (1967) who found that

rates of initiating and receiving positive interactions from peers

are predictive of peer acceptance and that rates of initiating







and receiving negative interaction from peers are predictive of peer

rejection.

Evidence for the concurrent validity of the classification system

utilized in this study is supplied in three studies. Employing third,

fifth, and eighth grade children (20% black and 49% male sample), Coie

et al. (1982) found significant main effects for sociometric group on

each of the following peer perceived behavioral descriptors: "Cooper-

ates," "disrupts group," "acts shy," "fights," "seeks help," and

"leads peers." The various distinctions between groups in terms of

the groups' score on each of the behavioral descriptors were discussed

previously. A limitation of these findings is that children's behavioral

descriptor nominations were not independent of their like most and like

least nominations.

Dodge (1983) investigated directly observed behavioral differences

between the five sociometric groups by assigning 48 previously un-

acquainted second grade boys to one of six play groups of eight

children each and observing behavior over eight 1-hour sessions.

Relative to children in the average group, (1) popular children con-

versed with the group leader more and were reprimanded less frequently,

(2) rejected children exhibited more inappropriate play, hostile

verbalizations, exclusion of peers, and hitting of peers, while

engaging in less social conversation, (3) neglected children spent

more time in solitary play, conversed with the group leader less, and

made fewer extraneous verbalizations, while displaying more inappropri-

ate play and fewer hostile verbalizations, and (4) controversial

children spent less time in solitary play and showed more social

conversation, conversation with leader, aggressive play, hostile







verbalizations, exclusions of peers, and extraneous verbalizations,

and had more reprimands by leader. All group differences were statis-

tically significant.

Coie and Kupersmidt (1983) investigated directly observed behavioral

differences between four of the sociometric groups (excluding contro-

versial children) by assigning one child of each type to each of eight

4-child groups. Behavioral group differences occurred on all four

observational categories. On active interaction, (1) popular children

were higher than neglected and average children,and (2) rejected

children were higher than neglected children. On parallel play,

popular, average, and neglected children were higher than rejected

children. On solitary appropriate activity, neglected children were

higher than popular children. Finally, on solitary inappropriate

activity, rejected children were higher than children in each of the

other three groups.

The directly observed and peer perceived behavioral group dif-

ferences--as well as the pattern of these differences--suggest that

the Coie et al. (1982) sociometric classification system is a valid

one.

Ethical concerns have been raised regarding the use of negative

peer nominations. While researchers are generally cautious in the

use of negative nominations (e.g., Asher & Hymel, 1981; Coie et al.,

1982; Hallinan, 1981; Richard & Dodge, 1982), have opted not to include

them (e.g., Ladd, 1981), or have advised against the use of negative

nominations in favor of a rating scale type of measure (e.g., Putallaz

& Gottman, 1983), the present study used negative nominations given

that "there is so much that cannot be learned about social relations







among children without introducing the negative choice issue" (Coie

et al., 1982, p. 567). The latter idea is especially true given that

"positive and negative scores are often not highly correlated"

(Hallinan, 1981, p. 101) and the fact that neglected children remain

undiscovered without the combined use of positive and negative nomina-

tions. In order to avoid these limitations, while still maintaining

ethical research standards, the negative nomination item used the phrase

"like least" instead of "dislike" or "don't like." In addition, the

negative nomination item was followed by information designed to

normalize the negative selection procedure.


Children's Depression Inventory (CDI)


The CDI (Covacs, 1981) allows children to rate each of 27 items on

a 0-2 scale which reflects the degree to which a particular symptom is

manifested. The scale has been used with children of ages 7 to 17,

including clinical and normative samples (Carlson & Cantwell, 1980;

Kovacs, 1981). According to Costello (1980), the CDI demonstrated

acceptable internal consistency (a = 0.86) for a sample of 875 children,

factor analysis resulted in only a single factor, and total scores did

not significantly correlate with either sex or age. Kazdin (1981) has

noted, regarding the CDI, that the internal consistency and interitem

and item-total score correlations for clinic and normative samples

resulted in significant correlations in the moderate range.

Moderately high test-retest reliability (r = 0.72) was demon-

strated over a 1-month interval with a nonclinic sample (Kovacs, 1981)

and, with a shortened version of the CDI over a 6-month interval,

(r = 0.44) (Tesiny & Lefkowitz, 1982).







In terms of convergent validity, Lefkowitz and Tesiny (1980) found

that the shortened CDI correlated significantly with the Peer Nomina-

tion Inventory of Depression (r = 0.24) and that this correlation

remained stable over a 6-month interval (r = 0.25).


State-Trait Anxiety Inventory for Children (STAIC)


The STAIC (Spielberger, Edwards, Lushene, Montuori, & Platzek,

1973) is similar in format to the State-Trait Anxiety Inventory which

serves as a measure of anxiety in adolescents and adults (Spielberger,

Gorsuch, & Lushene, 1970). The STAIC is intended for use with 9 to 12

year old children. It is composed of two scales--State and Trait--

which are designed to measure transitory and more chronically experi-

enced anxiety, respectively. The Trait scale was used in this study.

It consists of 20 items, each ranked on a 3-point severity scale, which

children are asked to respond to according to how they generally feel.

Reasonably good internal consistency was demonstrated with this scale

in terms of coefficient alpha, which was .78 for males and .81 for

females. The test, retest reliability coefficients obtained via

retesting fourth, fifth, and sixth graders after 6 weeks were .65 for

males and .71 for females. Concurrent validity was demonstrated via

its correlation with two other measures of trait anxiety in children--

.75 with the Children's Manifest Anxiety Scale (Castenada, McCandless,

& Palermo, 1956) and .63 with the General Anxiety Scale for Children

(Sarason, Davidson, Lighthall, Waite, & Ruebush, 1960).








Perceived Competence Scale for Children (PCS)


The PCS (Harter, 1982) is a 28-item questionnaire which is designed

to assess third through sixth grade children's self-reported motivational

level across four subscales: cognitive, social, athletic, and general.

The PCS utilizes a unique question format, the nature of which legiti-

mizes children's endorsements and, thereby, reduces their socially

desirable responding. Harter (1982) notes that whereas the Coopersmith

Self-Esteem Inventory, a widely used, global measure of self-esteem,

correlates .33 with the Children's Social Desirability Scale (Crandall,

Crandall, & Katkovsky, 1965), the PCS correlates only .09 with this

social desirability measure. Regarding the question of format, the

subject is first requested to decide which of the two types of children

that he or she is most like--those on the left side or those on the

right side of the questionnaire. The subject next determines whether

the description on that side is "really true" or "sort of true" for him

or her. Each item is scored from 1 to 4, where 1 indicates low per-

ceived competence and 4 indicates high perceived competence. Thus, a

subject's score for each of the four 7-item factors may range from 7 to

28. The structure of the test is such that in 14 of the test items,

the first part of the statement reflects high perceived competence,

and in the other 14, high competence is reflected in the second part

of the statement. Regarding item order, no more than two items in a

row are keyed in the same direction and no more than two consecutive

items are from the same subscale. For each subscale, four items are

keyed in one direction, and three are keyed in the other direction.








With regard to validity, the four-factor pattern of the PCS has

proven to be extremely stable and has been replicated on five samples

in addition to the standardization sample. Harter (1982) notes that

the average loading of items on their designated factors for a New York

cross-validation sample was .67, .61, .64, and .50 for the Cognitive,

Social, Physical, and General subscales. Very little (and inconsis-

tent) cross-loading of factors has occurred. Concurrent validity of

the Social subscale is reported by Harter (1982) in that the correla-

tion between Social subscale scores and the same sex peer ratings for

fourth, fifth, and sixth grade children was .59. The Cognitive subscale

received validational support in a study reported by Harter (1982) in

which sixth grade children in the top third of the sample in terms of

perceived cognitive competence, given their choice of anagrams of vary-

ing difficulty, chose significantly more difficult anagrams (i.e.,

length) than did children in the bottom third of the sample in terms of

perceived cognitive competence. Additional validational support for

the Cognitive subscale consists in the fact that, in a study reported

by Harter (1982) learning disabled children scored significantly lower

than nondisabled children on the Cognitive subscale but did not differ

in terms of the other subscales.

Regarding the internal consistency of the PCS subscales, Harter

(1982) notes that across the Cognitive, Social, Physical, and General

subscales, the coefficient alphas across samples have ranged from .75

to .83, .75 to .84, .77 to .86, and .73 to .82 for the four subscales,

respectively. Test-retest reliabilities over a 9-month interval for

the original sample were .78, .80, .81, and .70.







Physical Attractiveness Measure


The measure of physical attractiveness consisted of children's

ratings on a 1 to 5 scale of head-and-shoulders photographs of unfamiliar

peers. Children were asked "how handsome (pretty) is this boy (girl)?"

(see Langlois & Stephan, 1977). Photographs were taken by the experi-

menter at one school and presented to children at a different school.

