- Youth Justice State Plan
- Disability Screening Assessment Project
- Youth Justice in SA
- Youth Justice Services
- Youth Justice Partners
- Working in Youth Justice
- Young People and the Law (sa.gov.au)
- Contact Youth Justice
Prevalence of Needs (Discipline-Specific Results)
(Further discussion of assessment results, including functional impacts, are included in the Key Messages section.)
Weschler Abbreviated Scale of Intelligence (WASI)
- 26 young people completed at least 2 subtests of the WASI, enabling a valid estimate of their overall intellectual functioning (FSIQ) to be calculated.
- 18 young people completed all 4 possible subtests of the WASI.
- Results (see Table 3) indicated that a higher number (15%) of the young people who were assessed were in the ‘Extremely Low’ range for the Verbal Comprehension Subscale, than for the Perceptual Reasoning Subscale (4.8%), the Full Scale IQ (2 subtests) (7.7%) and the Full Scale IQ (4 subtests) (5.6%). Scores in the ‘Extremely Low’ range are indicative of significant impairment in that area.
- Results also indicated that a higher number (88.9%) of the young people were assessed as being in the ‘Extremely Low’, ‘Borderline’ and ‘Low Average’ ranges for the Full Scale IQ (4 Subtests) as opposed to 11.1% in the ‘Average’ range, indicating that the majority of the population had at least some difficulties with intellectual functioning.
See further discussion of these results in Key Messages section.
Table 3: WASI descriptive data
3 (15 %)
1 (4.8 %)
Full Scale IQ (2 Subtests)
Full Scale IQ (4 Subtests)
Youth Level of Service/Case Management Inventory (YLS/CMI)
- 18 young people were assessed using the YLS/CMI.
- Results (see Table 4) indicated that the large majority of scores were in the:
- ‘High’ risk classification ranges for Prior/Current Offences (76.5%)
- Education/Employment (77.8%)
- Peer Relations (83%)
- Substance Abuse (72.2%)
- Leisure/Recreation (94.4%)
- overall Total Score (50%).
- For the ‘Family Circumstances/Parenting’ category, scores were split fairly evenly across the ‘Moderate’ and ‘High’ classifications.
- The majority of scores for the Personality/Behaviour and Attitudes/Orientation categories were in the ‘Moderate’ risk classification ranges.
- Regarding the total score, 50% of the participants were identified with a risk of re-offending in the ‘High’ range and an additional 38.9% were in the ‘Moderate’ range, suggesting that the large majority of this population (88.9%) had a moderate to high risk of re-offending.
Table 4: YLS/CMI descriptive data
Prior and Current Offences
Family Circumstances and Parenting
Education and Employment
Leisure and Recreation
Personality and Behaviour
Attitudes and Orientation
Adolescent Psychopathology Scale - Short Form (APS-SF)
- The total number of participants assessed with the APS-SF was 23.
- Results (see Table 5) indicated that the categories with the highest percentage of responses in the ‘Severe’ and ‘Moderate’ ranges (that is, T score above 70) were:
- Conduct Disorder (60.9%)
- Substance Abuse (56.1%)
- Anger/Violence Proneness (40.9%)
- Academic Problems (34.7%)
- Post-Traumatic Stress Disorder (36.4%).
- This suggests that a significant portion of the population presented with psychopathology and/or maladaptive behaviour.
- The categories that returned the least responses in the ‘Severe’ and ‘Moderate ranges were:
- Generalised Anxiety Disorder (21.7%)
- Major Depression (18.2%)
- Eating Disturbance (4.5%)
- Suicide (17.9%)
- Self-Concept (4.8%)
- Interpersonal Problems (13.6%).
Table 5: APS-SF descriptive data
T scores below 60
Oppositional Defiant Disorder
Generalised Anxiety Disorder
Posttraumatic Stress Disorder
Occupational Therapy Assessments
Beery Visual-Motor Integration (VMI)
All young people who completed the Beery VMI assessment (n = 27) recorded scores outside the average range in at least one of the three subtests, with 16 young people (59%) scoring in the ‘Very low’ range for at least one subtest.
- Visual Motor Integration scores (Assesses the degree to which visual perception and finger-hand movements are well coordinated): 22 young people (81.5%) returned scores outside the average range, with 6 young people (25.9%) recording scores within the ‘Very Low’ range.
- Visual Perception subtest (Assesses the brain’s ability to make sense of what the eyes see): 20 young people (74.1%) returned scores outside the average range, with 7 young people (25.9%) recording scores within the ‘Very Low’ range.
