Model | Predictor | Est | CI | t | p (PI) | \(\overline{B}\) | χ2 (df) | ΔAIC (ER) | R2m/R2ca |
|---|
1 | Intercept | 4.20 | 19.24–27.64 | 2.49 | . |  |  |  |  |
log(mean VeDBA) | − 0.37 | − 6.74–5.99 | − 0.51 | (0.0001–1). |  | 0.03 (1) | 1.96 (2.66) | 0.01/0.10 |
2 | Intercept | 10.54 | 6.74–14.27 | 5.82 | *** |  |  |  |  |
log(mean RocRM1) | 2.59 | 0.73–4.28 | 2.87 | (7.11−6–0.56)* | 16.66 | 6.40 (1) | 4.4 (9.02) | 0.15/0.41 |
3 | Intercept | 7.61 | 6.10–9.19 | 9.27 | *** |  |  |  |  |
log(mean RocRM5) | 2.26 | 1.01–3.55 | 3.19 | (1.1−6–0.34)** | 53.26 | 8.46 (1) | 6.46 (25.28) | 0.18/0.40 |
4 | Intercept | 7.06 | 5.67–8.22 | 11.00 | *** |  |  |  |  |
log(mean RocRM10) | 2.32 | 1.03–3.54 | 3.84 | (6.909−8–0.14)*** | 83.86 | 11.79 (1) | 9.79 (133.62) | 0.23/0.46 |
5 | Intercept | 5.25 | 4.27–6.09 | 10.74 | *** |  |  |  |  |
log(mean RocRM60) | 1.68 | 0.95–2.37 | 4.57 | (2.04−9–0.04)*** | 685.31 | 16.28 (1) | 14.28 (1261.43) | 0.31/0.51 |
- Results of the final (following removal of terms through stepwise elimination) linear mixed models where VO2 mg ind−1 h−1 is predicted by either log-transformed mean VeDBA or log-transformed mean RocRM (over either 1 s, 5 s, 10 s or 60 s stepping value)
- PI refers to the prediction interval of the p value
- ΔAIC refers to the difference between the null and full models, as shown by ANOVA model comparison, and ER refers to the AIC evidence ratio
- \(\overline{B}\) refers to Bayes factor upper bound. A \(\overline{B}\) was not calculated for Model 1 due to the p value being  >  0.37
- aR2m (marginal R2) refers to the variance of the (final) linear mixed model explained by only the fixed-effect terms, whereas R2c (conditional R2) refers to the variance explained by both fixed- and random-effect model terms [49]
- ‘.’, ‘*’, ‘**’ and ‘***’ refer to p values of  >  0.05,  <  0.05,  <  0.01 and  <  0.001, respectively
