Measuring written linguistic accuracy with weighted clause ratios: A question of validity

•Weighted clause ratios which are based on the construct of communicative adequacy are introduced as a possible measure of linguistic accuracy for research purposes.•Weighted clause ratios are tested for validity against two widely used measures of linguistic accuracy: Error-free T-units, Error-free clauses.•Many-Facets Rasch Measurement was used to identify a number of advantages and disadvantages associated with these measures in terms of their validity.•Appropriateness of each measure seems to be highly context dependent.

Determining linguistic improvement in L2 writing requires a precise measure of linguistic accuracy. Although numerous metrics of linguistic accuracy have been used in L2 research, Wigglesworth and Foster (2008) proposed a new kind of measure—a weighted clause ratio—based on the adequacy of the writer's conveyed meaning. This paper evaluates the validity of this metric and compares it to two of the most similar measures of linguistic accuracy currently in use: the error-free T-unit ratio and the error-free clause ratio. The data collected and analyzed in this study were drawn from over 350 writing samples generated by 81 ESL writers whose language abilities range from low or intermediate to advanced. To provide baseline, comparative data, this study also analyzed writing samples from 16 native English-speaking students. This study utilized Many-Facet Rasch Measurement and other analyses to identify variables affecting the validity of the weighted clause ratio.