Evaluating estimators of species richness: the importance of considering statistical error rates

The performance of species richness estimators can be highly variable. Evaluating the accuracy and precision of different estimators for different assemblages is common in the ecological literature, but estimator performance is rarely measured in terms of research goals such as detecting patterns in diversity. We evaluated the efficacy of nonparametric richness estimators to detect changes (i.e. type-I and type-II error rates) in species richness using two experimental designs: a block design and a trend analysis. We also evaluated estimator efficacy across a variety of species-abundance distributions, species number and sample sizes. The evaluation was performed using simulated data that mimicked the qualities of real data to ensure real-world relevance. We found that the bias and precision of all estimators evaluated had high sensitivity to sample size and shifts in the species-abundance distribution. Importantly, all estimators demonstrated elevated type-I error rates when the species-abundance distribution varied. These inflated type-I error rates resulted in spurious conclusions about patterns in species richness. Results suggest that caution should be taken when using nonparametric estimators to detect pattern in species richness. Furthermore, estimator evaluations should always include measures of type-I and type-II error rates. These quantities can reveal the inference consequences of the dependency of estimator bias and precision on community and sampling characteristics.