The test of everyday attention for children: a confirmatory factor analysis approach
Date
2011
Authors
Passantino, Deborah, author
Davies, Patricia, advisor
Gavin, William, committee member
Diehl, Manfred, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
As the incidence of children diagnosed with Autism Spectrum Disorders (ASD) and Attention Deficit Hyperactivity Disorder (ADHD) continues to grow, the need for objective measures of attentional performance is clearly warranted for evaluating attentional differences and guiding intervention. This study examined the multidimensional nature of attention. Previous research suggests that there may be three types of attention: selective attention, control shift attention, and sustained attention. One hundred and eleven children age six to twelve completed the nine subtests of the Test of Everyday Attention for Children (TEA-Ch, Manly, Robertson, Anderson & Nimmo-Smith, 1999). Using a confirmatory factor analysis approach, this study sought to determine whether a three-factor model, as supported in a prior confirmatory factor analysis study with Australian children (Manly, Nimmo-Smith, Watson, Anderson, Turner, & Robertson, 2001), could be replicated with an American sample, or alternatively if a four factor model, with the addition of divided attention, would better explain the covariance structure of this study's data. An additional objective addressed in this study was whether the three-factor model could be improved by using raw scores while taking the effects of age and gender into account compared the three factor model using scaled scores. A two factor model was also explored due to high correlations between the latent factors in the three factor model. Confirmatory factor analysis indicated that a two-factor model using age-scaled scores best explained the covariance structure in this sample's data, χ2 (26, N=111) = 34.65, p = .120, NFI = .79, NNFI = .89, CFI = .92. Whereas, the three-factor model using age-scaled scores was less desirable, χ2 (24, N=111) = 34.63, p = .074, NFI = .79, NNFI = .86, CFI = .91. Although not as strong as some of the comparative fit indices of the Manly et al. (2001) normative study, overall the indices of fit of this study's two-factor model yielded a better solution than the three-factor model. These results suggest that selective attention and control shift attention may not reflect separate constructs of attention as shown in the Manly, et al. (2001) study. Additionally, the use of age-scaled scores in the three-factor model was superior to raw scores with age and gender controlled, χ2 (24, N=111) = 42.07, p = .013, NFI = .71, NNFI = .75, CFI = .83. Furthermore, the four-factor model using age-scaled scores, χ2 (21, N=111) = 34.25, p = .034, NFI = .79, NNFI = .81, CFI = .89 was also less desirable than the two-factor model using age-scaled scores. Because this study confirms the ability to assess multidimensional aspects of attention, the TEA-Ch may be a valuable tool for practitioners and researchers. However, one possible drawback of the TEA-Ch is the hour required for children to complete its nine subtests. A briefer screening tool of the first four subtests of the TEA-Ch is suggested when time constraints arise. However, further analysis is recommended to determine if the four subtests in the TEA-Ch screening tool are optimal. Thus, additional research is needed with respect to shorter multidimensional assessments of attention to inform intervention and consequently improve the quality of life for children with attentional differences.
Description
Rights Access
Subject
assessment
attention
attention deficit hyperactivity disorder
attention disorders
children
factor analysis