With their well-established predictive power for negative long-term
outcomes such as delinquency, poor social functioning, incarceration,
substance abuse, violence towards others, and school drop-out, the early
identification of social skills deficits and often co-occurring conduct
problems is crucial (Coie, 1990; Loeber & Farrington, 2000; Olson
& Hoza, 1993; Parker & Asher, 1987; Walker & Sprague, 1999).
Antisocial behavior results in antisocial attitudes, lack of school
readiness, peer and teacher rejection, lack of emotion regulation
skills, difficulty following school rules, and severe tantrum behavior
in the short term (Walker & Sprague, 1999). Walker and
Sprague's research in the school setting has found that
teacher's ratings of social skills are one of three important
measures to identify a child at-risk for delinquency, with the other two
being the discipline record of the child and negative behavior displayed
toward peers during recess. With these three measures, 80% of their
sample of antisocial fifth graders were correctly classified as being
arrested or not arrested five years later as tenth graders. Certainly
there are adolescents who are arrested that do not have this early
at-risk background, but these young adults do not tend to have the same
severely violent and negative outcomes into adulthood (Walker &
Sprague, 1999). These results point to the importance not only of early
identification and intervention, but of accurate teacher ratings of
social skills deficits.
In a review of the literature between 1991 and 2002 on behavior
problems in low-income children, Qi and Kaiser (2003) found that
behavior problems have been noted in 16-30% of children in Head Start.
In a transactional model proposed by the authors, social skills deficits
was one of the risk factors associated with behavior problems (Qi &
Kaiser, 2003). Likewise, Campbell (1994) found that boys who were
"hard-to-manage" during the preschool years also had behavior
problems two years later, and that these difficulties were also related
to poor social competence according to ratings by mother, father,
teacher, and the boys themselves. More recently, Keane and Calkins
(2004) found that parent ratings of two-year-olds' externalizing
behavior predicted both social skill level and externalizing behavior
rated by teachers and peer reports of social behavior in Kindergarten.
Clearly, the level of social skills and conduct problems exhibited early
on has lasting impact in later school years.
Preschoolers, however, present some major assessment obstacles due
to their limited ability to provide accurate self- and peer-reports
(Foster, Inderbitzen, & Nangle, 1993). Alternate methods are not
without their own particular concerns. For instance, the ability of
direct observation to gather representative portrayals of low occurrence
behaviors such as aggression has been questioned, and analogue measures
have come under increasing scrutiny as well (Bierman & Welsh, 2000;
Foster et al., 1993). As a result, standardized teacher ratings are
commonly recommended measures of social skills and conduct problems with
preschoolers.
Consequently, comprehensive teacher rating measures, such as the
Social Skills Rating System for Teachers, Preschool Level (SSRS-T;
Gresham & Elliot, 1990) have generated much research and clinical
interest because of their purported ability to assess multiple
dimensions of children's social behavior and conduct problems in a
cost efficient manner. Our review of the literature found a fair degree
of support for the psychometric soundness of the SSRS, but also raised
some unanswered questions. In general, the SSRS is considered to be a
comprehensive measure of social skills in children (Demaray et al.,
1995) that is time-efficient for teachers (Bramlett, Dielmann, &
Smithson, 1999). One advantage of the SSRS is that the items are
primarily behavioral, requiring a low level of inference on the part of
the teacher (Benes, 1994).
The manual for the SSRS-T cites adequate reliability for the
Preschool Level. Internal consistency using Cronbach's alpha was
.94 for the Social Skills scale (Cooperation = .90, Assertion = .90,
Self-Control = .91) and .82 for the Problem Behavior scale
(Externalizing = .85, Internalizing = .74) of the SSRS-T (Gresham &
Elliott, 1990). Test-retest reliability was not evaluated on the
preschool sample, however. Four-week test-retest reliability
coefficients were .85 for the Social Skills scale and .84 for the
Problem Behaviors scale of the SSRS-T at the elementary level (Gresham
& Elliott, 1990). It is important to note that the standardization
preschool sample is considered experimental, due to the fact that its
representation was lacking and the number of participants was relatively
low at approximately 200 (Furlong & Karno, 1994).
More direct efforts to validate the SSRS-T, Preschool Level have
also yielded supportive evidence. The SSRS-T was able to discriminate
between typically-developing preschoolers and those who were
developmentally delayed, diagnosed with Attention Deficit/Hyperactivity
Disorder, and enrolled in Head Start (Byrne, Bawden, DeWolfe, &
Beattie, 1998; Lyon, Albertus, Birkinbine, & Naibi, 1996; VonBrock
Treuting & Elliott, 1997), demonstrating concurrent validity. With
respect to evidence supporting convergent validity, the Social Skills
scale and the Problem Behaviors scale of the SSRS-T were found to
correlate in the expected directions with teacher ratings on the
Socialization Domain of the Vineland Adaptive Behavior Scales: Classroom
Edition, an author-constructed teacher questionnaire of prosocial
behavior, and the Revised Conners Ratings Scales teacher form (Lyon et
al., 1996; VonBrock Treuting & Elliott, 1997). Studies designed to
examine convergent validity across both domains of positive social
skills and conduct problems using empirically established measures,
however, are necessary in order to further ascertain the validity of
SSRS-T scores at the preschool level. Ideally, these studies would
extend methods beyond teacher ratings to include self-report and
parent-report ratings as well.
