• autism;
  • PDD;
  • psychometrics;
  • ADI-R;
  • ADOS;
  • SCQ;
  • TCI;
  • CBCL


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Social Responsiveness Scale (SRS) is a quantitative measure of autistic traits in 4- to 18-year-olds, which has been used in behavior-genetic, epidemiological and intervention studies. The US standardization demonstrated a single-factor structure and good to excellent psychometric properties. The cross-cultural validity of the German adaptation of the parent-report SRS in a sample of N=1,436 children and adolescents: 838 typically developing and 527 clinical participants (160 with autism spectrum disorders (ASDs)) was examined. Internal consistency (0.91–0.97), test–retest reliability (0.84–0.97), interrater reliability (0.76 and 0.95) and convergent validity with the Autism Diagnostic Observation Schedule as well as the Autism Diagnostic Interview—Revised and Social Communication Questionnaire (0.35–0.58) were satisfactory to good. The SRS total score discriminated between ASD and other mental disorders. SRS scores proved to be sufficiently independent of general psychopathology. Principal component analyses yielded single-factor solutions for the normative and clinical subsamples. In addition, construct validity was ensured by consistent correlations with the Vineland Adaptive Behavior Scales, the Child Behavior Checklist and the Junior Temperament and Character Inventory. Normative SRS total scores for girls and boys as well as values for ASD were lower in the German sample, while scores for conduct disorder and attention deficit hyperactivity/conduct disorder combined were higher. Generally, cross-cultural validity of the SRS seems to be sufficiently assured for a large European sample. However, some discrepancies regarding SRS normative and clinical raw score distributions, reliability and validity findings are critically discussed.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The autism spectrum disorders (ASDs) encompass autistic disorder, Asperger disorder and atypical autism/pervasive developmental disorder—not otherwise specified (PDD-NOS), conditions which are categorically defined in the International Classification of Diseases—Tenth Revision (ICD-10) and the Diagnostic and Statistical Manual of Mental Disorders—Fourth Edition, Text Revision (DSM-IV-TR) as pervasive developmental disorders. Recent research suggests that ASDs represent the severe end of a quantitative distribution of impairments that occur in nature and extend into normality. For instance, milder phenotypes of ASD can be observed in relatives of individuals with ASD [Bailey, Palferman, Heavey, & Le Couteur, 1998; Bölte, Knecht, & Poustka, 2007; Yirmiya & Shaked, 2005]. Moreover, a growing number of studies imply a continuous distribution of autistic traits in the general population [Constantino, Przybeck, Friesen, & Todd, 2000; Constantino & Todd, 2000, 2003; Skuse, Mandy, & Scourfield, 2005; Spiker, Lotspeich, Dimiceli, Myers, & Risch, 2002].

The Social Responsiveness Scale [SRS; Constantino & Gruber, 2005] is the first widely used quantitative parent/teacher-report measure of autistic traits for use in the general population, and in educational and clinical settings. It generates an index of deficiency in social reciprocity, rather than providing an “all-or-nothing” characterization about the presence of symptoms or a given disorder. The objective of the present article was to examine the cross-cultural validity of the parent-report SRS. This would add to the growing literature on the validity of the SRS measurement system and the larger concept of autism as a quantitative trait, when examined outside of the US samples. In this report, the properties of the German adaptation of the instrument were analyzed in a large normative and clinical sample. A summary of published research on the psychometric properties of the SRS original as applied in genetic, epidemiological and intervention research is provided in Table I.

Table I. Reliability and Validity Studies on the SRS
Author(s) (yr)AreaFindings
  1. a

    ADHD, attention deficit hyperactivity disorder; ADI-R, Autism Diagnostic Interview—Revised; AUC, area under the curve; CBCL, Child Behavior Checklist; DD, unspecific developmental disorders; PDD-NOS, pervasive developmental disorders—not otherwise specified; ROC, receiver operating characteristics; VABS, Vineland Adaptive Behavior Scales; SRS, Social Responsiveness Scale.

