Our cross-validation procedures produced out-of-sample estimates that were not quantitatively compromised, relative to the in-sample predictions reported in Ramsden et al. (2011), see Table 3.
The tree provides important confirmation for recent estimates of higher-level squamate phylogeny based on molecular data (but with more limited taxon sampling), estimates that are very different from previous morphology-based hypotheses.
In line with this finding, Xiang et al. 19 concluded that a reduced data sample produced estimates that were congruent with the larger data sample, while calibration information played a major role in affecting the time scale.
Having identified the best-fit clock and demographic models (see Table S4), we carried out a suitable number of independent MCMC runs to achieve convergence both with effective sample size estimates that usually exceeded 200 and with stable distributions for key model parameters (visualised with Tracer v1.5).
The five successful samples provided paleointensity estimates that were within 16% of the expected value of 100 µT (Table 2).
Based on the previous RCT using CFT, 39 a sample size calculation estimates that a sample size of 64 in each group will have 80% power to detect a difference in means of 5.0 (disability) and 1.0 (pain) between the two arms of the study, assuming that the common SD is 10.0 (disability) and 2.0 (pain), and using a two-sided 5% significance test.
Monte Carlo re-sampling provided standard error estimates that were about 10% lower for systematic sampling.
Second, smaller sample sizes result in estimates that are considerably less precise than those from the College Board data.
This method allowed us using the variability within a sample to estimate that sampling distribution empirically and this is done by randomly resampling with replacement.
Thus, random sampling gives AS rate estimates that are comparable between species.
We found that dividing regions into geographic sub-samples did not produce estimates that differed significantly from the original estimates (see Material and Methods).
