<p>This article presents the software <i>Spower</i>, an R package designed as a general-purpose Monte Carlo simulation experiment tool to perform power analyses. The package includes complete customization capabilities with support for five distinct (expected) power analysis criteria (prospective/post hoc, a priori, compromise, sensitivity, and criterion), each of which reports the sampling uncertainty associated with the resulting estimates. Researchers may choose to define their own population generating and analysis function for their tailored simulation experiments, or may choose from a selection of the predefined simulation experiments available within the package. To facilitate comparability and further extensibility, simulation counterparts of the subroutines from the popular stand-alone software <i>G*Power</i> 3.1&#xa0;(Faul et al., <i>Behavior Research Methods,</i> <i>41</i>(4), 1149–1160 <CitationRef CitationID="CR12">2009</CitationRef>) are included within the package, along with other useful simulation experiment subroutines for improving estimation precision and creating visualizations.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Spower: A general-purpose Monte Carlo simulation power analysis program

  • R. Philip Chalmers

摘要

This article presents the software Spower, an R package designed as a general-purpose Monte Carlo simulation experiment tool to perform power analyses. The package includes complete customization capabilities with support for five distinct (expected) power analysis criteria (prospective/post hoc, a priori, compromise, sensitivity, and criterion), each of which reports the sampling uncertainty associated with the resulting estimates. Researchers may choose to define their own population generating and analysis function for their tailored simulation experiments, or may choose from a selection of the predefined simulation experiments available within the package. To facilitate comparability and further extensibility, simulation counterparts of the subroutines from the popular stand-alone software G*Power 3.1 (Faul et al., Behavior Research Methods, 41(4), 1149–1160 2009) are included within the package, along with other useful simulation experiment subroutines for improving estimation precision and creating visualizations.