Psychological Assessment © 2010 American Psychological Association

2010, Vol. 22, No. 4, 809 – 815 1040-3590/10/$12.00 DOI: 10.1037/a0020255

Base Rates of Social Skills Acquisition/Performance Deficits, Strengths,

and Problem Behaviors: An Analysis of the Social Skills Improvement
System—Rating Scales

Frank M. Gresham Stephen N. Elliott and Ryan J. Kettler

Louisiana State University Vanderbilt University

Base rate information is important in clinical assessment because one cannot know how unusual or
typical a phenomenon is without first knowing its base rate in the population. This study empirically
determined the base rates of social skills acquisition and performance deficits, social skills strengths, and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

problem behaviors using a nationally representative sample of children and adolescent ages 3–18 years.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Using the national standardization sample of the Social Skills Improvement System—Rating Scales (N ⫽
4,550) across 3 informants (teacher, parent, and student) and across 3 broad age groupings (3–5 years,
5–12 years, and 13–18 years), these base rates were computed. Results showed that the base rates for
social skills acquisition deficits and problem behaviors are extremely low in the general population. Base
rates for social skills performance deficits and social skills strengths were considerably higher, with
students in the 5- to 12-year-old age group reporting fewer performance deficits and more social skills
strengths than older children (13–18 years). Teachers and parents reported more performance deficits and
fewer social skills strengths across all age groups than students in the 5- to 12-year-old age group. These
results are discussed in terms of the utility of base rate information in clinical decision making.

Keywords: base rates, acquisition deficits, performance deficits, problem behaviors, strengths

Children and youth with or at-risk for various behavior disorders portant because one must assess the nature of specific social skills
experience significant difficulties in the development and mainte- difficulties in order to more effectively design and deliver specific
nance of satisfactory interpersonal relationships, exhibition of intervention approaches in remediating social skills deficits. As
prosocial behavior patterns, and social acceptance by peers and such, social skills deficits have been broadly classified into two
teachers (American Psychiatric Association, 1994; Gresham, 1997, basic types: acquisition deficits and performance deficits.
1998; Maag, 2005, 2006; Walker, Ramsay, & Gresham, 2004). Acquisition deficits result from either the absence of knowledge
These social competence deficits often lead to short-term, inter- about how to perform a given social skill or difficulty in knowing
mediate, and long-term difficulties in domains of educational, which social skill is appropriate in specific situations (Gresham,
psychosocial, and vocational functioning (Kupersmidt, Coie, & 1981a, 1981b, 2002). Based on the above conceptualization, ac-
Dodge, 1990; Newcomb, Bukowski, & Pattee, 1993; Parker & quisition deficits can result from deficits in social-cognitive abil-
Asher, 1987). The fact that most children with or at-risk for ities or difficulties or deficits in appropriate discrimination of
behavior disorders exhibit severe social competence deficits man- social situations. Acquisition deficits can be characterized as
dates that professionals conduct a comprehensive, systematic as- “can’t do” problems because the child cannot perform a given
sessment of social competence difficulties. social skill under the most optimal conditions of motivation. In
short, social skills acquisition deficits result from faulty instruction
Classification of Social Skills Deficits of appropriate social behaviors and/or faulty learning of appropri-
ate social behaviors. Acquisition deficits require direct interven-
An important conceptual feature of social skills that has direct
tion approaches such as direct instruction, modeling, coaching, and
implications for their comprehensive assessment is the distinction
behavioral rehearsal (Elliott & Gresham, 2008).
between social skills acquisition deficits and social skills perfor-
In contrast, performance deficits are conceptualized as the fail-
mance deficits (Gresham, 1981a, 1981b). This distinction is im-
ure to perform a given social skill at acceptable levels even though
the individual knows how to perform the social skill (Gresham,
1981a, 1981b). These types of social skills deficits can be thought
This article was published Online First August 30, 2010. of as “won’t do” problems in that the child knows what to do but
Frank M. Gresham, Department of Psychology, Louisiana State Univer- does not want to perform a particular social skill. These types of
sity; Stephen N. Elliott and Ryan J. Kettler, Department of Special Edu-
social skills deficits are motivational or performance issues rather
cation, Vanderbilt University.
than learning or acquisition issues. Performance deficits require
This work was supported by Grant R324A090098 from the Institute of
Educational Sciences, U.S. Department of Education. reinforcement-based intervention approaches designed to increase
Correspondence concerning this article should be addressed to Frank M. the frequency of children’s social skills (Elliott & Gresham, 2008).
Gresham, Department of Psychology, Louisiana State University, Baton It is also important in the assessment of social skills to identify
Rouge, LA 70803. E-mail: [email protected] specific social skills strengths an individual possesses. A social
809
810 GRESHAM, ELLIOTT, AND KETTLER

