Computer-Based Assessment of Word Knowledge in Teens With Learning Disabilities

Computer-based and standard administrations of the Test of Word Knowledge (TOWK)-Level 2 core subtests (Word Definitions, Synonyms, Figurative Usage, Multiple Contexts) (Wiig & Secord, 1992) were compared for 30 subjects with learning disabilities. Half completed the computer-based version first and half the standard version first. Three weeks later, subjects were given either the standard or computer-based version in a counterbalanced design. The total, receptive, and expressive composites and three subtest means were highest for the standard administration (p<.05). Correlation coefficients (r) ranged from .81 to .88 for composites (p<.01). Confidence intervals (at 90% level) overlapped in 93% for the total, 97% for the receptive, and 93% for the expressive composite. Paired total scores were within the same diagnostic category (below or at/above normal) for 87%, receptive for 83%, and expressive for 90% of the subjects. Differenes in task formats influenced the results, indicating the need for independent norms for computer-based adaptations of standardized language tests.

KEY WORDS: computer-based, assessment, word knowledge, adolescents, learning disabilities

Submitted on March 10, 1994
Accepted on November 4, 1994

CiteULike    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?