Precision and reliability in paleocommunity comparisons based on cluster-confidence intervals: How to get more statistical bang for your sampling buck
Precision and reliability in paleocommunity comparisons based on cluster-confidence intervals: How to get more statistical bang for your sampling buck
复制标题
DOI:
10.2307/3515400
复制
发表时间:
1999-10-01
期刊:
影响因子:
1.6
通讯作者:
Rutherford, SD
中科院分区:
文献类型:
--
作者:
Bennington, JB;Rutherford, SD
Statistically meaningful comparisons of fossil assemblages can be made by calculating cluster-confidence intervals around mean species abundances using data from replicate bulk samples; However, different levels of precision and reliability are obtained fi om cluster-confidence intervals depending on how sampling effort is distributed when collecting replicate samples. If the species-abundance distributions ata locality are patchy, then both the precision and the reliability of cluster-confidence intervals are increased by distributing sampling effort among a larger number of bulk samples. Collecting larger numbers of samples does not significantly increase the sampling effort if the total! number of individuals collected remains the same. Cluster-confidence intervals (CIs) around mean species abundances at a locality provide a way to assess the magnitude of differences observed between localities and can be used to assess the significance of the absence of rare species. Conclusions derived from cluster-CIs regarding the significance of differences in observed species abundances are similar to those obtained using analysis of variance (ANOVA) Cluster-Cls provide a powerful and easily applied tool for assessing the significance of differences in species abundances and paleo-community composition observed in the fossil record.