Are there Errors in Error Analysis?

Publisher Website
Google Scholar
Abstract
This study examined (a) developmental changes from the end of kindergarten to the end of first grade and achievement group differences (top, middle, low) in errors on visual language tests (which involve letters or word-like stimuli, but do not require reading or spelling) and on auditory language tests; (b) relationships between these visual and auditory errors and spelling errors on a computerized test of written reproduction at the end of first grade; and (c) relationships between achievement group, stimulus type (phonically regular and irregular words, nonsense words, letter strings), and error types (omissions, substitutions, transpositions) on the computerized test (N = 27). Of 56 coded visual and auditory errors, 21 showed significant developmental changes and 11 showed significant achievement group differences. Only one error type - reversals - was correlated across two different tasks. Level of performance on visual and auditory language tests was correlated with level of performance on spelling achivement, but specific kinds of visual or auditory language errors were not, in general, related systematically to specific kinds of spelling errors on the computerized test. The interaction between achievement group and error type was not significant, but the three-way interaction among achievement group, error type, and stimulus type was significant. Although the computerized test showed a relationship between errors and information processing, the psychometric test results suggest that visual and auditory errors may not reflect stable, valid information processing constructs that ex- plain spelling errors. Clinicians are advised to use error analysis cautiously until its reliability and validity for psychoeducational assessment are demonstrated.