| Abstract: | Four types of study relating to the sensitivity of cloze to intersentential constraint are reviewed: (1) factor analytical and correlational studies; (2) studies which varied the length of available context; (3) studies which varied the mode of presentation; (4) studies which varied the quality of coherent context. Many of the investigations failed to provide convincing evidence that cloze is sensitive to intersentential constraint. However, this could be explained in part by weaknesses in the studies themselves. The factor analytical and correlational studies often failed to employ an adequate measure of global comprehension. While other studies demonstrated the importance of immediate context they did not specifically address the question of intersentential constraint. Quality of context needs to be considered separately from length of context. The quality of context studies provided conflicting results. Again, some of these investigations have serious weaknesses which tend to prejudice the results. The most reliable recent studies suggest that cloze is sensitive to intersentential constraint. Finally, tentative recommendations are made for establishing testing conditions which are most likely to encourage the use of intersentential constraint. |
