Percentile Norms for the Aberrant Behavior Checklist in ASD

Background:

The Aberrant Behavior Checklist (ABC) is one of the most commonly used rating scales among youth with intellectual and developmental disabilities, including autism spectrum disorder (ASD). It was originally developed with adolescents and adults in residential facilities and has since been revised to be suitable for children and for those living in the community. The original factor structure with five ABC subscales (Irritability, 15 items; Lethargy/Social Withdrawal, 16 items; Stereotypic Behavior, 7 items; Hyperactivity/Noncompliance, 16 items; Inappropriate Speech, 4 items) has recently been supported in children with ASD. The Irritability subscale was the primary outcome measure in several pivotal multi-site drug studies which resulted in FDA indications for risperidone and aripiprazole in the treatment of irritability in ASD.

Objectives:

The objective of this study is to present normative data on the ABC subscales among youth with ASD, overall and separated by educational level (preschool, school-age, or adolescent), IQ (dichotomized at 70), and gender.

Methods:

Data were obtained on 1,796 youth with ASD between the ages of 2 and 18 years who participated in the Autism Treatment Network (ATN). The ATN is a network of 17 children’s hospitals. The first parent-completed ABCs and demographic information were used for these analyses. Since the ATN database is clinical, IQ measures were not uniform across participants. Intelligence was dichotomized at 70 to capture children with or without intellectual disability. Descriptive statistics were calculated on ABC subscales, including measures of skew and kurtosis. Percentile norms are reported (due to the significant positive skew) based on various demographic characteristics.

Results:

Overall the mean Irritability subscale score was 12.8 (sd 9.6, skew .76, kurtosis -.04), Lethargy/Social Withdrawal was 10.0 (sd 7.8, skew .92, kurtosis .55), Stereotypic Behavior was 5.0 (sd 4.6, skew .887, kurtosis .07), Hyperactivity/Noncompliance was 18.7 (sd 11.4, skew .408, kurtosis -.67), and Inappropriate Speech was 3.7 (sd 3.1, skew .63, kurtosis -.43). A significant positive skew emerged on all subscales when separated into demographically-homogenous subgroups. As such, percentile norms and percentile-based T-scores were calculated overall and separated by educational level (preschool, school-age, or adolescent), IQ (dichotomized at 70), and gender.

Conclusions:

Normative information is useful for determining how prevalent an individual’s problem behavior score is. In situations where distributional score characteristics deviate from univariate or multivariate normality, percentile ranks may be more useful clinically. Many statistical analyses assume distributional normality but are robust against minor violations to this at a group level. Percentile rankings, however, allow clinicians to quantify the individual’s behavior more accurately. It also has potential research applications. For example, if a study seeks to investigate clinically-significant irritability and requires a baseline score two standard deviations above the ASD mean (which would hypothetically be the 98^th percentile), the researcher may use a cutoff of 32 on the Irritability subscale. However, the 98^th percentile actually coincides with a raw score of 36. This and other potential applications are explored.