Reliability and Validity of the Positive and Negative Affect Schedule (PANAS) for Individuals with Autism Spectrum Disorders

Background: Individuals with autism spectrum disorders (ASD) present a higher prevalence of internalizing symptomatology (i.e. depression and anxiety) than the general public (Mattila et al., 2010). The Positive and Negative Affect Schedule (PANAS; Watson et al., 1988) was created to measure affective states using a dimensional approach. The authors of the measure assert that positive items are specifically related to depression and not anxiety, and that negative items are highly related to both depression and anxiety (tripartite model). The PANAS has been widely used in both clinical and non-clinical populations (Crawford & Henry, 2004; Tuccitto et al., 2009), but very few researchers have utilized it in a population of individuals with ASD. 

Objectives: Because the measure has not previously been validated for the ASD population, this study investigated the reliability and validity of the PANAS for this population.

Methods: The PANAS consists of 10 negative and 10 positive mood questions, rated on a 5-point scale (0=not at all, 4=extremely). The two subscales have been shown to have sound psychometric properties, and to have high levels of internal consistency and convergent validity. Inclusion in an ongoing longitudinal study at the University of Michigan indicated that parents were able to complete this measure on their child, regardless of level of functioning. Participants for this study were selected from a database of an ongoing longitudinal study based at The University of Michigan Autism and Communication Disorders Center (UMACC).  In this study, parents completed the PANAS with regard to their child for up to 4 waves [Parent N=520; Proband M age=17.8 (SD=1.8); Proband M VIQ=53.9 (SD=40.9)]. In addition, more able (MA) probands (VIQ>=70) completed the PANAS about themselves at up to 5 waves [N=145; M age=18.2 (SD=1.4); M VIQ=99.8 (SD=17.0)].   

Results: Internal consistency of the scale was computed for both the parent (α=.80) and proband (α=.82) versions of the PANAS. Test-retest reliability on positive (Parent r=.62-.76; Proband r=.16-.57) and negative (Parent r=.56-.65; Proband r=.53-.90) items was good. To test convergent validity, correlations with the PANAS scores were computed for parent-report measures: Children’s Depression Inventory (CDI; Functional Pos r= -.57*, Neg r=.47*; Emotional Pos r= -.55*, Neg r=.65), Adult Manifest Anxiety Scale (AMAS-A; Pos r= -.07; Neg r=.53*), and the Quality of Life Questionnaire (QoL-Q; Pos r=.52*; Neg r= -.22*). Correlations for MA self-report scales were as follows: BDI-II (Pos r= -.30*; Neg r=.65*), AMAS-A (Pos r= -.24*; Neg r=.63*), QoL-Q (Pos r=.22*; Neg r= -.43*), and the Well-being Scale (Pos r=.10; Neg r=.18).

Conclusions: The tripartite model of the PANAS was supported for the parent-report PANAS (PANAS-P). Positive items on the PANAS-P correlated significantly with QoL and Well-being questionnaires in the expected direction. The tripartite model was mostly supported for the MA self-report PANAS, but the total anxiety score correlated significantly with the positive items of the PANAS-MA. QoL scores, but not Well-being scores, correlated significantly with positive items on the PANAS-MA. This measure also showed high internal reliability, and is appropriate to use with ASD populations. Factor analysis results will also be reported.