SEARCH

SEARCH BY CITATION

FilenameFormatSizeDescription
ece3615-sup-0001-SupportingInformation.docxWord document194K

Data S1. Estimation of the magnitude of the variance of individual random effects with a restrictive and penalizing prior.

Figure S1. Histograms representing the probability density of the variance (inline image) for the two different priors used for model H2: (a) & (c) uniform U(0,10) distribution on the corresponding standard deviation σα; (b) & (d) inversegamma (4,0.05) distribution on inline image. For better readability, we show the densities on two different x-axis scales: histograms (a) and (b) are scaled on [0;0.15], where the density of inline image values under the inversegamma prior (b) can more easily be interpreted; histograms (c) and (d) are scaled on [0;100], where the full extent of the density of inline image values under the uniform prior (c) can be better seen. We see that the density of inline image values is heavily shifted toward very small values in the case of the inversegamma (4,0.05) prior distribution compared with the case of the uniform U(0,10) prior distribution. On the [0;100] scale, the density of the inversegamma (4,0.05) distribution looks as a single peak at zero.

Table S1. Summary of the parameter posterior distributions from model H2 (“fixed heterogeneity” hypothesis; see main text) using two contrasted priors on the variance (inline image) of random individual effects: (i) the inversegamma (4,0.05) distribution on inline image which puts a lot of weight on very small values (close to zero), and thus strongly penalizes high values; (ii) the uniform U(0,10) distribution on the corresponding standard deviation σα, which is a very vague prior and has more weight on high values of variance. The mean, the standard deviation (SD), 2.5% and 97.5% percentiles (perc.) of the posterior distribution are shown. Symbols with a star (e.g., inline image) correspond to parameters back-transformed to the more interpretable scale of a probability of reproduction (i.e., in the interval [0,1]). See main text for more details.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.