Evidence of the validity of the SSRS-T Preschool Level is also
found in studies that have employed the SSRS-T as a dependent variable.
One study found that children exhibiting more complex play behaviors
during circle time had higher Social Skills scores, while children with
developmentally advanced behaviors had lower Problem Behaviors scores
(Farmer-Dougan & Kaszuba, 1999). The SSRS-T has also been found to
be sensitive to treatment effects (Fantuzzo et al., 1996; Hibbart &
Sprintall, 1995; McKinney & Rust, 1998) and has been used in
validation studies of new instruments due to its widely accepted
empirical validation (e.g., Bain & Pelletier, 1999; Fantuzzo,
Sutton-Smith, Coolahan, Holliday Manz, Canning, & Debnam, 1995).
This instrument has been widely used with culturally diverse
samples, particularly in the Head Start system (e.g., Bramlett et al.,
1999; Fantuzzo et al., 1996; VonBrock Treuting & Elliott, 1997). In
fact, several validation studies of the SSRS-T, Preschool Level have
been conducted on minority samples (Holliday Manz, Fantuzzo, &
McDermott, 1999; Fagan & Fantuzzo, 1999; Fantuzzo, Holliday Manz,
& McDermott, 1998). Fantuzzo et al. (1998) attempted to validate the
SSRS-T Preschool Level with 943 African-American Head Start children.
These authors raised legitimate concerns that the preschool level of the
SSRS was validated on the tryout sample with a longer version of the
questionnaire and the factor structure was never subsequently confirmed
once the shortened questionnaire was finalized. An exploratory factor
analysis revealed a different three-factor internal structure of the
Social Skills scale (i.e., Self-Control, Verbal Assertion, and
Interpersonal Skills) while the Problem Behaviors scale was replicated
(i.e., Externalizing and Internalizing). Two random subsamples were
drawn from the larger sample of children and similar exploratory factor
analyses revealed the same results. Follow-up confirmatory oblique
cluster analysis displayed that this factor structure yielded maximum
item variance without any items subsequently loading onto other factors.
Further, the three Social Skills scales and two Problem Behaviors scales
loaded on a single factor in a higher order factor analysis, but were
inversely related, indicating that these two domains represent opposite
poles of the overall construct of social competence. The degree of
overlap between sociometrics and the SSRS-T was less than six percent
(Fantuzzo et al., 1998), suggesting that measures of peer acceptance are
relatively independent from the measurement of social competence in
preschoolers. No other external validation was performed, however, such
as a comparison of the Externalizing Behavior subscale with an
established measure of conduct problems.
Further validation of the SSRS-T, Preschool Level is necessary. The
published version of the SSRS-T, Preschool Level is based on an
experimental sample of approximately 200 participants with adequate
internal consistency reported in the manual (Gresham & Elliott,
1990). This scale has demonstrated convergent validity with the
Socialization subscale of the Vineland Adaptive Behavior Scales and the
Revised Conners Ratings Scales teacher form (Lyon et al., 1996; VonBrock
Treuting & Elliott, 1997) and the ability to discriminate between
typically-developing children, developmentally delayed children,
children with ADHD, and children attending Head Start (Byrne et al.,
1998; Lyon et al., 1996; VonBrock Treuting & Elliott, 1997). In a
direct attempt to determine the internal structure of the SSRS-T,
Preschool Level with a sample of African-American preschoolers, Fantuzzo
et al. (1998) revealed an alternate factor structure for the Social
Skills scale. Although Fantuzzo and colleagues examined the relationship
between sociometrics and the SSRS-T, Preschool Level, relationships
between other measures of behavior and the scale were not investigated.
Therefore, what is largely missing from this literature is external
validation of the SSRS-T, Preschool Level (i.e., convergent and
discriminant validity)--either the original published SSRS-T, Preschool
Level factor structure or the Fantuzzo factor structure. An additional
question is the applicability of the Fantuzzo factor structure to a
low-income, predominantly Caucasian sample, given the fact that Fantuzzo
and colleagues (1998) conducted their factor analysis of the SSRS-T,
Preschool Level on an entirely African American Head Start sample.
The Present Study
In this study, we conducted an independent validation of one of the
more promising rating systems tailored for preschool use, the Social
Skills Rating System (SSRS; Gresham & Elliott, 1990), in a
predominantly Caucasian Head Start sample. The inclusion of maximally
different forms of assessment (e.g., teacher-report, peer-report, and
behavioral analogue interview) minimizes possible common method variance
contributions (Campbell & Fiske, 1959; Foster & Cone, 1995).
This study represents the first validation of the published SSRS-T,
Preschool Level factor structure against several empirically established
instruments in the literature and including both positive and negative
aspects of social behavior. Furthermore, the present study provides a
comparison between this published factor structure and the factor
structure proposed by Fantuzzo and colleagues in work with minority Head
Start preschoolers. Together, these analyses will further validate the
SSRS-T with respect to multiple domains of social behavior (i.e., both
social skills and conduct problems) by employing maximally different
assessment methods.
