School of Public Health, Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan Province, China.
State Key Laboratory of Grassland and Agro-ecosystem, College of Ecology, Lanzhou University, Lanzhou, Gansu Province, China.
PLoS One. 2022 Jul 8;17(7):e0271185. doi: 10.1371/journal.pone.0271185. eCollection 2022.
In the case of comparing means of various groups, data exploration and comparison for affecting factors or relative indices would be involved. This process is not only complex requiring extensive statistical knowledge and methods, but also challenging for the complex installation of existing tools for users who lack of statistical knowledge and coding experience. Like, the normal distribution and equal variance are crucial premises of parametric statistical analysis. But some studies reported that associated data from various industries violated the normal distribution and equal variance, parametric analysis still involved leading to invalid results. This is owing to that the normal distribution tests and homogeneity of variance test for different variables are time-cost and error-prone, posing an urgent need for an automatic and user-friendly analysis application, not only integrating normal distribution tests and homogeneity of variance test, but also associated the following statistical analysis. To address this, we developed a Shiny/R application, moreThanANOVA, which is an interactive, user-friendly, open-source and cloud-based visualization application to achieve automatic distribution tests, and correlative significance tests, then customize post-hoc analysis based on the considerations to the trade-off of Type I and Type II errors (deployed at https://hanchen.shinyapps.io/moreThanANOVA/). moreThanANOVA enables novice users to perform their complex statistical analyses quickly and credibly with interactive visualization and download publication-ready graphs for further analysis.
在比较各种组的均值时，需要进行影响因素或相对指标的数据探索和比较。这个过程不仅复杂，需要广泛的统计知识和方法，而且对于缺乏统计知识和编码经验的用户来说，现有的工具的复杂安装也是具有挑战性的。例如，参数统计分析的正态分布和等方差是至关重要的前提。但是，一些研究报告指出，来自不同行业的数据违反了正态分布和等方差，参数分析仍然涉及到，导致结果无效。这是因为对不同变量的正态分布检验和方差齐性检验既耗时又容易出错，因此迫切需要一个自动且用户友好的分析应用程序，不仅集成了正态分布检验和方差齐性检验，还可以进行相关的统计分析。为了解决这个问题，我们开发了一个 Shiny/R 应用程序 moreThanANOVA，这是一个交互式、用户友好的、开源的和基于云的可视化应用程序，用于实现自动分布检验和相关的显著性检验，然后根据对 I 型和 II 型错误的权衡考虑定制事后分析（部署在 https://hanchen.shinyapps.io/moreThanANOVA/）。moreThanANOVA 使新手用户能够通过交互式可视化快速、可信地执行他们的复杂统计分析，并下载可用于进一步分析的发布就绪图形。
