+ All Categories
Home > Documents > 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

Date post: 03-Jun-2018
Category:
Upload: karinadaparia
View: 218 times
Download: 0 times
Share this document with a friend

of 24

Transcript
  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    1/24

    School Psychology ReviewVolume 24, No. 4, 1995, pp. C,-1S-c>71

    SOCIAL SKILLS ASSESSMENT: A COMPARATIVEEVALUATION OF SIX PUBLISHEDRATING SCALESMichelle K. Demaray, Stacey L. Ruffalo, John Carlson, R. T. Busse,Amy E. Olson, Susan M. McManus, and Amy LeventhalUniversity of Wisconsin-Madison

    Abstract: A comparative and integrative review was conducted of six published rat-ing scales commonly used to assess the social skills of preschool and school-agedchildren. Four norm-referenced instruments are reviewed: SchoolSocial BehaviorScales (SSBS; Merrell, 1993), Social SkiUs Rating System SSRS; Gresham & Elliott,1990), Waksman ocialSkillsRating ScaleWSSRS; Waksman, 1985), and Walker-MC-ConnellScaleof Social Competencend SchoolAdjustment WMS; Walker & Mc-Connell, 1988). The SchoolSocial Skills Rating Scale S3; Brown, Black, & Downs,1984) and Social Behavior Assessmentnventory (SBAI; Stephens & Arnold, 1992)are included as criterion-referenced rating scales. Content and use, standardizationsample and norms, scores and interpretation, and psychometric properties were re-viewed. We concluded that the most comprehensive instrument is the SSRS becauseof its multi-source approach and intervention linkage. The SSBS and the WMS are use-ful tools for a more limited school scope. The remaining norm-referenced scale,WSSRS, is not recommended. Following initial screening, the S3 and SBAI are usefulfor a more specific examination of particular behaviors to target for change.

    The assessment of social skills is an im-portant issue in school settings. Researchhas indicated that children who persistentlyexhibit social skills deficits often experi-ence short- and long-term negative conse-quences (Elliott, Sheridan, & Gresham,1989). For example, peer relationship diffi-culties have been found to remain stableover time (Coie & Dodge, 1983), and to bepredictive of adult psychopathology(Parker & Asher, 1987). Further, social skillsdeficits have been related to poor academicadjustment for children with (McKinney &Speece, 1983) and without disabilities(Hoge & Lute, 1979), and poor peer accep-tance of students with disabilities (Gre-sham, 1988; Hazel & Schumaker, 1988;Stumme, Gresham, & Scott, 1982).Recognition of the importance of as-sessing social skills has given rise to a num-ber of rating scales specifically focused onchildrens and adolescents social behav-iors. Social skills rating scales have been de-veloped for use by parents, teachers, andother adults associated with the individualbeing rated. These measures can range incomposition from scales with two or moresubscales assessing different areas of a con-

    struct, to single scales measuring narrowlydefined constructs. Rating scales can pro-vide valuable information regarding chil-drens social skills; however, best practicesassessment calls for gathering informationfrom multiple sources, settings, and meth-ods. Elliott and colleagues (1989) provideda useful heuristic for the assessment of so-cial skills that includes: (a) teacher, parent,and student ratings; (b) teacher, parent, andstudent interviews; (c) observations; (d) be-havioral role-plays; and (e) sociometrics.Social behavior rating scales can facili-tate a variety of assessment decisions (Wil-son & Bullock, 1989) and can be used (a) aspart of the screening, referral, and identifi-cation process; (b) to compare behaviorsacross settings; (c) to monitor behaviorover time; and (d) in research. An importantadvantage of rating scales over other as-sessment methods is the ability to obtainjudgments about a variety of traits or be-haviors from many sources in a time effi-cient manner. Elliott, Busse, and Gresham(1993) suggested some additional advan-tages of rating scales. For example, ratingscales can be used to demonstrate a promptreaction to a teacher or parent referral,

    Address all correspondence concerning this manuscript to Michelle K. Danaray The University ofWisconsin-Madison, Educational Sciences Building, 1025 W. Johnson St., Madison, WI 53706.648

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    2/24

    Social Skills Assessment 649which helps communicate concern for thereferred child and respect for the adult whorefers the child, and reflects good organiza-tion skills. Also, having a completed ratingscale prior to meeting with a teacher or par-ent can facilitate precise, behavior-specificcommunications and more efficient use oftime during interviews. Further, well-con-structed social skill scales include polls ofitems representing the domain of interest(i.e., possess content validity). Thus, ratingscales can cover a broad range of poten-tially relevant problems which may extendbeyond a specific referral complaint (Mc-Conaughy, 1993).

    McConaughy (1993) summarized thelimitations of rating scales thereby requir-ing multiple methods of assessment. For ex-ample, rating scales measure current levelsof functioning, but do not assess etiology.Further, rating scales do not assess ecologi-cal conditions surrounding behavior (e.g.,antecedents, consequences), and do notdictate intervention strategies. Finally, rat-ing scales involve perceptions that may varywith informants and context/settings.Therefore, although the focus of this articleis on social skills rating scales, one shouldrealize that rating scales should not be usedin isolation to make decisions regardingclassification or intervention.Social skills rating scales can be poten-tially useful components in the assessmentprocess and, as such, consumers should beaware of the relative merits of a given in-strument. To that end, this article providescritical reviews and comparative evalua-tions of six social skills rating scales. Beforeengaging in the reviews and comparisons, abrief discussion is offered of guidelines forevaluating social skills rating scales and isfollowed by the procedures used for inclu-sion and for critical review of the scales.Guidelines for EvaluatingSocial Skills Rating Scales

    Evaluating social skills rating scales (in-deed, any rating scale) is a multifactoredprocess. We drew upon several sources (forfurther discussions see American Psycho-logical Association [APA], 1985; Edelbrock,1983; Elliott et al., 1993; McCloskey, 1990;Wilson & Bullock, 1989) to construct guided

    criteria from which to judge an instrumentsmerits. We have condensed the informationgleaned into four evaluative dimensions:1. Content and use of a scale. Impor-tant aspects to be considered for this di-mension are completeness and user friend-liness of material and manuals, and appro-priate format (e.g., anchor points, instruc-tions), and scoring procedures.2. Standardization sample andnorms. Norm-referenced rating scales areused throughout the country and for appro-priate interpretation must be developedwith representative standardization sam-

    ples (standardized norming procedures arenot a prerequisite for criterion-referencedscales). Norming procedures should beclearly delineated, including information onthe year norming transpired, descriptivestatistics, and the sampling procedure used.3. Scores/interpretation. Important as-pects on this dimension are detailed de-scriptions of scores, and appropriateness ofscores for the purposes of the scale. Inter-pretation of scores also should be de1i.n

    eated and not extend beyond the purposesof the scale.4. Psychometric properties, involves ascales reliability and validity. Interrater,test-retest, and internal consistency are im-portant reliability considerations for mostrating scales. Evaluating scale validation in-cludes consideration of a scales internal va-lidity (i.e., content validity, construct valid-ity) and external validity (i.e., criterion-re-lated validity, convergent/divergent validity,

    predictive validity).Procedural Guidelines

    The first four scales chosen for revieware norm-referenced, whereas the last twoscales are criterion-based (see Table 1 forscale summaries). The norm-referencedscales are: (a) School Social BehaviorScales (SSBS; Merrell, 1993); (b) SociaZSFcills Rating System (SSRS; Gresham & El-liott, 1990); (c) WaFcsman Social SMls Rat-ing Scale (WSSRS; Waksman, 1985); and (d)Walker-McConnell Scale of Social Compe-tence and School Adjustment (WMS;Walker & McConnell, 1988). The criterion-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    3/24

    650 School Psychology Review, 1995, Vol. 24, No.4based scales are: (a) SchooZ Social ScalesRating ScaZe(S3; Brown et al., 1984); and (b)Social Behavior Assessment Inventory(SBAI; Stephens & Arnold, 1992).Selection criteria for inclusion of thescales were: (a) a specific focus onchild/adolescent social skills; (b) a focus onuse in schools; and (c) whether the scalewas published at the time of this review?We focused on published scales because w edeemed it important to present readilyavailable scales. The six scales are pre-sented in alphabetical order, with the norm-referenced scales arbitrarily presented first.The reviews begin with brief descriptions ofeach instrument, followed by evaluationsbased on the four dimensions discussedabove as well as the authors critical re-views and judgments of the quantity andquality of summary information/data. Thecriteria of limited, adequate, and excellentare used to characterize data concerningthe technical properties of each scale (seeTables 2 and 3 for reliability and validitysummaries). Because there is no algorithmfor determining the technical acceptabilityof a given scale or test (APA, 19S5), thesecriteria and evaluations were guided by ourjudgments of converging evidence for eachmeasure. In general, the criterion of exceZ-Lent indicates strong converging evidencewithin a given dimension (e.g., internal con-sistency > .90, 2-week test-retest 2 .SO). Thecriterion of adequate is used to indicate theevidence is acceptable for scale usewhereas Limited is used to characterize anunacceptable level of evidence.

