UBC Theses and Dissertations
The color-coded cloze procedure : a method to assist adult ESL students in searching for clues to fill in cloze blanks Labrum, Howat Alan
Research by Shanahan and Kamil (1982, 1983, 1984) suggested that students did not use inter-sentential information to improve their doze test scores. Chihara and Oiler (1977) had discovered in their research with more proficient ESL students the contrary. Building on the idea that the issue of too-local reading depended on both the proficiency and motivation of the students along with the availability of beyond—sentence clues, the present research project was designed using color-coded blanks and parts of speech to highlight inter-sentential clues, especially re iterative—word clues. The purpose of the color—coding was to see if, by differentiating blanks and words, students could focus more sharply on the necessary information, primarily re—iterative words that occurred beyond the sentence of each doze blank, and use these words to fill the doze blanks. The students in the treatment group were given a key to the relationship between the color and the parts of speech but were not explicitly told to look for re-iterative-word clues. It was expected, however, that each colored blank would stimulate the students to look through the passage to find a word of the same color to fill the said blank. It was also the goal to maximize the number of inter—sentential re—iterative clues and see how cloze scores would be affected when color— coding was used as opposed to no color—coding (the major control). If the inter—sentential clues were present, then color—coding should have resulted in the treatment group having better scores. Randomly—colored words were used as another control to determine whether any positive effect had come from the coding or just from color as a motivational factor. Lastly, the pre- and post—tests were non—colored in order to see if the color—coding treatment had just a temporary means of help which led to no lasting gains once removed.. A pilot project was done with a class of twenty adult upper—intermediate ESL students using four forms of a standard doze test. Based on the results the colored—blank form was dropped so the three classes in the main project could each have one form. To suit the needs of the more advanced university students, new stories were chosen and prepared using a rationalized doze to maximize the number of re-iterative—word clues. In the main research project most of the random-color group (the least proficient group) dropped out after the pre—test. The non—colored rational doze group received higher scores than the color—coded treatment group on all of the tests. When the mean scores were graphed both these groups made steady progress from practice test to practice test, the treatment group appearing to almost catch up. Improvement was made from pre- to post—test by both groups but less by the treatment group, especially when only inter—sentential blanks were counted. Generally speaking, t—tests and a very sensitive statistical program (“One Between and One Repeated Measures Factor ANOVA”) confirmed this improvement but showed that the color—coded doze treatment group and the non-colored doze control group in most cases did not differ significantly. The graphical analyses of the results were more optimistic in favor of the color-coded treatment than the statistical analyses were but the small sample size (N = 13) made the statistical findings unclear at times. Improvement in some cases may have been because of the declining readability levels of subsequent passages. The number of blanks filled, the number filled correctly, and the relationship of these two were analyzed to determine the confidence and productive confidence levels of each group. Results showed the color—coded treatment group were less confident in filling blanks and made limited gains in productive confidence over the control group.
Item Citations and Data