Constantino et al. [2000]Scores in PDD-NOS, mood disorders, ADHD, conduct disorder and psychosis101.5 (SD=23.6), 59.4 (SD=30.1), 51.1 (SD=32.9), 48.4 (SD=18.7), 40.3 (SD=8.4)
 Discriminant validityHigher scores in PDD-NOS (F=11.7, P<0.000001)
 Latent class analysis in school childrenBest fit for three classes, with typical pattern for one underlying continuously distributed variable
 Factor analysis in school childrenOne-factor solution (explained variance: 70%)
Constantino and Todd [2000]Phenotypic concordance in twinsMonozygotic=0.73
Constantino and Todd [2003]Score distribution in the general populationNormally distributed, 1.4% of boys and 0.3% of girls above PDD-NOS cut-off
 Structural equation modelingSeventy-six percent genetic effects on scores
  No evidence for sex-specific genetic effects
Constantino, Davis et al. [2003]Correlation with ADI-R domain scores0.52–0.70
 Scores in autism/Asperger's117.4 (SD=29.9)/111.0 (SD=27.4)
Constantino, Hudziak, and Todd [2003]Relation between scores and general psychopathology (CBCL)CBCL accounts for 43% of the variance (above all syndrome scales: attention and social problems), 44% variance independent of CBCL
Constantino et al. [2004]Factor analysis in clinical sampleOne-factor solution (explained variance: 34.9%)
 Cluster analysisTen clusters (with two main clusters), high correlations between neighboring clusters
 Scores in DD and mixed child psychiatric diagnoses79.8 (SD=33.9)
  Girls: 45.17 (SD=33.7), boys: 34.6 (SD=23.0)
Constantino and Gruber [2005]Parent rating raw score means in general populationFemales: 27.6 (SD=18.1), males: 33.7 (SD=20.9)
 Correlation of scores with age0.02
 Interrater reliability0.91 (mother vs. father)
  0.82 (mother vs. teacher), 0.75 (father vs. teacher)
 Internal consistency0.93 (females), 0.94 (males),
  0.97 (clinical sample)
 Retest reliability0.77 (females), 0.85 (males)
 ROC analysis (autism spectrum vs. other clinical samples)AUC: 0.85, score of 85 has sensitivity of 0.70 and specificity of 0.90
Ho, Todd, and Constantino [2005]Scores in twins vs. non-twinsMale twins score higher than non-twins
Constantino and Todd [2005]Intergenerational transmissionCorrelation of scores between parents and children and between spouses around 0.4; scores in offspring elevated (effect size 1.5), if both parents score in the upper quartile; assortative mating explained ∼30% of parent score variance
Constantino et al. [2006]Scores in siblings of subjects with PDDAggregation of autistic traits in unaffected siblings, higher than that in siblings of subjects with psychopathology unrelated to autism
Pine, Luby, Abbacchi, and Constantino [2006]Evaluation of version for 3-yr-oldsInterrater reliability and retest reliability: 0.75
  Correlation with VABS: −0.86
  Correlation with ADI-R: 0.63
  Suitable for assessing treatment response
Tse, Strulovitch, Tagalakis, Meng, and Fombonne [2007]Usage as outcome measure in social skills trainingSensitive to treatment effects
Duvall et al. [2007]Genome-wide scan using SRS scoresLoci on chromosomes 11 and 17
Reiersen, Constantino, Volk, and Todd [2007]Scores in ADHDElevations of scores in children with ADHD
Constantino et al. [2007]Comparison of teacher ratings with parent ratings and expert ratingsTeacher/parent correlation: 0.72; teacher/ADI-R: 0.15–0.43; teacher/ADOS: 0.26–0.40
Reiersen, Constantino, and Todd [2008]Co-occurrence of motor problems and autism traits in ADHDBoys (not girls) with ADHD and motor problems show higher autism traits than those with ADHD alone
Virkud, Todd, Abbacchi, Zhang, and Constantino [2008]SRS scores in simplex and muliplex autism familiesIndication of different intergenerational transmissions of autistic traits in simplex and multiplex families
Constantino et al. [2008]Course of autistic traits in malesTest-retest correlation of 0.90 (1–5-yr interval)


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References


The standardization of the German SRS comprised N=1,436 individuals, n=838 normative and n=527 clinical children and adolescents. Of the latter, 160 had an ASD and 367 other psychiatric diagnoses (see Table II).

Table II. Sample Characteristics: Age and IQ by Sex
  1. a

    ASD, autism spectrum disorders; CLIN, other psychiatric diagnoses; NOR, typically developed controls.

Sex (m/f)12931252115442396
Age: M (SD)9.3 (3.8)9.6 (4.4)10.3 (3.4)11.1 (3.9)9.8 (3.4)9.9 (3.4)
IQ: M (SD)90.8 (27.0)86.1 (15.2)98.9 (16.1)99.8 (14.9)  

The normative subsample was recruited in the years 2005/2006 in 12 preschools (n=130), five primary (n=377) and one secondary school (n=331) in the city of Offenbach/M., federal state of Hesse. The sample included 396 girls and 442 boys with a mean age of 9.9 years (SD=3.4). There were 653 participants with German citizenship, 32 with double, 27 with Turkish, 9 with Greek and 8 with Italian nationality. The remaining 109 participants had 31 other nationalities. Maternal-report SRS was collected from 760 and paternal SRS report from 478 probands in the normative sample. In 427 cases the SRS report was available from both parents.

The clinical subsample consisted of 146 girls and 381 boys with a mean age of 10.3 years (SD=3.8). They were collected in the years 2005/2006 within the clinical routine at the Department of Child and Adolescent Psychiatry, Frankfurt/M. University Hospital. In addition, participants with ASD were ascertained within a long-standing autism research project at the same department as well as at the autism center of the Saint Mary Hospital in the city of Dresden, federal state of Saxony. Maternal SRS reports were available from 426 participants and paternal SRS report from 189. Ratings from both parents were collected in 152 cases.

Psychiatric diagnoses were based on ICD-10 research criteria. They were consensus diagnoses of the department's clinicians. The consensus diagnosis was reached by at least two of four experienced child psychiatrists or clinical psychologists and based on all available clinical evidence, including standardized diagnostic scales. According to this procedure, there were 160 children with ASD: 105 had autism (F84.0), 27 atypical autism (F84.1), 18 Asperger's syndrome (F84.5) and 10 PDD-NOS (F84.9). Thirty-one were female and 129 male with a mean age of 9.2 years (SD=3.6). The mean IQ for the ASD sample was computed from available data on a variety of cognitive tests (Wechsler Intelligence Scales for Children and Adults—Third Edition, Wechsler Adult Intelligence Scale—Revised, Raven's Progressive Matrices, Kaufman Assessment Battery for Children, Culture Fair Test); full scale IQ ranged from 35 to 134 with a mean of 89.8 (SD=25.1).

Furthermore, the clinical sample included 134 participants with attention deficit hyperactivity disorder (ADHD) (F90.0), 64 with neurotic or emotional disorders of childhood (F4 and F93), 55 with combined ADHD and conduct disorders (F90.1), 39 with conduct disorder (F91), 13 with circumscribed developmental disorders (F81), 12 with eating disorders (F50), 11 with personality disorders (F60), 11 with affective disorders (F3) and 28 with other disorders (enuresis, pica, selective mutism, attachment disorders, tics, mental retardation and schizophrenia). They had a mean age of 10.4 years (SD=3.8), consisted of 115 girls and 252 boys and their mean IQ was 99.3 (SD=15.6).

In addition to the normative and clinical samples, maternal-report SRS data were also available from 33 biological siblings of individuals with ASD. This sample included 19 girls and 14 boys with a mean age of 10.8 years (SD=4.3).