    SCHOOL SOCIALBEHAVIOR SCALESThe SchooZ Social Behavior Scales(SSBS; Merrell, 1993) is a 65-item, norm-ref-erenced rating scale designed specificallyfor school-based use. Ratings of students(Grades K-12) are obtained from teachersor other school personnel. Responses aremade on a 5-point Like&type scale. Twomajor scales make up this instrument: so-cial competence (Scale A) and antisocialbehavior (Scale B). Each scale consists ofthree subscales (A = interpersonal skills,self-management, and academic skills; B =hostile-irritable. antisocial-aggressive. and

    disruptive-demanding). The SSBS was de-veloped as: (a) a screening tool for identify-ing students who are behaviorally at-risk;(b) part of a multi-method, multi-source as-sessment battery for determining programeligibility and designing appropriate inter-vention programs; and (c) a research instru-ment for studying social competence andantisocial behavior.Content and Use

    Completion of the SSBS is easy and canbe done by most teachers in 5 to 10 min. Thecomponents of this instrument include aneasy to follow technical manual and a two-page rating booklet. The instructions areconcise and easy to follow. However, a po-tential problem is that the rater is instructedto circle 1, which indicates Never, if thestudent does not exhibit a specified behav-ior or if the rater has not had the opportu-nity to observe the behavior. Consequently,lower scores may misrepresent a studentsskills. Thus, scale use could be enhanced byproviding a rating method to distinguishwhether the rater has had the opportunityto observe the behavior. Scoring the SSBS isa simple process; however, a disadvantageis that all items must be completed to inter-pret the data.Standardization Sample and Norms

    The standardization sample consistedof 1,858 children from Grades K through 12.A number of limitations exist in the stan-dardization sample. A large part of the sam-ple (40 ) came from Washington and Ore-gon, limiting representativeness. Also, theethnic distribution is not representative ofthe U.S. population. The author noted, how-ever, that the ethnicity of subjects is not acritical factor in influencing scores on be-havior rating scales once the effects of so-cial class have been controlled (Achen-bath & Edelbrock, 1981 cited in Merrell,1993, p. IS). The representativeness of SESis questionable because: (a) teachers pro-vided information about parents occupa-tions; (b) parent occupation was gatheredfor only one-half of the sample; and (c) onlyprimary wage earners occupations were re-ported, which does not represent otherhousehold income. The regular education to

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    4/24

    TE1

    SeeCaescoRewSaBoRnSe

    GaLa

    TNmIemNmvSme

    Fm

    InepevPoe

    R

    Fm

    S (Me

    1 -sww

    ST

    (Gem&

    Eo1

    6GaKo1

    Nmvsmen=1

    Rae

    o(8

    Sae

    o(1

    Rehcy(ao

    Whe8

    Ba8

    Hsc3

    Oh2

    Me5Fme4

    5GaPeKo1

    Nmvsmen=41

    Rae

    o(8

    Sae

    o(1

    Rehcy(ao

    Whe7

    Ba1

    Hsc6

    Oh3

    Me5Fme4

    1om

    2mose

    Sacme

    Asabo

    3sacme

    se

    Ineps

    ss

    Sm

    m

    Amcss

    3asabose

    Herae

    Asa-aev

    Dsuvmn

    m

    3om

    3mose

    Paw

    Sass

    ga

    Poembos

    Amccme

    3sassse

    C

    ao

    Ao

    Sco

    3poembose

    Eenzn

    Inenzn

    Havy

    5pnLkyseb

    oe

    q

    obo

    _________

    E

    Sass3pLkyse

    zc

    b

    oe

    ama

    obo

    Poembos3pLky

    seb

    oe

    ob

    io Amccme

    5pLk

    tyseb

    opca

    cues

    (Te1cn

    .

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    5/24

    (Te1cn

    GaLa

    TNmIemNmvSme

    Fm

    InepevPoe

    R

    Fm

    S (Gem

    &Eo

    1

    5SmaST

    3om

    2mose

    Pa

    Sass

    ga

    Poembos(P

    4sassse

    C

    ao

    Ao

    Sco

    R

    by

    3Ps

    e(smaST

    Ss

    (Gem

    &Eo

    1

    3SmaST

    2om

    1mose

    Pa

    Sass

    ga

    4sassse

    C

    ao

    Sco

    Emh

    Ao

    Wwm

    1

    2GaKo1

    2om

    1mose

    Nmvsmen=3

    me

    Sass

    Rae

    o(1

    feme

    2se

    Sae

    o(0

    Aevdmn

    Rehcy

    pvdmn

    Nnomopod

    SmaS

    sassapo

    lembose

    SmaSTsassse

    Ima

    ranaeoanoy

    ohs

    yfomga

    71-______

    4pnLkyseb

    oe

    q

    obo (T

    e1cn

    .

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    6/24

    (Te1cn

    GaLa

    TNmIemNmvSme

    Fm

    InepevPoe

    R

    Fm

    WM (Wak&

    MC 1

    4GaKo6

    1om

    1mose

    5pnLkyseb

    oe

    Nmvsmen=18

    Sacme

    q

    obo

    Rae

    o(1

    3se

    Sae

    o(0

    T

    peeesabo

    Rehcy(ao

    Ppeeesabo

    Whe8

    S

    aum

    Ba1

    Hsc1

    Oh8

    Me5Fme4

    -________

    __

    S

    4GaKo1

    1om

    4mose

    (Bow

    A

    eao

    Ba&

    Peao

    Dw

    S

    ue

    1

    Caombos

    -_______

    S

    1GaKo9

    1om

    4mose

    (Se

    Eromabos

    &And

    Ineps

    bo

    1

    Seaebo

    Teaebo

    3se

    6pnLkyseb

    oe

    q

    obo

    4pnLkyseb

    oe

    q

    obo

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    7/24

    654 School PsychologyReview, 1995, Vol. 24, No.4TABLE 2Reliability Information for Commonly Used Social Behavior Rating Scales

    InternalTest Name Consistency Test-Retest InterraterSchool Social Behavior Scales .94-.96a .76-.82a (3 wks.) .72-.83a(SSBS; Merrell, 1993) .9 1-.96b .60-.73b (3 wks.) .53-.71Social Skills Rating Scales(SSRS; Cresham & Ell iott , 1990)

    l SSRS-Teacher .93-.94c .85c .84d .93e (4 wks.) .32 (teacher-student)c.82-.88dl SSRS-Parent .87-.90c .87c .65d (4 wks.) .31 @arent-teacher)c.73-.87dl SSRS-Student 83 C. .68c (4 wks.) .24 (parent-student)c

    Waksman Social Skills .92 .73f .64g (4 wks.) .57-.72fRating Scale -.09-.8Og(WSSRS; Waksman, 1985)Walker-McConnell Scaleof Social Competence andSchool Adjustment

    x90 .61-.97 (2 wks.-6 mos.) .53-.77

    (WMS;Walker & McConnell,1988)School Social SkillsRating Scale

    (S3; Brown, Black, & Downs,1984)

    .81-.93h (10-21 days) .70-,78h

    Social Behavior AssessmentInventory(SBAI; Stephens &Arnold, 1992)

    .90-.94 .91-.9gh(9 of 30 subscales)

    aSSBS - Social Competence ScalebSSBS - Antisocial Behavior ScaleCSSRS Social Skills ScaledSSRS - Problem Behavior ScaleeSSRs - Academic Competence ScaleAggressive Domain subscale (no information provided for total score)gPassive Domain subscale (no information provided for total score)hAgreement p ercentages = agreements / (agreements + disagreements)**Data unless noted are based on total scores.

    special education ratio of the sample ade-quately represented the general population.Despite statistically significant mean scoredifferences between male and female sub-jects, separate norms for males and femalesare not provided. The author chose not toseparate norms by sex because of weak cor-relations between sex and SSBS scores, andoverall scores more often are compared toschool rules than gender norms. Unfortu-nately, background information was not

    given on the 688 teachers who completedratings on students. In addition, it would bebeneficial to have information about therange and average number of children ratedby each teacher.Scores &nd Interpretation

    The manual presents raw scores, stan-dard scores (M = 100, SD = 15 for total scalescores), percentiles, and social functioninglevels. Social functioning levels indicate