Method
Participants
Eighty-two preschool children enrolled in four local Head Start
centers participated (n's = 17, 20, 20, and 25). The participation
rate was 100% in each of the four classrooms due to a blanket consent
achieved through the Parent Council to investigate the effects of a new
curriculum (Carpenter, 2002). All measures reported here were collected
at the baseline data collection point. Forty-two of the preschoolers
were male and 40 were female. The sample was predominantly Caucasian (n
= 77), with two African-American students, two non-Caucasian Hispanic
students, and one Native American student. The age of the preschoolers
ranged from 36 to 62 months (M = 48.4; SD = 7.2).
Measures
Teachers completed three behavior rating scales for each child in
their classroom. Additionally, each child met one-on-one with graduate
or advanced undergraduate research assistants to complete peer
acceptance ratings and the puppet interview. Please see Table 1 for
means and standard deviations of the measures described below.
Social Skills Rating System-Teacher Form. The Preschool Level (ages
3 to 5) of the SSRS-T consists of 40 items scored on a 3-point scale.
The SSRS-T is composed of two main scales: Social Skills (e.g., Makes
friends easily; Gives compliments to peers; Waits turn in games or other
activities) and Problem Behaviors (e.g., Has temper tantrums; Appears
lonely). The Social Skills scale has Standard Scores ranging from 40 to
130, whereas the Problem Behaviors scale scores range from 85 to 145.
Both scales have a mean of 100 and a Standard Deviation of 15. The
original published version of the Social Skills scale has three
subscales: Cooperation, Assertion, and Self-Control. The Problem
Behaviors scale has two subscales: Externalizing and Internalizing.
Subscales are not standardized. Average standard scores on the Social
Skills scale and the Problem Behaviors scale are reported in Table 1.
The psychometric properties of the SSRS-T indicate adequate reliability
and validity (e.g., Byrne et al., 1998; Gresham & Elliott, 1990;
Lyon et al., 1996; VonBrock Treuting & Elliott, 1997).
Aggressive Behavior Subscale of the Child Behavior Checklist
Caregiver-Teacher Report Form for Ages 2-5. Teachers were asked to
complete the 23-item Aggressive Behavior subscale of the Child Behavior
Checklist Caregiver-Teacher Report Form for Ages 2-5 (C-TRF; Achenbach,
1997) for each child in their classroom. The Aggressive Behavior
subscale provides a comprehensive assessment of general conduct
problems. Teachers are asked to rate on a 3-point scale the degree to
which the child is, for example, "Defiant," "Gets in many
fights," and "Teases a lot." As T-Scores on the
Aggressive Behavior subscale range from < 50 to 100, raw scores were
used in the present study to increase variability in the scores. In
other words, raw scores were able to capture the variability of scores
in those subjects classified as "< 50" with a T-score,
rather than each of those subjects receiving the same score. The average
sum is reported in Table 1. This subscale is appropriate to use
independently of the entire C-TRF based upon the reliability and
validity of its scores. In fact, during validation of the C-TRF, the
Aggressive Behavior subscale was found have the strongest psychometric
properties of all the C-TRF subscales (Achenbach, 1997).
Teacher ratings of relational aggression. The relational aggression
factor of the Preschool Social Behavior Scale--Teacher Form (PSBS-T;
Crick, Casas, & Mosher, 1997) was completed by teachers. The
six-item relational aggression factor taps into the construct of
relational aggression with items such as "Tells others not to play
with or be a peer's friend," rated on a one to five Likert
scale. The average sum is reported in Table 1. The items were derived by
face-validity standards and an exploratory factor analysis indicated
that the relational aggression items emerged as a separate factor
independent of the overt aggression, prosocial behavior, and depressed
affect factors. A Cronbach's alpha of .96 on the relational
aggression factor indicates high internal consistency of the subscale
(Crick et al., 1997).
Peer ratings of acceptance. A sociometric rating scale was
administered individually to each child to provide a general measure of
peer acceptance. The children rated how much they liked to play with
each of their classmates by assigning a happy, neutral, or sad face to
pictures of each child shown in random order (Asher, Singleton, Tinsley,
& Hymel, 1979). Adequate reliability and validity have been
demonstrated in preschoolers (e.g., Asher et al., 1979; Boivin &
Begin, 1986; Hymel, 1983; Olson & Lifgren, 1988; Poteat, Ironsmith,
& Bullock, 1986). Average ratings, with a possible range of one to
three, were determined for each child by creating a mean rating based on
each child's ratings by each of his or her classmates. These
average ratings provided a peer report of each child's global
social acceptance or like-ability.
Friendliness ratings from the Enactive Social Knowledge Interview.
(ESKI; Mize & Ladd, 1988). Mize and Ladd's (1988) enactive
social knowledge interview, an analogue behavioral measure, was
individually administered to each child by highly trained graduate or
advanced undergraduate research assistants. The child was given a puppet
and told to pretend that he or she was the puppet. Six hypothetical
social dilemmas were presented to the child who then acted out his or
her response to each situation with the puppet and various props. The
children's responses to the hypothetical social dilemmas were
audiotaped and then transcribed verbatim. Using a coding manual (Mize
& Ladd, 1988), the transcriptions were then coded on a one to five
Likert scale measuring degree of friendliness versus hostility,
resulting in a possible mean score ranging from one to five.