The order of presentation of each picture was rerandomized after each

rating. A subject's physical attractiveness scores consisted of the

average rating received by the same sex, opposite sex, and by both

sexes combined. The coefficient alpha reliability was .95, suggesting

that the physical attractiveness ratings were reliable.

In terms of validity, Kleck et al. (1974) and Langlois and Stephan

(1977) found that physical attractiveness is a significant predictor of

sociometric status.


Academic Achievement


The most recent nationally standardized achievement test composite

score was recorded from each child's school file. These composite

scores were drawn from one of the following three tests (depending upon

which test the school happened to use): Metropolitan Achievement Test,

Comprehensive Test of Basic Skills, or Science Research Associates

Achievement Tests.

Utilizing the Metropolitan Achievement Test, Green et al. (1980)

found that children who were high on academic achievement received

more positive nominations and fewer negative nominations than did

children who were low on achievement.







Teacher Rating of Social Competence


The literature on children's social competence contains few

measures of competence-related behavior (Achenbach, 1978), including

very few teacher rating measures of social competence (Gesten, 1976) and

no teacher measures which can be applied to middle-school children.

Given the present study's concern with children's social competence,

particularly in regard to ICPS skill, a teacher rating measure of

social competence which incorporates items concerning interpersonal

problem-solving behavior was constructed. Items were adapted from the

Child Behavior Checklist (Achenbach, 1978; Achenbach & Edelbrock, 1979)

and the Health Resources Inventory (Gesten, 1976). Teachers were

asked to rate a number of prosocial and negative/antisocial behaviors

of each child relative to peers.


Procedure


Informed consent was obtained from parents to permit their children

to participate in this study. Children were group administered, in

order, the Children's Depression Inventory (Kovacs, 1981), the State-

Trait Anxiety Inventory for Children (Trait Scale) (Spielberger,

Edwards, Lushene, Montuori, & Platzek, 1973), a peer nomination measure

consisting of both positive and negative nominations, and the Per-

ceived Competence Scale for Children (Harter, 1979). Prior to this

administration, children were given the following instructions:

I am interested in how children your age think and feel,
and I am interested in your ideas about friendship. I
am interested in your ideas--not what you think your
parents or other adults would say. It is important that
you be as honest as possible. I am interested in what
you think. There are no right or wrong answers. Different







people think and feel different things and have
different ideas about friendship. All information
will be totally confidential. That means that
nobody sees your answers except me. That means
that your parents and teachers will not see them.
So, the idea is to be as honest as possible about
your ideas. If you are not sure about your ideas,
it's okay to guess because, remember, there are no
right or wrong answers.

At a later date, children whose peer nomination scores qualified

them for inclusion into one of the five sociometric groups were indi-

vidually retested by the experimenter or one assistant. Because

children were identified on the basis of code numbers in the group

assignment process, the experimenter and his assistant were blind to

the sociometric status of each child. During retesting, each child was

administered an abbreviated version of the PSP and asked to make

physical attractiveness ratings of the head-and-shoulders, black-and-

white photographs of unfamiliar peers. Following these procedures, a

photograph of the child was taken for subsequent physical attractive-

ness ratings by other children. For the children who were retested,

teachers filled out a 23-item questionnaire designed to assess children's

social competence. In addition, the composite scores from each of the

children's most recently completed, nationally standardized achievement

test were collected from school files.

Of the 307 children initially sampled, 216 (70.4%) met the

criteria for inclusion into one of the five sociometric groups.

Absentees accounted for the loss of 23 subjects, resulting in a final

group of 193 subjects. The sociometric group assignment was as

follows: popular (N = 39) 20.2%; rejected (N = 38) 19.7%; neglected

(N = 14) 7.3%; controversial (N = 14) 7.3%; and average (N= 88) 45.6%.

A test for differential sex distribution across the five sociometric







groups was not significant: X2(4) = 0.855, p = .9310. This suggests

that sex distribution among these groups was relatively balanced:

popular (M = 22; F = 17); rejected (M = 21; F = 17); neglected (M = 17;

F = 7); controversial (M = 7; F = 7); and average (M = 43; F = 45).

Statistical analyses were conducted to ascertain the presence of

significant effects with regard to each of the following areas:

(1) demographic variables; (2) sociometric group differences; (3) socio-

metric group differences by sex; (4) prediction of sociometric status;

and (5) the proposed mechanism by which ICPS skill affects sociometric

status.

In order to examine the possibility of demographic group differ-

ences across each of the measures, potential sex differences were

assessed via a series of t tests, while potential race, age, and SES

effects were assessed via a series of analyses of variance.

Analyses of variance were conducted in order to examine whether

children of different sociometric status were discrepant in terms of

(1) potential predictors of sociometric status (i.e., ICPS, physical

attractiveness, and academic achievement), (2) potential affective and

motivational mediators of sociometric status (i.e., depression, anxiety,

and perceived competence), and (3) teacher ratings of social competence.

These sets of analyses were subsequently repeated separately for both

males and females in order to ascertain whether the pattern of varia-

tion among sociometric groups on each of these variables differed

with respect to sex.

In order to assess the relative predictive validity of the ICPS,

physical attractiveness, and academic achievement variables with

regard to sociometric status, a series of stepwise regressions was







conducted for both the total sample and for each sex. The following

four standardized sociometric variables were utilized, individually, as

criterion variables in the stepwise regression procedures: (1) positive

nominations, (2) negative nominations, (3) social preference, and

(4) social impact. In total, 12 stepwise regression procedures were

conducted--given the use of four criterion variables for the total

sample and for each sex. In addition, in order to assess the relative

extent to which children's problem-solving knowledge (i.e., ICPS skill),

problem-solving behavior (teacher-rated social competence), physical

attractiveness, and academic achievement predicted each of the four

standardized sociometric variables, Pearson product-moment correlations

were conducted for all pairwise combinations of these four variables

with the predictor variables, resulting in a 4 X 10 correlation

coefficient matrix.

Finally, in order to investigate the mechanism by which ICPS is

hypothesized to affect sociometric status, the following were assessed

via Pearson product-moment correlations: (1) the degree to which.

children's problem-solving knowledge (i.e., ICPS skill) was translated

into problem-solving behavior (i.e., teacher-rated social competence)

and (2) the degree to which children's problem-solving behavior was

reflected in sociometric status.













RESULTS


Demographic Variables


A number of significant demographic effects were obtained. In

terms of sex differences, females scored higher than males on each of

the physical attractiveness variables: physical attractiveness ratings

by the same sex, PASAME, t(169.6) = -5.51, p < .0001; physical attrac-

tiveness rating by the opposite sex, PAOPP, t(191) = -2.89, p < .01;

and physical attractiveness ratings by both sexes, PATOT, t(191) =

-4.48, p < .0001. (Given the significantly greater variability in

the PASAME scores of girls than boys, Satterthwaite's approximation for

the degrees of freedom was used, and an approximate t statistic was

computed.) These results suggest that females were perceived as more

physically attractive by peers than were males. In addition, males

scored significantly higher than females on the Social and Physical

subscales of the PCS: PCSS, 5(177.4) = 2.03, p < .05; and PCSP,

t(191) = 3.38, p < .001. (Again, given the significantly greater

variability in the PCSS scores of girls relative to boys, Satterthwaite's

approximation for the degrees of freedom was used, and an approximate

t statistic was computed.) These findings appear to indicate that males

view themselves as more competent in terms of social and physical

(athletic) abilities. No sex differences were obtained on the PSP

despite the fact that Richard and Dodge (1982) designed this measure,

the story content of which deals only with boys, for use with boys.







Racial differences emerged on four variables. However, analysis

via Duncan's Multiple Range Test revealed that, in each case, these

differences were between an unspecified ("Other") racial group, con-

sisting of a small number (N = 7) of children of largely Asian/Oriental

descent, and one or more of the other three racial groups. Given the

paucity of children in this racial group and the fact that the other

three groups (which comprised 96.4% of the sample) were not signifi-

cantly different from each other on any of the variables, the obtained

racial differences were deemed practically insignificant for the total

group and, consequently, were ignored in further analyses.

An age effect occurred with respect to one variable--ICPS2 (the

percentage of effective solutions offered after the initial solution):

F(2,190) = 3.22, p < .05. Subsequent analysis via Duncan's Multiple

Range Test indicated that fourth graders scored higher than fifth

graders. Sixth graders did not differ significantly on this variable

from either fourth or fifth graders.

Significant socioeconomic (SES) effects occurred on five variables.