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    8/24

    TE3

    VdyInomofoCmyUSaBoRnSe

    TNm

    Ce

    CeoRae

    Cu

    SMe

    1

    Ceuseoaeewo

    WmSaSsRnSeWm

    boadposbe

    1

    ps

    l

    Sacme

    r=7

    Soemoaceao

    l

    Asabor=8

    CsT

    RnSeCs1

    l

    Sacme

    r=6o8

    l

    Asabor=3o9

    Sgcdeeaobw

    su

    dswhawhlennem

    todse

    Sgcdeeaobw

    su

    dswwegeah

    c

    sdetobuonaaaa

    le

    ST

    (Gem&

    Eo1

    WakMC

    SeoSaCme

    (MC

    1

    l

    Sacme

    r=9

    l

    Asabor=7

    CeuseoaeewoSaBoAm

    (Se

    1

    boadposb

    l

    Sassr=6

    es

    l

    Poembosr=5

    Uoma

    ran

    l

    Amccme

    r=6

    dmaesavdy

    CdBoCsTR(A

    &

    Soemoaceao

    Ebo1

    l

    Sassr=6

    l

    Poembosr=8

    l

    Amccme

    r=5

    HeT

    RnSeHe1

    l

    Sassr=7

    l

    Poembosr=6

    l

    Amccme

    r=6

    Snomopodu

    WM

    Lencoodoma

    c

    aothsasss

    se

    Lgsxdee

    femeebn

    mesasbosmee

    hbnmepoembos

    Sgcdeeaobw

    su

    dswhawhlennds

    aeaohdse

    (Te3cn

    .

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    9/24

    (Te3cn

    TNm

    Ce

    CeoRae

    Cu

    S

    P

    (Cem&

    Eo1

    SmaS

    T

    CdB

    oC

    sPA

    &

    Ebo1

    l

    SassS

    whsa

    cme

    (C

    =5

    0Poembos(S

    wh

    bopoem(C

    =JO

    SmaS

    T

    S

    s

    (Gem&

    Eo1

    SmaS

    Twhh

    CdB

    oC

    sYA

    &

    e

    oohuom

    Ebo1

    taanbnu

    oy

    l

    Sacme

    (C

    whsa

    ohhgs

    om

    ssS

    =2

    l

    Sacme

    (C

    whpoem

    bos(S

    =3

    PesHsCdesSK

    Se

    (Pes1

    l

    Sass=3

    SmaS

    T

    WS

    CeuseoaeewoPaPoemB

    oC

    sRs

    (Wam1

    boadposb

    (Wam1

    =6

    ju

    SnomopodfoS

    Sgcdeeaobw

    su

    dswhawhemods

    ae

    Sdee

    hgaeod

    mnseomeahgp

    svdmnseogrs

    (Te3cn

    .*

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    10/24

    (Te3cn

    TNm

    Ce

    CeoRae

    Cu

    WMWak

    &MC

    1

    CeuseoaeewoSaSsRnSemT

    Gem

    Sgcdeeaobw

    su

    boadposb

    &Eo1

    =7

    dsasfoasabo

    es

    HmeCdesSCoSe

    andesus

    (Hme1

    r=7

    Sgcdeeaobw

    n

    SemcSenoBoDsds

    reanaeansus

    (Wak&Ss1

    Av

    Sgcdeeaobw

    su

    caomboser=7

    dswhawhdse

    WkPoemBodco

    CsWak1

    e

    ranr=8

    T

    sSCoRnSeHme

    1

    r=7

    SnomopodfoS

    -~~~~~~~~~~~~~~~~~~~~~

    -

    SBow

    CeuseoaeewoNdaaesetos

    thceo

    Nnomopodocu

    Ba&Dw

    boadposb

    reaevdyohsnum

    vdy

    1

    es

    L

    onmnpoepes

    cacodso

    -

    ~~~~~~~~~~~~~~~~

    ~~~~~~~~~~~~~~~~

    SSe

    CeuseoaeewoFe

    ooabodemn

    Sgcdeeaobw

    su

    &And1

    boadposb

    tho

    boaovo

    r=4

    dswhawhemo

    o

    es

    SnomopodfoST

    lenndse

    L

    onmnpoepes

    cacodso

    NeAnomoisomhevm

    ueohwsne

    anomoisomMe1

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    11/24

    658 School Psychology Review, 1995, Vol. 24, No.4general levels of social-behavioral compe-tence. There are two social functioning lev-els which indicate the need for further eval-uation (i.e., moderate problem and signifi-cant problem). For further interpretation,the manual provides means and standarddeviations by grade and sex for the majorscales and respective subscales.Psychometric Properties

    The manual presents internal consis-tency, test-retest, and interrater reliabilityevidence for the two major scales and re-spective subscales. Internal consistency isexcellent for all scales and subscales. Ade-quate to excellent test-retest reliabilities in-dicate the SSBS has adequate short-termstability over a 3-week interval. Evidencefor interrater reliability indicates agreementgenerally is adequate. This result is limited,however, in that the sample used in the reli-ability study only included 40 elementary-aged students with learning disabilities.The validity of the SSBS was exploredin several ways. Content validity was pro-vided through an examination of the devel-opment of the instrument. Behavioral de-scriptors in each of the two areas were ex-amined and reviewed by teachers, graduatestudents, and parents of children in theK-12 grade range. Adequate item-total cor-relations provided further evidence for con-tent validity. Three criterion-related validitystudies were conducted, all of whichdemonstrated adequate to excellent corre-lations between the two major scales of theSSBS and criterion measures. Two studieson construct validity provided evidence forthe ability of the SSBS to distinguish social-behavioral differences between students.Factor analysis (further evidence of con-struct validity) also provided justificationfor the use of the subscales on each of thetwo major SSBS scales.Limitations of the reliability and validityresearch include inadequate informationabout test-retest reliability over longer timeintervals, poor description of item selectionwhich limits conclusions about content va-lidity, and a lack of reliability and validityresearch conducted by independent re-searchers.

    SummaryThe SSBS has the potential to be a use-ful measure of social competence and anti-

    social behavior patterns of school-aged chil-dren. Due to its recent development, addi-tional research is needed to further sub-stantiate the reliability and validity of thisscale. Presently, positive features includethe comprehensiveness of the manual, theease with which the instrument can beused, and the simplicity of the interpretativeframework. Evidence provided in the man-ual indicates the SSBS has adequate to ex-cellent reliability, a solid factor structure,and adequate content, construct, and dis-criminant validity. Presently, the SSBS canbe used as a screening instrument to iden-tify problem behaviors which may warrantfurther evaluation.SOCIAL SKILLS RATING SYSTEMThe Social Skills Rating System (SSRS;Gresham & Elliott, 1990) is a norm-refer-enced rating scale comprised of three sepa-rate rating forms for teachers, parents, andstudents. Instrument length varies depend-ing upon the rater and grade level of the stu-dent. The SSRS is intended for use withpreschool (age 3-5), elementary (GradesK-6), and secondary (Grades 7-12) stu-dents. The number of items ranges from40-57 for the teacher form, 49-55 for theparent form, and 34-39 for the student form.Responses are completed on a 3-point Lik-e&-type scale. Three main scales make upthis instrument: social skills (teacher, par-ent, and student forms), problem behaviors

    (teacher and parent forms), and academiccompetence (teacher form). The socialskills scale includes a separate 3-point im-portance rating (teacher and parent formsonly). All SSRS forms consist of three sub-scales: cooperation, assertion, and self-con-trol. The parent version includes an addi-tional responsibility subscale, whereas anempathy subscale is unique to the studentversion. The purpose of the SSRS is to as-sist professionals in screening and classify-ing children suspected of having significantsocial behavior problems and aid in the de-velopment of appropriate interventions foridentified children (Gresham & Elliott,1990, p. 1).

