THE ABILITY OF SECOND AND THIRD GRADE STUDENTS TO DRAW INFERENCES AFTER READING SELECTIONS OF PROSE

Comprehension is difficult to investigate because it occurs at a number of cognitive levels which are integrated with the reader's prior knowledge. In particular, the ability of young readers to draw inferences after reading passsages of narrative prose has been inadequately researched. This study examined to what extent second and third grade children make inferences after orally reading two 400 word second grade stories. Variations of reading ability, grade, and sex were analyzed with subject responses to different types of inferential questions.;The results showed: (1) No differences exist on grade or sex variables and young readers' ability to make inferences, or in their use of prior knowledge. (2) Skilled readers were better than less skilled readers at making all of the types of inferences analyzed except those classified as motivational questions wherein no difference was found. (3) Approximately 70% of each type of inference was answered correctly except those classified as motivational which were answered correctly 49% of the time.;The primary finding was that young readers, who have their decoding skills intact, can answer most types of inferential questions at a satisfactory instructional level. Subjects experienced most difficulty with motivational inferences which specify a character's voluntary goals, thoughts, and actions. Motivational inferences might be further studied with interactions that help children focus on this level of comprehension.;Using a procedure developed by Warren, Nicholas, and Trabasso (1979), the test passages were diagrammed to control the formation of questions representing four types of logical inferences (motivational, physical causation, psychological causation, and enablement). Value, another type of inferential question, was included and studied through examiner interaction about characters and their actions. Sixty students, 30 from each of grades two and three, were screened to insure their competence in the oral reading of the test passages. Each subject was asked 24 inferential questions. In the analysis, these questions were reclassified on the basis of whether a correct response could be inferred from the text (text-connecting) or could be inferred from the reader's prior knowledge.