Analysis via Duncan's Multiple Range Test revealed that four of these

significant effects involved a small number (N = 6) of children in the

lowest socioeconomic class (Class=V) obtaining significantly lower

scores than children in Classes I-IV. Given the small number of

children in this group and the fact that the other four groups (which

comprised 96.9% of the sample) were not significantly different from

each other on any of the variables, the obtained SES differences were

viewed as practically insignificant and, consequently, were ignored in

further analyses. The fifth significant SES effect involved children

in the highest socioeconomic class (Class=I) scoring lower than children







of Classes III, IV, and V on ICPS2: F(4,188) = 4.40, p < .01. This

finding indicates that children of Classes III, IV, and V offered a

greater percentage of effective solutions following their initial

solution than did children of Class I.


Sociometric Group Differences


The pattern of sociometric group differences across all variables

is shown in Table 1. With respect to the ICPS, physical attractiveness,

and academic achievement measures, no sociometric group differences

emerged on the ICPS and academic achievement measures. The only

sociometric group differences occurred on PAOPP, F(4,188) = 3.29,

p < .05. Analysis of group differences via Duncan's Multiple Range

Test revealed that (1) controversial (X = 2.8) children were rated higher

than children in the average (X = 2.4), rejected (X = 2.4), and

neglected (7 = 2.2) groups (who did not differ from one another) and

(2) popular (X = 2.6) children were rated higher than neglected

children. These results suggest that popular and controversial children

were viewed by opposite sex peers as the most attractive followed in

descending order by children in the average, rejected, and neglected

groups. Marginally significant, similar effects occurred on the other

two physical attractiveness variables. On PASAME, F(4.188) = 1.93,

p < .11, popular (7 = 2.9) children were rated significantly higher

than neglected (7 = 2.5) children, with neither group differing from

the other three groups. Thus, same sex peers rated popular children as

more physically attractive than neglected children. On PATOT, F(4,188)

= 2.22, p < .07, popular (X = 2.7) and controversial (X = 2.8) children

were rated significantly higher than neglected (T = 2.4) children, with














Cf) C-
In -
0 0
0 1.0 to i- 0- M N 0 -- 0 C\j to O
0i in C) Nr- L C\ J- 0 to 0 0 0 0
*.* *.* *.* *.





-K
e- Ln to in L r- M CNj rN- U CO LAON O C 0 o C" c\j

0000 0 C .- Cr4 C\ C C'J 00 LO C,; CM 0.,






.a) .0 .0 .0 .0
03 0 0 0 0 L0 00 0a ( 0 0 o0 (0 0 0
00 C) zr 00 M L0 ru 44' to to to 00 0) C-

Cn LA to i- u 0) e'. cMJ C~J oo t 0 000 o-
r- to LO UO r- C; CN C; N CC









to 0 0 0 m m 0 m 00 m0 0 00 0.
0) 0 e- r- COO ^O) CO 0n O DOO ko L o

) 4--O' C0oI C;M Csj c0 toO ; Or- 0
-- N- L LA r- N N O C('4-









U .0 .0
0O 0 0 0 03 0.0 QQ .0 0 '0 0.0 a 0 0 0
00 .- o C) en N- U-) C. :) C, i- to h 9Z






Cr) to ) (a. Ct) ('4 (4 (a to LAO (c 0 0 u
) O TM ON M NN N, t M C ) C\O COO O M M
U- NC fl o C O C \- 'C4 CN (_4







.0 0 .0
0l TT 0a m C.0 MMt0 0 0 lO 0 00 0 -0 ON







00d ) ml to; 0 C; C L\ A 0)0 ('4 N-.- C; _: '
IDO o OO 0) o '4 C\4 C\1 0Ij N- 0) o 0) Or N.











O O 0 CO L 4 L LO N NU- 0
r- Nt Ln e- n- r- e- a-






.0 Q.0
(0 t0 0 ^0 0 ^0 0 0 f0 f0 l f0 *0 0 0 ^0
L. N- ') c 0) to N- tC C CO O V L 0 )

C L O OC O CL C L tO LA O V (-- (n
-- tO LA N- (N C4 C' (' N'












Lii Lii
i- M m LA S QL 1- () 3-

C t CA / /) /) CC C c 0 C. ) ? UC 0.( Lu
-. C. C. C. Q. OC L- V) CL C Q- CL C C-
L) C. .) L .) 0 0 < CC < : 0 C. C C.) .) Li
<- 0.0. 0. C. CA 0.0 0. 0. I--


0-


C)


F-


I--


Ca








C..J
I-






LU
i- CA


.J
-J
I -J





U-
CA
3

t./)
cc


1)
01

S..
a)








0


L)
>









0)

S-

C




0



-C
0)
4-1)


01



-01


C0
01
"'a)

4-
U
0)







5-
01





0.
0
0.


C 4-)
*r- .
e P
SL.. *-
C) o0


" 4-





v m
*0
(U


VC
(U,






0 0
C,,C
o 0










0 4-

4-
V r0



-)
4-* -0







a) t>




S-

[4- -c
0






4 I






a) 0
S -


4-) (00



*r- C
>4-







*-) 4- -
0>
0 )
4- :
*'- 3) 0
CL 0 a)




*- 0@ -~
100



*- 4








4-) 0 c
o. a> a)
*i 0 I-












S.. CD

)4 -
0.4- 3-



*r* r-
L4-' S
uC) 0
u- a) -^












C I- a
0 <4
0.4 S)






*I- E
>/ 01 L0
SC..- 51





0 0)



C 0) U-a
0- 4.)

(0 0) /

11i- 0)

S.1= i 'S







none of these groups differing from the other two groups. Thus, popu-

lar and controversial children were viewed in terms of combined sex

ratings as more attractive than neglected children. It is interesting

to note that in terms of mean scores for all three of the physical

attractiveness variables, the popular and controversial children re-

ceived the highest scores and were followed in descending order by

children in the average, rejected, and neglected groups. It is clear,

given the significant and marginally significant effects on the physi-

cal attractiveness variables and the absence of significant effects on

the ICPS and academic achievement variables that physical attractive-

ness is a better discriminator in terms of sociometric status than is

ICPS or academic achievement, which did not differ among the five

sociometric groups.

Among the depression, anxiety, and perceived competence measures,

significant sociometric group differences occurred on two of the four

PCS subscales. On the Social subscale, PCSS, F(4,188) = 5.29, p <

.001, (1) controversial (X = 22.6) and popular (7 = 21.8) children

rated their social competence higher than neglected (C = 18.9) and

rejected (C = 17.7) children, and (2) average (X = 20.8) children rated

themselves higher than rejected children. These results suggest that

the controversial and popular children viewed themselves as more in

possession of social competence-related skills than did the neglected

and rejected children, while average children were more competent than

rejected children. On the Physical subscale, PCSP, F(4,188) = 3.25,

p < .05, popular (X = 22.7) children rated their physical competence

higher than did rejected (C = 19.1) children. This finding indicates

that popular children feel more athletically competent than do rejected







children. A final,marginally significant effect occurred on the General

subscale, PCSG, F(4,188) = 2.39, p < .06, with controversial (7 = 21.8)

children scoring significantly higher than rejected (X = 18.3) children.

This result suggests that controversial children feel generally more

competent than do rejected children. An interesting aspect of the

findings on each of the PCS subscales (including the Cognitive subscale,

on which there were no sociometric group differences) was that the

rejected children obtained the lowest score on each subscale and were

significantly lower than at least one of the other sociometric groups

on three of these subscales. An additional, marginally significant main

effect for which there were no significant group differences on Duncan's

test occurred on the CDI, F(4,188) = 2.04, p < .10. On this measure

rejected children had the highest mean score (7 = 10.9) followed by

controversial (X = 9.9), average (X = 8.6), popular (X = 6.5), and

neglected (X = 6.4) children.

On the teacher rating of social competence, TEACHER, F(4,188) =

9.28, p < .0001, significant sociometric group differences emerged.

Duncan's test showed that neglected (X = 27.3), popular (N = 26.9), and

average (X = 24.1) children were rated significantly higher by teachers

than controversial (7 = 18.5) and rejected (X = 17.3) children. These

results suggest that teachers viewed rejected children and controversial

children (who are, by definition, rejected by many peers--in addition

to being liked by many peers) as less socially adept than children in

the other sociometric groups in terms of classroom behavior.







Sociometric Group Differences by Sex


Analysis of sociometric group differences by sex on the ICPS,

physical attractiveness, and academic achievement variables again

resulted in no group differences for either sex on the ICPS and

academic achievement variables. However, significant effects were

obtained on the physical attractiveness variables, and these effects

differed with respect to sex. For males, on the physical attractive-

ness measures, sociometric group differences occurred on PAOPP,

F(4,95) = 3.06, p < .05, and PATOT, F(4,95) = 3.01, p < .05. On both

PAOPP and PATOT, Duncan's test indicated that controversial children

received the highest ratings followed in order by popular, average,

rejected, and neglected children. On both variables, controversial,

popular, and average children received ratings which were significantly

higher than those of the neglected group, suggesting that opposite sex

peers as well as the combined sexes viewed neglected children as less

attractive than controversial, popular, and average children. In addi-

tion, controversial children were rated significantly higher than

rejected children, suggesting that opposite sex peers as well as the

combined sexes viewed rejected children as less attractive than con-

troversial children. A marginal effect was obtained on PASAME,

F(4,95) = 2.23, p < .08. On this variable popular and controversial

children received significantly higher scores than neglected children,

suggesting that same sex peers view neglected children as less attrac-

tive than popular and controversial children. The pattern of group

differences on these variables is very similar to the pattern for the

total sample, including an identical rank ordering of mean scores







obtained by the groups on these variables. Given that there were no

significant sociometric group differences for females on these vari-

ables, it appears that it was the pattern of variation in the males'

scores which accounted for the group differences for the total sample.

