Length of the writing samples Fig. 3. Seventy percent of words in common with source. D. Pecorari/Journal of Second Language Writing 12 (2003) 317-345 Fig. 5. Forty percent of words in common with source. it permits greater depth of response. Unattributed repetition from source by passage percentage score reported above was recalculated for each writing sample as @ whole. This ‘“‘broad context’ figure was arrived at by dividing the number of words in common by the total number of words in each sample. Once again. signaled quotations were excluded. Table 4 shows that there was considerable variance among the writers on this point. At the extremes, only 4% of the words not in quotation marks in Maria’s text came from her sources, while Ingrid’s writing sample, which contained no signaled quotations at all, took 92% of its language from sources. Unattributed repetition as a proportion of total text, by writer (quotation excluded) Unattributed repetition as a proportion of compared text, by writer (quotation excluded) Opaque source use features by writer