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    12/24

    Social Skills Assessment 659Content and Use

    The SSRS can be completed in 20 min.Materials are user friendly thus facilitatingaccurate completion of the instrument bynonprofessional rates. Particular strengthsinclude layout of the rating form, readabil-ity, clear and comprehensive examples inthe manual, and clearly phrased items. Themanual recommends that all SSRS forms becompleted in coordination with each otherto achieve the most thorough picture of achilds behavior. However, any SSRS formcan be administered and scored separately.The SSRS is easily scored by hand in about5 minutes. The manual includes all neces-sary information regarding scoring, andhelpful step-by-step cases demonstratingthe scoring and interpretation process. Amicrocomputer scoring program also isavailable.The Assessment Intervention Record(AIR), an eight-page integrative form, is abeneficial feature of the SSRS that providesa means to assist in the coordination of themulti-rater information. The AIR provides amethod to document background inforrna-tion and aids in analyzing and synthesizingmulti-rater assessment data that includessocial behavior strengths and weaknesses.Standardization Sampleand Norms

    The standardization sample included4,170 children from Grades 3 to 10, whowere rated by their teachers and parents.This sample is large and includes nearlyequal numbers of girls and boys. Thepreschool norms were developed from asmaller national try-out sample of childrenages 3 to 5 (N = 200). A limited number ofstudents were included in the 11th (N = 44)and 12th grade levels (N = SO). A weaknessof the standardization is an overrepresenta-tion of Whites and African Americans, andan underrepresentation of Hispanic Ameri-cans and other minorities. Norms are avail-able for broad age or grade level groupings(preschool, Grades K-6 and Grades 7-12),sex, and disability status (for Grades K-6).The inclusion of norms for elementary stu-dents with disabilities is a noteworthy fea-ture and may be helpful in facilitating clas-sification decisions. In addition, the manual

    provides information about the parents andteachers who completed ratings on stu-dents.Scores and Interpretation

    The manual presents raw scores, star+dard scores (M = 100, SD = 15), percentileranks, confidence bands, and behavior lev-els. Behavior levels (i.e., fewer, average,more> are descriptive methods for inter-preting social skills and problem behaviorsin comparison to the standardization sam-ple. The importance ratings are useful whendeveloping intervention plans. Designationof a social skills behavior level as fewernecessitates consideration of importanceratings in conjunction with frequency rat-ings. At this point, the authors recommendan item-level analysis be undertaken to de-termine if behaviors constitute social skillsacquisition deficits (lack of a skill) or per-formance deficits (skill exhibited inconsis-tently).Psychometric Properties

    The SSRS manual provides detailed in-formation about the psychometric qualitiesof all three forms of the instrument. The in-ternal consistency of the teacher form is ex-cellent. With regard to the parent and stu-dent forms, the internal consistencies areadequate overall; however, the reliability es-timate for the preschool problem behaviorscale is limited. Test-retest reliabilities areexcellent for the teacher form. Stability rat-ings also are excellent on the SSRS-parentsocial skills scale and limited for the prob-lem behaviors scale. Test-retest reliabilityfor the SSRS-student-social skiIls scale islimited.Interrater reliability coefficients for thetotal social skills subscales collapsedacross the three levels (preschool, elemen-tary, and secondary) are relatively low. It isimportant to note that childrens behaviormay vary from one setting to another, thusincreasing the likelihood that the reliabilitybetween raters will be lower. Although low,the SSRS interrater reliability coefficientsare slightly better than most cross-infor-mant ratings (e.g., Achenbach, Mc-Conaughy, & Howell, 1987).

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    13/24

    660 School Psychology Review, 1995, Vol. 24, No.4Evidence is presented in the manual forthe content, criterion, and construct validityof the SSRS. The content validation stems

    from item development based on empiricalliterature and childrens social behaviors. Amajor source of content validity involvedthe use of importance ratings to providemeaningful content validation in terms ofthe social significance for items included inthe scale. Criterion-related validity studiesfor the SSRS-Teacher and SSRS-Parentyielded adequate reliability coefficients.The SSRS-P currently is the only formal par-ent measure of childrens social skills. Thus,there is a lack of directly appropriate in-struments with which to compare. Crite-rion-related validity evidence for the stu-dent form is limited to adequate.The authors presented several types ofevidence for construct validity, focusing ondevelopmental changes, sex differences,and group separation. The validity of theSSRS has been examined in other studies, inaddition to those reported in the manual(e.g., Chewning, 1992; Stinnett, Oehler-Stin-nett, & Stout, 1989). Specifically, these stud-ies investigated the construct validity of theSSRS-T, and found it able to discriminateamong groups of children. Construct valid-ity also is supported through moderate tohigh correlations regarding developmentalchanges with the Walker-McConnell Scale ofSocial Competence and School Adjustment(Walker & McConnell, 1988) and the Mat-sons Evaluation of Social Skills withYoungsters (Matson, Rotatori, & Helsel,1983).Summary

    The SSRS is a helpful tool for assess-ment and intervention planning, and shouldbe considered by those looking for amethod to understand a childs behaviorfrom multiple perspectives. The SSRS sys-tem is user friendly; the materials and man-ual are clear and well-designed. A definingfeature of the SSRS is its capability for ob-taining multi-rater comparisons. However,due to the nature of item reduction thereare only a modest number of common so-cial skills items for comparison. Item reduc-tion completed on the SSRS consistently re-flects the unique and specific nature of so-

    cial skills across settings, as well as behav-iors that are valued by those making ratings.In fact, the items were derived from thosethat were considered important to specificraters (i.e., parents and teachers). Nonethe-less, caution must be used when comparingsocial skills ratings because fewer than 40of the items are common across the threeforms of the system. Hence, consumersmust be aware of the lack of complete itemconsistency and make appropriate compar-isons as recommended by the authors of theSSRS.The standardization sample is of excel-lent size and fairly representative, although11th and 12th grade norms are based on alimited sample size, which limits use of thescale with these grade levels. Overall, thepsychometric properties of the SSRS are ex-cellent. A few exceptions include somewhatlimited test-retest reliability for the studentform, relatively limited reliabilities on theparent form problem behavior scale, andlower criterion-related validity on the stu-dent form. Thus, the problem behaviorscale should be used primarily as ascreener.

    WMMAN SOCIAL SKILLSRATING SCALEThe Waksman So&at SkiZZs RatingScaZe(WSSRS; Waksman, 1985) is a brief,2 l-item norm-referenced scale designed forratings by teachers on students in GradesK-12. Responses are made on a 4-point Lik-ert-type scale. One major scale (socialskills) and two subscales (aggressive andpassive) comprise the measure. Separateforms exist for male and female students.The WSSRS was developed to assist psy-chologists, educators, and other cliniciansto identify specific and clinically importantsocial skill deficits in children and adoles-cents (Waksman, 1985, p. 1). The authorsuggested the scale has several uses whichinclude screening, identification and classi-fication, selecting students for social skillstraining or counseling programs, and pro-gram evaluation.

    Content and UseAdministration of the WSSRS is easyand can be completed quickly. Brief instruc-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    14/24

    Social Skills Assessment 661tions ask teachers to base ratings on first-hand knowledge of a student, and to use theaverage student in the regular educationclass as a basis for comparison. Potentialdifficulties exist with some aspects of theprotocol. For example, the aggressive andpassive domains are clearly titled on theprotocol, which may lead the witness bycreating a response bias. The author appro-priately suggested that multiple ratingsshould be obtained to assess cross-situa-tional behaviors. The manual is well laidout, although the brevity of the manual re-sults in the failure to include more informa-tion helpful to the consumer, specifically inthe areas of scale development. Hand scor-ing is easily accomplished on a single sheetprotocol.Standardization Sample and Norms

    Kindergarten through high school stu-dents (N = 331) from 10 schools in Portland,Oregon were rated by their teachers. Nei-ther ethnicity nor SES were reported for thestandardization sample. A further limitationis that demographic information is not pro-vided for the teachers who performed theratings, nor is information provided aboutthe date or year in which the scale wasnormed.Scores and Interpretation

    The interpretive framework of theWSSRS rests primarily with percentilescores for combined grade levels (i.e., K-5,6-8,9-12). No rationale was offered for theuse of the different grade level combina-tions. A major oversight in the manual is thelack of standard scores and standard errorsof measurement. Also, although separate fe-male and male protocols often are usedwith rating scales because of differing re-sponse patterns, a clear rationale was notprovided in the manual for the use of theseparate forms.Psychometric Properties

    The reported internal reliability esti-mate is evidence for high internal consis-tency. Test-retest reliabilities were reportedfor the WSSRS on a l-week and l-month in-terval. One month reliabilities were ade-

    quate for both the aggressive and passivedomains, with no information provided forreliability of the total score. One week test-retest reliabilities appear to be high, how-ever, the calculations are limited due to theuse of a sample of 30 seventh and eighthgraders. The manual presents three studiesinvestigating the interrater reliability of theWSSRS. These studies contained very smallsamples (N = 7 for two studies and N = 39for one study). They yielded adequate inter-rater reliability for the aggressive domainand inconsistent reliability coefficients forthe passive domain. Interrater reliabilitiesfor the total score were not reported. It isnoted in the manual that one teachers in-consistent ratings were responsible for rad-ically affecting the coefficient (p. 3). Itwould have been helpful and appropriate,however, if the author also had provided co-efficients that did not use those inconsis-tent ratings. Taken together, the inconsis-tent results and restricted samples limitconclusions about the reliability of theWSSRS.Content validation of the WSSRS con-sisted of importance ratings of scale itemsfrom 0 (low importance) to 5 (high impor-tance) of scale items. Judges were asked torate if the behaviors were important for achilds development (Waksman, 1985, p. 2).It was not stated whether the raters had anyexpertise in the areas of social skills, childdevelopment, and/or child behavior. Also, itis unclear how the items were perceived bythe judges, since the items describe nega-tive behaviors that should not be demon-strated. Concurrent and construct validitycoefficients provided in the manual are ade-quate. However, inconsistent fmdings indi-cate the need for further concurrent validitystudies. The construct validity of theWSSRS is supported by higher aggressivedomain scores for males and significantlyhigher scores for adolescents with emo-tional disabilities as compared to the nor-mative sample.Summary