The SRS [Constantino & Gruber, 2005] is a quantitative measure of autistic traits in 4- to 18-year-olds. It comprises 65 items using a “0” (not true) to “3” (almost always true) point Likert scale generating one total score (max. 195). The questions focus on the child's behavior during the past 6 months and can be completed in 15–20 min. The SRS can be used both as a screener and as an aid to clinical diagnosis of ASD, particularly less severe forms like PDD-NOS. Total raw scores can be transformed into T-scores in order to provide the relative normative position of any proband regarding autistic traits. An adult and a preschool version of the SRS have also been developed [Constantino, 2002; Pine et al., 2006]. For the German adaptation, the SRS was translated, retranslated and authorized.

The combined usage of the Autism Diagnostic Interview—Revised [ADI-R; Rutter et al., 2003] and the Autism Diagnostic Observation Schedule [ADOS; Lord, Rutter, DiLavore, & Risi, 2001] is viewed as a “gold standard” for diagnosing ASD. Both are thoroughly validated scales organized around ICD-10/DSM-IV-TR criteria and provide empirically derived diagnostic algorithms for autism. The ADI-R and ADOS algorithms generate scores for each of the three subdomains of autism: social interaction, communication and stereotyped behaviors—very recently, ADOS algorithms were proposed to integrate the scores for social and communicative deficits, in keeping with the results of factor analysis in large clinical samples. To meet ADI-R criteria for autism, a child must reach the cut-off in each domain and exhibit signs of abnormality prior to the third birthday. In the ADOS cut-offs must be met for the social and communicative domain and a combined cut-off for the sum of social and communication scores. There are two cut-offs: autism and autism spectrum. The latter distinction is only made in the ADOS; the ADI-R classifies as autistic or not autistic. The Social Communication Questionnaire [SCQ; Rutter, Bailey, & Lord, 2001] is a 40-item screener for autism providing a total score and cut-offs for autism and ASD. It has demonstrated good psychometric properties [Berument, Rutter, Lord, Pickles, & Bailey, 1999]. In this study, the German versions of the ADI-R [Bölte, Rühl, Schmötzer, & Poustka, 2006], ADOS (modules 1–4) [Rühl, Bölte, Feineis-Matthews, & Poustka, 2004] and SCQ [Bölte et al., 2000; Bölte & Poustka, 2006] were used, which have demonstrated good reliability and validity [Bölte, Holtmann, & Poustka, 2008; Bölte & Poustka, 2004; Poustka et al., 1996]. For the ADI-R, we expected correlations comparable to those reported for the original SRS. As the SCQ is derived from the ADI-R and both SRS and SCQ are parent-report questionnaires, we expected somewhat higher correlations here than those between the ADOS and the SRS. The ADOS is an expert rating based on a standardized clinical observation at a single point in time, while ADI-R, SCQ and SRS items are coded for observations over broader periods of time and in naturalistic contexts.

The Vineland Adaptive Behavior Scales [VABS; Sparrow, Balla, & Cicchetti, 1984] is an interview to register the functional behavior in four domains: motor (in children aged <6 years), socialization, communication and daily living. A summary score (adaptive behavior composite (ABC)) expressing the global adaptive functioning derived from the results of the subscales can be generated. In the present study we used the German screening version of the VABS [Bölte & Poustka, 2002], which comprises a selection of 60 items (15 for each domain) of the entire form of 261 items. There is high convergence of the screeners and the full VABS with correlations between r=0.87 and 0.98. Raw scores collected with the screener can be converted to normative values. In accord with the previous research using the VABS in autism [Carter et al., 1998], we expected a negative correlation between the SRS and the VABS domain and ABC scores.

The Child Behavior Checklist [CBCL; Achenbach, 1991] is a parent questionnaire to screen for psychopathology in 4- to 18-year-olds. The behavior problem scale contains 118 items. Of these, 85 constitute 8 syndrome scales: withdrawn, somatic complaints, anxious/depressed, social problems, thought problems, attention problems, delinquent behavior and aggressive behavior. In this study we used the German form of the CBCL, whose norms and psychometrics are highly comparable to the US original [Döpfner, Schmeck, & Berner, 1994]. The CBCL has demonstrated usefulness in identifying individuals with ASD. Syndrome scales of the CBCL have been reported to be significantly associated with autism [Bölte, Dickhut, & Poustka, 1999; Duarte, Bordin, de Oliveira, & Bird, 2003], particularly attention problems, social problems, thought problems and withdrawn behavior. We expected intercorrelations between the SRS and the CBCL consistent with these previous findings.

The Junior Temperament and Character Inventory [JTCI; Luby, Svrakic, McCallum, Przybeck, & Cloninger, 1999] is a test to assess basic personality dimensions in 12- to 18-year-olds. The temperament dimensions are novelty seeking, harm avoidance, reward dependence and persistence. The three character dimensions are self-directedness, cooperativeness and self-transcendence. In this study we used the German adaptation of the JTCI [Goth, Cloninger, & Schmeck, 2004]. In two independent studies with adult samples, autism has been demonstrated to negatively correlate with novelty seeking and reward dependence and to positively correlate with harm avoidance [Anckarsäter et al., 2006; Kunihira, Senju, Dairoku, Wakabayashi, & Hasegawa, 2006]. Therefore, we postulated a similar pattern of association in the current study. Recent developmental studies of autism have suggested that core components of the disorder may arise from a relative failure of the integration of social circuitry and reward circuitry early in life, and for this reason we were particularly interested in the association between quantitative autistic traits and reward dependence.

Data Analysis

Reliability was assessed by computing internal consistency (Cronbach's α) of all 65 items as well as test–retest reliability and interrater reliability of the SRS total score using intraclass correlation (ICC) and Pearson product moment (PPM) correlation in the normative and clinical samples.