Among the depression, anxiety, and perceived competence variables,

significant group differences were obtained for males on the PCSS,

F(4,95) = 4.26, p < .01. Further analysis via Duncan's test revealed

that popular (X = 23.2) and controversial (I = 22.0) boys scored

significantly higher than rejected (7 = 18.3) boys. These results

suggest that popular and controversial boys viewed themselves as more

socially competent than rejected males. For females, marginally

significant effects occurred on three of the four PCS variables. On

the PCSS, F(4,88) = 2.24, p < .08, Duncan's test revealed that con-

troversial (7 = 23.1) girls scored higher than both neglected (X =

17.9) and rejected (X = 17.1) girls, suggesting that the latter groups

view themselves as less socially competent than the former group. On

the PCSP, F(4,88) = 2.43, p < .06, popular (X = 21.4) children scored

significantly higher than rejected (X = 16.8) children, suggesting

that popular girls feel more athletically competent than do rejected

girls. On the PCSG, F(4,88) = 2.37, p < .06, controversial (X = 23.0)

girls scored higher than rejected (X = 16.6) girls, indicating that

controversial girls view themselves as more generally competent than

do rejected girls.

While it is apparent that there were not as many significant

sociometric group differences for each sex across the four PCS vari-

ables as there were for the total sample across these variables

(statistical artifact resulting from a higher N in the total sample?),







it is interesting to note that, for both the total sample and the sex-

specific sociometric group analyses, each significant or marginally

significant result involved the rejected children achieving the lowest

mean score and rating themselves significantly lower than the popular

and/or controversial group.

On the teacher rating measure, TEACHER, significant differences

were found among the sociometric groups for both males, F(4,95) =

5.38, p < .001, and females, F(4,88) = 4.18, p < .01. Duncan's test

indicated that (1) neglected (X = 26.0) and popular (X = 25.9) boys

scored significantly higher than controversial (7 = 18.6) and rejected

(I = 16.1) boys, and (2) average (7 = 24.5) boys scored higher than

rejected boys. These findings appear to indicate that teachers

endorsed more social competence-related behaviors for (1) neglected

and popular boys than for controversial and rejected boys and (2) aver-

age boys than for rejected boys. For females, neglected (X = 28.6) and

popular (X = 28.3) girls were rated higher than rejected (7 = 18.6) and

controversial (X = 18.4) girls, suggesting that teachers viewed

neglected and popular girls as more socially competent than rejected

and controversial girls. The significant sociometric group differ-

ences for each sex were very similar to the differences for the total

group. In each case, the ordering of mean scores shows that teachers

endorsed the most social competence-related behaviors for neglected

children, followed in order by popular and average children and then

the controversial or rejected children. The pattern of group differ-

ences suggests that children who have been found by others to perform

more socially disruptive behaviors (Coie & Kupersmidt, 1983; Dodge,

1983) and to be viewed more negatively by peers (Coie et al., 1982)







(i.e., children in the controversial and rejected groups) are viewed as

less socially competent by teachers.

An unfortunate and limiting aspect of the teacher ratings regard-

ing its validity as a measure of children's social competence is

revealed by the fact that neglected children, who have been described

by Dodge (1983), on the basis of directly observed behavior as exhibit-

ing "inept peer interaction in the absence of antfsocial behavior"

(p. 1397), received the highest scores from teachers. This finding

suggests that teachers may most highly value "obedient," nondisruptive

behavior in the classroom setting (see Winnett & Winkler, 1972).


Prediction of Sociometric Status


Given that the analysis of sociometric group differences by sex on

the ICPS, physical attractiveness, and academic achievement variables

resulted in (1) different effects with respect to sex and (2) no

group differences for females, a stepwise regression was conducted for

each of the four standardized sociometric standing scores for the total

group and for each sex in an effort to explain the obtained sex dif-

ferences and to ascertain the relative predictive validities (especially

for females) of ICPS, physical attractiveness, and academic achievement

for sociometric standing. Sociometric standing was defined as any and

each of the following previously defined scores--which were standardized

by classroom and which served as the criterion variables in the stepwise

regressions: (1) positive nominations, (2) negative nominations,

(3) social preference, and (4) social impact.

The results of the stepwise regression analyses for both the total

group and for each sex proved rather uninformative, with no variable







in any of the 12 regressions accounting for more than 9% of the variance

and 9 of the 12 regressions having no significant (p < .05) predictor

variables. Thus, the stepwise regression procedure resulted in no new

information regarding the relative predictive validity of the ICPS,

physical attractiveness, and academic achievement variables for either

sex or for the total sample.

Given the lack of information obtained by these procedures as well

as the necessity of gaining information regarding the predictive

validity of these variables for sociometric status, a correlational

analysis was conducted. The ICPS, academic achievement, and physical

attractiveness variables formed the vertical axis of the resulting

correlation matrix while the four standardized sociometric variables

constituted the horizontal axis. In addition, teacher-rated social

competence was added to the variables on the vertical axis. This

addition was done in order to ascertain whether physical attractiveness

or teacher-rated social competence (i.e., problem-solving behavior)--

both of which had significantly discriminated among sociometric

groups--was a better predictor of sociometric status.

Table 2 shows that, although physical attractiveness was a better

predictor of sociometric group status than was ICPS (i.e., knowledge

of appropriate social behavior), it was not as good a predictor as

teacher-rated social competence (i.e., performance of socially appropri-

ate behavior--including problem-solving behaviors). Although both

physical attractiveness and teacher-rated competence discriminated

among sociometric groups, the teacher measure was a significant and

better predictor of each of the four sociometric standing variables

than were the physical attractiveness variables, which significantly








TABLE 2

MATRIX OF CORRELATIONS BETWEEN STANDARDIZED SOCIOMETRIC VARIABLES
AND ALL PREDICTOR VARIABLES


PREDICTOR

ICPS1 .

ICPS2 .

ICPS3 .

ICPS4 .

ICPS5 .

AA. .

PASAME..

PAOPP .

PATOT .

TEACHER .


VARIABLES


STANDARDIZED SOCIOMETRIC VARIABLES

Positive Negative Social Social
Nominations Nominations Preference Impact

.02 .06 .11 -.02

-.08 .04 -.03 -.09

.01 -.06 -.06 .05

-.00 -.01 -.00 .00

-.02 .11 .11 -.07

.07 -.08 .02 .08

.17* -.09 .10 .14

.19** -.05 .13 .12

.16* -.08 .08 .12

.23** -.42**** -.18* .36***


= significant at
** = significant at
*** = significant at
**** = significant at


.05 level
.01 level
.001 level
.0001 level


k







predicted only the positive nominations variable. (It is unclear why

the teacher measure was negatively correlated with the social preference

score.) These results suggest that (1) children who received high

social competence ratings from teachers received more positive peer

nominations and fewer negative peer nominations and (2) children who

received low social competence ratings from teachers received more

negative and fewer positive peer nominations.


Translation of ICPS Skill into Sociometric Status


Further correlational analyses were conducted in order to derive

information regarding the mechanism by which ICPS skill affects socio-

metric status. Pearson product-moment correlations were conducted in

order to assess (1) the degree to which children's problem-solving

knowledge (i.e., ICPS skill) was translated into problem-solving

behavior (i.e., teacher-rated social competence) (see Table 3) and

(2) the degree to which children's problem-solving behavior was

reflected in sociometric status (i.e., the four standardized socio-

metric variables) (see Table 2). Among the teacher-rated competence

and ICPS variables, it was found that the correlation of teacher-rated

competence with ICPS3 (percentage of effective solutions following the

first solution) was .15 (p < .05) and with ICPS5 (percentage of aggres-

sive solutions following the first solution) was -.15 (p < .05).

