Developing the Japanese Version of the VABS-II (2): Examining the Validity by Confirmatory Factor Analyses

Background:

Currently, there is no comprehensive adaptive behavior scale available in Japan to help making a decision on a necessary level of support for individuals with disabilities. The unavailability of adaptive behavior scales has led to an underestimation on needs in adaptive functioning of individuals with disabilities, particularly those who exhibit a significant discrepancy between their cognitive ability and adaptive functioning (i.e., those who exhibit higher IQ but have deficits in adaptive functioning). The Vineland Adaptive Behavior Scales, Second Edition (VABS-II; Sparrow, Cicchetti, & Balla, 2005) is one of the most widely used adaptive behavior scale in the US and other countries and is used not only for identifying individuals with cognitive disabilities, but also assessing needs of individuals with pervasive developmental disorder (PDD) and other developmental disabilities.

Objectives:

The original VABS-II consists of four primary factors that include 11 subdomains: Communication, Daily Living Skills, Motor Skills and Socialization domains, as well as the Adaptive Behavior composite (overall score of the others; Sparrow et al., 2005). Additionally, these primary factors (i.e., second tier) compose of the adaptive behavior as a secondary factor (i.e., top tier). The purpose of this study was to examine the validity of the Japanese VABS-II by confirmatry factor analyses.

Methods:

The participants were 1,219 normally developing children/individuals (614 males, 605 females) aged 0 to 87 years in 28 prefectures (out of 47) in Japan. The selection of the participants was modeled after the original VABS-II and matched to the demographics of Japan regarding age and male-to-female ratio. For those from 0 months to two years old, the data were collected for each month up to one-year-old (e.g., 1-month-old, 2-month old), and every other month up to two-year-old because they usually exhibit rapid developmental change. For data analyses, missing values were estimated by the full information maximum likelihood (FIML).

Results:

Confirmatory factor analyses by age exhibited sufficient factor loading for each primary factor, indicating that the original factor model fit the present data for all age groups. Each primary factor also exhibited sufficient factor loading for the secondary factor. When evaluating the best fit for a structural model using fit indices, four-factor model (three-factor for seven years and older because there was no Motor domain) showed the best fit for all the age groups but one (i.e., 7-12 years) rather than independent model (i.e., all the subdomains were independent from each other) or one factor model (i.e., all the subdomains belonged to a single factor).

Conclusions:

The validity of the Japanese version of the VABS-II appeared to have a sufficient validity supported by an adequate model fit.