Project LISTEN Reading Tutor — Carnegie Mellon's Speech-Recognition Oral Reading Tutor Tested in a Randomised Trial

Project LISTEN's Reading Tutor listens to children read aloud using speech recognition and gives individualised oral reading practice and feedback.
Mostow, Nelson-Taylor and Beck (2013, Journal of Educational Computing Research) ran a randomised controlled trial in two elementary schools of one affluent suburban district: 193 pupils in grades 1-4 were randomly assigned and 178 contributed to the findings. Pupils averaged 19 hours on the tutor over 7 months, versus 20-25 minutes of daily sustained silent reading in class.
The What Works Clearinghouse rated the study as meeting standards without reservations because of low attrition, and rated the evidence Promising (Tier 3) for literacy achievement and phonics. Significant advantages were found on the Woodcock Reading Mastery Test total reading composite (480.01 vs 471.49), the basic skills composite (477.80 vs 463.25) and word identification, plus phonemic blending and spelling at one month. Differences in reading comprehension, oral reading fluency, word attack and reading attitude were not significant.
Limits: the rating rests on a single study in two high-resourced schools, so transfer to other contexts is unproven.

Read the full analysis: https://ri.cmu.edu/?p=8210