Oral Reading Skills of Children with Oral Language (Word-Finding) Difficulties

Abstract

We examined how children with and without oral language (word-finding) difficulties (WFD) perform on oral reading (OR) versus silent reading recognition (SRR) tasks when reading the same words and how lexical factors influenced OR accuracy, error patterns, and nature of miscues. Primary-grade students were administered an experimental reading measure. Words were controlled for lexical factors known to influence oral language, such as frequency, lexical neighborhood, familiarity, and phonotactic probability. For learners with WFD, SRR was superior to OR; lexical factors predicted OR success; WF error-patterns emerged in OR; and miscues were higher in frequency, more familiar, and from denser neighborhoods than targets.

Notes

1. Although a more typical protocol is to use Kucera and Francis's (1967) norms to determine frequency of occurrence, we thought adult-based norms might not be the best measure of the frequency with which a child encounters any given word. The Carroll et al. corpus is based specifically on reading material intended for children.

2. Originally a seventh reading analysis section was administered, syllable dividing. Learners earned low accuracy scores on both oral reading (58%) and corresponding SRR (66.6%) tasks. We therefore concluded that this skill was above the instructional level of our participants and removed it from the battery.

3. Note that while we wrote “koy” in the text, these were said aloud, so there was no apparent mismatch between the spelling of the sequence in isolation and in the target.

4. Two children had perfect scores in oral reading for their sections and thus were not tested on any items silently for that section; they were not included in the analysis.

5. Form-related blocked errors were studied only in the regression analysis, since while there were a substantial number of these errors, they were made by only a small number (3) of participants.

6. Although nonword miscues emerged, they were excluded from this substitution analysis because meaningful comparisons could not be conducted: familiarity and frequency values could not be calculated for items that did not result in real words. Further, miscues that were words in the adult lexicon but may not have been meaningful to the child (did not have a U-value of at least .1) were considered “nonwords” for the children and thus not considered in this analysis (candor for canary). Miscues of a purely syntactical nature, where the free morpheme was the same, were also excluded from these analyses as neighborhood values for the reading word (jumped) and miscue (jumps), determined on the base morpheme, were identical.