A comparison of the effects of oral and written teacher feedback with written teacher feedback only on specific writing behaviors of fourth grade children
The purpose of the study was to compare effects of two types of teacher feedback on two specific writing behaviors of fourth grade children.Experimental design of the study compared two groups using pretest/posttest measures: (1) Experimental Group I contained two intact classrooms with a total of forty-four fourth grade children. Treatment for this group involved having children write stories on a twice-weekly basis and receive a written and oral comment from the writing teacher about production of the stories. (2) Experimental Group II contained two intact classrooms totaling fifty-one fourth grade children. Treatment for this group involved having children write stories on a twice-weekly basis and receive only written comments from the teacher about production of the stories. No control group was used in this study.Objective data obtained for analysis were derived from the measurement instrument, a rating scale, specifically developed for this study. The instrument measured four writing behaviors--originality, vocabulary, content and mechanics. Only content and mechanics were used in treatment. Each of these behaviors were rated on a seven-point scale, with seven the highest and one the lowest. The highest rating indicated maximal use and the lowest minimal use of a specific writing behavior.The instrument appeared to have a high face validity. Examination of the items on the rating scale showed that they dealt with observable writing behaviors. To establish reliability of the instrument all compositions written were evaluated by a set of raters who were subjected to a series of training sessions in use of the instrument. Further, checks on reliability were made during seven of the fourteen writing sessions in the study. The reliability obtained by using the Cronback-Alpha Internal Consistency Index was obtained for each of the writing behaviors. Reliability coefficients were (1) Originality .99, (2) Vocabulary .84, (3) Content .84, and (4) Mechanics .62.A writing teacher was hired and trained by the experimenter to use a standard format for stimulating writing behavior. Once compositions were rated by the raters two controlled written comments were placed on compositions by the experimenter. Control over comments were achieved through development of a series of prepared feedback guides. A guide was prepared for the two writing behaviors subjected to experimental manipulation: content and mechanics. Comments were corrective in nature and contained information about content and mechanics of the composition. Comments also gave the reader suggestions for composition improvement.When compositions were returned in each experimental group treatment was administered by the writing teacher. In Experimental Group I each of the children was given an oral comment, in the form of a brief conference with the writing teacher, which accompanied the written comment. In Experimental Group II all of the children were told to look at written comments, no further reference was made about the compositions. A general statement of priase about the classes compositions as a whole was given to both groups.The following hypotheses were tested:Hypothesis I. There is no significant difference between the adjusted posttest meansof Experimental Group I and Experimental Group II for the writing behavior content.Hypothesis II. There is no significant difference between the adjusted posttest meansof Experimental Group I and Experimental Group II for the writing behavior mechanics.To test the hypothesis an analysis of covariance was used. The pretest served as the covariate and the posttest as the criterion. For each hypothesis the obtained F-ratio did not exceed the tabled F-value for significance at the .05 level of rejections. The null hypotheses were not rejected.