To address other aspects of validity, a number of procedures were performed [Cronbach & Meehl, 1955]: Convergent validity (correlation with theoretically very similar measures) was analyzed in the ASD sample by correlating the SRS with the ADI-R and ADOS domain scores and the total score of the SCQ. Here, correlations were corrected for range restrictions applying the case II procedure by Thorndike [1949], because ADI-R, ADOS and SCQ are categorical scales, not intended to generate variability. Discriminant validity (value for diagnostic classification: ASD vs. other mental disorders) was analyzed by t-test for independent samples, a discriminant analysis and a receiver operating characteristics (ROC) analysis in the clinical sample. In addition, item validities (discriminant power of single items) for the differentiation of ASD and other mental disorders were computed using t-tests in the clinical sample. Factorial validity was examined in the normative and clinical samples using exploratory principal component analyses, varimax rotation and factor extraction by the scree criterion. To explore external/concurrent validity (correlation with theoretically related measures), the SRS total score was correlated with the VABS-ABC score (standard score) in the ASD sample and the syndrome scales of the CBCL (T-scores) and the JTCI subscales (T-scores) in the clinical sample. Structural validity (behavior-genetic consistency) was examined using two approaches: first, in the clinical sample, by investigating the extent to which CBCL syndrome scales and SCQ scores contributed to variation in SRS scores (using a stepwise multiple linear regression analysis); second, by comparing SRS scores of siblings of children with ASD to normative values. For between-group comparisons effect sizes (Cohen's d) are provided.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

SRS Raw Scores

SRS total score means in the normative and clinical samples are given in Table III. There were no noteworthy correlations between age and SRS total scores in the normative or clinical sample (r=−0.06 and 0.00). In addition, correlations with IQ were low to negligible in clinically ascertained subjects with (r=−0.18) and without (r=−0.04) ASD. In the normative sample, the mean SRS total score differed moderately (d=0.26) but significantly (t=−5.9, P<0.0001) between mother and father ratings with a correlation between the ratings of r=0.76 (ICC). Therefore, mother and father ratings were analyzed separately here and separate norms have been generated for the German version for SRS use in practice [Bölte & Poustka, 2008]. Mean mother ratings were 25.3 (SD=16.7) for boys and 22.8 (SD=14.6) for girls. Father ratings were 27.4 (SD=14.9) for boys and 25.8 (SD=15.9) for girls. For mother ratings the differences in scoring the sexes were significant (d=0.16, t=2.2, P=0.03). Sex differences are reported here for the sake of establishing norms, but subsequent analyses on the instrument's psychometric properties were not conducted separately for boys and girls, since the gender differences were relatively modest.

Table III. SRS Total Raw Sore Distributions in the Normative and Clinical German Samples
 NMean (SD)
  1. a

    SRS, Social Responsiveness Scale; ASD, autism spectrum disorders; ADHD, attention deficit hyperactivity disorder; PDD-NOS, pervasive developmental disorders—not otherwise specified.

 Mother ratings  
  Females35322.8 (14.5)
  Males40725.3 (16.7)
 Father ratings  
  Females21225.8 (15.9)
  Males26627.2 (14.9)
 ASD160102.3 (31.8)
  Autism105107.3 (30.2)
  Asperger syndrome18100.2 (31.1)
  Atypical autism2789.7 (36.5)
  PDD-NOS1086.5 (26.9)
 ADHD13457.1 (27.6)
 Neurotic/emotional disorder6460.8 (29.8)
 ADHD and conduct disorder5574.0 (27.8)
 Conduct disorder3969.3 (25.7)
 Developmental disorders1339.3 (20.0)
 Eating disorders1246.9 (25.9)
 Personality disorders1164.7 (16.1)
 Affective disorders1152.1 (23.7)
 Other2861.9 (34.7)

In the clinical sample, parent ratings (mother vs. father) did not differ (d=0.06, t=1.1, P=0.34) and were highly correlated. Therefore, parental SRS ratings were aggregated in the clinical sample. If SRS ratings for a proband were available from both mother and father, mother ratings were used. Using this method, 426 SRS from mothers and 101 SRS from fathers were included for psychometric analyses. The mean SRS score was 102.3 (SD=31.8) in the ASD subsample and 61.0 (SD=28.7) in the sample of non-ASD clinical participants.


Reliability data are summarized in Table IV. Internal consistency in the normative sample was rtt=0.93 for mother ratings and rtt=0.91 for father ratings. In the clinical sample, Cronbach's α was rtt=0.97 for the combined parent ratings. Test–retest reliability (3 weeks to 4 months) in the normative sample was rtt=0.88 (ICC) and rtt=0.80 (PPM) in n=107 mothers and rtt=0.84 and rtt=0.72 in n=76 fathers. In the clinical sample, 49 tests and retests were collected (interval: 3–6 months). Here, retest reliability reached rtt=0.97 and rtt=0.95. Interrater reliability in the clinical sample between mother and father ratings in n=172 reached rtt=0.95 and rtt=0.91. All correlations reported here were significant on an α level of P<0.0001.

Table IV. Reliability Findings for the German Version of the SRS
  1. a

    SRS, Social Responsiveness Scale; ICC, intraclass correlation.

Internal consistency (Cronbach's α) 
 Normative sample 
  Mother ratings0.93
  Father ratings0.91
 Clinical sample0.97
Test–retest reliability (ICC/PPM) 
 Normative sample 
  Mother ratings (n=107)0.88/0.80
  Father ratings (n=76)0.84/0.72
 Clinical sample (n=49)0.97/0.95
Interrater reliability 
 Normative sample 
  Mother–father (n=427)0.76/0.61
 Clinical sample 
  Mother–Father (n=172)0.97/0.91


Findings for validity are given in Table V.

Table V. Validity Findings for the German Version of the SRS
  1. a

    SRS, Social Responsiveness Scale; ASD, autism spectrum disorder; ROC, receiver operating characteristics; CBCL, Child Behavior Checklist.

