Constructing and testing a disability index in a US sample of preschoolers with disabilities

Abstract

Purpose. To develop and test an index of disability severity in a heterogeneous population of preschoolers with disabilities.

Method. Using a nationally representative sample of 3104 children receiving special education services in the US, questions from a parent interview were used to develop an index of disability severity consisting of domains of functioning defined by gradients of severity. Regressions were used to examine the association between 15 functional domains and 8 cognitive, social/behavioural, and functional outcomes. Full and abbreviated versions of this disability index were compared. Correlations with proxy measures of disability in this population were used to establish concurrent validity, and the predictive ability of this measure was compared with that of the federally defined disability categories.

Results. Of the 15 domains examined, most were significant predictors of at least two outcomes. A shortened index of only six variables was found to be as effective as the longer version in characterising children's level of severity. The index was significantly correlated with intervention variables such as the age at which children began receiving special education or therapy service, r(2802) = −0.22, p ≤ 0.0001, the extent of modification of curriculum materials needed for the child, r(248) = 0.42, p ≤ 0.0001 and the number of services the child received in the school, r(2014) = 0.37, p ≤ 0.0001. Adding severity to special education category significantly increased the amount of variance in outcomes over special education category alone.

Conclusions. These results provide evidence for the importance of the construct of severity of disability and its measurement, particularly when conducting research with very young children.

Keywords:

• Severity
• disability
• preschool
• functioning
• ABILITIES Index
• ICF

Notes

1. During this initial stage, the domain ‘academic skills’ was dropped from inclusion because, unlike the other domains, this area reflects learned skills rather than functioning.

2. To preserve the range and maintain consistency with the other domains, values of ‘3’ were shifted to ‘4’ at this stage of the process.