To determine the relationship between oral and written language conventions at the macro and micro levels, a study analyzed various elements of stories generated by children. Subjects, nine above average readers between grades 4 and 9, were asked to tell a story and then to write a story. Sources of stories, overall plot organization, and number and kind of characters represent the macrostructures analyzed. Microstructures analyzed were story conventions, clausal connections, T-units, and type-token ratios. Results indicated that subjects' language mode sensitivity was reflected at the morphological level in the microstructures of their written stories. Observations of children's language usage suggested that a distinction between oral and written vocabularies may be superfluous and that the distinction lies, instead, between receptive and expressive language. Analysis showed that children differentiate between oral and written texts when using fillers. Results indicated that students seldom use the same macrostructures for both their oral and their written stories, suggesting that they develop a repertoire of macro story elements from which they select when generating stories. Results were consistent with M. Ice's longitudinal study. (Statistical tables and a breakdown of each story element are appended.) (JD)