The Relationship Between Predictive Reading and Predictive Spelling Strategies Using Cloze Tests

Abstract

The purpose of this research was to determine if there was a relationship between predictive strategies used in reading and predictive strategies used in spelling and to see if both could be measured using cloze tests. A secondary purpose was to see if there was a relationship between a spelling score on a standardized test and a score on a spelling cloze test; a relationship between a reading comprehension score on a standardized test and a score on a reading cloze test; a relationship between a standardized spelling test score and a score on a spelling word selection test; a relationship between a standardized reading comprehension test score and a standardized spelling test score; and a relationship between a spelling cloze score and a spelling word selection test score. The reading cloze test, spelling cloze test and the spelling word selection test were examiner-designed. The reading cloze test, consisting of forty-nine scored blanks, and the spelling cloze test, consisting of eighteen nonsense words with one or two-letter blanks per word, sought to determine if a subject used predictive strategies. The spelling word selection test consisted of twenty-five groups of three pseudo words and one nonsense word. The standardized spelling and reading comprehension test scores were taken from the Stanford Achievement Test. All tests were administered to a total of eighty-nine students: fifty-five regular sixth graders, fourteen gifted sixth graders and eleven gifted fifth graders who were in the same reading class, and nine learning disabled students who were not doing sixth grade work but were of sixth grade age. Class placement of subjects was determined by the school district. The reading cloze test was scored using synonyms as correct answers. The spelling cloze test was constructed using spelling rules and patterns but some answers which did not conform to these were also accepted if the nonsense word could be pronounced and if it looked like a real word. In each group in the spelling word selection test the nonsense word was the only correct answer. All test data (examiner-designed and standardized) were analyzed using raw scores. Results showed a significant linear correlation between all relationships studied. An informal analysis using averages was used on the time and the score from the examiner-designed tests. The examiner kept track of the time each test was begun and subjects recorded the time each test was completed. Class lists were used to break down the regular sixth graders into reading groups. The gifted fifth and sixth graders and the learning disabled class were already separate reading groups. The informal data analysis showed that on the average, the high reading group of regular sixth graders and the gifted fifth and sixth graders scored higher on all three tests. The medium reading group of regular sixth graders did not do as well and the low reading group of regular sixth graders and the learning disabled class scored even lower. The time factor appeared to have little impact on scores since often the poor readers took as much time as the better ones but still did not do as well.