- Motor Skills subtest (Assesses motor coordination, in particular, fine motor control): 23 young people (85.2%) returned scores outside the average range, with 7 young people (25.9%) recording scores within the ‘Very Low’ range.
- 24 young people (88.9%) returned scores outside the average range for multiple subtests, and 13 young people (48.1%) were outside average range for all three subtests. 16 young people (59%) returned scores outside the average range for both the visual perception and motor coordination subtests.
For all sub-tests of the Beery VMI, the proportion of participants that scored in the ‘Very Low’ range (equivalent to the lowest 2% of their age range) was equal to or greater than the proportion scoring in the ‘Average’ range (equivalent to the middle 68% of the age group):
- Visual-Motor Integration: 5 young people (18.5%) in the ‘Average’ range v 6 young people (22.2%) in the ‘Very Low’ range
- Visual Perception: 7 (25.9%) ‘Average’ v 7 (25.9%) ‘Very Low’
- Motor Coordination: 4 (14.8%) ‘Average’ v 7 (25.9%) ‘Very Low’.
A total of 29 young people completed a Sensory Profile assessment. Due to the age range of participants, the Sensory Profile 2 (normed for ages up to 14 years 11 months) was used for five participants, and the Adolescent and Adult Sensory Profile was used with the remaining 24 participants. Due to the individualised, highly-variable nature of sensory processing, it is difficult to extrapolate ‘trends’ from the data collected from the Sensory Profile results (Figure 3).
However, several notable findings are highlighted below:
- 6 young people (20.7%) returned scores in the ‘Similar to Most’ range (indicating scores within the average range) for all four quadrants (see Note 2). The remaining 23 young people (79.1%) returned scores outside the ‘Similar to Most’ range for at least one quadrant.
- 4 (13.8%) returned scores outside the ‘Similar to Most’ range for one quadrant; 8 (27.6%) in two quadrants, 6 (20.7%) in three quadrants and 5 (17.2%) returned scores outside the ‘Similar to Most’ range for all four quadrants.
- 10 young people (34.4%) returned scores in the Much More than Most (n = 9) or Much Less than Most (n = 1) ranges in at least one quadrant, indicating a high likelihood that their sensory processing is impacting their participation in daily tasks.
- Just under a third of participants (31.1%) returned scores in the ‘Less than Most’ or ‘Much Less than Most’ ranges for the Sensation Seeking quadrant, compared to 17.2% within the ‘More than Most’ or ‘Much More than Most’ ranges. It was the only quadrant with a greater proportion of scores in the lower ranges. Low scores in the Sensation Seeking quadrant indicates a young person may not seek additional sensory experiences. Functionally, this may lead to difficulties with task completion because these young people may lack the motivation required to complete daily life tasks.
- A higher proportion of the population returned scores in the ‘More than Most’ or ‘Much More than Most’ ranges for both quadrants associated with hyper-sensitivity or low-threshold quadrants (see Note 3) – Sensory Avoiding (48.2%) and Sensory Sensitivity (37.9%); compared with the high-threshold quadrants (see Note 4) – Low Registration (34.4%) and Sensation Seeking (17.2%).
- The majority of the population fell outside the ‘Similar to Most’ range in both the Sensory Avoidance (55.2% - 47.3% More or Much More than Most and 6.9% Less than Most) and Low Registration (51.7% - 34.4% More or Much More than Most and 17.2% Less or Much Less than Most) results.
Note 2: Scoring for the Sensory Profile assessments is structured according to the following quadrants:
- Sensation Seeking: The degree to which a young person OBTAINS sensory input.
- Sensory Avoiding: The degree to which a young person is BOTHERED by sensory input.
- Sensory Sensitivity: The degree to which a young person DETECTS sensory input.
- Low Registration: The degree to which a young person MISSES sensory input.
Note 3: Low-threshold items measure a person’s notice of or annoyance with sensory stimuli. Functional impacts of higher scores in these areas may include a young person withdrawing from particular environments or tasks (Sensory Avoiding) or becoming distracted by sensory input or reacting to stimuli in ways that might seem disproportionate to an observer (Sensory Sensitivity).
Note 4: High-threshold items measure an individual’s lack of response to sensory stimuli or suggests a need for more intense sensory stimuli.
Figure 3: Sensory profile results
Delis Rating of Executive Function (D-REF)
When compiling responses in which young people reported facing particular stressors either ‘daily’ or ‘weekly’, the five stressors most frequently reported by participants (n=28) were:
- “I do things without thinking” (22 young people: 16 responded that they face this daily, 6 responded that they face this weekly).
