Some Notes on Commercial Language Teaching Progams
S Krashen. April 9, 2020
As noted in “Amid coronavirus, students flock to Kahoot! and Duolingo. Is it the end of language teachers?” (USA Today, April 7) online language teaching companies are advertising their products aggressively these days. A look at the published research on the independent use of the most popular programs, however, shows mixed results.
Rosetta Stone appears to be is about as effective as traditional language instruction, which is nothing to brag about. The results of the studies do not support Rosetta Stone’s advertised claim that it is “The fastest way to learn a language.”
About 80% of former Duolingo students said they were “satisfied” with the course, but only 66 out the 156 who started the program finished it and responded to the questionnaire. I calculated that it would take Duolingo students about 60 hours to achieve what regular college students do in one semester .(45 hours of classtime).
Babbel did better. Based on the results of two months of Babbel study, researchers estimated. that Babbel users could accomplish about one semester’s progress in 21 hours of Babbel. Nearly all users said they liked Babbel.
In all the studies I examined, conclusions are based on performance of the WebCAPE examination, and unfortunately we don’t know much about this test, other than that it is multiple-choice: As far as I could tell, sample questions are not available.
Few other evaluations of commercial programs were available.
Rockman, et. al. S. 2009. Rosetta Stone Evaluation Report. resources.rosettastone.com/.../Rockman-Evaluation-Report.pdf
Vesselinov, R. and Grego, J. 2012. Duolingo effectiveness study. Final Report. static.duolingo.com/s3/DuolingoReport_Final.pdf
Vesselinov, R. and Grego, J. 2016. The Babbel efficacy study. https://press.babbel.com/en/releases/downloads/Babbel-Efficacy-Study.pdf
Vesselinov, R., 2008, Measuring the Effectiveness of Rosetta Stone. http://resources.rosettastone.com/CDN/us/pdfs/Measuring_the_Effectiveness_RS-5.pdf.