In a sample of preschoolers enrolled in publicly-funded daycare
(i.e., lower socioeconomic status), the enactive social knowledge
interview has demonstrated moderate internal consistency among the six
stories as measured by Cronbach's alpha's of .64 for
friendliness ratings (Mize & Ladd, 1988). Three-week test-retest
reliabilities were r = .77 for friendliness ratings. With respect to
evidence of validity, the friendliness ratings, measured on a continuum
from hostile to friendly, yielded statistically significant correlations
with teacher ratings of prosocial behavior (r = .56), teacher ratings of
aggressive behavior (r = -.46), directly observed prosocial behavior (r
= .32), and directly observed aggressive behavior (r = -.33). The
friendliness ratings did not correlate significantly with peer ratings
of acceptance in this sample (Mize & Ladd, 1988).
The enactive interview provided an objective behavioral and
cognitive assessment of children's social skills, balancing the
teacher ratings of behavior and peer ratings of acceptance in the
present study.
Results
Internal Consistency of the SSRS-T, Preschool Level
To examine the validity of the SSRS-T scores for this sample, we
first conducted an examination of the internal consistency of the
various factors, including both the factors of the published SSRS-T and
the Fantuzzo factor structure (see Table 2).
With .80 being the cutoff for good internal consistency (Henson,
2001), these results indicate that both the original factor structure
and Fantuzzo's factor structure of the Social Skills scale have
good internal consistency. Furthermore, of the Problem Behaviors
subscales, the Externalizing subscale has good internal consistency,
while the Internalizing subscale has only moderate internal consistency.
The more general Problem Behaviors Scale also just missed the criteria
for good internal consistency with a Cronbach's alpha of .78. It is
clear, however, that it is the Internalizing subscale that is lending
instability to the overall Problem Behaviors scale.
Internal Consistency of the Remaining Measures
Cronbach's alphas were also determined for the remaining
multi-item measures. For the Aggressive Behavior subscale of the Child
Behavior Checklist Caregiver--Teacher Report Form (C-TRF; Achenbach,
1997), a Cronbach's alpha of .94 was obtained. The relational
aggression factor of the Preschool Social Behavior Scale--Teacher Form
(PSBS-T; Crick, Casas, & Mosher, 1997) also had good internal
consistency in our sample, with a Cronbach's alpha of .93.
For the Mize and Ladd's (1988) enactive social knowledge
interview, interrater reliability was first calculated. A graduate
research assistant coded all of the data initially. In addition, two
advanced undergraduate research assistants double-coded the same data
(each coded half) for reliability and consensing purposes. Assessed
through Pearson correlation, the interrater reliability of the
friendliness ratings was .92 for Coder 1 and .86 for Coder 2. The two
undergraduate coders and the graduate research assistant consensed any
discrepant ratings and these consensed scores were utilized in the final
analyses. Cronbach's alpha for consensed friendliness ratings of
the six stories was .59.
Comparison of the Published and Fantuzzo Factor Structure
Correlational Patterns
To establish the external validity of the SSRS-T, we conducted an
examination of both the published version of the SSRS-T and that of
Fantuzzo and colleagues' factor structure. The nature of our
assessment data allows us to examine differential correlational patterns
between the two factor structures for convergent validity of the
scales' scores. We first correlated the published factor structure
of the SSRS-T, Preschool Level with the other measures of positive and
negative social behavior to look for the expected meaningful
relationships (see Table 3).
The three subscales had high positive correlations with each other
(range = .70 to .83) as well as high positive correlations with the
Social Skills scale (range = .89 to.94). In addition, all three social
skill subscales and the overall Social Skills scale were negatively
correlated both with the Externalizing subscale (range = -.26 to -.60),
and the C-TRF Aggressive Behavior subscale (range = -.32 to -.61). All
social scales were negatively correlated with the Internalizing
subscale. The correlations with relational aggression were more mixed,
with an overall positive correlation with the Social Skills scale and
with the Cooperation and Assertion subscales but no correlation with the
Self-Control subscale. Peer sociometrics are clearly positively
correlated with all three subscales and the overall Social Skills scale
(range = .30 to .45), as are the ESKI Friendliness Ratings (range = .25
to .30).
We next examined the correlation matrix for the Fantuzzo factor
structure (see Table 4). As with the published subscales of the SSRS,
the Fantuzzo factors also correlate highly with the published SSRS
Social Skills Scale (range = .81 to .91). The intercorrelations among
the three new factors were lower, however (range = .59 to .75). Again
the Internalizing subscale was negatively correlated with all of the
social scales. All but the Verbal Assertion subscale were negatively
correlated with both the Externalizing subscale (range = -.25 to -.61)
and the C-TRF Aggressive Behavior subscale (range = -.36 to-.63). Here,
the Interpersonal Skills and Verbal Assertion factors were also
positively correlated to relational aggression, whereas the Self-Control
factor was again not significantly correlated. Both peer sociometrics
and the ESKI Friendliness Ratings were positively correlated with the
new factors (ranges = .31 to .39 and .22 to .31, respectively).
Utility of the Externalizing Subscale
Results support the utility of the SSRS-T Externalizing subscale as
a measure of conduct problems in preschoolers. Raw scores of the
Externalizing subscale correlate in predicted directions with measures
of prosocial behavior (see Table 5). Specifically, the Externalizing
subscale is negatively correlated with measures of positive social
behavior (i.e., SSRS-T Social Skills and Peer Sociometrics) but not
correlated with Friendliness Ratings upon interview. It is positively
correlated with the well-established C-TRF Aggressive Behavior subscale
and relational aggression but not related to the SSRS-T Internalizing
subscale.
