There are various ways in which cut-off points can be established. The following method uses T-Scores and takes into account the age and gender to the young person. It also considers that high anxiety status may be reflected in both the total score and an elevated subscale score. Parents may report elevated scores on the SCAS-P in two ways: in terms of elevated total scores and high scores on one or more subscale scores. Although the majority of children who show a high total score also show a high score on one or more subscales, this is not always the case. Thus, for clinical assessments, we recommend examining the total and subscale scores. For screening purposes in community samples, it may be sufficient to use the total score for identification of children at risk.
Using the Parent scas T-Scores
The parent version of the SCAS is intended to provide an indicator of the number and severity of anxiety symptoms. It is not designed to be a diagnostic instrument for use in isolation although it provides important information to inform the assessment process. Where a clinical diagnosis is required the SCAS should be used as an adjunct to clinical interview. It may also be used for identification of children who have elevated symptoms of anxiety and for whom further assessment is recommended to determine whether there is a need for intervention. Similarly, it provides an indicator of response to treatment. It has also been used in several studies to identify children for whom early intervention or prevention is warranted on the basis of elevated anxiety symptoms being a risk factor for the development of future mental health problems.
using the t-scores
what are t-scores
T-Scores enable the comparison of a young person's scores against norms from an equivalent age and gender group from a representative normative sample. A T-score is a standardized score that is calculated from the total distribution of scores within the community sample. Scores are rescaled so that T-scores have a mean of 50 and a standard deviation of 10. Scores within one standard deviation (ie. a T-score of 10) above the mean on any dimension are regarded as being within the normal range on that dimension. This process ensures that all subscales and the total score can be interpreted along the same scale, with the same mean and standard deviation, even though they initially had different numbers of items and different non-transformed means.
The cut-off points used for T-scores depend on the purpose of the assessment and different authors suggest different cut-points that should be regarded as indicative of clinical or subclinical levels of psychopathology.
A T-score of 60 (ie. a T-score of 10 points above the mean T-score of 50 ) is approximately 1 standard deviation above the mean. This represents around the 84th percentile meaning that around 16% of children would be expected to show a score at this level and suggests elevated anxiety. As noted above the SCAS should not be used as a diagnostic instrument, in the absence of a clinical interview. We suggest using a T-score of 60 as indicative of sub-clinical or elevated levels of anxiety and justifies further investigation and confirmation of diagnostic status using clinical interview.
Some clinicians prefer to use a criterion of a T-score of 65, to indicate clinical status (or 1.5 standard deviations above the mean). A T-score of 65 represents around the top 6% of the population. However, given the variation in criteria used by researchers and clinicians to define clinical status, and that the SCAS-P is not intended as a clinical instrument when used in isolation, the score interpretation presented here uses the term “elevated†and is defined as T-scores above 60.
If preferred, percentile scores may be used instead of T-scores and the T-score tables also present the percentile value for each subscale and total score.
The following T-scores were developed based on the normative data reported on this site under the "norms" tab.
The cut-off points used for T-scores depend on the purpose of the assessment and different authors suggest different cut-points that should be regarded as indicative of clinical or subclinical levels of psychopathology.
A T-score of 60 (ie. a T-score of 10 points above the mean T-score of 50 ) is approximately 1 standard deviation above the mean. This represents around the 84th percentile meaning that around 16% of children would be expected to show a score at this level and suggests elevated anxiety. As noted above the SCAS should not be used as a diagnostic instrument, in the absence of a clinical interview. We suggest using a T-score of 60 as indicative of sub-clinical or elevated levels of anxiety and justifies further investigation and confirmation of diagnostic status using clinical interview.
Some clinicians prefer to use a criterion of a T-score of 65, to indicate clinical status (or 1.5 standard deviations above the mean). A T-score of 65 represents around the top 6% of the population. However, given the variation in criteria used by researchers and clinicians to define clinical status, and that the SCAS-P is not intended as a clinical instrument when used in isolation, the score interpretation presented here uses the term “elevated†and is defined as T-scores above 60.
If preferred, percentile scores may be used instead of T-scores and the T-score tables also present the percentile value for each subscale and total score.
The following T-scores were developed based on the normative data reported on this site under the "norms" tab.
scoring templates for PARENT scas t-SCORES
The scoring sheets can be downloaded and used to determine the Parent SCAS T-scores. Given the difference in means across age groups and gender, it is important to use the correct sheet based on the age and gender of the child.
The raw scores must first be calculated for each subscale and the total score. Then, the raw scores are circled on the sheet and the corresponding T-Score is identified from the T-score column. The raw score and corresponding T-score can be recorded at the bottom of the page.
This process can also be used to identify percentile scores.
The raw scores must first be calculated for each subscale and the total score. Then, the raw scores are circled on the sheet and the corresponding T-Score is identified from the T-score column. The raw score and corresponding T-score can be recorded at the bottom of the page.
This process can also be used to identify percentile scores.
how were the t-scores computed
The following section describes how the T-scores were calculated. The procedure for assigning T-Scores followed that outlined by Achenbach & Rescorla (2001) . This was performed separately for the total score and each subscale. For the parent SCAS this process was done separately for girls and boys and for ages 7-9 and 10-13 years. First the cumulative frequency distribution of each score was determined, with each raw score then assigned to the midpoint of the percentile range that it spanned. T scores were then allocated to these mid-point percentile scores.
