The effect of unknown words and the semantic structure of texts on reading comprehension

Abstract

The correlation between word knowledge and reading comprehension is well established but the nature of the relationship remains unclear. It involves complex interactions among factors such as the proportion of unfamiliar words in the passage, the extent of redundancy associated with these words, their relative salience to the semantic organization of the text and the readers‘ proficiency in responding to these text characteristics. This study is one of the few which have investigated these interrelationships. It sought to determine whether good readers differed from poor readers in utilizing redundancy in text to construct meanings for unknown words and whether good readers and poor readers were more likely to so use redundancy when unknown words were salient to the story line rather than peripheral. The study further investigated how the degree of salience and redundancy of unknown words affected recall of the gist of a story and the total story information recalled by good and poor readers. An initial test of fifth grade pupils' recall of the gist of an aural story was used to eliminate from the study those pupils whose scores indicated an inadequately developed schema for simple stories. This test revealed a significant difference between good and poor readers (categorized according to standardized reading test scores) and wide variation among the scores of poor readers, indicating support for previous research which suggested that certain subgroups of poor readers had inadequately developed story schemata. Data from pupils whose aural recall scores did not reach criterion were not included in the subsequent analyses. The good and poor readers later read experimental versions of two stories containing six percent unknown words of varying redundancy and salience. Tests of the meanings of these words revealed redundancy effects for only one story. In this story good readers and poor readers used redundancy to construct meanings for unknown words but good readers scored higher and poor readers did not utilize redundancy when the unknown words were salient. This was attributed to the increased processing demands of salient compared with peripheral unknown words and the greater flexibility of good readers in coordinating multiple processing goals in reading. Implications for the teaching of context clues were discussed. Recall scores for one story only were relevant since, in the other story, good readers proved to be familiar with target words which were assumed to be unknown. For the story which did contain unknown words, good readers' gist and total recall were impeded only when unknown words were salient and not redundant. Although poor readers recalled less of the gist and total story information, their pattern of gist recall was somewhat similar to that of good readers but, by contrast, the relative salience and redundancy of unknown words had no differential effects on poor readers' total recall scores. Implications for assessment of reading based on miscue analysis procedures were discussed.