Ottenbacher K J
Center for Health Sciences, University of Wisconsin, Madison 53706.
Arch Phys Med Rehabil. 1988 Mar;69(3 Pt 1):170-3.
Recent studies indicate that rehabilitation researchers frequently include multiple tests of significance in their investigations. This study examined the possibility that the Type I error rate in rehabilitation research is inflated due to the inappropriate analysis of data from multiple significance tests. The incidence of Type I error was computed for a sample of rehabilitation research. The results revealed a Type I error assumed level of 5% (p less than .05). The argument is made that methods to determine the rate of Type I error should be routinely used in rehabilitation research.
近期研究表明，康复研究人员在其调查中经常进行多次显著性检验。本研究探讨了康复研究中由于对多次显著性检验的数据进行不当分析而导致I型错误率膨胀的可能性。计算了一组康复研究样本的I型错误发生率。结果显示I型错误的假定水平为5%（p小于0.05）。有人认为，确定I型错误率的方法应在康复研究中常规使用。