    The WSSRS has several features thatusers may find attractive. The measure evi-dences excellent internal consistency, it canbe completed quickly, and the single-sheet

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    15/24

    662 School Psychology Review, 1995, Vol. 24, No.4protocol provides for simple scoring andquick processing. The manual provides sev-eral tables with thoughtful item and sum-mary information. Several aspects of thescale and development, however, mitigateagainst its use. Specifically, the protocol hassome questionable properties, and the un-representative nature of the sample limitsthe generalizability of the scale. Further-more, test-retest reliability and interrater re-liability are not sufficiently demonstrated,and validity information is lacking, espe-cially with regard to concurrent validity. l?i-nally, the scale measures social skillsdeficits rather than social skills per se,which we feel undermines the utility of thescale by focusing on negative rather thanprosocial behaviors.

    WALKER-MCCONNELL SCALE OFSOCIAL COMPETENCE ANDSCHOOL ADJUSTMENTThe Walker-McConnell Scale of SocialCompetence and School Adjustment (WMS;Walker & McConnell, 1988) is a brief, 43-item norm-referenced rating scale designed

    primarily for completion by teachers. Re-sponses are made on a &point Like&typescale. The WMS is intended for use with el-ementary students (Grades K-6). One majorscale (social competence) and three sub-scales (teacher-preferred social behavior,peer-preferred social behavior, and schooladjustment) comprise this instrument. Thepurpose of the WMS is the screening andidentification of social skills deficits amongelementary aged children in school(Walker & McConnell, 1988, p. 1). The WMSrecently was revised to extend its use toadolescents in Grades 7 to 12. (See Walker,Stieber, & Eisert [ 19911 for a comprehensivediscussion of the revision procedures.)However, this revision has yet to result inthe publication of a revised manual or newrating forms. Thus, the present review willfocus exclusively on the elementary versionof the WMS, which currently is a more es-tablished instrument.Content and UseThe authors estimate that administra-tion takes 5 minutes, which is an appealingaspect of this scale. The items are ordered

    randomly and are identified by a number ofrepresenting one of the subscales, thus pre-venting a possible labeling bias. A problemwith this format, however, is that it is noteasy to differentiate items when derivingsubscale raw scores. Some organizationaldevice could have been included on the pro-tocol to avoid this inconvenience and possi-ble score aggregation errors. The WMSmanual is comprehensive and generallywell-written, although a section addressingthe technical adequacy of the scale is some-what disjointed and lacks clarity. The tablesin this section are not labeled adequatelyand use descriptions that are, at times, dif-ferent from what is described in the narra-tive. For example, the authors reportedthere were significant mean differences fa-voring regular first graders over enteringtransition students on two of the three sub-scales and total scale score (p. 18). Withinthe manual, a table contradicts this reportby showing all three subscales to be signify-cant at p c .05.Standardization Sample and Norms

    The WMS was standardized during 1985and 1986 on 1,812 children adequately dis-tributed by sex and grade. The sample wasrepresentative of children from four majorU.S. census zones, however, the normswere not representative of the U.S. popula-tion. States from the Northwest represented49 of the sample, with a disproportionatepercentage coming from Alaska (18 ).Within the sample, Whites were slightlyoverrepresented, whereas African Ameri-cans and Hispanic Americans were slightlyunderrepresented. Also, the norm grouphad a substantial percentage of children inthe unspecified ethnic/racial category ofOther.Scores and Interpretation

    Standard scores (2M = 100, SD = 15 fortotal scale; m = 10, SD = 3 for subscales)and percentile ranks are provided for thetotal scale and the three subscales. The au-thors suggest that all items with a low ratingof 1 or 2 may be indicative of poor socialskills and potentially should be targeted forremediation. In addition, one of the pre-scribed uses of the WMS is for developing

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    16/24

    Social Skills Assessmentindividualized educational programming(IEP) goals by looking at item means andstandard deviations. These item level inter-pretations should be considered with cau-tion.Psychometric Properties

    The manual provides test-retest, inter-nal consistency, and interrater reliability ev-idence for the total scale and subscales. Ex-cellent internal consistency and adequate toexcellent test-retest reliabilities are re-ported. Interrater reliability is in the ade-quate range. Evidence for interrater reliabil-ity was based on only two studies, each ofwhich involved a small number of teachersand their instructional aides. Overall, theWMS appears to possess adequate reliabil-ity, but further studies with larger samplesare necessary to build a strong case for thereliability of the scale.The WMS appears to possess adequatecontent validity as demonstrated by item re-views by the authors for content relevanceand item clarity. Construct validity wasdemonstrated by a second order factoranalysis with varimax rotation, which ade-quately specified a three-factor solution. Allitems exhibited high commonality valuesand typically loaded more strongly on oneof the three factors. To demonstrate dis-criminant validity, numerous studies arepresented in the manual as support for thescales ability to differentiate between iden-tified groups. Many of these studies hadsmall sample sizes, used only a portion ofthe WMS, were described inadequately and,more critically, were not reported in a clear,integrated fashion. Thus, it is difficult todraw conclusions about the discriminantvalidity of the WMS. In a recent study byMerrell, Johnson, Merz, and Ring (1992),however, students with mild disabilities re-ceived significantly lower ratings overall onthe WMS as compared to average, generaleducation students. The authors providedcomparisons of the WMS to other measuresas support for criterion-related validity Thefindings also provided support for constructvalidity by comparing the WMS to otherscales measuring similar or divergent con-structs

    663Summary

    The WMS is a quick and easy to usemeasure that is appropriate for screeningchildren for social skills deficits. The sub-scales are highly intercorrelated, thus thedifferentiation is questionable betweenteacher-preferred and student-preferred so-cial behavior. The WMS possesses adequateto excellent overall reliability and appearsto have a substantial theoretical basis. Theauthors provide an array of data as evi-dence for validity, yet much of the evidenceis difficult to interpret given the manner inwhich the studies are presented in the man-ual. However, the WMS does appear to pos-sess adequate content and criterion-relatedvalidity.

    SCHOOL SOCIAL SKILLS@a RATING SCALEThe School Social SIXls Rating Scale(S3; Brown et al., 1984) is a 40-item, crite-rion-referenced rating scale. Responses aremade on a 6-point Likert-type scale. Thescale is intended for use with children from

    Grades K to 12. The S3 consists of four sub-scales: adult relations, peer relations,school rules, and classroom behaviors. It isdesigned to identify strengths and deficitsin school-related social behaviors of stu-dents. Sweetland and Keyser (1991) statedthat the S3 scale can be used for social skillsinstruction, discussion of student behaviorwith parents and other school personnel,and development and monitoring of socialbehavior goals and objectives for individualeducation plans (IEPs). In addition, an ac-companying curriculum can be used to im-prove identified social skills deficits.Content and Use

    It takes approximately 10 minutes forschool personnel to complete the scale forone student. The authors recommendedmultiple administrations of the S3 to iden-tify childrens specific social skills deficitsand to monitor progress. The authors statedthat a rater should be thoroughly familiarwith the scale and accompanying S3 cur-riculum before completing the scale Thisrequirement could involve a considerableamount of the raters time which is a draw-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    17/24

    664 School Psychology Review, 1995, Vol. 24, No.4back of this measure. The manual and ma-terials are clear and specific regarding di-rections for administration and scoring. Inaddition, the manual provides useful exam-ples that demonstrate scoring procedures.Scores and Interpretation

    Ratings on the S3 do not result in totalscores or summation of area scores. Sincethe scale was not standardized on a norma-tive sample, there are no standard scoresnor percentile ranks. The instrument is sim-ply a list of well-written, objective items.The scale does not allow for classificationor diagnostic decisions. However, it does fa-cilitate problem conceptualization and maybe useful in planning and/or selecting inter-ventions.Psychometric Properties

    Evidence for the reliability of the S3scale was provided in the form of test-retestand interrater reliability. The interrater reli-abilities ranged from adequate to excellent,and were based on the level of agreementbetween ratings conducted by independentobservers. The authors did not specifywhether these raters were parents, teach-ers, or other school personnel. Test-retestreliability ranged from adequate to excel-lent. A limitation of the test-retest and inter-rater reliabilities is that they were calcu-lated using a percentage agreement ratherthan correlational approach. The authorspresented adequate evidence for contentvalidity, but did not provide data to supporttheir claim for criterion-related or constructvalidity. The content validity of the instru-ment consisted of three steps: (a) creatingan item pool from several behavior ratingand social skills instruments; (b) a review ofthe scale by the authors, whose back-grounds consist of regular and special edu-cation training, teaching, and administra-tion; and (c) surveying teachers about theclarity and importance of the items.SummaryThe S3 is a simple to use measure, with cleardirections and useful examples provided inthe manual. It is relatively quick and easy touse, however, it can take extensive time tostudy the S3 curriculum before using the

    scale. The S3 has adequate to excellentoverall reliability; however, no evidencewas provided for internal consistency. Thevalidity of the instrument was not clearlydemonstrated by the data presented in themanual. The instrument is not appropriatefor classification purposes, however, usersmight consider using it if the primary goal isto identify target behaviors for classroom-based interventions.