- “I can’t seem to concentrate on something for very long” (21: 15 daily, 6 weekly).
- “I find it hard to keep doing a boring task like homework” (20: 14 daily, 6 weekly).
- “No matter how hard I try, I can’t seem to sit still for very long” (19: 16 daily, 3 weekly).
- “My mood can change from happy to mad or sad very quickly” (19: 14 daily, 5 weekly).
The highest prevalence of challenges reported by participants related to:
- Attention and Working Memory (Denotes the ability to store information in one’s head and retrieve it in an effective and efficient manner): 85.7% of population outside normal limits
- Executive Functioning (Conceptualised as a young person’s higher-level cognitive ability to effectively adapt and function within the demands of the environment.): 78.6% outside normal limits
- Behavioural Functioning (Denotes the ability to regulate one’s behaviour): 75%
The mean T-Score for all sub-sections was above 60, indicating that population averages were in the ‘Elevated’ range for all sections. Mean T-Scores for Executive Functioning and Attention/Working Memory both fell within the ‘Severely Elevated’ range (77.9%, 83.9% respectively; Table 6).
Table 6: D-REF assessment results
T-Scores above 60 are considered ‘Elevated’ and outside Normal Limits. T-Scores above 70 are considered ‘Severely Elevated’.
Mean T- score
% of participants outside
Twenty-four participants (85.7%) recorded scores outside the ‘Average/Borderline’ range on at least one subtest and 14 (50%) scored outside the average range on all sub-sections.
Fifteen participants (53%) scored within the ‘Severe’ range for at least one sub-section, 13 (46%) scored in the ‘Severe’ range on multiple sub-sections and seven (25%) scored in the ‘Severe’ range on all sub-sections.
In each of the D-REF’s sub-sections, a higher proportion of participants scored within the ‘Severely Elevated’ range (>98th percentile) than the ‘Within Normal Limits’ range (1st to 69th percentile).
- Behavioural Functioning – 39.3% severely elevated v 25% within average range
- Emotional Functioning – 35.7% v 32.1%
- Executive Functioning – 28.6% v 21.4%
- Total Composite Score – 35.7% v 25%
- Attention/Working Memory – 32.1% v 14.3%
- Activity/Impulsivity – 39.3% v 28.6%
- Compliance/Anger Management – 39.3% v 32.1%.
Twenty-nine young people completed the Cognistat assessment.
Of those, 26 (89.7%) returned scores outside the average range for at least one sub-section. The number of areas of impairments for individuals ranged from a single area of impairment (n = 10), to two individuals recording scores outside the average range in six sub-tests (Figure 4).
Figure 4: Cognistat participants by number of sub-test scores outside average range
- All participants (n = 29) of the Cognistat achieved scores within the average range for Orientation (Assesses one’s awareness of current day, date, time-of-day, birth date along with general knowledge questions associated with current and recent Prime Ministers and US Presidents)
- and 26 participants (89.7%) scored within the average range for Constructional Ability (Assesses visuo-constructional capacity (a combination of visuo-spatial, executive and motor capacity)
- In terms of other sub-tests that indicated the highest prevalence of difficulty, 15 young people (55.2%) returned scores outside the average range for Calculations (Assesses the participant’s ability to complete simple mental arithmetic problems)
- 9 individuals (31%) achieved scores outside the average range for both the Similarities (Assesses ability to determine abstract interpretations) and Judgement sub-tests (Assesses the ability of a participant to identify and understand consequences of situations and/or actions).
Speech Pathology Assessments
Clinical Evaluation of Language Fundamentals – 5 (CELF-5) Screening Test
Of the total 36 participants involved in the screening project, 29 young people (80.5%) participated in the CELF-5 screening test to identify ‘risk of language disorder/difficulty’.
Only three of these young people passed, meaning that almost nine out of every 10 young people assessed (89.7%) did not meet the expected criterion level for their age, and may be at risk of Language Disorder (LD). See Figure 5. This indicates that the majority of participants may be experiencing oral language difficulties.
Figure 5. Young people identified as being at risk of language disorder based on CELF-5 screening test
Understanding Paragraphs (USP, subtest of the CELF-5 full language assessment)
Of the total 36 participants involved in the screening project, only 30% (11 young people) completed the USP subtest. Eight of the 11 young people (72.7%) had moderate or severe difficulties understanding paragraph level verbal information (n = 3 and n = 7 respectively).
