| This study examines a corpus of elementary student spoken and written language. Seven major categories of texts are included in the corpus: Spoken interactions, spoken monologues, science textbooks, social studies textbooks, basal readers, children's literature, and student writing across a range of topics. The texts were entered into the computer and tagged for 67 different linguistic characteristics, then a second computer program compiled frequency counts for each of the linguistic features in each text. A factor analysis was then used to identify the salient co-occurrence patterns among linguistic features; that is, the ways that linguistic features function together as underlying 'dimensions' in student language. Based on the factor analysis, five dimensions of fifth grade language were identified. This model of fifth grade student language then served as a basis for investigating writing development in grades three, five, and six, examining the influence of task, first language, and demographic variables on student writing in these grades.;The study identified five major dimensions of student language: 'Edited informational' versus 'on-line informational'; 'Lexically elaborated narrative' versus 'Non-narrative'; 'Involved personal opinion' versus 'Non-personal uninvolved discourse'; 'Projected scenario'; and 'Other-directed idea justification'. Similarities and differences between student spoken and written language and the language of textbooks are also explored. Using this model of student language, patterns of writing development in third, fifth, and sixth grade across various writing tasks are described. Results also indicate that school serves to equalize many of the linguistic differences as students progress through the grades. |
