Are STR & CAT-Generated texts useful for comprehension of lecturing content in a foreign language?

In this study, we applied Speech-to-Text Recognition (STR) and Computer-Aided Translation (CAT) technologies in combination during lectures in English to facilitate Russian speaking students' comprehension of lecture content. The instructor lectured in English, the STR system generated texts from the voice input, and the CAT system then simultaneously translated the STR-texts into Russian. We aimed to test the feasibility of applying STR and CAT technologies in combination to support learning of Russian speaking students during lectures in English. To this end, we designed an experiment. Three groups with twenty students in each were formed and all students attended the same lectures. Students in a control group attended lectures without any support, students in an experimental group 1 attended lectures with STR support (i.e. they were presented with texts in English generated from the instructor's speech), and students in an experimental group 2 attended lectures with STR and CAT support (i.e. they were presented with texts in Russian translated from English). We compared language ability and post-test results of students in three groups and we surveyed perceptions of students in the experimental group 2 about usefulness of translated texts for their learning. Our results showed that applying STR and CAT technologies was a useful approach; translated texts helped improve learning performance of the experimental group 2 students compared to the control condition significantly. More specifically, translated texts were beneficial both for low and high ability students in the experimental group 2, so that their scores of the tests were much better compared to those of students in the control condition. When students got used to translated texts and found their strength and limitations, they better utilized these texts for their learning. This is why learning progress of students was increasing from Lecture 1 to Lecture 2. Our results also showed that low ability students benefit from translated texts the most; their language ability was lower than that of high ability students before the experiment but after it, learning performance had no significant difference. Finally, most of the students in the experimental group 2 perceived translated texts as useful for their learning and they intended to use them in the future for learning.