    THE SOCIAL BEHAVIORASSESSMENT INVENTORY @BAI)The Social Behavior Assessment In-

    ventory (SBAI; Stephens & Arnold, 1992) isa 135-item criterion-referenced rating scale.The SBAI is a revision of Stephens (1979)original Social Behavior Assessment (SBA).Responses are made on a 5-point Likert-type scale. The SBAI is intended for usewith students from Grades K to 9. The SBAIyields four areas of social behaviors: envi-ronmental, interpersonal, self-related, andtask-related. There are 30 subscales thatcomprise each of these areas. The authorsidentified four uses of the SBAI: (a) as a cur-riculum-based assessment for children ex-periencing social behavior problems; (b) asa screening tool to identify potential behav-ior problems; (c) to differentiate betweenstudents with and students without learningdisabilities, and between students with andwithout emotional disturbances; and (d) tomeasure success of an intervention overtime when repeated measurements are ob-tained. In addition, the SBAI was designedto be used in conjunction with Stephens(1992) Social SMlls in the Classroom inter-vention guide.Content and Use

    Time required to complete the SBAIrecord form is 3045 minutes per student.The SBAI can be administered in two ways.A teacher or other school personnel cancomplete the record form based on eithercumulative, retrospective knowledge of atarget students behavior, or while directlyobserving the target child. The record formallows the rater to compare behavior rat-ings across scales and within scales to de-termine the problem area(s).

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    18/24

    Social Skills Assessment 665Scores and Interpretation

    The SBAI essentially measures discrep-ancies between observed and reported be-havior and a teachers expectations for thebehavior. Item scores are summed withinsubscales to allow for visual analysis of stu-dents needs. The SBAI is not a norm-refer-enced scale, therefore, whether a studentsbehavior falls within expectation or ele-vated for any of the subscales will be basedsolely on the opinion of a single rater. If adiscrepancy is obtained, teachers or otherschool personnel can intervene by using thecompanion book, SociaZ SkiUs in the CZass-room (Stephens, 1992) to implement a so-cial skills curriculum. Thus, the use of theSBAI is limited primarily to selecting targetbehaviors for individual corrective training.Psychometric Properties

    The manual presents evidence concern-ing the internal consistency and interraterreliability of the SBAI. Overall, studies citedin the manual indicate high internal consis-tency of item content. However, internal re-liabilities were not reported separately bycontent area or by subscale which limitsclaims for the reliability of the scale. Dataregarding interrater agreement for the SBAIare excellent, however the study samplewas small (n = 40) and only investigatednine of the subscales. In addition, there isno report in the manual of test-retest relia-bility. Thus, more research is needed to es-tablish the reliability of the SBAI.Content validity was demonstratedthrough the development of the SBAI itemswhich included classroom observations, re-view of the social skills literature, and acontent analysis of behavior rating scales.Items were included on the SBAI if the spe-cific social skills correlated highly withschool success, were rated by teachers asimportant for school success, or appearedfrequently in published behavior ratingscales. Special education teachers (N = 200)and general education teachers (N = 200)rated the remaining 136 items on a 6-pointscale of importance for success in the class-room. The authors state that .a11 136 itemswere found to be sufficiently important andwere retained on the final version of theSBAI. In actuality, there are 135 items on the

    SBAI. The fact that one item was dropped.from the original SBA is not addressed. Wecan have confidence in the content validityof the SBAI given that a large number ofteachers of students with emotional distur-bances, learning disabilities, and cognitivedisabilities, as well as regular educationteachers rated the importance of items(Drabman, 1985). Convergent validity is ad-equate and was calculated to determine theextent to which the purported constructsmeasured by the SBAI are related to similarconstructs. The authors inappropriatelyconcluded that a correlation coefficient of.46 between frequency of on-task behaviorsand SBAI teacher ratings indicates moder-ately high convergent validity for the instru-ment with Hispanic populations. Althoughthe authors do not formally address the is-sue of construct validity, the informationpresented regarding sensitivity and speci-ficity of the SBAI provided some evidencethat the instrument was able to discriminateamong groups. However, the lack of con-struct validation information is a seriousomission. In addition, although an attemptwas made to test for bias (against Hispan-ics), the results were ambiguous. Thus,there is a clear need for more research tovalidate the constructs assumed to be mea-sured by the SBAI and to determinewhether it is culturally biased.Summary

    It appears the data used to develop andsupport the SBAI in the 1990s are the samedata used to support the SBA in the late1970s. Although the two measures are verysimilar, the manual does not clearly differ-entiate the SBAI from the SBA. Finally, be-cause the information that supports thetechnical properties of the SBAI apparentlyis the same information used to support theSBA, there has been no reported improve-ment in psychometric properties of the in-strument.The SBAI does appear to serve itsstated purpose of identifying target behav-iors for intervention. Teachers who rate tar-get children as displaying specific socialskills at an unacceptable level are able toact upon their concerns by implementingthe instructional or training process out-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    19/24

    666 School Psychology Review, 1995, Vol. 24, No.4lined in the companion book entitled SociaZSh Us in the Classroom (Stephens, 1992).The manual reported excellent internal con-sistency, however, the data are not pre-sented separately for the content areas orsubscales. Another study of limited sizedemonstrated excellent interrater reliabilityon nine of the subscales. Adequate contentand convergent validity were demonstrated.No data are presented for test-retest relia-bility or construct validity. A literaturesearch resulted in no additional evidencesupporting the psychometric properties ofthe SBAI. Additional research therefore isneeded to establish the reliability and con-struct validity of the SBAI.

    DISCUSSIONAfter reviewing the social skill scales itis apparent that many similarities and sev-eral significant differences exist amongthem. Although the scales seem to have acommon purpose - assessing social skillsin children - they demonstrate differencesin the way they attempt to achieve theirgoal. The distinguishing characteristics and

    similarities can be highlighted by consider-ing the construct being assessed, appropri-ate use, user-friendliness issues, and techni-cal properties.Construct of Social Skills

    The construct of social skills has beenconsidered from many different perspec-tives as demonstrated with the scales re-viewed in this article. For example, the So-cial SkiUs Rating System (SSRS; Gresham& Elliott, 1990) includes a measure of prob-lem behaviors and academic competencebecause of their interplay with a studentseffective social interactions. The School So-cial Behavior Scales (SSBS; Merrell, 1993)also includes problem behaviors as a majorscale to cover both positive and negative so-cial behavior. Both the SSBS and the SSRShighlight social skills as an integral compo-nent in evaluating the social functioning ofchildren at school. The Social Behavior As-sessment Inventory (SBAI; Stephens &Arnold, 1992) and the Walker-McConnellScale of Social Competence and School Ad-justment (WMS; Walker & McConnell,1988) focus on Dotential skills that demon-

    strate social competence, with the SBAIsemphasis on target behaviors for trainingand the WMSs emphasis on adtjustmentsnecessary for school success. The Wahs-man Social SiMls Rating Scale (WSSRS;Waksman, 1985) approaches the constructof social skills in terms of deficits and con-tains negatively worded items. The SchooZSocial Skills Rating Scale (S3; Brown et al.,1984) concentrates on social skills deficits,as well as strengths.Appropriate Use