Although not robust, these statistically significant correlations

suggest that relative to children who receive lower ratings of social

competence from teachers, children who receive higher ratings tend

to offer (1) a higher percentage of effective solutions following the

first solution and (2) a lower percentage of aggressive solutions










TABLE 3

MATRIX OF CORRELATIONS BETWEEN TEACHER-RATED
SOCIAL COMPETENCE AND ICPS VARIABLES


ICPS1 ICPS2 ICPS3 ICPS4 ICPS5


TEACHER -.04 -.13 .15* -.02 -.15*


* = significant at .05 level







following the first solution. None of the correlations between the

four sociometric standing variables and the five ICPS variables

approached significance. Thus, two of the five ICPS skill variables

significantly predicted teacher ratings of social competence (although

only minimal variance was accounted for) and teacher ratings signifi-

cantly predicted sociometric standing across all four variables

(although the direction for one variable was opposite to that expected

and, again, only a small percentage of variance was accounted for by

each of the variables).













DISCUSSION


The present study investigated the external validity of ICPS in

terms of sociometric status in elementary school children. Previous

investigations of this important relationship have been lacking or

methodologically flawed. In addition, the strength of the relation-

ship was evaluated against two other factors which have been previously

demonstrated to correlate with sociometric status: physical attractive-

ness and academic achievement. The results suggest that ICPS skill

(in terms of both quantitative and qualitative aspects) bears no

significant relationship to sociometric standing--in terms of either

sociometric group differences or standardized sociometric scores for

individual subjects. Academic achievement was also found to be unre-

lated to sociometric standing. Physical attractiveness, however, was

found to be significantly different across sociometric groups for

(1) the total sample--as rated byoppositesex peers and (2) for boys--

as rated by girls and by the combined sexes.

Main effects for several variables, including the sociometric,

ICPS, and physical attractiveness variables, with respect to age,

race, and/or SES were not incorporated into further analyses because

these main effects stemmed from either a small N for one of the groups

or were uninterpretable. The facts that fourth graders provided a

higher percentage of effective solutions than fifth graders, but not

sixth graders, and that fifth and sixth graders did not differ in







this respect were counterintuitive and may reflect difficulties with

the PSP. Perhaps the better performance of fourth graders relative

to that of fifth graders on this variable was the product of a higher

interest level in the testing procedure and/or in the test itself.

Unfortunately, this explanation fails to explicate the lack of obtained

differences on this variable between fourth and sixth graders and

between fifth and sixth graders. Similarly, on this same variable,

the fact that children from the lowest three classes obtained higher

scores than children in the highest class is counterintuitive given

that, although no SES differences were predicted, it would be expected

that, if SES differences did occur, these differences would show

higher SES children outperforming lower SES children. A lower interest

level of Class I children in the testing procedure and/or in the test

itself may have accounted for the poorer performance of these children.

The results of the present study suggest that children's physical

attractiveness shows a significant, positive relationship to their

sociometric status, while ICPS and academic achievement have little

if any relation to sociometric status. The negative findings regard-

ing the external validity of ICPS suggest that with children similar

to the sample utilized in the present study (1) this construct is not

a useful one (as conceptualized and measured in this study) in terms

of predicting the degree to which children are liked and/or disliked

by peers and, given the previously cited evidence regarding the

predictive validity of sociometric data, in terms of predicting the

degree to which they may be at risk for future interpersonal maladjust-

ment, and (2) ICPS training efforts are at best premature and may

potentially be wasted effort. In addition, it appears that physical







attractiveness, over which children as well as adults have only a

limited amount of control (i.e., typically within the boundaries set

by one's innate physical endowment), is a more potent determiner of

children's friendships than is ICPS--over which there is potentially

a great deal of control in terms of training children in better ways

of dealing with peer relationship problems. Two alternative explana-

tions may prevent one from becoming discouraged at these interpreta-

tions of the results. First of all, the results are limited in that

(1) the means-ends conceptualization of ICPS in this investigation

represents only one facet (albeit an important one) of ICPS as

proposed by Shure and Spivak, (2) the ICPS measure used in this study,

although apparently reasonably reliable, as well as valid (see Richard

& Dodge, 1982) is certainly not the definitive measure of ICPS or

means-ends thinking. Secondly, the fact that the teacher rating

measure of social competence (which included a number of items con-

cerning interpersonal problem-solving behavior with peers) was

better at discriminating among sociometric groups and better at pre-

dicting the four standardized sociometric variables than was physical

attractiveness suggests that interpersonal problem-solving is an

important concept and potentially a more potent and remediable con-

tributor to children's status among peers than is the relatively

unalterable physical attractiveness.

The finding that children in different sociometric groups did not

differ in terms of ICPS (i.e., problem-solving knowledge) but did

differ with respect to teacher-rated social competence (i.e., problem-

solving performance) suggests the operation of an inhibiting factor

regarding the translation of ICPS knowledge into socially competent







behavior. This explanation is supported by the finding that socio-

metric groups differed on two motivational variables--the PCS'

Social and Physical subscales (and nearly so on the PCS' General

subscale as well as the CDI)--and did so in a fashion which resembled

the pattern of sociometric group differences obtained on teacher-

rated social competence. Perhaps the most compelling findings in

this regard is that rejected children--who did not differ from

children in the other sociometric groups in terms of ICPS skill but

received the lowest ratings of social competence from teachers--were

lowest on self-rated perceived social, physical, and general competence

as well as highest on self-rated depression. These findings suggest

that depression and a lack of motivation may have interfered with

rejected children's ability to translate their ICPS knowledge--which

was equivalent to that of children in the other sociometric groups--

into socially competent behavior. At the same time, popular children,

who were equivalent to rejected children in terms of ICPS,but were

rated by teachers as significantly more competent than rejected

children, were significantly higher than rejected children in terms

of self-rated social and physical competence and general competence

(a nonsignificant trend) and lower in terms of self-rated depression

(again, a nonsignificant trend). This group comparison finding again

supports the idea that rejected children were deficient in terms of

socially competent behavior relative to popular children because of a

significantly lower motivational level and somewhat higher depression.

Unfortunately, the data on which this hypothesis is based do not

permit a determination as to whether the depression and low feelings

of competence or low social standing occurred first. The findings are







limited to describing the concurrent differences among these two

sociometric groups in terms of some specific variables.

Regarding the proposed mechanism by which ICPS skill affects

sociometric status, the results indicate that (1) ICPS skill had a

significant, but small, relation to children's teacher-rated social

competence, (2) children's teacher-rated social competence had a

significant, but small, relation to children's sociometric status, and

(3) ICPS skill had no statistically discernible relation to socio-

metric status. At least three interpretations of these results can

be made. First, ICPS skill is not related to sociometric status.

Second, the means-ends conceptualization of ICPS skill is lacking in

external validity. For instance, previous authors have criticized the

idea of the "open-middle" test (in which the interpersonal problems

which are presented to subjects have already been solved) given that

children's problems are not typically automatically solved before the

children have a chance to propose the means to solve the problems,

and given that not all problems are solvable. Third, ICPS skill may

be important only to the extent to which it is effectively translated

into behavior or--perhaps more appropriate to the present investigation

--to the extent to which it is not translated into behavior--a finding

which appeared to typify the rejected group vis-a-vis the popular

group. The results suggest that depression and low motivation may

account for the failure of rejected children to translate ICPS

knowledge into socially competent behavior.

The present study represents the first attempt to compare ICPS to

physical attractiveness and academic achievement in terms of the

relative strength of their relationships to sociometric status. Given








this fact, no findings exist which would permit a direct assessment of

the validity of the present study's findings regarding the relative

strength of the relationship among these variables. However, there

does exist some data regarding the validity of the obtained correla-

tions of physical attractiveness and academic achievement with socio-

metric status--against which the correlation between ICPS and socio-

metric status was evaluated in this study. Regarding the strength of

the relationship between academic achievement and sociometric status,

the obtained results disagree with those of Green et al. (1980) who

found that academic achievement was significantly positively related

to positive nominations and negatively related to negative nominations.

Several differences between the present study and that of Green et al.

(1980), which may account for this discrepancy, include variations in

(1) the computation of positive and negative nomination scores,

(2) the choice and computation of achievement scores--as well as the

tests from which the scores were obtained, and (3) age and cultural

background of the samples. The magnitude of the correlations obtained

by Green et al. (1980) suggest that, relative to the nonrural sample

of the present study, academic achievement may be an important pre-

dictor of sociometric status in rural samples--and perhaps relative to

ICPS and physical attractiveness (which were not evaluated by these

authors). As such, other investigators may wish to examine the

relative abilities of ICPS and academic achievement to predict socio-

metric status in rural samples relative to nonrural samples.