Importantly, the SSRS-T Externalizing subscale has the same
correlation pattern with the other measures as the well-established,
23-item C-TRF Aggressive Behavior subscale.
Discussion
The goal of the present study was to provide evidence for the
validation of the SSRS-T, Preschool Level as an instrument for measuring
the two extremes of social behavior, social skills and conduct problems,
in a predominantly Caucasian Head Start sample. To this end, it was
necessary to examine the original published factor structure of the
Social Skills scale versus the Fantuzzo factor structure. Given the
factor analysis findings of Fantuzzo and colleagues with minority Head
Start children (Fagan & Fantuzzo, 1999; Fantuzzo et al., 1998;
Holliday Manz et al., 1999), researchers and clinicians may be left
wondering if the entire Social Skills scale is even appropriate for use.
The results of the present study clearly demonstrate that both factor
structures correlate highly with the published Social Skills scale.
Further, the original published Social Skills scale correlates in
meaningful and theoretically expected directions with the other measures
of social behavior, rendering these scores to be a valid measure of
social skills. Critics might argue that one possible exception is the
positive correlation with relational aggression. However, recent
research has found the construct of relational aggression to be
multifaceted at the preschool level, consisting of both positive and
negative aspects of social behavior (e.g., Carpenter & Nangle, 2006;
Hawley, 2003; and see Underwood, Galen, & Paquette, 2001 for a
review). Therefore, regardless of which factor structure one prefers,
the more global and standardized Social Skills scale can also be
utilized for ease of interpretation and point of reference across
studies.
The results of the present study do not invalidate the factor
structure of the published SSRS-T, Preschool Level (e.g., Cooperation,
Assertion, and Self-Control). The factors correlate highly with each
other and with the other instruments using maximally different methods
in theoretically meaningfully ways. Evidence of discriminant validity
was found through a small negative correlation between each of the
social skills scales and the internalizing subscale of the SSRS-T,
constructs that are not expected to be positively related. These
relationships were found for both the original factor structure and the
Fantuzzo factor structure. However, researchers may prefer the Fantuzzo
factor structure for several reasons. First, there is less overlap among
the factor intercorrelations, although the three factors are still
moderately and positively correlated with each other and highly
correlated with the more global Social Skills scale. This suggests that
the three factors are measuring different aspects of social competence,
rather than all measuring the same construct three times. Second, the
Verbal Assertion factor is particularly intriguing in that it bears no
association with the Externalizing subscale or the Aggressive Behavior
subscale of the C-TRF, unlike the Assertion subscale of the published
SSRS-T. This indicates that Verbal Assertion may be a positive social
trait independent of behavioral problems. Third, the Interpersonal
Skills factor has less overlap with the conduct problem measures than
does the Cooperation subscale. Again, this is desirable in that it
indicates that the Interpersonal Skills factor is not merely measuring
the inverse of conduct problems, but rather capturing a unique social
construct.
The pattern of correlations is virtually the same for the
Self-Control subscale in both factor structures, which is not surprising
given that they have nine items in common. And, in both factor
structures, the Self-Control subscale is somewhat problematic. The
reason for the difficulty is the high negative correlation with the
conduct problems measures, indicating that the Self-Control measure is
the opposite of externalizing, conduct problem behavior. The salvaging
aspect of the correlation pattern is the lack of a correlation with
relational aggression. An examination of the items on these factors
reveals that Self-Control may also be measuring the inverse of
impulsivity with items such as, "waits turn," "controls
temper with peers," and "follows game rules." Relational
aggression may be entirely unrelated to impulsivity, whereas past
research has clearly found links between impulsivity and aggression
(e.g., Farrington, 1989; Luengo, Carrillo-de-la-Pena, Otero, &
Romero, 1994; Solanto et al., 2001). Future research examining
relational aggression may find the Self-Control subscale useful in
unmasking the relationship, or lack thereof, between relational
aggression and impulsivity.
In order to establish the SSRS-T as an empirically validated
instrument for the assessment of both social skills deficits and conduct
problems, we also investigated the utility of the Externalizing
subscale. Overall, our psychometric analysis provided strong support for
the internal consistency and validity of teacher-rated conduct problems
in a Head Start sample. The SSRS-T Externalizing subscale and the C-TRF
Aggressive Behavior subscale converged with one another and were
inversely associated with maximally different prosocial behavior
indices. Of note, neither instrument was associated with the SSRS-T
Internalizing subscale or the Friendliness Ratings from the analogue
behavioral interview, providing evidence of discriminant validity. In
our view, it is not necessarily expected that preschool children with
externalizing behavioral problems would have internalizing problems as
well, although this is sometimes the case. With respect to friendly
responses on the interview, this may indicate that children with
behavioral difficulties are capable of being friendly under typical
circumstances, which may not be related to the extreme behavior ratings
culled by the Externalizing subscale.