Convergent validity 
 Autism Diagnostic Interview—Revised (n=113) 
  Social interaction0.46
  Stereotypic behavior0.38
 Social Communication Questionnaire (n=107)0.58
 Autism Diagnostic Observation Schedule (n=119) 
  Communication/social deficits0.35
Concurrent/external validity 
 Vineland Adaptive Behavior Scales (n=93) 
  Adaptive behavior composite (ABC)−0.36
  Daily living−0.35
  Motor skills (n=19, aged <6 yr)−0.34
 Child Behavior Checklist (n=301) 
  Social problems0.64
  Attention problems0.61
  Thought problems0.48
  Aggressive behavior0.45
  Somatic complaints0.11
 Temperament and Character Inventory (n=49) 
  Novelty seeking0.38
  Harm avoidance0.33
  Reward dependence−0.50
Discriminant validity 
 Sixty four SRS items discriminate ASD from other clinical conditions (t>1.8, P<0.04, d>0.19)
 SRS total score discriminates ASD from other clinical conditions (t=14.1, P<0.001, d=1.31)
 Classification accuracy according to discriminant analysis: 89.4% (Wilk's λ=0.37, χ2=263.6, P<0.0001)
 ROC curves show 83% area under the curve for SRS total score as predictor for ASD vs. other clinical conditions; sensitivity of 0.73 and aspecificity of 0.81 for any ASD
Factorial validity 
 One-factor solution for normative mother ratings (explained variance: 17.9%, eigenvalue 11.6). Sixty of the SRS items had loadings on the firstfactor of r>0.30
 One-factor solution for normative father ratings (explained variance: 16.5%, eigenvalue 10.6). Sixty-one of the SRS items had loadings onfirst factor of r>0.30
 One-factor solution for clinical ratings (explained variance: 34.9%, eigenvalue 22.7). Except for 1 item (no. 43), all other 64 items had factor loadings between r=0.38 and 0.75 on the first factor
Structural validity
 Multiple linear regression model shows SRS data to be sufficientlyindependent of general psychopathology as measured by CBCL syndromes (R2=54%)
 Mean SRS total scores increased in brothers of individuals with ASD (50% in the mild to moderate clinical SRS range, T>60)
Convergent validity

All correlations with established autism scales were positive, significant (P<0.01) and moderate to good. Correlation of the SRS with the SCQ was r=0.58 in n=107 probands. In 119 cases data from the SRS and ADOS were available; the association between SRS total score and ADOS social score was r=0.35. Convergent validity of the SRS with the ADI-R domains in n=113 individuals was r=0.46 for social interaction, r=0.40 for communication and r=0.38 for stereotyped behavior.

Discriminant validity

Except for one item (no. 43, “Separates easily from caregivers”), all remaining SRS items discriminated between ASD and the other psychiatric participants (d>0.19, t>1.8, P<0.04). The SRS total score discriminated significantly between ASD and other mental disorders (d=1.36, t=14.6, P<0.0001). Discriminant analysis revealed that a linear combination of the SRS items reaches a classification accuracy (ASD vs. other mental disorders) of 89.4% (Wilk's λ=0.37, χ2=263.6, P<0.0001). ROC curves showed that the SRS total score as a predictor for the classification to the groups had a performance of 83% area under the curve (AUC). For the classification of core autism vs. other mental disorders, the AUC was slightly larger (88%). An SRS total score of 85 was associated with a sensitivity of 0.73% and a specificity of 0.81% for any ASD. An SRS score of 56 had a sensitivity of 90% for any ASD (specificity 50%). An SRS score of 100 had a specificity 90% (sensitivity 56%) for any ASD.

Factorial validity

All three principal component analyses showed single-factor solutions. In the clinical sample, the first factor explained 34.9% of the total variance with an eigenvalue of 22.7 and all remaining factors having eigenvalues <3.2. Except for 1 item (no. 43), all other 64 items had factor loadings between r=0.38 and 0.75 on the first factor (Table VI).

Table VI. Factor Loadings of SRS Items on the First Factor in the Normative and Clinical Samples
 Normative sample 
ItemMother—SRSFather—SRSClinical sample
  1. SRS, Social Responsiveness Scale.


In the normative sample, the first extracted factor explained 17.9% of the mother SRS data variance with an eigenvalue of 11.6. All other factors had eigenvalues <3.3 and all 65 SRS items correlated positively with the first factor (r=0.19–0.54). Sixty of the SRS items had loadings on the first factor of r>0.30 (Table VI). In the father SRS data set, the first factor accounted for 16.5% of the total variance, with an eigenvalue of 10.6 and factor loadings between r=0.07 and 0.55. Sixty-one of the SRS items had loadings on the first factor of r>30 (Table VI). The subsequent factors had eigenvalues <3.5.

External/concurrent and structural validity

For n=93 individuals with ASD, SRS and VABS data were available. The correlation between the total SRS and the ABC of the VABS was r=−0.36. The correlation with the single VABS domains was r=−0.43 for communication, r=−0.41 for socialization, r=−0.35 for daily living and r=−34 for motor skills (n=19, only in subjects aged >6 years). In the clinical sample, CBCL and SRS data were collected in 301 probands. All intercorrelations between the SRS and the CBCL syndrome scales were significant (P<0.0001) and ranged between r=0.11 for somatic complaints and r=0.64 for social problems. Multiple linear regression showed that the eight syndrome scales of the CBCL explained 54.5% of the variation in total SRS scores (F8,300=43.6, P<0.0001). In a smaller sample (n=87), a model also including the SCQ as predictor explained 82% of the SRS variance (F9,86=38.8, P<0.0001). In 49 cases of the clinical sample, concurrent validity of the SRS with JTCI was calculated. There were significant (P<0.05) negative correlations between the SRS total score and reward dependence (r=−50), self-directedness (r=−0.41) and cooperativeness (r=−0.39) and positive associations with novelty seeking (r=0.38) and harm avoidance (r=0.33).

