Measuring Written Language Ability in Narrative Samples

Abstract

A language-levels model was used to analyze written narratives. The results have implications for instruction and for documenting change in the story-writing performance of students with and without special needs in the area of written language development. This study was based on stories written at midyear by 277 students (224 typical learners and 53 children with special difficulty in the area of written language) across grade levels one through five. Measures that differentiated students by grade level and special difficulty are reported by language level. Discourse-level measures were total words, total T-units, and story scores. Sentence-level measures were sentence codes (simple or complex, correct or incorrect), mean length of T-unit, and number and types of conjunctions. Word-level measures were number of different words and proportions of words spelled correctly. Gender differences were found for productivity measures in favor of females; no differences were found when comparing stories written by African American and Caucasian students.

This work was supported by grant H324R980120 from the U.S. Department of Education, Office of Special Education Programs, to N. W. Nelson and C. M. Bahr, and grant H325H010023 to Y. Hyter, N. W. Nelson, M. Clark, and A. M. Van Meter. The authors wish to thank the numerous teachers and graduate students who assisted with this work, particularly Kylee Biddle, Karey Hill, Kristen Kopacz, Carrie Kopitzki, and Brandi Newkirk.

Notes

∗T significantly different from S at p < .05.

†T significantly different from S at p < .01.

∗T significantly different from S at p < .05.

†T significantly different from S at p < .01.

∗T significantly different from S at p < .05.

†T significantly different from S at p < .01.

Abbreviations: SI = simple incorrect; SC = simple correct; CI = complex incorrect; CC = complex correct.

∗T significantly different from S at p < .05.

†T significantly different from S at p < .01.