S*****U
发帖数: 99 | 1 在家自学的统计，看Little SAS book, 229页解释scheffe options 不是很清楚， 象
下面;
Means with the same letter are not significantly different.
Scheffe Grouping Mean N Team
A 54.833 12 pink
A
B A 50.500 12 gold
B A
B A 50.333 12 gray
B
B 49.833 12 blue
B
B 49.500 12 red
可以理解为分三个group， A, AB, B, each group significantly different from
another,
还是只有两个group: A, B . AB算不算？
不好意思我问菜鸟问题， google也没有答案, |
|
s*******2
发帖数: 499 | 2 bonferroni太保守，不如其他两种方法。线性模型就可以用tukey和scheffe.
scheffe的使用范围比tukey广泛。
scheffe |
|
|
W*****B
发帖数: 4796 | 4 克拉克盖博的孙子克拉克盖博三世今日家中猝死，年仅30岁。
看着一身纹身，多半是嗑药
不过人家估计12-3岁就开始操逼了，操过的白牛到现在没有上千也有几百个了。按锁男
的话说，死而无憾了。
Clark Gable's grandson, Clark Gable III, dead at 30, says report
Clark Gable's grandson Clark Gable III has died at his Texas home at age
30, his family told TMZ Friday.
According to the celebrity gossip site, the actor, model and TV host was
found dead in his bed by his fiancé Friday morning. The cause of death is
unclear. Gable's relatives didn't immediately respond to Fox News&#
39; request for comme... 阅读全帖 |
|
a*****g
发帖数: 19398 | 5 Student suspended for sexual harassment after hugging teacher
Editor's note: Kelly Wallace is CNN's digital correspondent and editor-at-
large covering family, career and life. Read her other columns and follow
her reports at CNN Parents and on Twitter.
(CNN) -- To Sam McNair, a 17-year-old high school senior in Duluth, Georgia,
it was an innocent hug.
"You never know what someone's going through," McNair told CNN affiliate
WGCL in Atlanta. "A hug might help."
It didn't in this case because afte... 阅读全帖 |
|
g****g
发帖数: 1828 | 6 In probability theory, the normal (or Gaussian) distribution, is a
continuous probability distribution that is often used as a first
approximation to describe real-valued random variables that tend to cluster
around a single mean value. The graph of the associated probability density
function is “bell”-shaped, and is known as the Gaussian function or bell
curve:[nb 1]
f(x) = \tfrac{1}{\sqrt{2\pi\sigma^2}}\; e^{ -\frac{(x-\mu)^2}{2\sigma^2}
},
where parameter μ is the mean (location of the pe... 阅读全帖 |
|
a*****g
发帖数: 19398 | 7 Student suspended for sexual harassment after hugging teacher
Editor's note: Kelly Wallace is CNN's digital correspondent and editor-at-
large covering family, career and life. Read her other columns and follow
her reports at CNN Parents and on Twitter.
(CNN) -- To Sam McNair, a 17-year-old high school senior in Duluth, Georgia,
it was an innocent hug.
"You never know what someone's going through," McNair told CNN affiliate
WGCL in Atlanta. "A hug might help."
It didn't in this case because afte... 阅读全帖 |
|
z**********e
发帖数: 22064 | 8 http://bbs.chinanews.com/picview-185-5218998-1.shtml
为大家整理一些近现代灭绝生物的图文资料，希望大家记住这些曾出现在这蔚蓝星球上
的美丽生灵也希望同样的悲剧不要再继续上演，图文资料均转自网络，如有错误之处欢
迎大神们指出我会尽快予以修改！
渡渡鸟，或作嘟嘟鸟（Dodo），拉丁学名：Raphus cucullatus又称毛里求斯渡渡鸟、
愚鸠、孤鸽，是仅产于印度洋毛里求斯岛上一种不会飞的鸟。这种鸟在被人类发现后仅
仅176年的时间里，便由于人类的捕杀和人类活动的影响彻底绝灭，堪称是除恐龙之外
最著名的已灭绝动物之一。也是毛里求斯唯一被定为国鸟的已灭绝鸟类。
牛津大学自然史博物馆（Oxford University Museum of Natural History）中，曾经
陈列有渡渡鸟最后一个完整的填充标本（标本号OUM 11605），可惜在1755年由于腐烂
而被馆长下令焚毁了。侥幸保存下来的一块头骨和一块脚骨，现在仍陈列在该馆中。
在早期的图片中，渡渡鸟一直是一种肥胖，愚蠢的形象。直到2002年2月，国家地理学
会的新闻报道，苏... 阅读全帖 |
|
D******n
发帖数: 2836 | 9 what multiple comparison appraoches do you use? bonferroni, scheffe, or tuke
y?
