Table 2. Generalized additive mixed model results of regression cardinal Elo score on age and personality smooths.
Estimated degrees of freedom (df) indicate the penalized number of regression terms associated with this smooth. Reference dfs indicate the unpenalized, maximum possible terms. Fixed effects (ti) are main and interaction effects of tensor product smooths using cubic regression splines with shrinkage. Random effects (s) smooths are parametric terms penalized by a ridge penalty.
| Model term | Estimated df | Reference df | χ 2 | p |
|---|---|---|---|---|
| ti (Age) | 5.970 | 6 | 7,696.943 | <0.001 |
| ti (Dominance) | 1.172 | 4 | 330.975 | <0.001 |
| ti (Conscientiousness) | 3.690 | 4 | 322.321 | 0.658 |
| ti (Extraversion) | 0.617 | 4 | 79.524 | 0.098 |
| ti (Agreeableness) | 0.001 | 4 | 0.000 | 0.551 |
| ti (Openness) | 1.350 | 4 | 418.057 | 0.001 |
| ti (Neuroticism) | 0.000 | 4 | 0.000 | 0.639 |
| ti (Age × Dominance) | 13.318 | 16 | 1,165.512 | 0.027 |
| ti (Age × Conscientiousness) | 15.869 | 16 | 2,914.363 | 0.019 |
| ti (Age × Extraversion) | 11.327 | 16 | 2,086.150 | <0.001 |
| ti (Age × Agreeableness) | 14.916 | 16 | 3,441.257 | 0.118 |
| ti (Age × Openness) | 14.961 | 16 | 1,494.961 | 0.021 |
| ti (Age × Neuroticism) | 3.141 | 16 | 140.932 | 0.051 |
| s (ID) | 0.009 | 27 | 0.008 | 0.241 |
| s (ID × Age) | 4.323 | 28 | 73.749 | 0.026 |
| s (ID × Dominance | 0.000 | 27 | 0.000 | 0.409 |
| s (ID × Conscientiousness) | 10.260 | 27 | 945.512 | <0.001 |
| s (ID × Extraversion) | 0.000 | 27 | 0.000 | 0.207 |
| s (ID × Agreeableness) | 3.358 | 27 | 54.724 | 0.054 |
| s (ID × Openness) | 0.002 | 27 | 0.002 | 0.206 |
| s (ID × Neuroticism) | 0.000 | 27 | 0.000 | 0.224 |