skills strength can be defined as an individual knowing how to and pattern of responding, or cut score depends on the intrinsic validity
performing specific social skills in a consistent and appropriate of a test in discriminating among categories and the base rate of the
manner. Social skills strengths are important because they can be phenomenon one is trying to predict (see Meehl & Rosen, 1955;
used as building blocks for the improvement of individuals’ less Mullins-Sweatt & Widiger, 2009). As such, the current investiga-
well-developed social skills (Elliott & Gresham, 2007; Gresham & tion sought to empirically establish the base rates of social skills
Elliott, 2008). acquisition/performance deficits, social skills strengths, and prob-
Another important component in the conceptualization of social lem behaviors using a large, representative sample of children and
skills deficits is the notion of competing problem behaviors (Gre- adolescents.
sham & Elliott, 1990, 2008). Competing problem behaviors effec- The base rate phenomenon is considered to be a part of Bayesian
tively compete with, interfere with, or “block” either the acquisi- statistics or logic in which the probability of any given hypothesis
tion or performance of a given social skill. Competing problem being true is changed or updated based on the accumulation of
behaviors can be broadly classified as either externalizing behavior additional information. Bayesian statistics use two types of prob-
patterns (e.g., noncompliance, aggression, or coercive behaviors) abilities: prior or antecedent probability (which is the probability
or internalizing behavior patterns (e.g., social withdrawal, anxiety, that a hypothesis is true prior to collecting data) and posterior
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

or depression). For example, a child with a history of noncompli- probability (which is the probability that a hypothesis is true after
ant, oppositional, and coercive behavior may never learn prosocial the collection of data). Bayesian statistics or logic is used in
behavioral alternatives such as sharing, cooperation, and self- decision theory, in which decisions about phenomena are based on
control because the competing function of these aversive behaviors the collection of additional information about those phenomena in
(Eddy, Reid, & Curry, 2002) causes the absence of opportunities to order to make accurate decisions (Robert, 2007). In the current
learn these behaviors. Similarly, an individual with a history of article, we present base rates of various types of social skills
social anxiety, social withdrawal, and shyness may never learn deficits, strengths, and problem behaviors in order to facilitate
appropriate social behaviors because of withdrawal from the peer accurate decisions about children and youths’ social competencies.
group, thereby creating an absence of the opportunities to learn
peer-related social skills (Gresham, Van, & Cook, 2006). Method

Purpose of the Present Study Participants

The purpose of the present study was to empirically determine Participants were 4,550 children and adolescents ages 3–18
the base rates of social skills acquisition and performance deficits, years who were assessed at 115 sites in 36 states as part of the
social skills strengths, and problem behaviors using a nationally national standardization of the Social Skills Improvement Sys-
representative sample of children and adolescents ages 3–18 years tem—Rating Scales (SSIS–RS; Gresham & Elliott, 2008). During
using the Social Skills Improvement System—Rating Scales (Gre- standardization, there were two versions of the teacher form. The
sham & Elliott, 2008). To date, we are not aware of any studies Teacher Preschool Form had 140 items and the Teacher Elemen-
using nationally representative samples to establish these base tary/Secondary Form had 144 items. There was one parent form
rates. Knowing the base rates of phenomena is important because with 197 items. Two versions of the student form were used; the
one cannot know how unusual or typical a phenomenon is without Student Elementary Form had 137 items and the Student Second-
first knowing its base rate in the population. That is, accurate ary Form had 143 items.
diagnostic and prognostic statements can frequently be made on Demographic targets for the SSIS–RS norm sample were based
the basis of extant actuarial data, thereby not requiring the use of on the Current Population Survey, March 2006 (U.S. Census
a psychometric device (Meehl & Rosen, 1955). Bureau, 2006) and were applied to the three norm groups: a
Several of the points made in Meehl and Rosen’s (1955) classic preschool group (ages 3–5) and two school-age groups (ages 5–12
article are particularly germane to the purpose of the current and ages 13–18). Each age group sample was designed to have
investigation. First, base rates are rarely reported in the clinical equal numbers of males and females and to match the U.S. pop-
assessment literature; therefore, it is impossible to know whether a ulation with regard to race/ethnicity, socioeconomic status, and
given psychological assessment device will produce a greater geographic region. More detailed information on the normative
number of correct decisions than simply using extant actuarial data sample can be found in the SSIS–RS manual (see Gresham &
(i.e., base rates). Second, data provided by most psychological Elliott, 2008). Table 1 shows the breakdown of the standardization
assessment tools are not sufficient to determine the efficiency of sample across the three age groups by rater and sex.
that tool in other settings where base rates are substantially differ-
ent. This is because sample sizes are frequently too small to Measures
determine the optimal cutting lines for valid decision making.
Third, the type of population to which a given assessment instru- The SSIS–RS is a multirater series of rating scales that includes
ment will be applied is often unclear because base rates will ratings from teachers, parents, and students. The SSIS–RS assesses
necessarily vary depending upon the clinical population to which three domains: social skills, problem behaviors, and academic
one is interested in generalizing. Fourth, the results of psycholog- competence. All forms include common social skills in the fol-
ical assessment research are often reported only in terms of sig- lowing domains: communication, cooperation, assertion, respon-
nificance tests reflecting group differences rather than the number sibility, empathy, engagement, and self-control. The Problem Be-
of correct decisions (true positives and true negatives) for individ- havior subscales include the domains of externalizing, bullying,
uals within those groups. Thus, the practical value of a sign, hyperactivity/inattention, internalizing, and autism spectrum. The
BASE RATES OF SOCIAL SKILLS DEFICITS 811