Given the methodological obstacles inherent in assessing conduct
problems during this developmental period, teacher ratings appear to be
an invaluable alternative to other more time- and labor-intensive
assessments, such as observations or peer ratings. The 6-item SSRS
Externalizing subscale also places little burden on teachers. Therefore,
the SSRS-T is capable of assessing both positive and negative social
skills in one instrument, providing clinical utility to researcher and
clinician alike. Future research should standardize the Externalizing
subscale on a large, heterogeneous sample in order to maximize this
utility.
The limitations of the present study include the relatively small
sample size that precluded a confirmatory factor analysis of either the
original published SSRS-T factor structure or the Fantuzzo factor
structure of the SSRS-T. It would be useful for future research to
address this caveat. The major goal of the present study, however, was
to support the validation of the SSRS-T using multiple methods, such as
behavioral analogue interview, rendering a large sample size much more
difficult to attain. An additional limitation is the predominantly
Caucasian sample. Given that the Fantuzzo factor structure was derived
entirely on a sample of African-American Head Start preschoolers, this
study suggests an application to Caucasian Head Start preschoolers as
well, closing this gap in the literature. Further validation of the
Fantuzzo factor structure with maximally different, independent
measures, like the ones used in the present study, is still necessary
with a minority population. By this same token, should the published
version of the SSRS-T, Preschool Level continue in its original format,
it too requires further validation studies with minorities. It will be
important to also consider the ethnicity of the raters, as teachers
certainly bring their own biases into ratings of their students. For
example, a behavior may be rated as aggressive by a teacher of one
cultural background and assertive by a teacher of another cultural
background.
We hope that this study clarifies the use of the SSRS-T, Preschool
Level as a psychometrically sound instrument. Both the original SSRS-T
Social Skills factor structure and the Fantuzzo factor structure
correlate in theoretically meaningful ways with measures employing
maximally different methods utilized in the present study. The Fantuzzo
factor structure appears to be preferable by virtue of distinguishing
different aspects of the global construct of social competence,
demonstrated by less factor intercorrelation overlap for the Fantuzzo
factor structure than the original SSRS-T factor structure. Moreover, as
assessed here, the Verbal Assertion Fantuzzo factor is a particularly
nice measure of uniquely positive social skills, independent of
externalizing behavior. Clinicians would be aided by the identification
of clinical cut-off scores for these factors and by the standardization
of the Externalizing Behavior subscale. Overall, the SSRS-T, Preschool
Level is a valuable method of assessing both positive and negative
aspects of social behavior in an efficient manner.
References
Achenbach, T. M. (1997). Guide for the Caregiver-Teacher Report
Form for Ages 2-5. Burlington, VT: University of Vermont Department of
Psychiatry.
Asher, S. R., Singleton, L. C., Tinsley, B. R., & Hymel, S.
(1979). A reliable sociometric measure for preschool children.
Developmental Psychology, 15, 443-444.
Bain, S. K., & Pelletier, K. A. (1999). Social and behavioral
differences among a predominantly African American preschool sample.
Psychology in the Schools, 36, 249-259.
Benes, K. M. (1994). Review of the Social Skills Rating System. In
J.C. Conoley & J.C. Impara (Eds.), The supplement to the eleventh
mental measurements yearbook (pp. 234-236). Lincoln, NE: The University
of Nebraska Press.
Bierman, K. L. & Welsh, J. A. (2000). Assessing social
dysfunction: The contributions of laboratory and performance-based
measures. Journal of Clinical Child Psychology, 29, 526-539.
Boivin, M. & Begin, G. (1986). Temporal reliability and
validity of three sociometric status assessments with young children.
Canadian Journal of Behavioral Science, 18, 167-172.
Bramlett, R. K., Dielmann, K. B., & Smithson, M. (1999).
Screening for behavior problems in preschool: A comparison of the Social
Skills Rating System (SSRS) and a structured teacher interview. Research
in the Schools, 6, 47-52.
Byrne, J. M., Bawden, H. N., DeWolfe, N. A., & Beattie, T. L.
(1998). Clinical assessment of psychopharmacological treatment of
preschoolers with ADHD. Journal of Clinical and Experimental
Neuropsychology, 20, 613-627.
Campbell, S. B. (1994). Hard-to-manage preschool boys:
Externalizing behavior, social competence, and family context at
two-year followup. Journal of Abnormal Child Psychology, 22, 147-158.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and
discriminant validation by the multitrait-multimethod matrix.
Psychological Bulletin, 56, 81-105.
Carpenter, E. M. (2002). A curriculum-based approach for
social-cognitive skills training: An intervention targeting aggression
in Head Start Preschoolers. Dissertation Abstracts International:
Section B: The Sciences and Engineering, 63, 3001.
Carpenter, E. M., & Nangle, D. W. (2006). Caught between
stages: Relational aggression emerging as a developmental advance in
at-risk preschoolers. Journal of Research in Childhood Education, 21,
177-188.
Coie, J. D. (1990). Toward a theory of peer rejection. In S. R.
Asher & J. D. Coie (Eds.), Peer Rejection in Childhood (pp.
365-401). Cambridge, England: Cambridge University Press.
Crick, N. R., Casas, J. F., & Mosher, M. (1997). Relational and
overt aggression in preschool. Developmental Psychology, 33, 579-588.
Demaray, M. R., Ruffalo, S. L., Carlson, J., Busse, R. T., Olson,
A. E., McManus, S. M., & Leventhal, A. (1995). Social skills
assessment: A comparative evaluation of six published rating scales.
