Psychometric Properties of the Danva-2 in High-Functioning Children with ASD

Poster Presentation
Saturday, May 4, 2019: 11:30 AM-1:30 PM
Room: 710 (Palais des congres de Montreal)
A. J. Booth, C. J. Rajnisz, J. D. Rodgers, S. L. Andrews, M. Thomeer and C. Lopata, Institute for Autism Research, Canisius College, Buffalo, NY

Impairment in facial emotion recognition (FER) has been documented in high-functioning individuals with autism spectrum disorder (HFASD) across a variety of emotions, research methodologies, and assessment tools (Harms, Martin, & Wallace, 2010). Some have proposed that these deficits contribute to the social difficulties that define ASD (Schultz et al., 2003). One of the most frequently used FER measures in HFASD studies (McMahon, Lerner, & Britton, 2013) is the Diagnostic Analysis of Nonverbal Accuracy – Second Edition (DANVA-2; Nowicki & Carton, 1993). However, despite its use in this population, no studies examining the psychometric characteristics within a homogenous HFASD sample have been conducted. Examination of the validity and reliability of the DANVA-2 within samples of children with HFASD would add to the interpretability of previous research utilizing the measure and support future use.


This study assessed the (1) internal consistency, (2) test-retest reliability, and (3) near and broad concurrent validity (comparisons with other established FER and social functioning measures) of the DANVA-2 faces subtests. In addition, this study also explored the relationships between performance on the DANVA-2 and child age, intelligence, and language abilities.


Participants. One-hundred twenty-one children, ages 6-13 years, with HFASD were included in this study. Analysis subgroups were formed from the total sample (N = 121) for concurrent validity (n = 36) and test-retest reliability analyses (n = 21 for 5-week and n = 21 for 12-week stability). Each child had a prior clinical diagnosis of ASD, a WISC-IV abbreviated IQ >70 (VCI or PRI >80), and a CASL expressive or receptive language score >75. All diagnoses were confirmed using the ADI-R.

Measures. DANVA-2; Cambridge Mindreading Face-Voice Battery for Children (CAM-C); Social Emotional Evaluation (SEE); Behavior Assessment System for Children, Second Edition-Parent Rating Scales (BASC-2-PRS); Social Responsiveness Scale (SRS); Wechsler Intelligence Scale for Children – Fourth Edition (WISC-IV); Comprehensive Assessment of Spoken Language (CASL).

Procedures. All data were collected as a part of the screening and baseline assessments for psychosocial treatment trials. Test-retest reliability samples for the 5-week and 12-week intervals were obtained by using participants from no-treatment (wait-list) conditions.


Internal consistency estimates were .70 for the DANVA-2-AF and .75 for the DANVA-2-CF subtests. For the DANVA-2-AF, reliability coefficients for immediate, 5-week, and 12-week retest intervals were .84, .75, and .68, respectively. For the DANVA-2-CF, reliability coefficients for the immediate, 5-week, and 12-week retest intervals were .78, .90, and .43, respectively. DANVA-2 scores were strongly associated with the CAM-C and SEE but unrelated to parent BASC-2 and SRS ratings. Significant correlations (small to moderate) were found between DANVA-2 scores and child age, IQ, and language ability.


Internal consistency and test-retest reliability of the DANVA-2 in an HFASD sample were found to be adequate for research purposes. Likewise, the measure strongly converged with other tests of emotion recognition. However, further research is needed to determine whether the DANVA-2 adequately predicts social competence and ASD symptomology and to what extent participant demographic characteristics may impact DANVA-2 performance.

See more of: Emotion
See more of: Emotion