Table 1.
Trends, offsets, and goodness-of-fit values from linear fitting of correlations between metabolic rates from model estimates and metabolic measurements.
| Experimental Parameter (x) | Trend a | Offset b | R2 | |||
|---|---|---|---|---|---|---|
| Var. | Const. | Var. | Const. | Var. | Const. | |
| Step Width19 (m) | 0.758 | 0.998 | 0.981 | 1.150 | 1.000 | 1.000 |
| Added Mass20: waist (kg) | 0.278 | 0. 404 | 2.097 | 2.516 | 1.000 | 1.000 |
| Added Mass20: thigh (kg) | 0.547 | 0.716 | 1.451 | 1.763 | 1.000 | 1.000 |
| Added Mass20: shank (kg) | 2.038 | 2.360 | −2.016 | −2.053 | 1.000 | 1.000 |
| Added Mass20: foot (kg) | 1.378 | 1.490 | −0.504 | −0.045 | 1.000 | 1.000 |
| Extra foot lift21 (m) | 0.497 | 0.611 | 1.425 | 1.858 | 1.000 | 1.000 |
| Reduced Gravity22 (g) | 1.145 | 1.796 | −0.293 | −1.157 | 1.000 | 0.998 |
| Flat Walking23: Flat (m/s) | 1.074 | 1.228 | −0.936 | −0.959 | 0.986 | 0.989 |
| Obesity24: Obese (m/s) | 0.897 | 1.250 | 0.303 | 0.021 | 0.989 | 0.995 |
For the measurements, we used polynomial equations reported in the original papers since we did not have access to individual subject data. The goodness-of-fit values are close to one, indicating that model estimates follow similar polynomial trends found in the original papers. Both constant efficiency estimates and variable efficiency estimates (abbreviated “Const” and “Var” respectively) are shown. A slope of unity with zero bias means perfect agreement with empirical data (see Fig. S1 for visualization).