Regarding the strength of the relationship between physical

attractiveness and sociometric status, the present results support

previous global findings (i.e., Kleck et al., 1974; Langlois &







Stephan, 1977) which suggest that accepted children receive independent

ratings of physical attractiveness which are higher than those of low-

accepted children. The present findings are somewhat discrepant,

however, from the results of Dodge (1983) who found, anong a sample of

56 second grade boys whose physical attractiveness was judged from

head-and-shoulders photographs by undergraduates, that neglected,

average, and popular children were rated as significantly more attrac-

tive than controversial children, who were more attractive than

rejected children. The major similarities between Dodge's (1983) study

and the present study is the rank order position of the popular

children who received the highest or next-to-highest ratings in each

study and the rejected children who received the lowest or next-to-

lowest ratings in each study. In general, however, the agreement

between this study and that of Dodge (1983) is certainly not compelling

and appears even less noteworthy given that there was only a single

agreement in terms of significant group differences, with controversial

children viewed as more attractive than rejected children in both

studies. The primary differences between Dodge's (1983) findings and

those of this investigation were in terms of the rank order positions

of the controversial and neglected groups, which were most and least

attractive, respectively, for the present study and fourth and third

most attractive, respectively, in Dodge's (1983) study. In addition,

this study's obtained group differences in terms of attractiveness

were not as clear-cut as those found by Dodge (1983), with the only

clearly significant differences including (1) controversial children

being viewed as more attractive than average, rejected, and neglected







children and (2) popular children being viewed as more attractive than

neglected children.

Potential limitations of the Dodge (1983) study include the facts

that (1) physical attractiveness ratings were made by undergraduate

students as opposed to the children's peers and (2) sociometric group

assignment was not statistically-derived but was apparently based on

a visual inspection of the obtained sociometric data. In addition,

Dodge's (1983) results may have differed from those of the present

study because of (1) a smaller number of raters (and, thus, poten-

tially decreased reliability of ratings?), (2) younger (second grade)

subjects, (3) totally male subject sample, and (4) the fact that

subjects in Dodge's (1983) study had known each other for only a

short period of time.

In general, the sociometric group differences obtained by Dodge

(1983), although somewhat at variance with the pattern of group dif-

ferences in the present study--which was potentially due to differ-

ences in the studies, generally support the finding of this investi-

gation that physical attractiveness is an important predictor of

children's sociometric status.

In any case, the results of this study, notwithstanding these

post hoc explanations, do not support the broadly-stated position of

Shure and Spivak that ICPS skill is an important mediator of social

adjustment for all children. Given the present results--which sug-

gest an apparent lack of validity of the ICPS construct, it appears

that researchers engaged in ICPS treatment research with children are

somewhat premature in their endeavors. Recent research by Kendall

and Fishler (1984) which focused on validational analyses of








ICPS measures, appears to be appropriate and much-needed. Utilizing

a correlational format, these authors investigated several concep-

tualizations of ICPS (i.e., means-ends thinking, identified obstacles,

alternative solutions, and consequential thinking) and their relation

to behavior assessed in an analogue family problem-solving session and

found some evidence for the ecological validity of children's alter-

native solutions and consequential thinking in terms of corresponding

problem-solving behaviors, but none for children's means-ends thinking

or identified obstacles. In addition, these authors noted that neither

children's problem-solving thinking nor problem-solving behavior were

related to behavioral adjustment ratings made by parents (Child

Behavior Checklist--Achenbach, 1970; Achenbach & Edelbrock, 1970) or

by teachers (Devereaux Elementary School Behavior Scale--Spivak et al.,

1976). They also found that ICPS skills were "related to other

problem-solving behaviors besides the ones to which they should most

closely correspond, thereby suggesting a lack of discriminant

validity." Limitations of the findings, which were discussed pre-

viously, included the analogue nature of the assessment setting and

the fact that children were required to solve the problems in an

interaction with their parents instead of with their peers.

Aside from stressing the necessity of validational studies, other

authors have criticized current conceptulaizations of ICPS on two

counts. First, the interpersonal problems with which children are

typically presented are all successfully solved, whereas real-life

problems are not always solvable. It has been suggested that the

manner in which children respond to their failure to attain a given

goal and deal with the potentially resulting frustration may be a more







potent predictor of sociometric status than are differences in the

manner in which they respond to a solvable problem. This hypothesis

is intuitively appealing given the author's personal bias that the

quality of persons' interpersonal relationships may often depend

upon their ability to continue to effectively utilize their cognitive

resources in the face of frustration, anxiety, and other affective/

motivational disruption. Given that the affective and motivational

differences among sociometric groups found in this study is consis-

tent with this hypothesis, it appears that the hypothesis certainly

bears testing.

Secondly, Renshaw and Asher (1982) have criticized the ICPS

literature for its focus on children's strategy selection in inter-

personal problem-solving situations to the exclusion of attention to

their goal selection in problem-solving situations. "Presenting

children with goals ignores the fact that social situations are ill

defined and that the selection of social goals may be a particularly

important aspect of children's social problem-solving" (p. 387). These

authors hypothesize that "part of the social skill deficit of unpopu-

lar children consists of their selection of goals that are inappropri-

ate to particular social situations (Asher & Renshaw, 1981). If

children pursue inappropriate goals in social situations, then their

strategies will likewise appear inappropriate, even though they may

know appropriate strategies for achieving particular goals" (p. 386).

Support for this hypothesis is provided by Renshaw and Asher (1983)

who found, in a sample of middle-class third through sixth grade

children, that children who received high scores on a peer play

rating measure suggested significantly more positive-outgoing goals







on an ICPS measure than did children who received low scores on the

peer play rating measure.

These criticisms of current conceptualizations of ICPS suggest

that the ICPS construct be altered or at least expanded to include

measures which are sensitive to qualitative differences in (1) the

manner in which children deal with unsolvable problems and the

affective/motivational disruption often inherent therein and (2) the

types of goals they select in problem-solving. These additional ICPS

measures, along with existing measures, should be subjected to rigor-

ous validational testing. Such expansion of the ICPS construct and

validational testing of the measures which operationalize the various

conceptualizations of this construct should provide researchers with

a better understanding of the construct. Specifically, their under-

standing will be expanded to include knowledge of which, if any, ICPS

conceptualizations and measures are clinically significant predictors

of present and future, sociometrically diagnosed interpersonal

maladjustment.

An implication of the present finding that physical attractive-

ness significantly predicted sociometric status, whereas ICPS did not,

is that researchers in the social skills training and social compe-

tence areas who are interested in enhancing the sociometric status of

unpopular children should attend to the physical attractiveness

dimension and evaluate its contribution to therapeutic outcome.

Although the physical attributes of each child allow only a limited

amount of alteration, perhaps researchers should incorporate grooming

behavior and education regarding the importance of appearance into

their treatment regimens. Unfortunately, the particular aspects of







physical attractiveness to which children were responding in making

their ratings were not assessed in this study. However, it seems

likely that instruction of unpopular children who have a limited

natural physical endowment regarding neatness of appearance in terms

of such variables as hair, clothing, and personal cleanliness may

help them to make maximum use of their physical attributes.

In summary, the results of the present study failed to support the

position of Shure and Spivak that ICPS is an important mediator of

social adjustment in children. Although teacher ratings of socially

competent behavior were significantly correlated with two of the five

ICPS variables, these correlations were extremely small and indicated

only minimal shared variance. None of the five ICPS variables was

significantly correlated with the sociometric variables or discrimi-

nated among relatively homogeneous sociometric groups derived from an

elaborate classification system. It was found that physical attrac-

tiveness was better at discriminating among groups and a more powerful

predictor of sociometric variable scores than was ICPS. This study's

finding that physical attractiveness is a significant correlate of

sociometric status is supported by the work of other authors (Dodge,

1983; Kleck et al., 1974; Langlois & Stephan, 1977). Potential

reasons for the lack of relationship between ICPS and sociometric

status were offered, including an affect- and motivation-based

explanation of rejected children's inability to translate ICPS

knowledge into socially competent behavior. The results of the

present study and the Kendall and Fishler (1984) study suggest that

future validational efforts are necessary with regard to the ICPS

construct. These efforts should focus on the validity of the various







ICPS measures currently being utilized by researchers in order to

obtain more information regarding the validity of each of the con-

ceptualizations of ICPS in addition to the validity of the ICPS

construct. In addition, the ICPS construct should be expanded to

include conceptualizations regarding (1) the manner in which children

deal with unsolvable interpersonal problems and (2) the types of goals

that they select in interpersonal problem-solving situations. It was

also suggested that the importance of physical attractiveness in

terms of children's sociometric status continue to be evaluated and

that perhaps physical attractiveness should be incorporated into

social skills treatment regimens for unpopular children.













REFERENCES


Achenbach, T. M. (1978). The Child Behavior Profile: I. Boys
aged 6 through 11. Journal of Consulting and Clinical Psychology,
46, 478-488.

Achenbach, T. M., & Edelbrock C. S. (1979). The Child Behavior
Profile: II. Boys aged 12-16 and girls aged 6-11 and 12-16.
Journal of Consulting and Clinical Psychology, 47, 223-233.

Allen, G. J., Chinsky, J. M., Larcen, S. W., Lochman, J. E., Selinger,
H. V. (1976). Community psychology and the schools: A
behaviorally oriented multilevel preventive approach. Hillsdale,
N.J.: Erlbaum.

