Table 8 The impact of control variables on stance construction (extended DiD model).
From: Reconstructing stance in EFL doctoral thesis writing through generative artificial intelligence
Variable | Hedges (β) | Boosters (β) | Attitude markers (β) | Self-referential expressions (β) |
|---|---|---|---|---|
Core effect | ||||
High-AI×Post2022 | –1.75** | 0.85 | –0.62* | +0.18 |
Control variables | ||||
Corpus linguistics theme weight | –0.43* | 0.12 | –0.09 | +0.05 |
Empirical research methods (Yes=1) | –0.21 | 0.34 | 0.17 | –0.11 |
Interaction effects | ||||
High-AI×corpus linguistics theme | –0.38* | +0.25 | –0.14 | +0.19 |
High-AI×empirical methods | +0.12 | –0.18 | +0.07 | –0.05 |
