THE PREDICTION OF END-OF-YEAR READING ACHIEVEMENT THROUGH CASE GRAMMAR CONSTRUCTS FOUND IN THE SPOKEN DISCOURSE OF BEGINNING FIRST-GRADE CHILDREN
Abstract
This study was done in order to investigate the relationship between the spoken discourse of beginning first-grade children and their reading achievement at the end of grade one. Specifically, the study addressed the following objectives: (1) Identification of patterns of case (verb/noun) relationships found in the spoken discourse of beginning first-grade children. (2) Determination of the reliability of a system of propositional analysis (Kintsch, 1974) for use in the study of children's spoken discourse. (3) Determination of the correlations of patterns of case relationships and related variables, found in the spoken discourse of beginning first-grade children, with end-of-year reading achievement. (4) Identification of differences in the spoken discourse of beginning first-grade children across three socioeconomic status groups. (5) Determination of the validity of the case-related variables examined for predicting reading achievement at the end of grade one. (6) Comparison of the semantically-based, case-related variables with measures used in prior studies of the relationship of spoken discourse and reading achievement in grade one. Data collection involved administration of the Recognition of Letters subtest of the Clymer Barrett Prereading Inventory (Clymer and Barrett, 1966), completion of an individual language skill rating scale by classroom teachers of the 82 subjects in the study, and an individual interview of each subject. A syntactic analysis was performed by computer and a semantic, propositional analysis was performed manually by the investigator. Variables used in prior syntactically-related studies were, corrected type-token ratio and average sentence/utterance length. Semantic variables involved in statistical analysis were; the nine highest frequency patterns of case relationships, a ratio of pattern type to its number of occurrences, average pattern length, and number of different patterns per subject. Additionally, the score of the Recognition of Letters subtest and teacher rating of language skill were examined. Reading achievement was measured with the Total Reading raw score of the Stanford Achievement Test: Reading (Madden, et al., 1973). Nineteen different patterns of case relationships were observed in the total of 2460 predication propositions. Five patterns, Agent, Agent-Object, Agent-Object-Goal, and Experiencer-Object, accounted for 97 percent of the total number of patterns analyzed. Results of the reliability check done on an augmented version of Kintsch's (1974) system of propositional analysis revealed agreement on 85 percent of 300 propositions analyzed by the investigator and a trained assistant. Results of correlational analyses showed that none of the case-related variables was significantly related to first-grade reading achievement and that of the spoken discourse measures, only average sentence/utterance length was related to reading achievement. One-way analysis of variance procedure revealed that the three socioeconomic status groups were not significantly different from one another with respect to any of the variables examined in the study. Regression analyses showed that the best predictors of first-grade reading achievement were knowledge of letter names and teacher ratings of language skill, which together accounted for 66 percent of the variance in reading achievement. Case pattern Agent-Object-Instrument, which added two percent to the variance explained, was the only other significant contributor to the regression model. Finally, only average sentence/utterance length was related to the case-related variables examined in the study.
Degree
Ph.D.
Subject Area
Literacy|Reading instruction
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.