School Psychology Review, 24, 648-671.
Fagan, J., & Fantuzzo, J. W. (1999). Multirater congruence on
the Social Skills Rating System: Mother, father, and teacher assessments
of urban Head Start children's social competence. Early Childhood
Research Quarterly, 14, 229-242.
Fantuzzo, J., Holliday Manz, P., & McDermott, P. (1998).
Preschool version of the social skills rating system: An empirical
analysis of its use with low-income children. Journal of School
Psychology, 36, 199-214.
Fantuzzo, J., Sutton-Smith, B., Coolahan, K. C., Holliday Manz, P.,
Canning, S., & Debnam, D. (1995). Assessment of preschool play
interaction behaviors in young low-income children: Penn Interactive
Peer Play Scale. Early Childhood Research Quarterly, 10, 105-120.
Fantuzzo, J., Sutton-Smith, B., Atkins, M., Meyers, R., Stevenson,
H., Coolahan, K., Weiss, A., & Manz, P. (1996). Community-based
resilient peer treatment of withdrawn maltreated preschool children.
Journal of Consulting & Clinical Psychology, 64, 1377-1386.
Farmer-Dougan, V., & Kaszuba, T. (1999). Reliability and
validity of play-based observations: Relationship between the PLAY
behavior observation system and standardized measures of cognitive and
social skills. Educational Psychology, 19, 429-440.
Farrington, D. P. (1989). Early predictors of adolescent aggression
and adult violence. Violence and Victims, 4, 79-100.
Foster, S. L., & Cone, J.D. (1995). Validity issues in clinical
assessment. Psychological Assessment, 7, 248-260.
Foster, S. L., Inderbitzen, H. M., & Nangle, D. W. (1993).
Assessing acceptance and social skills with peers in childhood: Current
issues. Behavior Modification, 17, 255-286.
Furlong, M. & Karno, M. (1994). Review of the Social Skills
Rating System. In J. C. Conoley & J. C. Impara (Eds.), The
supplement to the eleventh mental measurements yearbook (pp. 236-238).
Lincoln, NE: The University of Nebraska Press.
Gresham, F. M. & Elliott, S. N. (1990). Social Skills Rating
System: Manual. Circle Pines, MN: American Guidance Service.
Hawley, P. H. (2003). Strategies of control, aggression, and
morality in preschoolers: An evolutionary perspective. Journal of
Experimental Child Psychology, 85, 213-235.
Henson, R. K. (2001). Understanding internal consistency
reliability estimates: A conceptual primer on coefficient alpha.
Measurement & Evaluation in Counseling & Development, 34,
177-189.
Hibbart, M. T., & Sprinthall, N. A. (1995). Promoting social
and emotional development of preschoolers: Inclusion and mainstreaming
for children with special needs. Elementary School Guidance and
Counseling, 30, 131-142.
Holliday Manz, P. H., Fantuzzo, J. W., & McDermott, P. A.
(1999). The parent version of the preschool Social Skills Rating Scale:
An analysis of its use with low-income, ethnic minority children. School
Psychology Review, 28, 493-504.
Hymel, S. (1983). Preschool children's peer relations: Issues
in sociometric assessment. Merrill-Palmer Quarterly, 29, 237-260.
Keane, S. P. & Calkins, S. D. (2004). Predicting Kindergarten
peer social status from toddler and preschool problem behavior. Journal
of Abnormal Child Psychology, 32, 409-423.
Loeber, R. & Farrington, D. P. (2000). Young children who
commit crime: Epidemiology, developmental origins, risk factors, early
interventions, and policy implications. Development and Psychopathology,
12, 737-762.
Luengo, M. A., Carrillo-de-la-Pena, M. T., Otero, J. M., &
Romero, E. (1994). A short-term longitudinal study of impulsivity and
antisocial behavior. Journal of Personality and Social Psychology, 66,
542-548.
Lyon, M. A., Albertus, C., Birkinbine, J., & Naibi, J. (1996).
A validity study of the Social Skills Rating System-Teacher Version with
disabled and nondisabled preschool children. Perceptual and Motor
Skills, 83, 307-316.
McKinney, E. P., & Rust, J. O. (1998). Enhancing preschool
African American children's social skills. Journal of Instructional
Psychology, 25, 235-241.
Mize, J., & Ladd, G. W. (1988). Predicting preschoolers'
peer behavior and status from their interpersonal strategies: A
comparison of verbal and enactive responses to hypothetical social
dilemmas. Developmental Psychology, 24, 782-788.
Olson, S. L. & Hoza, B. (1993). Preschool developmental
antecedents of conduct problems in children beginning school. Journal of
Clinical child Psychology, 22, 60-67.
Olson, S. L. & Lifgren, K. (1988). Concurrent and longitudinal
correlates of preschool peer sociometrics: Comparing rating scale and
nomination measures. Journal of Applied Developmental Psychology, 9,
409-420.
Parker, J. G., & Asher, S. R. (1987). Peer relations and later
personal adjustment: Are low-accepted children at risk? Psychological
Bulletin, 102, 357-389.
Poteat, G. M., Ironsmith, M., & Bullock, J. (1986). The
classification of preschool children's sociometric status. Early
Childhood Research Quarterly, 1, 349-360.
