Validity and Reliability of the Transactional EXchanges in Texting for Aphasia Rating Scale

Text messaging is an increasingly relevant rehabilitation target for people with aphasia (PWA). Given the chronicity of aphasia and the likelihood that communication through this modality will remain a mainstay of daily life, it is critical to develop assessment tools to capture the texting abilities of PWA. This study aimed to evaluate the validity and reliability of the Transactional EXchanges in Texting for Aphasia (TEXT-A) rating scale.

Participants included 60 individuals with poststroke aphasia (40 females, 20 males). Performance on a texting script was evaluated using the TEXT-A's three-category rating in which turns elicited from participants with aphasia received a score of 0 (no transaction of message), 1 (partial transaction), or 2 (successful transaction). We assessed (a) interrater reliability using Krippendorff's alpha, (b) internal consistency using greatest lower bound and Cronbach's alpha, and (c) criterion validity using Spearman rank correlation coefficient to evaluate the relationship between scores yielded from the scale and Western Aphasia Battery-Revised Aphasia Quotient (WAB-R AQ), Language Quotient (LQ), and the Writing subscale score.

The average Krippendorff's alpha between each of the raters and the gold standard (scores from first author) was .902 (min α = .855). The adjusted greatest lower bound for both the expert rater and the rater associated with the lowest reliability estimate from the previous analysis was .736 while Cronbach's alpha was equal to .73 (95% confidence interval [.61, .85]). Spearman correlations between the total TEXT-A score and the WAB-R AQ, LQ, and Writing subscale scores were .58, .73, and .73, respectively (for all correlations, p < .001).

This study extends our initial evaluation of the TEXT-A, applying the scale to a larger, more heterogeneous sample. Findings suggest excellent interrater reliability, acceptable internal consistency, and moderate-to-strong criterion validity of the TEXT-A measure based on the relationship between TEXT-A and WAB-R scores.