Table 1 chetti, & Balla, 2005) are .65 and .44 for the teacher and parent
Representation of the Normative Sample by Sex, Age, and Form forms, respectively.

Female Male Combined

Age band by Establishing Base Rates
form n % n % n Base rates for social skills acquisition deficits, performance
Teacher form deficits, strengths, and problem behaviors were calculated by using
3–5 years 100 50 100 50 200 various combinations of frequency and importance ratings of the
5–12 years 275 50 275 50 550 SSIS–RS social skills items and the frequency ratings of the
13–18 years 100 50 100 50 200 problem behavior items for the teacher, parent, and student forms.
Parent form
3–5 years 200 50 200 50 400 For social skills acquisition and performance deficits and social
5–12 years 100 50 100 50 200 skills strengths, each item was a weighted product of its frequency
13–18 years 200 50 200 50 400 and its importance rating. For the identification of problem behav-
Student form iors, each item was identified only on the basis of its frequency
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

5–12 years 250 50 250 50 500

rating. These base rates were calculated across the three broad age
This document is copyrighted by the American Psychological Association or one of its allied publishers.

13–18 years 150 50 150 50 300

groupings from the SSIS–RS standardization sample (3–5 years,
5–12 years, and 13–18 years) and across three raters (teachers,
parents, and students) using data from the SSIS–RS national stan-
teacher form includes an Academic Competence scale on which dardization sample (Gresham & Elliott, 2008). The base rates
teachers rate student performance in reading, math, motivation, across the three age groups and three raters were selected because
parental support, and general cognitive functioning. This scale was the SSIS–RS provides normative data for these same three age
not used in the current investigation. groups and three raters.
Teachers and parents indicate the frequency with which the This strategy allowed for the calculation of different base rates
student exhibits each social skill and problem behavior on a across age and rater and was based on the definitions of acquisition
4-point scale of never, seldom, often, and almost always. Students and performance deficits, social skills strengths, and problem
indicate how true a statement about each social skill and problem behaviors in the SSIS–RS manual (Gresham & Elliott, 2008).
behavior is for them using a 4-point scale of not true, a little true, Social skills acquisition deficits were operationally defined as any
a lot true, and very true. In addition, teachers, parents, and students social skills item with a frequency rating of 0 (never) and an
indicate the importance of each social skill to the student’s devel- importance/belief rating of 1 (important) or 2 (critical). Social
opment and classroom success using a 3-point scale of not impor- skills performance deficits were operationally defined as any so-
tant, important, and critical. As is described later in this article, cial skills item with a frequency rating of 1 (seldom) and an
these frequency and importance ratings were used to establish the importance rating of 2 (critical). Social skills strengths were op-
base rates of social skills acquisition and performance deficits, erationally defined as any social skills item with a frequency rating
social skills strengths, and problem behaviors. Table 2 shows the of 3 (almost always) and an importance rating of 1 (important) or
number of items on each of the social skills and Problem Behavior 2 (critical). Problem behaviors were operationally defined as any
subscales for each of the three raters (teacher, parent, and student). problem behavior item having a frequency rating of 3 (almost
The SSIS–RS manual provides extensive validity evidence always).
based on test content, internal structure, intercorrelations among These operational criteria can be traced back to the precursor of
scales and subscales, item-total correlations, and relations with the SSIS–RS (the Social Skills Rating System or SSRS; Gresham
other variables (Gresham & Elliott, 2008). Item development of & Elliott, 1990) and prior to that to the conceptual work of
the SSIS–RS was based on a broad review of the empirical Gresham (1981a, 1981b). It could be argued that a rating of an
literature on social skills deficits in special populations, reviews of item as never occurring might not necessarily mean that the
published empirical studies using an earlier version of the scale individual does not know how to perform that behavior or that
(Social Skills Rating System; Gresham & Elliott, 1990), and there is not a substantial difference in an importance rating of 1
research on the relationship between specific social behaviors and
important social outcomes for children and youth. Intercorrelations
among scales and subscales for each form are moderate to high for Table 2
the Social Skills and Problem Behavior subscales. Item-total cor- Number of Items by Scale and Form to Meet 50%, 33%, 25%,
relations across forms by age tend to be moderate to high, many of and 15% Criteria
which exceed .70 –.80.
Item type by form ⬎50% 33%–49% 25%–32% 15%–24% Total
Correlations between the SSIS–RS and the Behavioral Assess-
ment System for Children, Second Edition (BASC–2; Reynolds & Teacher form
Kamphaus, 2004) are moderate to high, depending on the scales Social skills ⬎23 15–22 12–14 7–11 46
and subscales. For example, the median correlations between the Problem behaviors ⬎15 11–14 8–10 4–7 41
Parent form
SSIS–RS total social skills score and the teacher form of the Social skills ⬎23 15–22 12–14 7–11 46
BASC–2 social skills score are .78 and .69 for the teacher and Problem behaviors ⬎17 12–16 8–11 5–7 49
parent forms, respectively. Correlations between the SSIS–RS Student form
total social skills scores and the socialization scores of the Vine- Social skills ⬎23 15–22 12–14 7–11 46
Problem behaviors ⬎15 10–14 7–9 4–6 34
land Adaptive Behavior Scales, Second Edition (Sparrow, Cic-
812 GRESHAM, ELLIOTT, AND KETTLER