To take into account the positively skewed distributions and need for greater differentiation between high scores rather than low scores (for clinical use) we used to following procedure :-
- we assigned T score of 40 to all raw scores also close as possible to the 15.7th percentile and below for each subscale. This point reflects 1 SD below the mean.
- To facilitate comparisons across subscales, T scores at the lower end were collapsed into T = 40 or below.
- A T-score of 50 was allocated to the score most closely approximating to the 50th midpoint percentile.
- At the high end of scores T scores the size of the sample a T-score of 70 was allocated to the score closest to the 98th midpoint percentile. T-scores above the 98th percentile (T score =70) are less accurate given the relatively low prevalence of such high scores in the community sample. Thus scores above the 98th percentile were collapsed into T=70 and above.
- T scores were then allocated for 45, 55, 60, 65 and 70 based on midpoint percentiles:
T = 70: 98th %ile
T = 65: 94th %ile
T = 60: 84th %ile
T = 55: 70th %ile
T = 50: 50th %ile
T = 45: 32nd %ile
T = 40: 16th %ile
To take into account the positively skewed distributions and need for greater differentiation between high scores rather than low scores (for clinical use) we used to following procedure :-
- we assigned T score of 40 to all raw scores also close as possible to the 15.7th percentile and below for each subscale. This point reflects 1 SD below the mean.
- To facilitate comparisons across subscales, T scores at the lower end were collapsed into T = 40 or below.
- A T-score of 50 was allocated to the score most closely approximating to the 50th midpoint percentile.
- At the high end of scores T scores the size of the sample a T-score of 70 was allocated to the score closest to the 98th midpoint percentile. T-scores above the 98th percentile (T score =70) are less accurate given the relatively low prevalence of such high scores in the community sample. Thus scores above the 98th percentile were collapsed into T=70 and above.
- T scores were then allocated for 45, 55, 60, 65 and 70 based on midpoint percentiles:
T = 70: 98th %ile
T = 65: 94th %ile
T = 60: 84th %ile
T = 55: 70th %ile
T = 50: 50th %ile
T = 45: 32nd %ile
T = 40: 16th %ile
In keeping with Achenbach & Rescorla (2001) we emphasize that raw scores rather than T-scores should be used for statistical analyses, where the full range of scores is important.
The sample used for comuting the t-scores
The data for the T-Scores were computed using community samples from Australia, the Netherlands, the USA and the UK, obtained from the following sources:
a) Nauta, Scholing, Rapee, Abbott, Spence and Waters. (2004). A parent report measure of children's anxiety.Behaviour Research and Therapy, 42(7), 813-839. Dutch and Australian Samples: N=221
b) Brown-Jacobsen, A. M., Wallace, D. P., & Whiteside, S. P. H. (2011). Multimethod, Multi-informant Agreement, and Positive Predictive Value in the Identification of Child Anxiety Disorders Using the SCAS and ADIS-C. Assessment, 18(3), 382-392. USA Sample: N= 54
c) Whiteside, S. P. H., Gryczkowski, M. R., Biggs, B. K., Fagen, R., & Owusu, D. (2012). Validation of the Spence Children's Anxiety Scale's obsessive compulsive subscale in a clinical and community sample. Journal of Anxiety Disorders, 26(1), 111. USA Sample: N=423
d) Data from 2 studies conducted by Prof Cathy Creswell and colleagues at the University of Reading in the UK. UK Samples: N=708 and N = 361
b) Brown-Jacobsen, A. M., Wallace, D. P., & Whiteside, S. P. H. (2011). Multimethod, Multi-informant Agreement, and Positive Predictive Value in the Identification of Child Anxiety Disorders Using the SCAS and ADIS-C. Assessment, 18(3), 382-392. USA Sample: N= 54
c) Whiteside, S. P. H., Gryczkowski, M. R., Biggs, B. K., Fagen, R., & Owusu, D. (2012). Validation of the Spence Children's Anxiety Scale's obsessive compulsive subscale in a clinical and community sample. Journal of Anxiety Disorders, 26(1), 111. USA Sample: N=423
d) Data from 2 studies conducted by Prof Cathy Creswell and colleagues at the University of Reading in the UK. UK Samples: N=708 and N = 361
I am extremely grateful to these researchers for allowing me to use their data in order to provide a sample size large enough to compute T scores.
Only data from parents of youth aged 7 to 13 years were included from these studies in computing the T-scores as there were insufficient youth aged 14 years and older to permit the extraction of valid T-scores. The designs of all the studies included were intended to recruit children and parents who were representative of the general population in the countries concerned.
Only data from parents of youth aged 7 to 13 years were included from these studies in computing the T-scores as there were insufficient youth aged 14 years and older to permit the extraction of valid T-scores. The designs of all the studies included were intended to recruit children and parents who were representative of the general population in the countries concerned.
Mean Scores and Standard Deviations for the T-Score Sample
The values for means and standard deviations for each gender and age group are shown in the attached table. The results revealed significant differences in mean scores across genders and age levels, necessitating separate calculation of T-scores according to age and gender of the child. Please note that the means and SDs shown here differ slightly from those reported in the original paper by Nauta et al as the age groups are different.