Asher, S. R., & Hymel, S. (1981). Children's social competence in
peer relations: Sociometric and behavioral assessment. In
J. D. Wine & M. D. Smye (Eds.), Social Competence. New York:
Guilford Press.

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

Butler, L., & Meichenbaum, D. (1981). The assessment of interpersonal
problem-solving skills. In P. C. Kendall and S. D. Hollon
(Eds.), Assessment strategies for cognitive-behavioral interven-
tions. New York: Academic Press.

Carlson, G. A., & Cantwell, D. P. (1980). Unmasking masked depres-
sion in children and adolescents. American Journal of Psychiatry,
137, 445-449.

Castenada, A., McCandless, B. R., & Palermo, D. A. (1956). The
Children's Form of the Manifest Anxiety Scale. Child Develop-
ment, 27, 317-326.

Charlesworth, R., & Hartup, W. W. (1967). Positive social reinforce-
ment in the nursery school peer groups. Child Development, 38,
993-1002.

Coie, J. D., Dodge, K. A., & Coppotelli, H. (1982). Dimensions and
types of social status: A cross-age perspective. Developmental
Psychology, 18, 557-570.








Coie, J. D., & Kupersmidt, J. B. (1983). A behavioral analysis of
emerging social status in boys' groups. Child Development, 54,
1400-1416.

Conger, J. C., & Keane, S. P. (1981). Social skills intervention
in the treatment of isolated or withdrawn children. Psychologi-
cal Bulletin, 90, 478-495.

Costello, C. G. (1980). Childhood depression. In E. J. Mash &
L. G. Terdal (Eds.), Behavioral assessment of childhood
disorders. New York: Guilford Press.

Cowen, E. L., Pederson, A., Babigian, H., Izzo, L. D., & Trost, M. A.
(1973). Long-term follow-up of early detected vulnerable
children. Journal of Consulting and Clinical Psychology, 41,
438-446.

Crandall, V. C., Crandall, V. J., & Katkovsky, W. (1965). A
children's social desirability questionnaire. Journal of
Consulting Psychology, 2j9, 27-36.

Dodge, K. A. (1983). Behavioral antecedents of peer social status.
Child Development, 54, 1386-1399.

Gesten, E. L. (1976). A health resources inventory: The development
of a measure of the personal and social competence of primary-
grade children. Journal of Consulting and Clinical Psychology,
44, 775-786.

Gottman, J. (1977). The effects of a modeling film on social isola-
tion in preschool children: A methodological investigation.
Journal of Abnormal Child Psychology, 5, 69-78.

Gottman, J., Gonso, J., & Schuler, P. (1976). Teaching social skills
to isolated children. Child Development, 4, 179-197.

Green, K. D., Forehand, R., Beck, S. J., & Vosk, B. (1980). An
assessment of the relationship among measures of children's
social competence and children's academic achievement. Child
Development, 51, 1149-1156.

Gresham, F. M., & Nagle, R. J. (1980). Social skills training with
children: Responsiveness to modeling and coaching as a function
of peer orientation. Journal of Consulting and Clinical Psychology,
48, 718-729.

Hallinan, M. T. (1981). Recent advances in sociometry. In S. A.
Asher & J. Gottman (Eds.), The development of children's
friendships. New York: Cambridge University Press.

Harter, S. H. (1979). Manual for the Perceived Competence Scale for
Children (Form 0). Denver: University of Denver (Colorado
Seminary).







Harter, S. H. (1982). The Perceived Competence Scale of Children.
Child Development, 53, 87-97.

Hartup, W. W., Glazer, J. A., & Charlesworth, R. (1967). Peer
reinforcement and sociometric status. Child Development, 38,
1017-1024.

Hollingshead, A. B. (1957). Two-factor index of social position.
New Haven, Conn.: Department of Sociology, Yale University.

Hops, H., & Greenwood, C. R. (1980). Social skills deficits. In
E. J. Mash & L. G. Terdal (Eds.), Behavioral assessment of
childhood disorders. New York: Guilford Press.

Kane, J. S., & Lawler, E. E., III. (1978). Methods of peer assessment.
Psychological Bulletin, 85, 555-586.

Kazdin, A. E. (1981). Assessment techniques for childhood depression:
A critical appraisal. Journal of the American Academy of Child
Psychiatry, 20, 358-375.

Kendall, P. C., & Fischler, G. L. (1984). Behavioral and adjustment
correlates of problem-solving: Validational analyses of inter-
personal cognitive problem-solving measures. Child Development,
55, 879-892.

Kleck, R. E., Richardson, S. A., & Ronald, L. (1974). Physical
appearance cues and interpersonal attraction in children.
Child Development, 45, 305-310.

Kovacs, M. (1981). Rating scales to assess depression in school-aged
children. Acta Paedopsychiatrica, 46, 305-315.

Ladd, G. W. (1981). Effectiveness of a social learning method for
enhancing children's social interaction and peer acceptance.
Child Development, 52, 171-178.

Ladd, G. W., & Mize, J. (1983). A cognitive-social learning model
of social skill training. Psychological Review, 90, 127-157.

La Greca, A., & Santogrossi, D. (1980). Social skills training with
elementary school students: A behavioral group approach.
Journal of Consulting and Clinical Psychology, 48, 220-228.

Langlois, J. H., & Stephan, C. (1977). The effects of physical
attractiveness and ethnicity on children's behavioral attribu-
tions and peer preferences. Child Development, 48, 1694-1698.

Lefkowitz, M. M., & Tesiny, E. P. (1980). Assessment of childhood de-
pressions. Journal of Consulting and Clinical Psychology, 48,43-50.

McClure, L. F., Chinsky, J. M., & Larcen, S. W. (1978). Enhancing
social problem-solving performance in an elementary school
setting. Journal of Educational Psychology, 70, 504-513.







O'Conner, R. D. (1969). Modification of social withdrawal through
symbolic modeling. Journal of Applied Behavior Analysis, 2,
15-22.

Oden, S., & Asher, S. R. (1977). Coaching children in social skills
for friendship making. Child Development, 48, 495-506.

Putallaz, M., & Gottman, J. (1983). Social relationship problems
in children: An approach to intervention. In B. B. Lahey &
A. E. Kazdin (Eds.), Advances in clinical child psychology
(Vol. 6). New York: Plenum.

Renshaw, P. D., & Asher, S. R. (1982). Social competence and peer
status: The distinction between goals and strategies. In K. H.
Rubin & H. S. Ross (Eds.), Peer relationships and social skills
in childhood. New York: Springer-Verlag.

Renshaw, P. D., & Asher, S. R. (1983). Children's goals and
strategies for social interaction. Merrill-Palmer Quarterly,
29, 353-374.

Richard, B. A., & Dodge, K. A. (1982). Social maladjustment and
problem-solving in school-aged children. Journal of Consulting
and Clinical Psychology, 50, 226-233.

Rickel, A. U., Eshelman, A. K., & Loigman, G. A. (1983). Social
problem solving training: A follow-up study of cognitive and
behavioral effects. Journal of Abnormal Child Psychology, 11,
15-28.

Ridley, C. A., & Vaughn, S. R. (1982). Interpersonal problem
solving: An intervention program for preschool children.
Journal of Applied Developmental Psychology, 3, 177-190.

Roff, M. (1961). Children's social interactions and young adult bad
conduct. Journal of Abnormal Social Psychology, 63, 333-337.

Roff, M., Sells, B., & Golden, M. M. (1972). Social adjustment and
personality development in children. Minneapolis: University
of Minnesota Press.

Sarason, S. B., Davidson, K. S., Lighthall, F. F., Waite, R. R., &
Ruebush, B. K. (1960). Anxiety in elementary school children.
New York: Wiley.

Sharp, K. C. (1981). Impact of interpersonal problem-solving
training on preschoolers' social competence. Journal of
Applied Developmental Psychology, 2, 129-143.

Shure, M. B., & Spivak, G. (1972). Means-ends thinking, adjustment,
and social class among elementary-school-aged children.
Journal of Consulting and Clinical Psychology, 38, 348-353.







Shure, M. B., & Spivak, G. (1978). Problem-solving techniques in
childrearing. San Francisco: Jossey-Bass.

Shure, M. B., Spivak, G., & Jaeger, M. (1971). Problem-solving
thinking and adjustment among disadvantaged preschool children.
Child Development, 42, 1791-1803.

Spielberger, C. D., Edwards, C. D. Lushene, R. E. Montuori, J., &
Platzek, D. (1973). Manual for the State-Trait Anxiety Inven-
tory for Children ("How I Feel Questionnaire"). Palo Alto:
Consulting Psychologists Press, Inc.

Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual
for the State-Trait Anxiety Inventory (Self Evaluation Ques-
tionnaire). Palo Alto: Consulting Psychologists Press.

Spivak, G., Platt, J. J., & Shure, M. B. (1976). The problem-solving
approach to adjustment. San Francisco: Jossey-Bass.