(important) versus 2 (critical). It might also be argued that a Table 3

problem item receiving a rating of almost always might not nec- Base Rates for Social Skills Performance Deficits
essarily translate into a high frequency of that behavior measured
by other methods, such as systematic direct observations. None- Form type by
criterion 3–5 years 5–12 years 13–18 years
theless, we believe that our operationalization of acquisition def-
icits, performance deficits, social skills strengths, and problem 50% or more of items
behaviors using the above criteria was reasonable given the nature Teacher form 7.0% 4.2% 5.0%
of behavior rating scales in general, the goals of this study in Parent form 5.3% 2.3% 3.0%
Student form — 0.0% 3.0%
particular, and extensive research using the SSRS over the past 20
33%–49% of items
years (Gresham & Elliott, 1990, 2008). Teacher form 16.5% 16.9% 14.0%
Base rates were based on the percentage of items receiving these Parent form 16.0% 10.7% 11.8%
various ratings which were arbitrarily set as follows: 50% or more Student form — 9.0% 4.7%
of the items, 33%– 49% of the items, 25%–32% of the items, and 25%–32% of items
Teacher form 26.0% 26.5% 16.0%
15%–24% of the items. These arbitrary ranges were selected to
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Parent form 28.8% 19.7% 15.3%

This document is copyrighted by the American Psychological Association or one of its allied publishers.