Analysis of the SRS in the sample of siblings of children with ASD showed that the mean SRS total score was 28.6 (SD=25.7). There was a significant effect of sex (d=1.06, t=2.8, P<0.012). Male siblings (M=43.5, SD=32.5) scored significantly higher than female siblings (M=17.7, SD=10.7), and significantly higher than males in our normative sample. These mean scores corresponded to a T-score of 47 for the sisters and a T of 64 for the brothers. Half of the undiagnosed male siblings of ASD probands had scores in the mild to moderate clinical range of the SRS (T>60).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Generally, reliability and validity findings for the German adaptation were satisfactory to excellent and strikingly similar to the US data on the original version of the SRS. Internal consistency and retest reliability were fair to high in the normative and clinical samples. Although interrater reliability (mother vs. father) was extremely high in the clinical sample, the discrepancy between parental mean scores in the normative sample was large enough to warrant separate consideration of reports. Therefore, separate analyses for normative mother and father SRS were indicated as well as specific norms for individual assessment. One explanation for the current findings might be that parent's perception of internalizing symptoms differ more among mothers and fathers in the German than in the US general population. It has been reported that mothers are more sensitive observers of internalizing problems [Phares, 1997].

Factorial validity yielded essentially the same pattern of results as reported for the SRS original [Constantino et al., 2000, 2004; Constantino & Todd, 2003]: analyses were consistent with a single-factor structure in both the normative and clinical samples. All subsequent factors explained markedly less variance than the first, which is typical for a single dimension solution. Nevertheless, while in the clinical samples the primary factor accounted for an identical amount of variance (34.9%), in the German and US samples there was a discrepancy in the explained variance by the first factor in the German (16.5 and 17.9%) vs. American (70%) school samples which may have occurred on the basis of differences in the ages of the subjects in the respective samples and the fact that in the published American school samples, informants were teachers (rather than parents) who each provided ratings on more than one student (15 on average). In addition, one item in the clinical and four and five items, respectively, in the normative samples did not load sufficiently on the first extracted factor.

As shown in the US studies [Constantino & Gruber, 2005; Constantino et al., 2000], the SRS total score distinguished significantly between ASD and other psychiatric groups. In ROC analyses, a score of 85 discriminated best between ASD and other mental disorders. However, neither in the US-standardization sample nor in the German data set did this value allow sufficient sensitivity (0.70 and 0.73) specificity (0.90 and 0.81) for a solid clinical classification. Thus, use of separate cut-offs on the SRS for screening purposes (optimizing sensitivity) vs. clinical confirmation (optimizing specificity) are well worth considering, as discussed in the SRS Manual.

Comparable to the findings by Constantino, Davis et al. [2003] and Constantino, Hudziak et al. [2003] and consistent with previous research on CBCL scores in autism [Bölte et al., 1999], the CBCL syndromes accounted for only about half of the variance (54.5%) in SRS scores, with the scales social problems and attention problems contributing the most. Convergent validity with ADI-R and ADOS was robust, although lower than that reported for the original [Constantino, Davis et al., 2003; Constantino, Hudziak et al., 2003]. The latter likely arises from the fact that in the US sample, ADI-R and ADOS data were assessed across a broader range of autistic symptomatology, while in the German sample a substantially higher proportion of subjects (for whom ADI-R and ADOS data were available) carried full diagnoses of autistic disorder. Therefore, the variance was restricted to the more severe range of ADI-R/ADOS ratings, resulting in artificially lower estimates for convergent validity in the German sample.

The SRS has been used extensively for behavior-genetic autism research. In this study, we found substantially increased SRS scores in brothers of subjects with autism. This is consistent with findings of an excess of milder autism phenotypes in male first-degree relatives of autistic probands [Constantino et al., 2006; Virkud et al., 2008] and offers the first cross-cultural support for the presence of such patterns of familial aggregation, as measured using the SRS system.

There were moderate but noteworthy differences in the SRS raw score distributions between the German and the US samples. In the German normative sample, SRS total scores were about 0.25 SD lower for girls and 0.4 SD lower for boys than in the US-standardization sample. Also, sex differences were less pronounced in the normative German sample. In this regard, it could be particularly of importance that a substantial part of the US-standardization sample consisted of twins. Research on autism in general and autism traits on the SRS has shown that autism scores in twins, especially males, could be higher than those in the general population [Greenberg, Hodge, Sowinski, & Nicoll, 2001; Ho et al., 2005].

Just as raw SRS scores were lower in the normative German sample than in comparable US samples, mean SRS values for ASD syndromes were also somewhat lower in the German sample, indicating that it will be most appropriate for norms used in clinical and educational settings to be specific to the culture and language version of the SRS that is employed.

In contrast, SRS raw scores for some other psychiatric disorders reached higher values in the German sample than those expected by US findings. This is particularly true for conduct disorder and combined ADHD/conduct disorder. Here SRS scores were up to 1 SD higher than those reported in some US studies [Constantino et al., 2000, 2004], while scores for non-comorbid ADHD or mood disorders were comparable in the German and US standardization. Higher SRS scores for several mental disorders in the German sample might be attributable to the fact that the SRS was typically completed by parents at the child's first visit to the psychiatric department. Therefore, the symptoms were perhaps acute and there was more of a tendency for inflation of reports of social deficiency on the basis of the current level of functioning that led the family to seek treatment for the child. Also, diagnostic thresholds for some mental disorders (e.g. ADHD) are lower in DSM-IV-TR compared to ICD-10 (used in the German health-care system), leading to less severe cases being diagnosed in the US.

