| 查看: 221 | 回复: 0 | ||
[求助]
求助
|
|
It is frequently observed that percentage data are not normally distributed. If the data are not normally distributed, then they must be transformed prior to ANOVA. The usual transformation for proportional data is arcsine square root, although other transformations e.g. logarithms may give the data a normal distribution. When data have been transformed, the SEM must not be back-transformed. Consequently, it is customary to present the untransformed means and untransformed SEM and say that the statistical analysis was done on the transformed data. Alternatively, the back-transformed means can be presented with confidence limits that have been calculated on the transformed data and back-transformed for presentation. 上面是审稿人给的，看不明白什么意思。对统计学知识欠缺啊。 明白的朋友给讲讲。 |
回复此楼» 猜你喜欢
“人文社科而论，许多学术研究还没有达到民国时期的水平”
已经有4人回复
-
过年走亲戚时感受到了所开私家车的鄙视链
已经有11人回复
-
什么是人一生最重要的？
已经有4人回复
-
版面费该交吗
已经有3人回复
-
今年春晚有几个节目很不错，点赞！
已经有12人回复
-
体制内长辈说体制内绝大部分一辈子在底层，如同你们一样大部分普通教师忙且收入低
已经有12人回复
-
基金正文30页指的是报告正文还是整个申请书
已经有5人回复