represent a wide range in the distribution of social skills deficits, Student form — 1.0% 16.9%
strengths, and problem behaviors. Because this appears to be the 15%–24% of items
first study to investigate the base rates of social skills deficits, Teacher form 39.0% 43.6% 38.0%
Parent form 57.7% 40.5% 42.8%
strengths, and problem behaviors, we had no research literature to
Student form — 4.2% 21.9%
guide our selection of these cut points.
Based on the above criteria, for example, if a student received
frequency ratings of 0 and an importance/belief rating of 1 or 2 on
50% or more of the items, this would be the base rate for that (rater) ⫻ 3 (age group) chi-square analysis was not significant,
student’s acquisition deficits. Similarly, if a student received fre- ␹2(2) ⫽ 0.79, p ⫽ .675. A 3 (rater) ⫻ 3 (age group) chi-square
quency ratings of 3 on 50% or more of the problem behavior items, analysis was significant, ␹2(2) ⫽ 99.68, p ⬍ .001, with far fewer
this would be that student’s base rate for problem behaviors. Based students in the 5- to 12-year-old age group reporting performance
on the above definitions, this is how the various base rates were deficits (⬍1%) than the remaining age groups as rated by teachers
calculated for the remaining percentage intervals (i.e., 33%– 48%, and parents.
25%–32%, and 15%–24%). Thus, the base rate reflects the per- Table 3 depicts the base rates for performance deficits for
centage of students receiving the various combinations of ratings 25%–32% of the items by rater and age group. A 2 (rater) ⫻ 2 (age
at each of the percentage intervals. Table 2 shows the number of group) chi-square analysis was not significant, ␹2(2) ⫽ 2.69, p ⫽
items required to meet each of the definitions for social skills and .260. A 3 (rater) ⫻ 3 (age group) chi-square analysis was signif-
problem behavior items. icant, ␹2(2) ⫽ 203.48, p ⬍ .001, with far fewer students in the 5-
to 12-year-old age group reporting performance deficits compared
Results to students in the 13- to 18-year-old age group as well as teachers
and parents across all age groups.
Social Skills Acquisition Deficits Table 3 shows the base rates for performance deficits for 15%–
24% of the items by rater and age group. A 2 (rater) ⫻ 3 (age
The base rates for social skills acquisition deficits across the group) chi-square analysis was significant, ␹2(2) ⫽ 11.27, p ⬍
three raters and three age groups were all less than 1% for all .004, with parents reporting more children as having performance
percentage ranges of items (i.e., 50% or more, 33%– 49%, 25%– deficits in the 3- to 5-year-old age group compared to parents of
32%, and 15%–24%). As such, social skills acquisition deficits children in the 5- to 12- and 13- to 18-year-old age groups, as well
appear to be a rare phenomenon in a normative sample of children as teachers of children across all age groups.
and adolescents ages 3–18 years.
Social Skills Strengths
Social Skills Performance Deficits
Table 4 shows the base rates for social skills strengths for 50%
Table 3 shows the base rates for social skills performance or more of the items by rater and age group. A 2 (rater) ⫻ 3 (age
deficits for 50% or more of the items by rater and age group. A 2 group) chi-square analysis was significant, ␹2(2) ⫽ 25.15, p ⬍
(rater) ⫻ 3 (age group) chi-square analysis was not significant, .001, with parents of 3- to 5-year-old children identifying fewer
␹2(2) ⫽ 1.08, p ⫽ .583, indicating no differences in the proportion numbers of children as having social skills strengths than parents
of students having performance deficits as rated by teachers and of children in the 5- to 12- and 13- to 18-year-old age groups and
parents. A 3 (rater) ⫻ 2 (age group) chi-square analysis was compared to teachers across all three age groups. A 3 (rater) ⫻ 3
significant, ␹2(2) ⫽ 22.55, p ⬍ .001, with no students ages 5–12 (age group) chi-square analysis was significant, ␹2(2) ⫽ 122.92,
reporting performance deficits compared to the number of students p ⬍ .001, with far fewer students in the 5- to 12-year-old age group
in the 13–18 age group as well as the proportions of students reporting social skills strengths relative to students in the 13- to
identified as having performance deficits by teachers and parents 18-year-old age group and compared to teachers and parents across
across the other age groups. all three age groups.
Table 3 also shows the base rates for social skills performance Table 4 reports base rates for social skills strengths for 33%–
deficits for 33%– 49% of the items by rater and age group. A 2 49% of the items by rater and age group. A 2 (rater) ⫻ 3 (age
BASE RATES OF SOCIAL SKILLS DEFICITS 813

Table 4 small. For 50% or more of the items, the base rate for problem
Base Rates for Social Skills Performance Strengths behaviors were less than 1% across all raters and all age groups.
For the remaining percentages of items (33%, 25%, and 15%), all
Form type by chi-square analyses were not significant, indicating no rater or age
criterion 3–5 years 5–12 years 13–18 years
differences and no Rater ⫻ Age interaction effect. This was
50% or more of items primarily due to the low sample sizes, thereby decreasing the
Teacher form 29.0% 24.0% 26.0% power of our chi-square analysis. It should be noted, however, that
Parent form 16.5% 20.5% 28.3% there was a trend for students in the 5- to 12- and 13- to 18-year-
Student form — 2.6% 21.3%
old age groups to report more problem behaviors than teachers and
33%–49% of items
Teacher form 39.0% 39.0% 44.5% parents across all age groups.
Parent form 33.3% 41.5% 48.5%
Student form — 6.6% 42.3% Discussion
25%–32% of items
Teacher form 48.0% 46.2% 54.5% The current article reported base rates for social skills acquisi-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Parent form 46.5% 53.9% 58.8%

tion deficits, performance deficits, strengths, and problem behav-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Student form — 8.2% 51.7%