In summary, intercultural validity of the SRS as a parent-report measure of quantitative autistic impairment is supported in this large European sample. Further research involving teacher reports and continued exploration of the finding of a unitary factor structure for autistic traits appears warranted.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors thank Helge Sickmann, Daniel Löffler, Evelyn Herbrecht, MD, Gabriele Schmötzer, MD, Katja Albertowski, MD, Diana Schulze, Anna Seiverth, Kirstin Goth, PhD, Bernd Meyenburg, MD, and Olaf Lückfeld for their support in data collection. The authors also gratefully acknowledge the kindergartens and schools involved in this study (evangelische Kitas, Beethovenschule, Goetheschule, Waldschule, Buchhügelschule, Uhlandschule, Leibnizgymnasium), their representatives and all families who contributed.

The authors of this article disclose the following potential conflicts of interests: Sven Bölte and Fritz Poustka receive royalties for the German version of the Social Responsiveness Scale (“Skala zur Erfassung sozialer Reaktivität”) from Hans Huber Publishers. John N. Constantino receives royalties for the Social Responsiveness Scale from Western Psychological Services.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • Achenbach, T.M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 profile. Burlington, VT: Department of Psychiatry, University of Vermont.
  • Anckarsäter, H., Stahlberg, O., Larson, T., Håkansson, C., Jutblad, S.B., et al. (2006). The impact of ADHD and autism spectrum disorders on temperament, character, and personality development. American Journal of Psychiatry, 163, 12391344.
  • Bailey, A., Palferman, S., Heavey, L., & Le Couteur, A. (1998). Autism: The phenotype in relatives. Journal of Autism and Developmental Disorders, 28, 369392.
  • Berument, S.K., Rutter, M., Lord, C., Pickles, A., & Bailey, A. (1999). Autism Screening Questionnaire: Diagnostic validity. British Journal of Psychiatry, 175, 444451.
  • Bölte, S., Crecelius, K., & Poustka, F. (2000). Der Fragebogen über Verhalten und soziale Kommunikation: psychometrische Eigenschaften eines Autismus-Screening-Instruments für Forschung und Praxis. Diagnostica, 46, 149155.
  • Bölte, S., Dickhut, H., & Poustka, F. (1999). Patterns of parent-reported problems indicative in autism. Psychopathology, 32, 9498.
  • Bölte, S., Holtmann, M., & Poustka, F. (2008). The Social Communication Questionnaire (SCQ) as a screener for autism spectrum disorders: Additional evidence and cross-cultural validity. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 719720.
  • Bölte, S., Knecht, S., & Poustka, F. (2007). A case–control study of personality style and psychopathology in parents of subjects with autism. Journal of Autism and Developmental Disorders, 37, 243250.
  • Bölte, S., & Poustka, F. (2002). The relation between general cognitive level and adaptive behavior domains in individuals with autism with and without co-morbid mental retardation. Child Psychiatry & Human Development, 33, 165172.
  • Bölte, S., & Poustka, F. (2004). Diagnostische Beobachtungsskala für Autistische Störungen (ADOS): erste Ergebnisse zur Zuverlässigkeit und Gültigkeit. Zeitschrift für Kinder- und Jugendpsychiatrie und Psychotherapie, 32, 4550.
  • Bölte, S., & Poustka, F. (2006). Fragebogen zur sozialen kommunikation (FSK). Bern: Huber.
  • Bölte, S., & Poustka, F. (2008). Skala zur Erfassung sozialer Reaktivität (SRS). Bern: Huber.
  • Bölte, S., Rühl, D., Schmötzer, G., & Poustka, F. (2006). Diagnostisches Interview für Autismus—Revidiert (ADI-R). Bern: Huber.
  • Carter, A.S., Volkmar, F.R., Sparrow, S.S., Wang, J.J., Lord, C., et al. (1998). The Vineland Adaptive Behaviour Scales: Supplementary norms for individuals with autism. Journal of Autism and Developmental Disorders, 28, 287302.
  • Constantino, J.N. (2002). Social Responsiveness Scale–Adult Version (SRS). Los Angeles, CA: Western Psychological Services.
  • Constantino, J.N., Abbacchi, A.M., LaVesser, P.D., Reed, H., Givens, L., et al. (2008). Developmental course of autistic social impairment in males. Development and Psychopathology, in press.
  • Constantino, J.N., Davis, S.A., Todd, R.D., Schindler, M.K., Gross, M.M., et al. (2003). Validation of a brief quantitative measure of autistic traits: Comparison of the Social Responsiveness Scale with the Autism Diagnostic Interview—Revised. Journal of Autism and Developmental Disorders, 33, 427433.
  • Constantino, J.N., & Gruber, C.P. (2005). Social Responsiveness Scale (SRS). Los Angeles, CA: Western Psychological Services.
  • Constantino, J.N., Gruber, C.P., Davis, S., Hayes, S., Passanante, N., & Przybeck, T. (2004). The factor structure of autistic traits. Journal of Child Psychology and Psychiatry, 45, 719726.
  • Constantino, J.N., Hudziak, J.J., & Todd, R.D. (2003). Deficits in reciprocal social behavior in male twins: Evidence for a genetically independent domain of psychopathology. Journal of the American Academy of Child and Adolescent Psychiatry, 42, 458467.
  • Constantino, J.N., Lajonchere, C., Lutz, M., Gray, T., Abbacchi, A., et al. (2006). Autistic social impairment in the siblings of children with pervasive developmental disorders. American Journal of Psychiatry, 163, 294296.
  • Constantino, J.N., Lavesser, P.D., Zhang, Y., Abbacchi, A.M., Gray, T., & Todd, R.D. (2007). Rapid quantitative assessment of autistic social impairment by classroom teachers. Journal of the American Academy of Child and Adolescent Psychiatry, 46, 16681676.
