Table 3. Estimated average effect of model parameters, standard errors, and variance between classes for antipathy relations.
| Model parameter | Estimated average | SE | Estimated variance |
|---|---|---|---|
| Structural effects | |||
| Reciprocity | 0.52** | 0.16 | 0.00 |
| 2-in-star | 0.64** | 0.16 | 0.30 |
| 2-out-star | 2.07** | 0.12 | 0.15 |
| Transitive closure | 0.02 | 0.11 | 0.14 |
| Multiple connectivity | 0.01 | 0.04 | 0.02 |
| Control attribute effects | |||
| Male sender | 0.66** | 0.11 | 0.05 |
| Male received | 1.05** | 0.13 | 0.11 |
| Male similarity | -1.19** | 0.22 | 0.37 |
| Non-respondent receiver | 0.33** | 0.08 | 0.02 |
| Weight status effects | |||
| Overweight sender | 0.18** | 0.06 | 0.00 |
| Overweight receiver | 0.53** | 0.11 | 0.00 |
| Overweight similarity | -0.41 | 0.34 | 0.00 |
* p < .05.
** p < .01.
Note. Model parameters were estimated for 21 classes where acceptable ERGM model convergence was obtained, except reciprocity (19 classes), male similarity (19), overweight sender (20 classes), and overweight similarity (11 classes).