Assessing Verbal Ability in Children with ASD: Convergent Validity of the Ppvt-IV

Background: The Peabody Picture Vocabulary Test-IV (PPVT-IV; Dunn & Dunn, 2007) is a one-word receptive vocabulary assessment. Recent studies of individuals with Autism Spectrum Disorder (ASD) have used PPVT-IV scores as estimates of verbal or cognitive ability (e.g., Peterson et al., 2013; ). To date, the extent to which PPVT-IV scores represent a valid estimate of overall verbal or cognitive ability in individuals with ASD has not been examined.

Objectives: The current study examined the convergent validity between the PPVT-IV and standardized cognitive assessments (Mullen Scales of Early Learning (Mullen) and Diagnostic Ability Scales (Early Years (DAS EY) and School Age Forms (DAS SA)). Subject characteristics influencing the relationship between cognitive assessment scores and PPVT-IV scores were also examined.  

Methods: Participants included 2643 (86.7% male) children between 47-211 months (mean = 82.5, SD = 31.5) who participated in the Simons Simplex Collection. Age equivalents of full-scale IQ (FSIQ) and verbal IQ (VIQ) scores on the Mullen and DAS were correlated with PPVT-IV age equivalents using Pearson’s correlations. The effect of subject characteristics (age, gender, ADOS CSS score, and FSIQ) on the relationship between the PPVT-IV and VIQ was examined using Linear Mixed-Effects Models. 

Results: High positive correlations were found between the PPVT-IV and the full scale and verbal age equivalents of the Mullen (r = 0.83 and r = 0.82 respectively). Analyses also indicated high correlations between the PPVT-IV and the Mullen receptive and expressive verbal subscale age equivalents (r = .80 and r = .74).  Age was associated with differences between the PPVT-IV and Mullen verbal IQ (p < 0.05), such that PPVT and VIQ scores were more similar in younger participants.

For the DAS EY, results indicated moderate and high positive correlations between the PPVT-IV and FSIQ and VIQ scores (r = 0.59 and r = 0.76, respectively). FS IQ was a significant predictor of the difference between PPVT-IV and VIQ Z scores (p = 0.002), such that children with lower FSIQ scores had a smaller difference between PPVT-IV and VIQ scores.

For the DAS SA, results indicated moderate and high positive correlations between the PPVT-IV and DAS SA FSIQ and VIQ scores (r = .79 and r = .87, respectively).  FSIQ, ADOS CSS, and gender were significant predictors of differences in the PPVT-IV and the DAS School Age VIQ Z scores (p < 0.001).

Conclusions: Results indicated moderate and high positive correlations between FSIQ and VIQ scores and PPVT-IV scores, indicating a strong relationship between PPVT-IV scores and estimates of verbal and cognitive ability. The agreement between PPVT-IV and VIQ scores is affected primarily by level of functioning. This indicates that the PPVT-IV is a better estimate of overall verbal ability in lower functioning children with ASD. Researchers and clinicians who use the PPVT-IV to estimate overall verbal functioning should be aware that this practice is most valid for lower functioning children with ASD.