Update on the Child's Challenging Behaviour Scale following evaluation using Rasch analysis
Version of Record online: 5 MAR 2013
© 2013 John Wiley & Sons Ltd
Child: Care, Health and Development
Volume 40, Issue 2, pages 242–249, March 2014
How to Cite
Bourke-Taylor, H. M., Pallant, J. F. and Law, M. (2014), Update on the Child's Challenging Behaviour Scale following evaluation using Rasch analysis. Child: Care, Health and Development, 40: 242–249. doi: 10.1111/cch.12035
- Issue online: 13 FEB 2014
- Version of Record online: 5 MAR 2013
- Manuscript Accepted: 25 NOV 2012
- childhood disability;
- maternal health;
- Rasch analysis
The Child's Challenging Behaviour Scale (CCBS) was designed to measure a mother's rating of her child's challenging behaviours. The CCBS was initially developed for mothers of school-aged children with developmental disability and has previously been shown to have good psychometric properties using classical test theory techniques. The aim of this study was to use Rasch analysis to fully evaluate all aspects of the scale, including response format, item fit, dimensionality and targeting.
The sample consisted of 152 mothers of a school-aged child (aged 5–18 years) with a disability. Mothers were recruited via websites and mail-out newsletters through not-for-profit organizations that supported families with disabilities. Respondents completed a survey which included the 11 items of the CCBS. Rasch analysis was conducted on these responses using the RUMM2030 package.
Rasch analysis of the CCBS revealed serious threshold disordering for nine of the 11 items, suggesting problems with the 5-point response format used for the scale. The neutral midpoint of the response format was subsequently removed to create a 4-point scale. High levels of local dependency were detected among two pairs of items, resulting in the removal of two items (item 7 and item 1). The final nine-item version of the scale (CCBS Version 2) was unidimensional, well targeted, showed good fit to the Rasch model, and strong internal consistency.
To achieve fit to the Rasch model it was necessary to make two modifications to the CCBS scale. The resulting nine-item scale with a 4-point response format showed excellent psychometric properties, supporting its internal validity.