Attempt to automate the Self-explanation strategy instruction described in McNamara, D. (2004). SERT: Self-explanation reading training. Discourse Processes, 38(1), 1–30, via a computer-based training module, iSTART. Students who received training (N=17) produced more and higher-quality self-explanations than those who didn’t (N=21). Treatment improved performance on text-based test questions for students with lower knowledge of reading strategies, and on inference questions for those with higher knowledge of reading strategies.
Unlike (McNamara, 2004), no analysis is made based on students’ prior knowledge of the subject matter. The improvement seen to bridging performance is much higher than in the 2004 paper.
Having now read the questionnaire (Schmitt, 1990) they used to assess awareness of reading strategies, I find myself somewhat skeptical. The questionnaire is only partly aligned with the SERT strategies, and many of the items strike me as contrived and questionable. I haven’t dug into the validity literature for this questionnaire, but I suspect that a valid interpretation might be that it’s at least in part detecting “conscientious students” rather than anything specific about reading strategies. It’s also interesting that 7 of the 25 items pertain to prediction, which (McNamara, 2004) finds doesn’t correlate with comprehension.
Q. Main finding?
A. Students who received iSTART produced more and higher-quality self-explanations than control. iSTART improved text-based question performance for students with lower knowledge of reading strategies, and inference question performance for those with higher knowledge of reading strategies.
Schmitt, M. C. (1990). A Questionnaire to Measure Children’s Awareness of Strategic Reading Processes. The Reading Teacher, 43(7), 454–461. ~http://www.jstor.org/stable/20200439~