You are here
Home > Blog > General Topics > Spence Children’s Anxiety Scale Validation

Spence Children’s Anxiety Scale Validation

Validation of Spence Children’s Anxiety Scale


In Europe, anxiety disorders are highly prevalent and impairing conditions among children and adolescents. Up to 18 – 20% of youth in the general population presenting to primary care clinics report some internalizing problem. Previous studies showed that anxiety disorders start at an early age and social phobia is the earliest type of anxiety
disorder. Costello et al. found that 50% of anxiety disorders that manifested at 21 years of age had actually developed between the ages of 6 and 12. Also, a few studies confirm that girls tend to be more anxious than boys.

Cognitive-behavioral therapy and exposure-based treatment are the most effective anxiety interventions for children and adolescents. However, the percentage of children accessing mental health services is worrisome (only 50.6% in the U.S and 25.6% in European countries). Merikangas et al. reported that psychological help was sought majorly for children who presented behavioral problems or attention deficit hyperactivity disorder. A Spanish study found that only 33.3% of children (9–12 years) with anxiety problems accessed mental health services. Reasons might be that children with anxiety issues do not show as much difficulties as children with behavioral problems, which makes it difficult for the teachers and parents to recognize the problem.


In light of the high prevalence of anxiety disorders and their effect on children’s social, familial, and academic functioning, attempts to identify early risk factors or predictors of anxiety are needed. It is also critical for proposing etiological models, facilitating early detection, and developing preventive interventions.

Several anxiety assessment tools including structured interviews, self-report questionnaires and scales have been described in the literature.  Structured interviews require clinical expertise, are time-consuming and not practical for screening in community settings. In contrast, self-report questionnaires are a time- and cost-effective alternative in community settings (e.g in schools).

The Spence Children’s Anxiety Scale (SCAS)

The Spence Children’s Anxiety Scale (SCAS) is a widely used self-report measure designed to evaluate symptoms of anxiety disorders among children and adolescents. The scale has been cross-culturally adapted to different community samples after its original Australian version was proposed. The scale has already been cross-culturally adapted to many languages (e.g German, Dutch, Spanish, Greek, Japanese, Mexican, Swedish, and Italian), countries, and cultures after its original Australian version was proposed.

In 2018, Reardon et al. developed a short self-reported version (Spence Children’s Anxiety Scale–Children; SCAS-C-8).  They found that the SCAS-C-8 is unifactorial and the factor structure and psychometric properties of this modified scale have not yet been studied in Spanish children.


The objective of the study is to conduct a confirmatory factor analysis (CFA) on the SCAS-C-8, to evaluate its psychometric properties in pre-adolescent Spanish Children. It also aims to determine gender differences.



Participants included 824 pre-adolescent children from Southeastern Spain (98.8% Spanish), ranging from 8 to 12 years old ( M = 9.64, SD = 1.20). The age distribution of the sample was 8 years (22.6%), 9 years (23.8%), 10 years (25.8%), 11 years (22.7%), and 12 years (5.1%). About 24.9% of the participants were in third grade, 24.3% in fourth grade, 25.8% in fifth grade, and 25% in sixth grade. The average number of siblings of the participants was 1.18 (SD = 0.08).


Brief version of the Spence Children’s Anxiety Scale (SCAS-C-8)

The questionnaire included 8 items rated on a 4-point Likert scale (0 to 3; never to always). The SCAS-C-8 was developed from the full 38-item scale to assess generalized anxiety, separation anxiety, and panic/agoraphobia in 7- to 11-year-olds. Subjects that presented moderate item-total correlations and addressed symptoms of anxiety disorders included in the current DSM-5 were selected. A conscious attempt was also made to reduce overlap between items. The internal consistency in the current samples was adequate (α = 0.75).

The Strengths and Difficulties Questionnaire (SDQ)

The SDQ asks about 25 personality attributes which are divided between 5 scales: conduct problems, hyperactivity-inattention, peer problems, and prosocial behavior. The items are then rated on a 3-point Likert scale (0 to 2; from not true, somewhat true to certainly true).

  • The Externalizing Problems subscale = Conduct problems + Hyperactivity Problems subscales
  • Internalizing Problems subscale = Peer problems + Emotional Problems subscales

Combining the subscales minus the prosocial scale gives a total difficulties score, indicating the severity of the psychosocial problems. Whereas a higher score on Prosocial Behavior reflects fewer difficulties. The internal consistency in the current samples was good (α = 0.83).

The reliability indices obtained were: Conduct Problems (α = 0.65), Emotional Problems (α = 0.75), Peer Problems (α = 0.59), Hyperactivity/Inattention (α = 0.72), Prosocial Behavior (α = 0.70), Externalizing Problems (α = .0.78), and Internalizing Problems (α = 0.77).


Following the adaption of the scale to Spanish, the study was conducted on students of the four primary schools in southeastern Spain. All trials were approved by the Ethics Committee of Miguel Hernández University. Parents and principals of the schools provided consent. To evaluate the test-retest reliability of the SCAS-C8, a random subsample of children (n = 434) filled the form again 10 weeks after the initial assessment.