15%–24% of items iors across three informants and three broad age groups. These
Teacher form 64.5% 61.5% 66.0% base rates were based on a large, representative sample of children
Parent form 64.8% 71.9% 70.5% and adolescents ages 3–18 years. To our knowledge, this is the first
Student form — 9.8% 66.7% study to investigate base rates of social competencies using a
nationally representative sample of children and youth.
Based on the current data, the base rates for social skills acqui-
sition deficits are extremely low, with less than 1% of the stan-
group) chi-square analysis was significant, ␹2(2) ⫽ 33.44, p ⬍
dardization sample showing these types of deficits. This finding
.001, with parents of 13- to 18-year-old children reporting a larger
held true across teacher, parent, and student raters and across the
proportion of these students as having social skills strengths than
three age groups studied (3–5 years, 5–12 years, and 13–18 years).
parents of 3- to 5- and 5- to 12-year-old children. A 3 (rater) ⫻ 3
Practically speaking, these data suggest that it would be unusual
(age group) chi-square analysis was significant, ␹2(2) ⫽ 253.72,
for a given child or adolescent between the ages of 3 and 18 years
p ⬍ .001, with substantially fewer students in the 5- to 12-year-old
to be rated by any rater as having more than one social skills
age group reporting social skills strengths than students in the 13-
acquisition deficit, based on the fact that on average 0.25% of
to 18-year-old age group as well as teachers and parents across all
children and adolescents across all raters and all age groups have
three age groups.
acquisition deficits. In short, social skills acquisition deficits ap-
Table 4 shows the base rates of social skills strengths for
pear to be a rare phenomenon in a representative, normative
25%–32% of the items by rater and age group. There was a
population of children and adolescents. This finding was unex-
significant 2 (rater) ⫻ 3 (age group) chi-square analysis, ␹2(2) ⫽
pected; however, there are no previous studies against which the
44.88, p ⬍ .001, with teachers and parents of 13- to 18-year-old
current findings might be compared, because this is apparently the
children reporting greater proportions of them as having social
first study to investigate the base rates of social skills acquisition
skills strengths than teachers and parents of 3- to 5- and 5- to
deficits versus performance deficits. One hypothesis is that the
12-year-old children. A 3 (rater) ⫻ 3 (age group) chi-square
SSIS–RS does not have sufficient “floor” to reliably detect base
analysis was significant, ␹2(2) ⫽ 329.16, p ⬍ .001, with consid-
erably fewer students in the 5- to 12-year-old age group reporting
social skills strengths compared to 13- to 18-year-old students and
teachers and parents across all age groups. Table 5
Table 4 shows the base rates of social skills strengths for Base Rates for Problem Behaviors
15%–24% of the items by rater and age group. A 2 (rater) ⫻ 3 (age
Form type by
group) chi-square analysis was significant, ␹2(2) ⫽ 56.65, p ⬍
criterion 3–5 years 5–12 years 13–18 years
.001, with parents of 5- to 12- and 13- to 18-year-old children
reporting larger proportions of them as having social skills 50% or more of items
strengths than parents of 3- to 5-year-old children and teachers Teacher form 0.0% 0.4% 0.0%
across all three age groups. A 3 (rater) ⫻ 3 (age group) chi-square Parent form 0.3% 0.3% 0.0%
Student form — 1.8% 1.0%
analysis was significant, ␹2(2) ⫽ 474.17, p ⬍ .001, with substan- 33%–49% of items
tially fewer students in the 5- to 12-year-old age group reporting Teacher form 0.5% 1.1% 0.5%
social skills strengths compared to 13- to 18-year-old children and Parent form 0.8% 1.0% 0.5%
teachers and parents across all three age groups. Student form — 6.2% 2.7%
25%–32% of items
Teacher form 0.5% 2.0% 1.0%
Problem Behaviors Parent form 1.0% 1.4% 1.3%
Student form — 9.6% 4.3%
Table 5 presents the base rates for problem behaviors for 50% or 15%–24% of items
more, 33%– 49%, 25%–32%, and 15%–24% of the items by rater Teacher form 2.5% 4.2% 1.5%
and age group. Similar to the findings for acquisition deficits, the Parent form 4.3% 3.7% 2.3%
Student form — 19.6% 8.0%
base rates of problem behaviors in the normative sample are very
814 GRESHAM, ELLIOTT, AND KETTLER

rates of social skills acquisition deficits across ages and raters. all levels and ages was 0.046%, meaning that one could expect, on
Although this may be a plausible hypothesis, we are unaware of average, a child or adolescent to have only one to two items rated
any previous studies in the literature that have investigated the as being problem behaviors. There was a tendency for more
base rate phenomenon for social skills deficits. students (ages 5–12 and 13–18) to identify themselves as having
This finding most certainly cannot be generalized to other problem behaviors (M ⫽ 1.4%) compared to teachers and parents
clinical populations of children and adolescents. For example, (M ⫽ 0.014%). This pattern continued across the remaining cri-
social competence difficulties are either part of the diagnostic terion levels.
criteria or associated features specified in the Diagnostic and This investigation consistently found differences in the reported
Statistical Manual of Mental Disorders (4th ed.; American Psy- base rates of 5- to 12-year-old children versus those of the remain-
chiatric Association, 1994). This is true for the diagnoses of ing informants. Children in this age group reported fewer perfor-
attention-deficit/hyperactivity disorder (ADHD), conduct disorder, mance deficits, fewer social skills strengths, and more problem
oppositional defiant disorder, and autistic disorder. As such, one behaviors than older children as well as teachers and parents.
would expect a greater number of acquisition deficits for children These data suggest that children in this age group will underreport
in these diagnostic categories. For example, children diagnosed
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

their social skills performance deficits and social skills strengths

This document is copyrighted by the American Psychological Association or one of its allied publishers.

