TABLE 1.
Studies | Sample size | Interaction style | Interaction mode | Interaction content | Assessment of learning outcomes | Effect size (r) |
Dikker et al., 2017 | 12 | FTF&NFTF | HTF<F | Conceptual knowledge | IQ | 0.595 |
Pan et al., 2018 | 12 | FTF | HTF | Song learning | post | 0.636 |
Pan et al., 2018 | 12 | FTF | LTF | Song learning | post | −0.200 |
Zheng et al., 2018 | 60 | FTF&NFTF | HTF<F | Mathematical knowledge | post vs. pre | 0.510 |
Cohen et al., 2018 | 18 | / | / | Conceptual knowledge | post vs. pre | 0.494 |
Liu et al., 2019 | 17 | FTF | LTF | Mathematical knowledge | IQ | 0.610 |
Liu et al., 2019 | 17 | NFTF | LTF | Mathematical knowledge | IQ | −0.080 |
Liu et al., 2019 | 17 | FTF | LTF | Mathematical knowledge | post | 0.730 |
Liu et al., 2019 | 17 | NFTF | LTF | Mathematical knowledge | post | −0.004 |
Nozawa et al., 2019 | 32 | FTF | LTF | Conceptual knowledge | IQ | 0.330 |
Bevilacqua et al., 2019 | 12 | FTF&NFTF | LTF | Conceptual knowledge | IQ | 0.382 |
Bevilacqua et al., 2019 | 12 | FTF&NFTF | LTF | Conceptual knowledge | post | 0.130 |
Davidesco et al., 2019 | 31 | FTF | LTF | Conceptual knowledge | post | 0.520 |
Zhu et al., 2019 | 16 | / | / | / | IQ | 0.622 |
Pan et al., 2020 | 24 | FTF | HTF | Conceptual knowledge | post vs. pre | 0.655 |
Pan et al., 2020 | 24 | FTF | LTF | Conceptual knowledge | post vs. pre | −0.210 |
Nguyen et al., 2021 | 20 | / | / | Conceptual knowledge | post vs. pre | 0.673 |
Sun et al., 2020 | 34 | / | / | / | / | 0.650 |
Sun et al., 2020 | 34 | / | / | / | / | 0.580 |
Zheng et al., 2020 | 60 | FTF&NFTF | HTF<F | Mathematical knowledge | IQ | 0.310 |
Pan et al., 2021 | 16 | FTF | HTF | Mathematical knowledge | post vs. pre | 0.510 |
Meshulam et al., 2021 | 24 | / | / | Conceptual knowledge | post vs. pre | 0.270 |
Zhu et al., 2021 | 24 | FTF | HTF | Conceptual knowledge | post vs. pre | 0.570 |
FTF, face-to-face; NFTF, non-face-to-face; HTF, High turn-taking frequency; LTF, low turn-taking frequency; Post, post-testing scores; Post vs. Pre, comparison between post-testing and pre-testing scores; IQ, interaction quality evaluation scores.
“/” indicates the coding is not applicable for the sample.
Interaction style and mode for offline scanning studies were not coded.
In addition, the study of Sun et al. (2020) was only conducted in the overall effect size analysis because the experiment setting is different from others and the moderators are not applicable, but the effect size (r) reflects the relationship between INS and teacher-learner interaction outcome.