Qi, C.H., & Kaiser, A.P. (2003). Behavior problems in preschool
children from low-income families: Review of the literature. Topics in
Early Childhood Special Education, 23, 188-216.
Solanto, M. V., Abikoff, H., Sonuga-Barke, E., Schachar, R., Logan,
G.D., Wigal, T., Hechtman, L., Hinshaw, S., & Turkel, E. (2001). The
ecological validity of delay aversion and response inhibition as
measures of impulsivity in AD/HD: A supplement to the NIMH Multimodal
Treatment Study of AD/HD. Journal of Abnormal Child Psychology, 29,
215-228.
Underwood, M. K., Galen, B. R., & Paquette, J. A. (2001). Top
ten challenges for understanding gender and aggression in children: Why
can't we all just get along? Social Development, 10, 248-266.
VonBrock Treuting, M. V. & Elliott, S. N. (1997). Social
behavior ratings of at-risk preschool children: Comparisons with typical
preschool children by parents and teachers. Canadian Journal of School
Psychology, 13, 68-84.
Walker, H. M., & Sprague, J. R. (1999). The path to school
failure, delinquency, and violence: Causal factors and some potential
solutions. Intervention in School and Clinic, 35, 67-73.
Erika Carpenter Rich
Harbor Regional Center, Torrance, California
Elizabeth J. Shepherd and Douglas W. Nangle
University of Maine
Correspondence to Douglas W. Nangle, Ph.D., Dept. of Psychology,
University of Maine, 5742 Clarence Cook Little Hall, Orono, ME 04469;
e-mail: doug.nangle@umit.maine.edu.
Table 1 Table of means
Mean SD N
SSRS Social Skills 95.13 17.14 n = 79
SSRS Problem Behaviors 98.7 13.0 n = 79
C-TRF Aggressive Behavior 7.16 8.41 n = 79
Relational Aggression 7.57 3.25 n = 82
Peer Acceptance 2.41 .25 n = 77
ESKI--Friendliness Ratings 2.80 .61 n = 77
Table 2 Internal consistency and average sum scores of the SSRS-T
factors
Cronbach's Average Sum
Alpha Score (SD)
Scale N = 79 N = 79
Social Skills Scale (30 Items) .95 35.6 (12.6)
Cooperation (10 Items) .85 13.4 (3.9)
Assertion (10 Items) .92 10.3 (5.5)
Self-Control (10 Items) .90 12.0 (4.4)
Fantuzzo's Interpersonal Skills (9 Items) .90 11.6 (4.4)
Fantuzzo's Verbal Assertion (5 Items) .84 3.8 (2.8)
Fantuzzo's Self-Control (11 Items) .90 13.6 (4.7)
Problem Behaviors Scale (10 Items) .78 4.1 (3.4)
Externalizing (6 Items) .87 3.0 (3.0)
Internalizing (4 Items) .61 1.1 (1.3)
Table 3 Correlation matrix of the published SSRS-T factor structure
1. 2. 3. 4.
1. SSRS-Social Skills --
2. SSRS-Cooperation .93*** --
3. SSRS-Assertion .94*** .83*** --
4. SSRS-Self-Control .89*** .77*** .70*** --
5. SSRS-Internalizing -.29** -.21[dagger] -.32** -.25*
6. SSRS-Externalizing -.45*** -.42*** -.26* -.60***
7. C-TRF Aggressive Behavior -.49*** -.46*** -.32** -.61***
8. Relational Aggression .22* .30** .33** -.03
9. Peer Sociometrics .39*** .28** .45*** .30**
10. ESKI Friendliness Ratings .30** .29** .28** .25*
Note. Pairwise sample sizes range from 75 to 79.
*p < .05. **p < .01. ***p < .001. [dagger]p = .07.
Table 4 Correlation matrix of the SSRS-T Fantuzzo factor structure
1. 2. 3. 4.
1. SSRS-Social Skills --
2. SSRS-Interpersonal Skills .91*** --
3. SSRS-Verbal Assertion .81*** .75*** --
4. SSRS-Self-Control .90*** .69*** .58*** --
5. SSRS-Internalizing -.29** -.26* -.28* -.33**
6. SSRS-Externalizing -.45*** -.25* -.15 -.61***
7. C-TRF Aggressive Behavior -.49*** -.36*** -.14 -.63***
8. Relational Aggression .22* .35*** .37*** -.02
9. Peer Sociometrics .39*** .34** .37*** .31**
10. ESKI Friendliness Ratings .30** .31** .24* .22*
Note. Pairwise sample sizes range from 75 to 79.
*p < .05. **p < .01. ***p < .001.
Table 5 Correlations with the SSRS-T Externalizing Subscale versus the
C-TRF Aggressive Behavior Subscale
SSRS-T Externalizing C-TRF Aggressive
Subscale Behavior Subscale
C-TRF Aggressive Behavior .89*** --
SSRS-T Social Skills -.45*** -.49***
SSRS-T Internalizing .10 .20
Relational Aggression .20[dagger] .24*
Peer Sociometrics -.24* -.26*
ESKI Friendliness -.09 -.16
Note. Pairwise sample sizes range from 77 to 79.
*p < .05. **p < .01. ***p < .001. [dagger]p = .07.