Table 2.
Odds ratios obtained by logistic regression.
| A | B | C | D | E | F | G | H | |
|---|---|---|---|---|---|---|---|---|
| Adj. Pseudo-R2 | 0.00305 | 0.00446 | 0.00720 | 0.00935 | 0.01643 | 0.01780 | 0.02050 | 0.02265 |
| Intercept | 0.010*** | 0.011*** | 0.011*** | 0.011*** | 0.010*** | 0.010*** | 0.010*** | 0.010*** |
| Support | – | – | – | – | 1.096*** | 1.095*** | 1.097*** | 1.098*** |
| Knowledge | – | – | – | – | 1.217*** | 1.216*** | 1.216*** | 1.216*** |
| Conflict | – | – | – | – | 1.024*** | 1.026*** | 1.025*** | 1.024*** |
| Power | – | – | – | – | 1.097*** | 1.099*** | 1.097*** | 1.096*** |
| Similarity | – | – | – | – | 1.110*** | 1.108*** | 1.110*** | 1.112*** |
| Status | – | – | – | – | 0.930*** | 0.929*** | 0.934*** | 0.937*** |
| Trust | – | – | – | – | 1.143*** | 1.143*** | 1.144*** | 1.144*** |
| Identity | – | – | – | – | 1.085*** | 1.086*** | 1.086*** | 1.086*** |
| Sentiment Pos. | 1.174*** | 1.173*** | 1.175*** | 1.177*** | 1.110*** | 1.110*** | 1.110*** | 1.112*** |
| Sentiment Neg. | 1.080*** | 1.082*** | 1.081*** | 1.081*** | 1.056*** | 1.058*** | 1.058*** | 1.057*** |
| Comment Left-Wing | – | 1.014 | – | – | – | 1.005 | – | – |
| Comment Right-Wing | – | 0.790*** | – | – | – | 0.795*** | – | – |
| Post Left-Wing | – | 0.867*** | – | – | – | 0.873*** | – | – |
| Post Right-Wing | – | 0.852*** | – | – | – | 0.849*** | – | – |
| Both Polarized | – | – | 0.321*** | – | – | – | 0.324*** | – |
| Both Polarized & Diff. Side | – | – | 2.555*** | – | – | – | 2.510*** | – |
| Diff. Side | – | – | 1.022 | – | – | – | 1.023 | – |
| Shared Group | – | – | – | 0.286*** | – | – | – | 0.287*** |
Each column corresponds to a model with a specific set of variables. A description of each confounder is given in Table 3. We indicate with asterisks the statistically significant correlations (with one, two or three asterisks corresponding to P < 0.05, P < 0.01 and P < 0.001 respectively). P-values are corrected according to the Benjamini-Hochberg procedure29, to reduce the chance of spurious correlation emerging because of the high number of factors we consider.