Statistical analysis

  • Ordinal data analysis and Multiple Group Confirmatory Factor Analysis (MG-CFA) were performed in the framework of R 3.5.2 with RStudio 1.1.453.
  • Hu and Bentler’s criteria were used to determine cutoff criteria for fit indexes. Values greater than 0.90 for comparative fit index (CFI) and Tucker–Lewis index (TLI) and less than 0.08 for root mean square error of approximation (RMSEA) indicated an acceptable fit of the model.
  • Intraclass correlation coefficient (ICC) for test-retest was calculated using baseline and 10-week post-assessment. An ICC value of 0.60 or above was regarded as good reliability.
  • A path diagram of the SCAS-C-8 and SDQ subscales was used to determine the evidence of construct validity.
  • Factorial invariance was established, following Timmon’s suggestions.


Confirmatory Factor Analysis (CFA)

  • The results of the Confirmatory Factor Analysis (CFA) indicated a good fit to the one factor structure. χ2(20) = 111.96, CFI = 0.94, TLI = 0.92, RMSEA = 0.07, 95% CI (0.062, 0.080).
  • The factorial loadings of all items exceeded 0.45, ranging from 0.49 to 0.68. Though items 5 and 6 showed factorial loadings of 0.38 and 0.29, respectively, the researchers did not eliminate any of them.

Psychometric properties and reliability

  • Except for items 5 (“Trouble going to school in the mornings”) and 6 (“Suddenly start to tremble or shake”), the corrected item-total correlations were above 0.30.
  • The overall internal consistency of the SCAS-C-8 was adequate (α = 0.75)
  • The subsample (n = 434) showed no statistically significant differences as a function of gender (χ2(1) = 0.97, p = 0.32).
  • Children who answered the retest were slightly older than children who did not answer the second evaluation [t (792) = −3.25, p ≤ 0.001, d = 0.23].
  • Test–retest stability of the SCAS-C-8 was adequate (ICC = 0.77, 95% CI [0.72, 0.87]).
  • Test reliability was similar between 8- to 9-year-olds vs. 10- to 12-year-olds groups. ICC = 0.76, 95% CI (0.68, 0.82) for 8 to 9-year-olds;  ICC = 0.75, 95% CI (0.68, 0.80) for for 10 to 12-year-olds


  • The significant and direct correlations between the SCAS-C-8 and items of the SDQ confirmed the evidence of convergent validity.
  • A high correlation was found between the SCAS-C-8 and item 24 (“Emotional Problems subscale'”), providing evidence of validity.
  • Low correlations were found between the SCAS-C-8 and items of the Externalizing Problems subscales, including Conduct Problems and Hyperactivity items.

Factorial invariance

  • The sample was divided into boys (n = 393) and girls (n=431), to determine the effect of gender on the measurement model.
  • In the first one-factor model, fit indices meeting the benchmarks for acceptable fit supported configural invariance. χ2(40) = 80.68, p < 0.001, RMSEA = 0.05, CFI = 0.94, TLI = 0.92.
  • Weak invariance (Δχ2(40) = 8.36, Δdf = 7, ΔCFI ˂ 0.01) suggested equivalence of the weight of each item regarding its factor.
  • ΔCFI > 0.01 during the strong invariance test indicated a poorer fit of the model.
  • Estimation of the intercepts of item 6 (“I suddenly start to shake or move for no reason”) showed ΔCFI ˂ 0.01, supporting the model with an acceptable fit, χ2(53) = 98.39, p < 0.001, RMSEA = 0.04, CFI = 0.94, TLI = 0.93.


The SCAS is used in clinical contexts for both assessment and evaluation purposes. It is also used to identify children at risk of developing anxiety problems and for monitoring outcome intervention.

The costs of anxiety extend past the obvious emotional, mental, and physical burden in children. Accurate anxiety assessment tools facilitate early detection and reduce barriers associated with early intervention.

The results reported a good fit to the unifactorial structure, adequate internal consistency, and validity. Factorial invariance obtained confirmed that the psychometric properties of the SCAS-C-8 are independent of gender. The fit to a one-factor structure, following Reardon et al, showed good fit indices.

The internal consistency and the test-retest reliability of the Spanish version of the SCAS-C-8 were adequate both in the community (α = 0.84) and clinical samples (α = 0.77).
Items 5 and 6 showed slightly lower factor loadings and correlations compared to others.

SDQ results revealed that higher rates of emotional problems and internalizing problems cause greater anxiety. This indicates that, despite the brevity, the SCAS-C-8 accurately assesses emotional problems in children. The validity evidence data coincided with those found by Reardo et al, where there was convergent and divergent validity between the SCAS-C-8 and the SDQ subscales.

Regarding gender, girls in this study showed higher levels of anxiety than boys. The perception of the parents and teachers also confirmed the gender difference. Moreover, the data was in line with previous studies done on community samples from other countries.


  • The sample was selected by convenience, which could hinder the results.
  • The psychometric properties and factor structure of the SCAS-C-8 were studied in the general population. So its applicability in clinical samples needs further studies.
  • The moderate and low reliability of the subscales of the SDQ with the sample must be considered when interpreting the results.


The SCAS-C-8 is suitable for assessing anxiety symptoms in pre-adolescent Spanish children. When applied in community settings, the scale serves as an effective screening tool to identify children at risk, assisting in preventive interventions.

Oncology Related Tools



Latest Research


About Author

Similar Articles

Leave a Reply