Table 4.
Estimates for a restricted SAOM model of the Cardiff ASSIST data (Campbell et al. 2008)
| Effect name | θ̂ | se(θ̂) | t. conv. | θ̂* | p value |
|---|---|---|---|---|---|
| Rate (1) | 16.762 | 1.290 | – | – | – |
| Rate (2) | 10.834 | 0.681 | – | – | – |
| Outdegree | −2.892 | 0.029 | −0.034 | −2.910 | – |
| (2) | – | – | – | 0.042 | 0.467 |
| Reciprocity | 1.958 | 0.065 | 0 | 1.935 | – |
| (2) | – | – | – | 0.056 | 0.673 |
| Transitive triplets | 0.415 | 0.015 | −0.025 | 0.457 | – |
| (2) | – | – | – | −0.078 | 0.012 |
| Age similarity | 0.318 | 0.149 | 0.034 | 0.037 | – |
| (2) | – | – | – | 0.112 | 0.902 |
| Smoking sim. | 0.634 | 0.117 | 0.002 | 0.457 | – |
| (2) | – | – | – | 0.101 | 0.384 |
| Same sex | 0.876 | 0.133 | −0.001 | 0.923 | – |
| (2) | – | – | – | −0.044 | 0.492 |
| Same form | 0.566 | 0.042 | −0.040 | 0.767 | – |
| (2) | – | – | – | −0.450 | 0 |
θ̂ corresponds to the estimate obtained by conditional method of moments, and θ̂* corresponds to the one step estimates
Rows marked (2) indicate the corresponding values for a dummy term interacted with the preceding effect
Estimates for the time dummy terms are given below their respective base effects
The p value is for the score test of the hypothesis that the dummy term is equal to zero