with ADHD often exhibit temper outbursts, bossiness, insistence and overreport their frequencies of problem behaviors relative to
that requests be met, and social rejection by peers. Much of this older students as well as parents and teachers.
behavior pattern can be attributed to the fact that these children, One hypothesis is that the SSIS–RS student form for 5- to
perhaps because of their impulsivity and inattention, have not 12-year-old children is not a valid instrument. This, however, can
learned socially appropriate ways of interacting with others. It is be refuted by looking at the correlations between this instrument
well established that most children and youth with ADHD have and other instruments that have extensive validity evidence. For
deficient social skills (DuPaul & Stoner, 2003; MTA Cooperative example, Gresham and Elliott (2008) showed moderate correla-
Group, 1999). What is not currently known is what proportion of tions between the total social skills and total problem behavior
these children has acquisition versus performance deficits. scores of the SSIS–RS student form (8- to 12-year-olds) and the
Base rates for performance deficits were far higher in the scales on the BASC–2, respectively (Reynolds & Kamphaus,
normative sample than acquisition deficits. This finding held true 2004). The SSIS–RS student form has been shown to reliably
for teacher and parents informants across all three age groups and differentiate children with ADHD, specific learning disabilities,
for student informants ages 13–18 years. Using the criterion of and emotional disturbance from the normative sample (see Gre-
50% or more of the items identified as performance deficits, the sham & Elliott, 2008).
base rates ranged from 0% (students, ages 5–12) to 7% (teachers, It is unclear why self-reports of this age group are questionable.
ages 3–5) with a median base rate of 3.5%. Overall, one can expect However, there are some data from other researchers addressing
that only about 3.5% of children and youth will have 23 of the 46 the validity of self-reports in children. Pelham, Fabiano, and Mas-
SSIS–RS social skills items rated as performance deficits. Obvi- setti (2005) reported that there is presently no evidence supporting
ously, these base rates increase when using less stringent criteria of the validity of child self-report of ADHD symptoms. Similarly,
33%, 25%, and 15% of the items, respectively. Klein, Dougherty, and Olino (2005) cautioned against the use of
A consistent finding across all criterion levels was that students self-report measures in the assessment of childhood depression,
ages 5–12 years identified themselves as having fewer perfor- particularly in settings with very low or very high base rates for
mance deficits compared to older students (ages 13–18) as well as depression. In contrast, self-report measures of childhood anxiety
teachers and parents. Using the 50% or more criterion, no students using measures such as the Revised Children’s Manifest Anxiety
in the 5- to 12-year-old group identified themselves as having Scale (Reynolds & Richmond, 1985) and the Multidimensional
performance deficits, whereas teachers and parents identified 4.3% Anxiety Scale for Children (March, Parker, Sullivan, Stallings, &
and 2.25% of children with performance deficits, respectively. Conners, 1997) have received much stronger empirical support
This pattern persisted across the 33%, 25%, and 15% criterion (see Silverman & Ollendick, 2005). It may well be the case that
levels. younger children are not as accurate as older children, teachers,
Base rates for social skills strengths were substantially higher and parents in identifying their social competencies. This finding
than the base rates for performance deficits. For example, the mean requires further research to evaluate whether it holds true for
base rate for social skills strengths using the 50% or more criterion clinical samples.
(excluding students ages 5–12) was 28.3% compared to the mean
base rate for performance deficits of 4.24%. This pattern continued
for the remaining criterion levels for social skills strengths. References
Whereas students ages 5–12 years identified themselves as
having no social skills performance deficits, they were also less American Psychiatric Association. (1994). Diagnostic and statistical man-
likely to identify themselves as having social skills strengths. In ual of mental disorders (4th ed.). Washington, DC: Author.
DuPaul, G., & Stoner, G. (2003). ADHD in the schools (2nd ed.). New
fact, using the 50% or more criterion level, only 2.6% of students
York, NY: Guilford Press.
in this age group identified themselves as having social skills
Eddy, J., Reid, J., & Curry, V. (2002). The etiology of youth antisocial
strengths compared to 20.5% and 24.0% of teachers and parents of behavior, delinquency, and violence and a public health approach to
this age group, respectively. prevention. In H. Walker, M. Shinn, & G. Stoner (Eds.), Interventions
Similar to the base rates for social skills acquisition deficits, the for academic and behavior problems: II. Preventive and remedial ap-
base rates for problem behaviors are rare in a normative sample. proaches (pp. 27–51). Bethesda, MD: National Association of School
Using the 50% or more criterion level, the mean base rate across Psychologists.
BASE RATES OF SOCIAL SKILLS DEFICITS 815