Test of Narrative Language 2nd ed. (TNL-2)
Of the total 36 participants involved in the screening project, 23 young people (63%) had their narrative-level language skills assessed by the TNL-2. Results suggested that for this group of young people, their ability to tell or retell a story (Narrative Production) was stronger than their ability to understand story-level language (Narrative Comprehension). For Narrative Production 82.7% participants scored at or above average compared to 65.3% scoring at or above average range for Narrative Comprehension. Of the 34.7% (n = 8) who scored below the average range for narrative comprehension, two scored in the very poor range, and three scored in the poor and below average ranges respectively.
Discussion of Language Assessment Results
The discrepancy between results on the CELF-5 language screener and the TNL-2 are of interest; whilst almost 90% of the 29 young people who completed the CELF-5 language screener failed this test, hence indicating a risk of language disorder (see Note 5) only 34.7% and 17.3% of the of the 23 young people who completed the TNL-2 fell below the average range in Narrative Comprehension and Production respectively.
This could be due to the fact that 16 out of 23 (69.6%) of young people assessed on the TNL-2 were aged older than the test’s normative data maximum age (which stopped at 15 years 11 months). This could have led to participants’ results looking better than their actual skill level (that is, false-negative results), because 16 to18 year olds’ ‘severity rating’ was measured against the normative data for 14;0 to 15;11 year old age range. However, it was felt that the TNL-2 offered valuable information about functional narrative abilities, which is why a decision was made to proceed with the use of this assessment in the screening project.
For many young people, narrative (story telling) skills were a relative strength compared to their results in the CELF-5 screener and the USP, which look at foundational language skills. This could reflect the propensity for young people in youth justice being skilled at participating in everyday conversation while simultaneously masking underlying gaps in their language abilities, and struggling with the language required for academic success. The CELF-5 screener is geared more towards assessing foundation language skills and the language skills required for success in an educational environment. Difficulties with foundational oral language skills such as syntax (sentence structure), morphology (grammar), semantics (word meaning) can impact on literacy attainment in an education context and engagement in educational programs, which may contribute to pathways into offending behaviour.
Differences were also found in the performance of young people on the TNL-2 comprehension component and the USP, which both examine the ability to comprehend large amounts of verbal information (one to three paragraphs in length). Although 65.3% (of 23) young people scored at or above average range on the TNL-2 Narrative Comprehension, 72.7% (of 11) young people had moderate or severe difficulties of impairment identified by the USP. A key difference between these two tests is that the TNL-2 provides high amounts of visual information (that is, pictures) in addition to the verbal story that the young person is asked to understand and/or retell. The discrepancy between narrative skills (as assessed by the TNL-2) versus the CELF-5 screener and USP may reflect a testing anomaly, given the use of visual information in the TNL-2. This visual scaffolding within the TNL-2 may be a reason why many young people performed better in the TNL-2 than the USP and CELF-5 screener; consequently highlighting the benefit of using visual supports to aid receptive and expressive oral language.
Note 5: As indicated by the test manual, Wiig, E. H., Secord, W. A., Semel, E., 2013
Comparative analyses and Interpretation
A number of correlational and comparative analyses were conducted to gain insight into the concurrent validity between different measures which measure similar domains of functioning, and to identify any differences between groups of interest. Due to the small sample size and lack of statistical power necessary for detecting significant differences between groups, a limited number of analyses could be interpreted with confidence.
The following key questions were examined.
Are risk assessments correlated with other measures?
- The VONIY score was significantly correlated with the YLS/CMI total score r(12) = 0.68, p = .004. This means that, for this population, when a young person has a higher level of criminogenic need as assessed by the VONIY, then it is likely that they will have a high risk of re-offending as assessed by the YLS/CMI. In addition, the risk of re-offending category was found to be primarily in the ‘moderate’ or ‘high’ categories for the majority of the participants, thereby indicating the need for moderate/high intensity criminogenic intervention.
- The DREF Compliance/Anger Management score was significantly correlated with the DREF Activity/Impulsivity Score r(26) = 0.75, p = .000. This means that for young people, in this population, who experience difficulties with compliance and/or anger management, they are also likely to experience difficulties with activity and/or impulsivity.
- The DREF Compliance/Anger Management score was correlated with the VONIY total score r(20) = 0.31, p = .082 and the YLS/CMI r(13) = 0.41, p = .066, however the correlation was not found to be statistically significant. This means that young people who had greater difficulties complying with direction and managing their anger tended to have higher overall scores for criminogenic need and risk, but this tendency was indicative, not definitive.
Are scores on measures assessing cognitive ability correlated with scores on measures assessing oral language skills?
