Suner Aslı, Karakülah Gökhan, Koşaner Özgün, Dicle Oğuz
Department of Biostatistics and Medical Informatics, School of Medicine, Ege University, 35100 Bornova, Izmir Turkey.
Department of Medical Informatics, Health Sciences Institute, Dokuz Eylül University, 35340 Inciraltı, Izmir Turkey ; Neurobiology-Neurodegeneration and Repair Laboratory, National Eye Institute, National Institutes of Health, Bethesda, MD 20892 USA.
Springerplus. 2015 Oct 22;4:633. doi: 10.1186/s40064-015-1421-9. eCollection 2015.
The improper use of statistical methods is common in analyzing and interpreting research data in biological and medical sciences. The objective of this study was to develop a decision support tool encompassing the commonly used statistical tests in biomedical research by combining and updating the present decision trees for appropriate statistical test selection. First, the decision trees in textbooks, published articles, and online resources were scrutinized, and a more comprehensive unified one was devised via the integration of 10 distinct decision trees. The questions also in the decision steps were revised by simplifying and enriching of the questions with examples. Then, our decision tree was implemented into the web environment and the tool titled StatXFinder was developed. Finally, usability and satisfaction questionnaires were applied to the users of the tool, and StatXFinder was reorganized in line with the feedback obtained from these questionnaires. StatXFinder provides users with decision support in the selection of 85 distinct parametric and non-parametric statistical tests by directing 44 different yes-no questions. The accuracy rate of the statistical test recommendations obtained by 36 participants, with the cases applied, were 83.3 % for "difficult" tests, and 88.9 % for "easy" tests. The mean system usability score of the tool was found 87.43 ± 10.01 (minimum: 70-maximum: 100). A statistically significant difference could not be seen between total system usability score and participants' attributes (p value >0.05). The User Satisfaction Questionnaire showed that 97.2 % of the participants appreciated the tool, and almost all of the participants (35 of 36) thought of recommending the tool to the others. In conclusion, StatXFinder, can be utilized as an instructional and guiding tool for biomedical researchers with limited statistics knowledge. StatXFinder is freely available at http://webb.deu.edu.tr/tb/statxfinder.
在生物医学科学领域，分析和解释研究数据时，统计方法的不当使用屡见不鲜。本研究旨在开发一种决策支持工具，通过整合和更新现有的用于选择合适统计检验的决策树，涵盖生物医学研究中常用的统计检验方法。首先，仔细审查了教科书中、已发表文章以及在线资源中的决策树，并通过整合10个不同的决策树设计出一个更全面统一的决策树。决策步骤中的问题也通过简化并举例丰富问题进行了修订。然后，将我们的决策树应用到网络环境中，开发了名为StatXFinder的工具。最后，向该工具的用户发放了可用性和满意度调查问卷，并根据从这些问卷中获得的反馈对StatXFinder进行了重新整理。StatXFinder通过引导44个不同的是非问题，为用户在选择85种不同的参数和非参数统计检验时提供决策支持。36名参与者在实际案例应用中获得的统计检验推荐准确率，“困难”检验为83.3%，“简单”检验为88.9%。该工具的系统可用性平均得分为87.43±10.01（最低分：70 - 最高分：100）。系统总可用性得分与参与者属性之间未发现统计学显著差异（p值>0.05）。用户满意度调查问卷显示，97.2%的参与者对该工具表示赞赏，几乎所有参与者（36人中有35人）都认为会向他人推荐该工具。总之，StatXFinder可作为统计学知识有限的生物医学研究人员的教学和指导工具。可通过http://webb.deu.edu.tr/tb/statxfinder免费获取StatXFinder。