    The authors of the four norm-refer-enced scales (i.e., SSBS, SSRS, WSSRS, andWMS) claim their instruments can be usedfor assisting the process of screening andidentifying children with social behaviordifficulties. The criterion referenced scales(S3 and SBAI) have an intervention focus,and are useful for a more specific examina-tion of particular behaviors to change. TheSSBS and SSRS are comprised of objective,specific items and also are well-suited foridentifying target behaviors.User FriendlinessIn determining the user friendliness ofthe measures, administration time, types ofscores, and clarity of the manual were con-sidered. Most scales take between 5 to 10minutes to complete, although the SSRS-Ttakes about 20 minutes due to the additionof problem behavior and academic compe-tence scales, and the SBAI takes about 35 to40 minutes because of the large number ofitems. The scales all were consistent in theirease of administration and scoring, which isan advantage of using rating scales. How-ever, the S3 was more difficult to administerinitially due to the amount of time it takes tolearn the S3 curriculum.The four norm-referenced scales variedin the interpretive data they provided. TheWSSRS provided the least; only presentingpercentile information. An advantage of theSSBS, WMS, and the SSRS is their use ofstandard scores. Inclusion of standardscores is helpful because they provide amethod for interpretation in relation topeers and other instruments. The WMS alsoprovided percentiles and standard errors ofmeasurement. The SSBS and the SSRS in-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    20/24

    Social Skills Assessment 667eluded the most interpretive information,with the inclusion of social functioning lev-els and behavior levels, respectively. The S3and SBAI as criterion referenced scales donot provide standardized scores or per-centile ranks. These two instruments re-quire evaluating raw scores in relation tosatisfactory acquisition levels.The manual also is an important part ofthe user friendliness and comprehensive-ness of an instrument. It is a users guideand the primary, if not only, information re-garding development and use of the instru-ment. We found the most comprehensivemanual was the SSRS because of the com-plete coverage concerning the developmentand technical information, as well as theguided scoring examples for all forms. TheSSBS also included a useful, well-writtenmanual. The WSSRS and WMS manualswere lacking in some areas. Specifically, theWSSRS did not include important scale de-velopment information that would be bene-ficial for the user. The WMS manual alsowas not adequate because of the inconsis-tencies between the narrative and tableswhen discussing technical information. Wefound the manual for the S3 to be useful,and the inclusion of examples was helpful.The SBAI manual was very clear and userfriendly. However, it did not provide an ex-planation of the differences between theoriginal Social Behavior Assessment (SBA;Stephens, 1979) and the SBAI, or a rationalefor the revision.Technical Properties

    The technical properties we evaluatedincluded standardization samples, reliabil-ity, and validity. Often a scale may be ap-pealing because of attractive protocols,ease of use, or familiarity. However, the in-strument also needs to be based uponsound psychometric properties to yield themost reliable and valid information.Sample size, demographics, and re-gional representation are factors to con-sider when evaluating standardization sam-ples. The standardization sample was excel-lent for the SSRS. The SSBS and WMS alsowere based on large, diverse samples; how-ever, both were limited in regional repre-sentativeness. The WSSRS is based on an in-

    adequate standardization sample from onlyone city in Oregon.With regard to the reviewed instru-ments reliabilities, we evaluated internalconsistency, test-retest, and interrater relia-bilities. All scales reported excellent inter-nal consistency, with two exceptions beingthe SSRS-Student and the S3. The SSRS-Stu-dent did evidence adequate internal consis-tency, however, it was not as strong as theteacher and parent versions. It also must benoted that the internal consistency informa-tion for the SBAI was not reported for con-tent area or subscale which limits its inter-pretation. Unfortunately, no internal consis-tency information was provided for the S3.It is necessary to evaluate test-retest re-liabilities very carefully because samplesize for this evidence often is quite smalland the time from the first administration tothe second is not consistent across thescales. The SSRS evidenced excellent sta-bility ratings on its teacher and parent ver-sion but the student version was limited.Adequate to excellent test-retest reliabili-ties were reported for the WMS. The highreliabilities were from a relatively short Z-week time interval, whereas the lower relia-bilities were those derived from a 6-monthinterval. Adequate test-retest reliability wasreported for Scale A of the SSBS, whereasScale B reliability was limited. The WSSRSdemonstrated limited reliability Moderateto high test-retest reliabilities also were re-ported for the S3. Technically, however,these are not reliability coefficients, butrather agreement percentages. Unfortu-nately, no test-retest information was re-ported for the SBAI.Adequate interrater reliability was re-ported for both the WMS and SSBS Scale A,however, Scale B of the SSBS evidencedlimited to adequate interrater reliability Ex-cellent interrater reliability was reportedfor the SBAI, and those for the S3 were ade-quate to excellent. Yet, both are agreementpercentages and not technically interraterreliabilities. In addition, only a limited num-ber of subscales were used as evidence forthe SBAIs reliability. Only limited conclu-sions can be drawn concerning the inter-rater reliability of the WSSRS because of in-consistent results and restricted samples.The interrater agreement for the SSRS ap-

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    21/24

    668 School PsychologyReview, 1995, Vol. 24, No.4pears low when compared to the other re-viewed scales. One must realize that thesereliabilities are based upon cross-informantratings (teacher-parent, student-parent,student-teacher) and, as such, the inter-rater reliabilities reported for the SSRS areadequate and consistent with research find-ings on interrater reliabilities between dis-similar raters (see Achenbach et al., 1987).Validity evidence was provided for thereviewed scales through content, criterion,and construct validity. Overall validity eval-uations take into account these three areas.Substantial validity evidence was providedto support the use of the SSRS and SSBS. Astrong feature of the SSRS is its use of im-portance ratings as a meaningful social val-idation. One exception to the overall amplevalidity evidence for the SSRS is the low cri-terion related validity for the student ver-sion of the scale. Further research is neededto substantiate validity evidence for the re-cently developed SSBS. The WMS reportedgood content and criterion related validityHowever, much of the validity evidence forthe WMS was difficult to interpret whichlimits its usefulness. Adequate validity evi-dence is reported for the WSSRS, howevervalidity information is lacking and inconsis-tent in comparison to the other measures.The criterion referenced instruments (S3and SBAI) provided only limited validity ev-idence.A final and important aspect for consid-eration rests within the construct of treat-ment validity. Treatment validity of an as-sessment measure is demonstrated whenthe measure enhances treatment outcome(Nelson & Hayes, 1979). Linking socialskills assessment to intervention throughthe use of specific target behaviors andtreatment guides/curricula potentially addsto the validity and usefulness of a scale. Thequestion remains, however, whether a givensocial skills instrument serves to improvetreatment outcomes. Although the topic oftreatment validity is beyond the scope ofthis article, future reviews and research onsocial skills rating scales are warrantedwhich focus specifically on this issue.

    ConclusionsConsumers of rating scales that assesssocial skills need to identify the type of in-

    formation they require, as well as the valid-ity of the instrument. Thus, potential socialskills rating scale users need to be informedconsumers. Through evaluating the areason which we chose to focus (content anduse, standardization sample and norms,scores and interpretation, and reliabilityand validity) the following recommenda-tions are offered.The most comprehensive instrument isthe SSRS. We recommend using the SSRSbecause of its multi-source approach, inter-vention linkage, and overall strong reliabil-ity and validity. For a more limited schoolscope, the SSBS and the WMS are recom-mended. These instruments are primarilydesigned for use in the school setting byteachers. Concentration on social skills asperceived only by school based personnel,however, is a limitation of both scales con-sidering best practices necessitates evaluat-ing behaviors in multiple settings using mul-tiple sources. Good technical evidence isdemonstrated by both the SSBS and theWMS scales. However, we must add a cau-tionary note to the WMS because of the un-clear presentation in the manual of its valid-ity evidence. The remaining norm refer-enced scale, the WSSRS, offers some attrac-tive features (e.g., brevity and good internalconsistency). This scale is not recom-mended, though, because there are morepsychometrically sound alternative ratingscales.

    Finally, the two criterion-referencedscales (S3 and SBAI) are useful for identify-ing specific target behaviors for interven-tion. These two instruments, as with all cri-terion-referenced scales, are not appropri-ate for making screening or classificationdecisions. This is a limitation for profes-sionals who need to make a range of deci-sions. The SBAI and S3 also have limitedpsychometric support. A weakness of the S3versus the SBAI is the extensive time ini-tially required to study the S3 curriculum. Astrength of the SBAI is the availability of acompanion intervention guidebook, SocialSh ZZsn the CZassroom (Stephens, 1992)which outlines step-by-step procedures for

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    22/24

    Social Skills Assessment 669teaching social skills found to be lacking ac-cording to ratings on the SBAI.In conclusion, many important aspectsmust be considered when evaluating thequality and usefulness of rating scales.Users must be knowledgeable about thequalities of the instrument for appropriateuse and interpretation. Overall, ratingscales provide a means to evaluate the so-cial skills of a specific child in a short pe-riod of time. Because assessment of chil-drens social skills requires a multi-methodapproach, rating scales are an effectivemeans of information gathering to be usedin conjunction with other types of assess-ment such as observations, interviews, self-reports, and sociometric techniques.