Elliott, S. N., & Gresham, F. M. (2007). SSIS Classwide Intervention Meehl, P., & Rosen, A. (1955). Antecedent probability and the efficiency
Program Teacher’s Guide. Minneapolis, MN: Pearson Assessments. of psychometric signs, patterns, or cutting scores. Psychological Bulle-
Elliott, S. N., & Gresham, F. M. (2008). Social Skills Improvement System tin, 52, 194 –216.
Intervention Guide manual. Minneapolis, MN: Pearson Assessments. MTA Cooperative Group. (1999). 14-month randomized clinical trial of
Gresham, F. M. (1981a). Assessment of children’s social skills. Journal of treatment strategies for attention deficit hyperactivity disorder. Archives
School Psychology, 19, 120 –134. of General Psychiatry, 56, 1073–1096.
Gresham, F. M. (1981b). Social skills training with handicapped children: Mullins-Sweatt, S., & Widiger, T. (2009). Clinical utility and DSM–V.
A review. Review of Educational Research, 51, 139 –176. Psychological Assessment, 21, 302–312.
Gresham, F. M. (1997). Social competence and students with behavior Newcomb, A., Bukowski, W., & Pattee, L. (1993). Children’s peer rela-
disorders: Where we’ve been, where we are, and where we should go. tions: A meta-analytic review of popular, rejected, neglected, controver-
Education and Treatment of Children, 20, 233–249. sial, and average sociometric status. Psychological Bulletin, 113, 99 –
Gresham, F. M. (1998). Social skills training: Should we raze, remodel, or 128.
rebuild? Behavioral Disorders, 24, 19 –25. Parker, J., & Asher, S. (1987). Peer relations and later personal adjustment:
Gresham, F. M. (2002). Teaching social skills to high-risk children and Are low-accepted children at risk? Psychological Bulletin, 102, 357–
youth: Preventive and remedial approaches. In M. Shinn, H. Walker, & 389.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

G. Stoner (Eds.), Interventions for academic and behavior problems: II. Pelham, W., Fabiano, G., & Massetti, G. (2005). Evidence-based assess-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Preventive and remedial approaches (pp. 403– 432). Bethesda, MD: ment of attention deficit hyperactivity disorder in children and adoles-
National Association of School Psychologists. cents. Journal of Clinical Child and Adolescent Psychology, 34, 449 –
Gresham, F. M., & Elliott, S. N. (1990). Social Skills Rating System. 476.
Minneapolis, MN: Pearson Assessments. Reynolds, C. R., & Kamphaus, R. W. (2004). Behavioral Assessment
Gresham, F. M., & Elliott, S. N. (2008). Social Skills Improvement Sys- System for Children—Second Edition. Minneapolis, MN: Pearson As-
tem—Rating Scales manual. Minneapolis, MN: Pearson Assessments. sessments.
Gresham, F. M., Van, M., & Cook, C. R. (2006). Social skills training for Reynolds, C. R., & Richmond, B. O. (1985). Revised Children’s Manifest
teaching replacement behaviors: Remediation of acquisition deficits for Anxiety Scale: Manual. Los Angeles, CA: Western Psychological Ser-
at-risk children. Behavioral Disorders, 32, 32– 46. vices.
Klein, D., Dougherty, L., & Olino, T. (2005). Toward guidelines for Roberts, C. P. (2007). The Bayesian choice: From decision-theoretic
evidence-based assessment of depression in children and adolescents. foundations to computational implementation. New York, NY: Springer.
Journal of Clinical Child and Adolescent Psychology, 34, 412– 432. Silverman, W., & Ollendick, T. (2005). Evidence-based assessment of
Kupersmidt, J., Coie, J., & Dodge, K. (1990). The role of peer relationships anxiety and its disorders in children and adolescents. Journal of Clinical
in the development of disorder. In S. Asher & J. Coie (Eds.), Peer Child and Adolescent Psychology, 34, 380 – 411.
rejection in childhood (pp. 274 –308). New York, NY: Cambridge Uni- Sparrow, S., Cicchetti, D., & Balla, D. (2005). Vineland Adaptive Behavior
versity Press. Scales: Second Edition. Minneapolis, MN: Pearson Assessments.
Maag, J. W. (2005). Social skills training for youth with emotional and U.S. Census Bureau. (2006). Current Population Survey, March 2006
behavioral disorders and learning disabilities: Problems, conclusions, [Data file]. Washington, DC: Author.
and suggestions. Exceptionality, 13, 155–172. Walker, H. M., Ramsay, E., & Gresham, F. M. (2004). Antisocial behavior
Maag, J. W. (2006). Social skills training for students with emotional and in school: Evidence-based practices. Belmont, CA: Wadsworth/
behavioral disorders: A review of reviews. Behavioral Disorders, 32, Thomson Learning.
155–172.
March, J., Parker, J., Sullivan, K., Stallings, P., & Conners, K. (1997). The
Multidimensional Anxiety Scale for Children (MASC): Factor structure, Received January 8, 2010
reliability, and validity. Journal of the American Academy of Child & Revision received April 5, 2010
Adolescent Psychiatry, 36, 554 –565. Accepted May 3, 2010 䡲