- The WASI (Full Scale 4 subtests score) was significantly correlated with the TNL-2 Comprehension Percentile r(11) = 0.75, p = .001 and with the TNL-2 Narrative Language Ability Percentile r(9) = 0.65, p = .015 but not the TNL-2 Production Percentile. This means that for young people who have cognitive difficulties they will also have difficulties with understanding story-level oral language (i.e. large amounts of oral language presented at once).
Are there differences in verbal comprehension scores according to level of risk of language disorder?
- There was a significant difference in WASI Verbal Comprehension Scores according to the level of risk of language disorder as assessed by the CELF-5 Screener. WASI Verbal Comprehension Scores were significantly higher in the ‘at/above criterion group’ of the CELF-Screener (M=101.50, SD=10.61) than in the ‘below-criterion group’ (M=78.36, SD=8.57), t(14) = t(-3.51), p =0.003. These results suggest that young people who are identified as ‘at risk of language difficulties’ by the CELF-5 language screener also exhibit more difficulties in understanding verbal information (as assessed by the WASI).
- Moreover, more young people scored lower on the verbal subtests of the WASI than the non-verbal subtests. These results suggested that these difficulties with verbal information in this population may be more reflective of limited educational opportunities and/or speech and language disorder or differences, than actual intellectual disability, and highlight the need for targeted interventions to improve young people’s educational experiences. These interventions may need to occur earlier, in the community. For Youth Justice, it also highlights the importance of recognising the differing levels of verbal capacity in this population and ensuring that staff are sensitive and responsive to this.
- Analyses involving the CELF-5 Screener need to be interpreted with caution because the majority of young people failed the test (only 3 young people scored at or above the score expected for their age).
- Both the TNL Production and the TNL NLA Index scores were not significantly related to whether Young People failed or passed the CELF-5 Language Screener.
Was there a relationship between areas of need and amount of time spent in custody?
- It should be noted that the number of nights spent in custody figure was not stratified for age, and therefore young people who were of older age had a greater duration of involvement with Youth Justice and a higher ‘nights in custody overall’ figure than younger adolescents. A such, while the following results are of note, they must be interpreted with some caution.
- There was not a statistically significant difference between how many nights a young person spent in custody in the low/moderate risk categories (M=352.89, SD=422.98) compared with the high risk category on the YLS/CMI (M=547.50, SD=491.44); t (16)=-.90, p = 0.38. Although the number of nights spent in custody did not differ ‘significantly’ between young people scoring in the low/moderate risk categories (M, SD) compared with the high risk category (M, SD) of the YLS/CMI, it is clear from the absolute values that young people in the high risk category spent, on average, considerably more time in custody than young people assessed as lower risk.
- Similarly, there was not a statistically significant difference between how many nights a young person spent in custody for the moderate need category (M=78.23, SD=27.36) compared with the high/intensive need categories on the VONIY (M=575.59, SD=414.88); t (24)=-1.66, p = 0.11. However, it is clear from the absolute values that young people in the high need range on the VONIY spent, on average. considerably more time in custody than young people assessed as having a lower level of criminogenic need.
- There was not a significant difference between how many nights a young person spent in custody for those with below criterion scores on the CELF-5 Screener.
- Composite variables (see Note 6) were created to analyse the possible impact of verbal functioning (see Note 7). The difference between how many nights a young person spent in custody for the average or above average verbal functioning composite (M=2.50, SD=0.58) and the low verbal functioning composite (M=2.34, SD=0.55) was not statistically significant, t (31)=0.52, p = 0.60.
- A composite variable was also generated to analyse the potential impact of complexity of disability-related need (see Note 8) — this was defined as below average on 3 or more assessments vs above average on 3 or more assessments. The difference between how many nights a young person spent in custody for the below average category (M=2.33, SD=0.52) compared to the average or above average category (M=2.40, SD=0.56) was not statistically significant, t (34)=-0.27, p = 0.79. This indicates that the composite variable for ‘complexity’ that was created, did not sufficiently discriminate (differentiate) between groups. However, given the small numbers of young people in each group, the results should be interpreted with caution.
Note 6: The verbal functioning composite for average to above average functioning was based on WASI Verbal Comprehension subscale in average range and CELF-5 Screener at or above criterion.
Note 7: The verbal functioning composite for low functioning was based on the WASI Verbal Comprehension subscale in the borderline range or lower and the CELF-5 Screener below criterion.
Note 8: The complexity category was divided into two groups: scores below average on 3 or more assessment measures and scores above average on 3 or more assessment measures.
---------Page last updated : 31 Aug 2021