    FOOTNOTE&ice the initial writing of this article the originalresearch version of the Matson Evaluation of SocialSkills with Youngsters (MESSY; Matson et al., 1983)was published. Space constrains precluded its inclu-sion for review.

    Edelbrock, C. (1983). Problems and issues in using rat-ing scales to assess child personality and psy-chopathology. School Psychology Review, 12,293-299.Elliott, S. N., Busse, R. T., & Gresham, F. M. (1993). Be-havior rating scales: Issues of use and development.School Psychology Review, 22,313-321.Elliott, S. N., Sheridan, S. M., & Gresham, E M. (1989).Assessing and treating social skills deficits: A casestudy for the scientist-practitioner. Journal ofSchool PsychoLogy, 27, 197-222.Gresham, E M. (1988). Social competence and motiva-tional characteristics of learning disabled students.In M. Wang, M. Reynolds, & H. Walberg (Eds.),Handbook of special education: Research andpractice (Vol. 2; pp. 283-302). Oxford, England:Pergamon Press.Gresham, E M., &Elliott, S. N. (1990). Z eSocial SkiUsRating System. Circle Pines, MN: American Guid-ance Service.Harter, S. (1985). Manual for the SewPerception Pro-file for chi&&~. Denver, CO: University of Denver.Hazel, S., & Schumaker, J. (1988). Social skills deficits.In J. Kavanagh & T. Truss (Eds.), Learning disabil-ities: Proceedings of the national conference (pp.293-366). Parkton, MD: York Press.

    REFERENCES Hoge, R. D., & Lute, S. (1979). Predicting academicAchenbach, T., & Edelbrock, C. (1983). Manual for theChild Behavior Checklist and Revised Child Be-havior Profile. Burlington: Universi ty of Vermont,Department of Psychiatry .Achenbach, T. M., McConaughy, S. T., & Howell, C. T.(1987). Child/adolescent behavioral and emotionalproblems: Implications of cross-informant correla-tions for situational specificity . Psychological Bul-letin, 101, 213-232.American Psychological Association. (1985). Stan-dards for educational and psychological testing.Washington, DC: Author.Brown, L. J., Black, D. D., & Downs, J. C. (1984). SchoolSocial Skills Rating Scale. New York: Slosson Edu-cational Publications.Chewning, T. G. (1992). An investigation of the dis-criminant and concurrent validity of the SocialSkills Rating System-Teacher form. DissertationAbstracts International, 53, 1843A.Coie, J. D., & Dodge, K. A. (1983). Continuities andchanges in childrens social status: A five-year lon-gitudinal study. Merrill-Palmer Quarterly, 29,261-282.Conners, C. K. (1990). Manual for the Canners Rating

    ScaZes. Toronto: Multi-Health Systems.Drabman, R. S. (1985). [Review o f Social Behavior As-sessment.] In J. V Mitchel (Ed.), T e ninth mentalmeasurements yearbook (p. 1410). Lincoln, NE:Buros Institute of Mental Measurements

    achievement from classroom behavior.Educational Research, 49,479496.

    ReVieW ofHumphrey, L. L. (1982). Childrens and teachers per-spectives on childrens self-control: The develop-ment of two rating scales. Journal of Consultingand Clinical Psychology, 50, 624-633.Matson, J ., Rotatori, A. F., & Helsel, W. J. (1983). De-velopment of a rating scale to measure social skillsin children: The Matson Evaluation of Social Skillsin Youngsters (MESSY). Behavioral Research andTherapy, 21,335-340.McCloskey, G. (1990). Selecting and using early child-hood rating scales. Topics in Early Childhood Spe-

    cial Education, 10, 39-64.McConaughy, S. H. (1993). Advances in empiricallybased assessment of childrens behavioral andemotional problems. School Psychology Review,Z&285-307.McKinney, J. D., & Speece, D. C. (1983). Classroom be-havior ad the academic progress of learning dis-abled students. Journal of Applied DevelopmentalPsychoLogy, 4, 149-161.Merrell, K. W. (1989). Concurrent relationships be-tween two behavioral rating scales for teachers: Anexamination of self-control, social competence,

    and school behavioral adjustment. Psychology inthe Schools, 26(3), 267-271.Merrell, K. W. (1993). School Social Behavior Scales.Bradon, VT: Clinical Psychology Publishing Com-

    Pany.

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    23/24

    670 School Psychology Review, 1995, Vol. 24, No.4Merrell, K. W., Johnson, E. R., Merz, J. M., &Ring, E. N.(1992). Social competence of students with mildhandicaps and low achievement: A comparativestudy. School Psychology Review, 21(l), 125-137.Nelson, R. O., & Hayes, S. C. (1979). Some current di-mensions of behavi0ra.I assessment. Behavioral As-sessment, 1, 1-16.Parker, J. G., & Asher, S. R. (1987). Peer relations andlater personal adjustment. Are low-accepted chiI-dren at risk? Psychological Bulletin, 102, 357-389.Piers, E. V. (1984). Piers-Harris Childrens Self-Con-cept Scale (revised manual>. Los Angeles, CA: West-ern Psychological Services.Stephens, T. (1992). Social skiUs in the classroom.Columbus, OH: Cedars Press.Stephens, T. (1981). Technical manual: Social Behav-ior Assessment. Columbus, OH: Cedars Press.Stephens, T. M., & Arnold, K. D. (1992). Social Behuv-ior Assessment Inventory: Professional manual.Odessa, FL: Psychological Assessment Resources,Inc.Stinnett, T. A., Oehler-Stinnett, J., & Stout, L. J. (1989).Ability of the Social SkiUs Rating System-Teacherversion to discriminate behavior disordered, emo-tionally disturbed, and nonhandicapped students.School Psychology Review, 18(4), 526-535.Stumme, I?, Gresharn, F. M., & Scott, N. (1982). Validityof Social Behavior Assessment in discriminatingemotionally disabled from nonhandicapped stu-dents. Journal of Behavioral Assessment, 4,327-342.

    Sweetland, R. C., & Keyser, D. J. (Eds.).(1991). Tests: Acomprehensive reference for assessments in psy-chology, education, and business (3rd ed.). Austin,TX: Pro-Ed.Waksman, S. A. (1985). Waksman Social SkillsRating Scale. Portland, OR: ASIEP Education.Waksman, S. A. (1980). The portland problem behaviorchecklist preliminary manual. Portland, OR: En-richment Press.Walker, H. (1983). Walker problem behavior ickmtifi-cation checklist. Los Angeles: Western Psychologi-cal Services.Walker, H. M., & McConnell, S. R. (1988). The Walker-McConnell Scale ofSocial Competence and SchoolAdjustment. Austin, TX: Pro-Ed.Walker, H., & Severson, H. (1992). Systematic screen-ing for behavior disorders. Longmont, CO: SoprisWest, Inc.Walker, H. M., Stieber, S., & Eisert, D. (1991). Teacherratings of adolescent social skills: Psychometriccharacteristics and factorial replicabUy acrossage-grade ranges. School Psychology Review,20(2), 301-314.

  • 8/12/2019 1995 Social Skills Assessment a Comparative Evaluation of Six Published Rating Scales

    24/24

    Social Skills Assessment 671

    Michelle Kilpatrick Demaray, MS, is a doctoral student in the School Psy-~ chology Program at the Universi ty of Wisconsin-Madison. She is currentlyon internship at Meyer Rehabilitation Institute at the University of Ne-~ braska Medical Center. Her interests include Attention Deficit Hyperactiv-ity Disorder and assessment issues.

    Stacey L. Ruffalo, MS, is a doctoral student in the School Psychology Pro-gram at the University of Wisconsin-Madison. Her research interests in-clude performance-based assessment and rating scale development.John Carlson, MS, is a doctoral student in the School Psychology Programat the University of Wisconsin-Madison. He is on internship at PrimaryChildrens Hospital in Salt Lake City, Utah. His research interests include in-ternalizing and externalizing disorders of childhood.R. T. Busse, MS, is a doctoral student in the School Psychology Program atthe Universi ty of Wisconsin-Madison and the senior project associate on aconsultation research project. Research interests include the effectivenessof consultation, verbal interactions, and consultee/client readiness forchange.Amy E. Olson, MS, received her degree f rom the School Psychology Pro-gram at the Universi ty of Wisconsin-Madison. She is currently working as aschool psychologist in Milwaukee, WI. Her interests include working withbilingual children.Susan M. McManus, MS, received her degree from the School PsychologyProgram at the University of Wisconsin-Madison. She is currently workingas a school psychologist for the West Aurora School District in Illinois. Herinterests include development of intervention assistance teams and con-sultation.Amy Leventhal, MS, is a doctoral student in the School Psychology Pro-gram at the Universi ty of Wisconsin-Madison. Her research interests are inadolescent peer relations.


Recommended