  • Constantino, J.N., Przybeck, T., Friesen, D., & Todd, R.D. (2000). Reciprocal social behaviour in children with and without pervasive developmental disorders. Journal of Developmental and Behavioral Pediatrics, 21, 211.
  • Constantino, J.N., & Todd, R.D. (2000). Genetic structure of reciprocal social behavior. American Journal of Psychiatry, 157, 20432045.
  • Constantino, J.N., & Todd, R.D. (2003). Autistic traits in the general population: A twin study. Archives of General Psychiatry, 60, 524530.
  • Constantino, J.N., & Todd, R.D. (2005). Intergenerational transmission of subthreshold autistic traits in the general population. Biological Psychiatry, 57, 655660.
  • Cronbach, L.J., & Meehl, P.E. (1955). Construct validity in psychological tests. Psychological Bulletin, 52, 281302.
  • Döpfner, M., Schmeck, K., & Berner, W. (1994). Handbuch: Elternfragebogen über das Verhalten von Kindern und Jugendlichen. Forschungsergebnisse zur deutschen Fassung der Child Behavior Checklist (CBCL/4-18). Köln: Arbeitsgruppe Kinder-, Jugend- und Familiendiagnostik (KJFD).
  • Duarte, C.S., Bordin, I.A., de Oliveira, A., & Bird, H. (2003). The CBCL and the identification of children with autism and related conditions in Brazil: Pilot findings. Journal of Autism and Developmental Disorders, 33, 703707.
  • Duvall, J.A., Lu, A., Cantor, R.M., Todd, R.D., Constantino, J.N., & Geschwind, D.H. (2007). A quantitative trait locus analysis of social responsiveness in multiplex autism families. American Journal of Psychiatry, 164, 656662.
  • Goth, K., Cloninger, C.R., & Schmeck, K. (2004). Das Junior Temperament und Charakter Inventar für Jugendliche—JTCI/12-18 R. Frankfurt/M.: Klinik für Psychiatrie und Psychotherapie des Kindes- und Jugendalters.
  • Greenberg, D.A., Hodge, S.E., Sowinski, J., & Nicoll, D. (2001). Excess of twins among affected sibling pairs with autism: Implications for the etiology of autism. American Journal of Human Genetics, 69, 10621067.
  • Ho, A., Todd, R.D., & Constantino, J.N. (2005). Brief report: Autistic traits in twins vs. non-twins—a preliminary study. Journal of Autism and Developmental Disorders, 35, 129133.
  • Kunihira, Y., Senju, A., Dairoku, H., Wakabayashi, A., & Hasegawa, T. (2006). ‘Autistic’ traits in non-autistic Japanese populations: Relationships with personality traits and cognitive ability. Journal of Autism and Developmental Disorders, 36, 553566.
  • Lord, C., Rutter, M., DiLavore, P., & Risi, S. (2001). Autism Diagnostic Observation Schedule (ADOS). Los Angeles, CA: Western Psychological Services.
  • Luby, J.L., Svrakic, D.M., McCallum, K., Przybeck, T.R., & Cloninger C.R. (1999). The junior temperament and character inventory (JTCI): Preliminary validation of a child self report measure. Psychological Reports, 84, 11271138.
  • Phares, V. (1997). Accuracy of informants: Do parents think that mother knows best? Journal of Abnormal Child Psychology, 25, 165171.
  • Pine, E., Luby, J., Abbacchi, A., & Constantino, J.N. (2006). Quantitative assessment of autistic symptomatology in preschoolers. Autism, 10, 344352.
  • Poustka, F., Lisch, S., Rühl, D., Sacher, A., Schmötzer, G. & Werner, K. (1996). The standardized diagnosis of autism: Autism Diagnostic Interview—Revised: Inter-rater reliability of the German form of the ADI-R. Psychopathology, 29, 145153.
  • Reiersen, A.M., Constantino, J.N., & Todd, R.D. (2008). Co-occurrence of motor problems and autistic symptoms in attention-deficit/hyperactivity disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 662672.
  • Reiersen, A.M., Constantino, J.N., Volk, H.E., & Todd, R.D. (2007). Autistic traits in a population-based ADHD twin sample. Journal of Child Psychology and Psychiatry, 48, 464472.
  • Rühl, D., Bölte, S., Feineis-Matthews, S., & Poustka, F. (2004). Diagnostische Beobachtungsskala für Autistische Störungen (ADOS). Bern: Huber.
  • Rutter, M., Bailey, A., & Lord, C. (2001). Social Communication Questionnaire (SCQ). Los Angeles, CA: Western Psychological Services.
  • Rutter, M., LeCouteur, A., & Lord, C. (2003). Autism Diagnostic Interview—Revised (ADI-R). Los Angeles, CA: Western Psychological Services.
  • Skuse, D.H., Mandy, W.P., & Scourfield, J. (2005). Measuring autistic traits: Heritability, reliability and validity of the Social and Communication Disorders Checklist. British Journal of Psychiatry, 187, 568572.
  • Sparrow, S., Balla, D., & Cicchetti, D. (1984). Vineland Adaptive Behavior Scales. Circle Pines, MN: American Guidance Service.
  • Spiker, D., Lotspeich, L.J., Dimiceli, S., Myers, R.M., & Risch, N. (2002). Behavioral phenotypic variation in autism multiplex families: Evidence for a continuous severity gradient. American Journal of Medical Genetics, 114, 129136.
  • Thorndike, R.L. (1949). Personnel selection: Test and measurement techniques. New York: Wiley.
  • Tse, J., Strulovitch, J., Tagalakis, V., Meng, L., & Fombonne, E. (2007). Social skills training for adolescents with Asperger syndrome and high-functioning autism. Journal of Autism and Developmental Disorders, 37, 19601968.
  • Virkud, Y.V., Todd, R.D., Abbacchi, A.M., Zhang, Y., & Constantino, J.N. (2008). Familial aggregation of quantitative autistic traits in multiplex versus simplex autism. American Journal of Medical Genetics: B Neuropsychiatric Genetics, July 10 (Epub ahead of print).
  • Yirmiya, N., & Shaked, M. (2005). Psychiatric disorders in parents of children with autism: A meta-analysis. Journal of Child Psychology and Psychiatry, 46, 6983.