标题: 样本量太少做了多元逻辑回归分析,审稿人建议做Univariate analys [打印本页] 作者: dellysu 时间: 2012-12-24 22:09 标题: 样本量太少做了多元逻辑回归分析,审稿人建议做Univariate analys 小弟最近投了一篇文章,研究的样本量是26人,考虑影响该疾病情况的因素可能为性别,年龄,高血压病,糖尿病,抽烟,喝酒,教育程度,性格等,用逻辑回归分析出了独立危险因素。审回要大修。其中一个审稿人的意见是:I think that multiple logistic regression is not applicable here because of the small sample size of the groups. Univariate analysis, eventually with adjustment for age and sex, may be more appropriate.