Wagner, Keith Anthony2024-08-152024-08-1519831983https://hdl.handle.net/1828/20017This study examined the relationship between the level of abstraction from raw experience in children's written composition and the complexity and types of syntactic structures used. Level of abstraction was defined in terms of five writing modes. The modes, from least to most abstract, were as follows: description, factual narration, fictional narration, exposition, and argument. It was hypothesized that each mode would elicit a higher level of syntactic complexity, as measured by I-Unit length and by the number of sentence-combining transformations per I-Unit, than the next lower mode on the hierarchy. It was also hypothesized that the frequency of occurrence of specific types of transformations per I-Unit would be different in each mode. Sex was predicted not to be a significant moderator variable. Six Grade Six teachers in four elementary schools in the Public School District of Grande Prairie, Alberta, Canada volunteered to participate in this study. Each teacher had his students write two compositions in each of the five modes of writing, in accordance with plans developed by the investigator. The order of modes was randomized for each school. Those students who did not write all compositions were eliminated from the study. A random sample of three boys and three girls from each of the six classes was chosen. The investigator conducted a detailed syntactic analysis of each of the three hundred sixty compositions. The resultant research design was a 2 x 5 (sex x mode) factorial design. Two way analysis of variance with repeated measures on the mode variable was used to determine significance, and the Scheffé technique was used to determine significance between cells. The results confirmed that different modes of writing do yield differing levels of complexity, but not consistent with the directions hypothesized. Using T-Unit length as the measure, the result was as follows: Argument > Exp. > Desc = Factual Narr. > Fictional Narr. Using sentence-combining transformation per T-unit as the measure, the result was as follows: Argument = Exp. = Desc > Factual Narr. > Fictional Narr. It was also found that different modes of writing do yield differing frequencies of specific transformations per T-Unit. Sex was found not to be a significant moderator variable. It was concluded that there may not be a direct relationship between level of abstraction and syntactic complexity. However, since the relationship between level of abstraction and modes of writing was only a theoretical assumption, it was suggested that further research which uses more direct measures of abstraction level should be done. The study was significant in that it suggests that if one wishes students to grapple with a full range of syntactic structures, they should be assigned writing in various modes. It also added further confirmation to other studies which have shown that developmental research on syntactic complexity must take mode of writing into account. This study also demonstrated that fictional narration and factual narration yield differing levels of syntactic complexity; no previous studies have examined the differences between those modes. This study used more extensive pre-writing experiences than did other similar studies. The subjects produced more complex sentences than what previous studies would lead one to expect. Therefore, the relationship between the nature of pre-writing experiences and syntactic complexity generally and within each of the modes is suggested as a fruitful area for further research.261 pagesAvailable to the World Wide WebThesis