Table 2.
Language choice effect sizes between key dates
| Language | Sample Eeffects | |||
|---|---|---|---|---|
| Start—Aggression | Aggression—War | War—End study | Aggression—End Study | |
| UA over RU | +66.13% | +13.00% | +65.72% | +87.25% |
| UA over EN | +21.43% | −52.08% | +41.96% | −31.98% |
| RU over EN | −19.01% | −61.74% | −29.33% | −72.96% |
| Behavioural effects | ||||
|---|---|---|---|---|
| UA over RU | +128.69% | +52.08% | +129.24% | +248.63% |
| UA over EN | +64.14% | −33.61% | +92.663% | +27.90% |
| RU over EN | −38.23% | −38.69% | −20.659% | −51.36% |
Effect sizes for both sample and behavioural changes extracted from the language model described in Section “Language Modelling” between key dates. All effect sizes are relative increases in the odds between the two respective dates. For the start date calculation, we drop the first two weeks of the study period. Start: start of the study period—27th January 2020. Aggression: first official US report of a mobilization of the Russian troops along the Ukrainian border—11th November 2021. War: outbreak of the war—24th February 2022. End Study: end of the study period—10th October 2022.