Skip to main content
. 2020 May 20;23(1):207–222. doi: 10.1093/biostatistics/kxaa015

Fig. 1.

Fig. 1.

A comparison of nonparametric and parametric bootstrap approaches to estimating the variance of diversity estimates under a model that incorporates microbial co-occurrence patterns. The parametric bootstrap has lower variance than the nonparametric bootstrap (left panel), and the median difference with true variance close to zero (right panel). No approach to covariance estimation consistently outperforms other approaches.