Spivak, G., & Shure, M. B. (1974). Social adjustment of young
children: A cognitive approach to solving real-life problems.
San Francisco: Jossey-Bass.

Spivak, G., & Spotts, J. (1966). Devereaux child behavior rating
scale manual. Devon, Pa.: Devereaux Foundation.

Strain, P. S., & Fox, J. J. (1981). Peer social initiations and the
modification of social withdrawal: A review and future perspec-
tive. Journal of Pediatric Psychology, 6, 416-433.

Tesiny, E. P., & Lefkowitz, M. M. (1982). Childhood depression:
A 6-month follow-up study. Journal of Consulting and Clinical
Psychology, 50, 778-780.

Ullman, C. A. (1957). Teachers, peers, and tests as predictors of
adjustment. Journal of Educational Psychology, 48, 257-267.

Urbain, E. S., & Kendall, P. C. (1980). Review of social cognitive
problem-solving interventions with children. Psychological
Bulletin, 88, 109-143.

Vosk, B., Forehand, R., Parker, J. B., & Rickard, K. (1982). A
multimethod comparison of popular and unpopular children.
Developmental Psychology, 18, 571-575.

Weissberg, R. P., Gesten, E. L., Rapkin, B. D., Cowen, E. L.,
Davidson, E., Flores de Apodaca, R., & McKim, B. J. (1981).
Evaluation of a social problem-solving training program for
suburban and inner-city third-grade children, Journal of Con-
sulting and Clinical Psychology, 49, 251-261.

Winett, R. A., & Winkler, R. C. (1972). Current behavior modifica-
tion in the classroom: Be still, be quiet, be docile. Journal
of Applied Behavior Analysis, 5, 499-504.

























APPENDIX
INFORMED CONSENT FORM








INFORMED CONSENT TO PARTICIPATE IN RESEARCH


J. HILLIS MILLER HEALTH CENTER

UNIVERISTY OF FLORIDA

GAINESVILLE, FLORIDA 32610



You are being asked to volunteer as a participant in a research study.
This form is designed to provide you with information about this study
and to answer any of your qeustions.


1. TITLE OF RESEARCH STUDY



A Developmental Study of Children's Friendships




2. PROJECT DIRECTOR


Name: Micheal L. Kieffer, M.S.

Telephone Number: (904) 371-2693



3. THE PURPOSE OF THE RESEARCH


This study is designed to investigate the extent to which several
factors predict how well children are liked by classmates. These
factors are children's academic achievement, physical attractiveness,
and ability to solve interpersonal problems with classmates. In
particular, this study seeks to discover which of these three factors
best predicts how well children are liked by classmates.







Page 2

4. PROCEDURES FOR THIS RESEARCH

This study will be conducted at your child's school. With your per-
mission, your child will be given six questionnaires to fill out.
These questionnaires ask about your child's thoughts, moods, concerns,
and ideas about how to make friends and how to handle problems with
classmates. An example of a question is "Name three classmates that
you like the most and name three that you like the least." In addi-
tion, a photograph of your child will be taken by me. Other children
of the same age who do not attend you child's school and who do not
know your child will rate how attractive they find the picture. The
purpose of this rating is to see whether the ratings which children's
pictures receive has to do with how well-liked they are. Your child
will also be asked to rate the photographs of children who do not
attend his/her school and whom he/she does not know. All children
who participate will be instructed not to discuss their answers with
anyone. Finally, your child's most recent standardized achievement
test scores will be obtained by me from the school files. Achievement
test scores will allow me to see to what extent these test scores
predict how well-liked children are. The entire procedure will be
conducted in one or two sessions and will take a total of no more
than one hour. All information obtained about your child will be held
strictly confidential and your child's name will not be associated
with any of this information. Your child can withdraw from this project
at any time should you desire or should your child choose to do so.
I have enclosed a general information questionnaire for you to com-
plete. This information will help me a great deal. I would appreciate
your completing this form and returning it to me with the signed
consent form.


5. POTENTIAL RISKS OR DISCOMFORTS

There are no known risks associated with the procedures involved.


6. POTENTIAL BENEFITS TO YOU OR TO OTHERS

For participants, there do not appear to be any specific benefits
associated with the procedures involved. However, you child's partici-
pation in this study will help me to learn whether children's academic
achievement, their physical attractiveness, or their ability to solve
interpersonal problems with their classmates is most predictive of how
well-liked they are. If you are interested in the results of this
study, I would be happy to send them to you. Thank you for your
cooperation.


7. ALTERNATIVE TREATMENT OR PROCEDURES, IF APPLICABLE


Not applicable







Page 3


8. GENERAL CONDITIONS

I understand that I will / will not X receive money
for my participation in this study. If I am compensated, I will
receive



I understand that Iwill /will not X be charged addi-
tional expenses for my participation in this study. If I am charged
additional expenses these will consist of



I understand that I am free to withdraw my/my child's consent and
discontinue participation in this research project at any time
without this decision affecting my/my child's medical care.

In the event of my/my child sustaining a physical injury which is
proximately caused by this experiment, no professional medical
care received at the J. Hillis
Miller Health Center exclusive of hospital expenses will be provided
me without charge. This exclusion of hospital expenses does not
apply to patients at the Veterans Administration Medical Center
(VAMC) who sustain physical injury during participation in VAMC-
approved studies. It is understood that no form of compensation
exists other than those described above.

All data obtained from this research will remain confidential. The
University of Florida will protect the confidentiality of this
document and your records from this research to the extent provided
by law.


9. SIGNATURES

I have fully explained to
the nature and purpose of the above-described procedure and the
benefits and risks that are involved in its performance. I have
answered and will answer all questions to the best of my ability.
I may be contacted at telephone number (904) 371-2693



2/1/84
Signature of Person Obtaining Consent Date







Page 4



I have been fully informed of the above-described procedure with its
possible benefits and risks and I have received a copy of this
description. I have given permission of my/my child's participation
in this study.




Signature of Patient or Subject or Date
Relative or Parent or Guardian (specify)




Signature of Child (7 to 17 yrs. of age) Date


Signature of Witness


Date







General Information


Relation to child


Your child's name _

Race (check one):

( ) Black
( ) White Hispanic
( ) White non-Hispanic
( ) Other (specify ____)

Your marital status (check one):

( ) Never married
( ) Separated, widowed, divorced
( ) Married


What is the highest year of school you
have completed?

Mother

( ) Less than 8th grade (
( ) 8th-llth grade (
( ) 12th grade (high school graduate)(
( ) Some college (
( ) College graduate (
( ) Graduate or professional degree (


Sex Age


and your child's other parent


Father

) Less than 8th grade
) 8th-llth grade
) 12th grade (high school graduate)
) Some college
) College graduates
) Graduate or professional degree


In what type of job are you and your child's other parent currently
employed and what is (are) your job titless?






PLEASE RETURN THIS PAGE AND THE SIGNED CONSENT FORM. THANK YOU FOR
YOUR COOPERATION.


Your name













BIOGRAPHICAL SKETCH


Michael L. "Mike" Kieffer was born in Iowa City, Iowa, on

September 20, 1957. He grew up primarily in South Bend, Indiana, and

Lakeland, Florida. He received his B.A. in psychology from the

University of Notre Dame in 1979 and his M.S. in clinical psychology

from the University of Florida in 1983. He plans to complete his

doctoral training via a one-year predoctoral internship in clinical

psychology at the Indiana University School of Medicine in Indianapolis

for the 1984-85 year. Mr. Kieffer's professional interests are pri-

marily clinical in nature and include psychotherapy, family and

marital therapy, interpersonal conceptualizations of personality

development and psychopathology, childhood depression, and children's

social competence. Extracurricular interests include tennis, golf,

fishing, running, and sailing. Mr. Kieffer lives with his wife of

three years, Fi, and his 10-month old son, Christopher, and credits

Fi with a large share of the responsibility for his success in his

graduate school work.










I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




Lawrence J. Siege Cairman
Associate Professor of Clinical
Psychology






I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.




Jam s H. Johnson
Associate Professor of Clinical
Psychology






I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.



Rudy E. Vuchinich
Assistant Professor of Psychiatry









I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.


Drew Bradlyn
Assistant Professor of Clini6a
Psychology


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is
fully adequate, in scope and quality, as a dissertation for the degree
of Doctor of Philosophy.


This dissertation was submitted to the Graduate Faculty of the College
of Health Related Professions and to the Graduate School, and was
accepted as partial fulfillment of the requirements for the degree of
Doctor of Philosophy.

August 1985


Dean, College of Health Related
Professions


Dean for Graduate Studies and Research








































UNIVERSITY OF FLORIDA
II 111111111111111111111111 1 111111 111111111i1 I
3 1262 08554 5019




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EUSOU3UXX_7YPGDW INGEST_TIME 2012-02-20T21:18:01Z PACKAGE AA00009101_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES