Comparing Measures From Computer-Administered and Examiner-Administered Narrative Retells in Spanish: A Pilot Study.

Abstract

Narrative language sample analysis (LSA) is a recommended best practice for the assessment of monolingual and bilingual children. With business-as-usual narrative LSA, examiners are actively involved in all aspects of the elicitation. Software advancements have shown multiple benefits of computer-administered language assessments, some of which may be beneficial for narrative assessments, particularly for bilingual children. The goal of this pilot study was to test the feasibility of computer-administered narrative retells in bilingual children. Ten English-Spanish bilingual children, kindergarten to fourth grade, completed two narrative retells using wordless picture books (Frog Goes to Dinner and Frog on His Own) in two conditions: examiner-administered and computer-administered. Five narrative measures were generated from these 20 transcripts. Significant, strong correlations were observed between the two elicitation methods for four of the five measures. We completed a series of Wilcoxon signed-ranks tests and found no significant differences in measures across the elicitation methods. Follow-up descriptive analyses revealed few large differences across elicitation methods for the individual participants. This study provides preliminary evidence on the use of a computer-administered narrative procedure and motivates further research on the method to confirm its validity and to document its effectiveness within clinical practice. https://doi.org/10.23641/asha.20346648.