Table 2. Mixed-Effects Multivariate Regression Model.
| Variable and Covariate | Coefficient or Estimatea | SE (95% CI) | P Value |
|---|---|---|---|
| Fixed Effects | |||
| Estimated ageb | |||
| Surgery | −4.61 | 0.18 (−4.97 to −4.25) | <.001 |
| Constant | 58.64 | 0.22 (58.21 to 59.07) | <.001 |
| Attractivenessc | |||
| Surgery | 6.72 | 0.38 (5.96 to 7.47) | <.001 |
| Constant | 45.29 | 0.57 (44.17 to 46.42) | <.001 |
| Successfulnessc | |||
| Surgeryc | 3.85 | 0.37 (3.12 to 4.57) | <.001 |
| Constant | 58.14 | 0.42 (57.31 to 58.97) | <.001 |
| Healthc | |||
| Surgery | 7.65 | 0.40 (6.87 to 8.42) | <.001 |
| Constant | 55.62 | 0.46 (54.73 to 56.52) | <.001 |
| Random Effects | |||
| Variance, aged | |||
| Observer | 15.72 | 1.25 (13.46 to 18.36) | |
| Residual | 47.56 | 0.91 (45.82 to 49.36) | |
| Variance, attractivenessd | |||
| Observer | 129.82 | 9.26 (112.87 to 149.32) | |
| Residual | 210.21 | 4.10 (202.33 to 218.39) | |
| Variance, successd | |||
| Observer | 56.48 | 4.66 (48.04 to 66.40) | |
| Residual | 194.55 | 3.84 (187.16 to 202.23) | |
| Variance, healthd | |||
| Observer | 67.26 | 5.50 (57.30 to 78.94) | |
| Residual | 222.61 | 4.40 (214.15 to 231.40) | |
| Covariance, residual | |||
| Age, attractiveness | −15.79 | 1.37 (−18.48 to −13.11) | <.001 |
| Age, success | 0.92 | 1.31 (−1.64 to 3.48) | .48 |
| Age, health | −8.67 | 1.40 (−11.42 to −5.92) | <.001 |
| Attractiveness, success | 100.01 | 3.15 (93.82 to 106.19) | <.001 |
| Attractiveness, health | 117.58 | 3.44 (110.84 to 124.33) | <.001 |
| Health, success | 129.09 | 3.45 (122.31 to 135.86) | <.001 |
a
Coefficients are given for fixed effects; estimates for random effects.
b
Observer-estimated age coded on a visual analog scale from 30 to 80 years with 1-year increments.
c
Observer-perceived attractiveness, success, and health coded on a visual analog scale from 0 to 100 with 1-unit increments.
d
The variance components are estimates of the spread of the data based on the modeling. The standard reporting of the mixed-effects model does not report a P value for variance as these estimates are not hypothesis tested.