yuanyuanlu
[求助]问个苯苯的统计问题
734
9
2005-08-16 04:08:00
大家知道怎么把means分类啊？就是哪几个properties重要性差不多，分为一组。
比如说有1，2，3，4，5个property需要rank，然后求mean 分别得到5个means。请问怎样知道这些means significantly different? 我想要这些means分成几个clusters。每个cluster里的means 不sig different from each other；but one property in one cluster is sig different from the property in another cluster.
能用repeated measurement ANOVA吗？我用了里面的pairwise comparison, 只能得到一对对的结果，不能分成几个clusters 。
不知道大家知不知道我的问题，我现在表达能力越来越差了，脑子也不好使
[此贴子已经被作者于2005-8-16 4:11:28编辑过]
ttest is only good for paired comparison. For comparison among three or more than three populations, i guess you might want to use MANOVA or GLM? But both have very strict assupmtions.
hehe. I am curious right now. Hope some statistics guru can solve the problem very soon.
post-hoc?
我也想用这个阿，可是repeated meausre ANOVA不能用post hoc，反正执行不了。
ttest is only good for paired comparison. For comparison among three or more than three populations, i guess you might want to use MANOVA or GLM? But both have very strict assupmtions.
hehe. I am curious right now. Hope some statistics guru can solve the problem very soon.
shanshan, are u in statistics
ttest is only good for paired comparison. For comparison among three or more than three populations, i guess you might want to use MANOVA or GLM? But both have very strict assupmtions.
I don't have dependent variables ya...
shanshan, are u in statistics
不是阿。赫赫
hehe. i see .
you might have panel data.
are you using spss? I guess stata might have some useful function. Can not recall the details. I need to go back and check.
lalalalala
if you want to use post hoc in three+ group comparison, stata can do this with ANOVA.
not quite sure about other statistics softwares.
[此贴子已经被作者于2005-8-16 11:45:16编辑过]
lalalalala
if you want to use post hoc in three+ group comparison, stata can do this with ANOVA.
not quite sure about other statistics softwares.
really? I don't have stata software. Let me go back and check if I can find some trial version! Thank you so much, shanshan
到底了
HOT DEALS