the multiple comparison is supposed to be more "accurate" because it uses mo
re data, more data the better. One of the reasons the two groups have no sig
nificant difference in multiple comparison is that the error estimation is i
nflated because u have also considered the error from other groups.
simple example is, u have a device to measure ur height, first time 1.7m, se
cond time 1.8m, third time 1.7m. u do the sa |
|
h******e
发帖数: 1791 | 10 有什么根据吗？在什么情况下选择tukey，什么时候用bonferroni，什么时候用scheffe
？谢谢。 |
|
p*****0
发帖数: 3104 | 11 基本上是这样
如果同时比较的term比较少，例如1个，两个，可以用bonf
如果比较多，要用scheffe
具体用schefft还是bonfferoni，可以比较critical value
tukey只能比较equal sample size,如果sample size相等，tukey是比较号的选择 |
|
y***i
发帖数: 11639 | 12 【 在 peoplem (我爱我家) 的大作中提到: 】
玉整
http://graphpad.com/faq/viewfaq.cfm?faq=1081
补充一下，prism的网站上说是，哪怕anova不显著，依然可以用Tukey, Scheffe,
bonferroli等 test检查是否有某个是显著的。
cor
来比
方法
alpha
种
tte
（
违反
查了一下书
中心极限定理说：任何分布，在样本数趋于无穷时，其均值都符合正态分布。
我的理解是：这说明我样本数非常大时，无论分布如何，哪怕是严重违背正态分布，都可
以用ttest。不仅仅是有少数outlier的情况.
科班的同学评价一下这个理解对不对？什么样的样本数算是非常大？
te |
|
b******l
发帖数: 1632 | 13 哦。有种预感，我们说得不是同一个东西。。。
等我找到书再说。。。
以做
Scheffe |
|
|
s*****0
发帖数: 357 | 15 Step 1: one-way ANOVA
| |-->如果不显著，STOP.
|
|-->如果显著,说明至少有两组间存在显著性差异，GOTO Step 2。
Step 2: Multiple comparison, 必须控制overall significance level (假设0.05）
。很多选择，比如Bonferroni，Scheffé, Tukey等等。各自都有各自的适用情况和优
缺点，懒得费笔墨，自己翻翻书或者网上搜搜吧，比如Bonferroni在comparison增多的
情况下会太严苛，Tukey最适合pairwise comparison(在你的例子中即4选2,会有6个对
照).
Step 2会得到你想要的答案，既那些组存在显著性差异。
设每组100只老鼠。然后所有检测值都满足equal variance和normal distribution。
（2
。这种设计应该用one-way ANOVA pairwise comparison吧？有些文章用的是t test。
2vs1, 3vs1, 4vs1,然后 |
|
n*********0
发帖数: 22 | 16 请问你说的multiplicity adjusted p-value是simultaneously testing, adjusted
family p-value吗？就是同时进行多项比较，控制family p-value小于一定值，而不是
控制单项p-value。
如果是的话，可以用sas软件计算。有三种procedures， Tukey， Scheffe，和
bonferroni。 具体的用法和code网上有，可以搜索到。 |
|
L**********A
发帖数: 7115 | 17 只有两个GROUPS 啊 A 和B
AB 是说和A不sig 不同 和B 也不sig不同 |
|
S*****U
发帖数: 99 | 18 太感谢啦， 我是只上过一门统计课，没有足够training
可惜伪币太少， 没办法发包子给你。 |
|
g**r
发帖数: 425 | 19 如果基本的统计概念不清楚，从这个开始：
1. Statistics Inference (Casella, Berger)
如果对1中的内容比较熟悉，但平常用的多是建线性模型:
2. Analysis of Variance (Scheffe) |
|
|
h******e
发帖数: 1791 | 21 【 以下文字转载自 Biology 讨论区 】
发信人: hehehehe (入了大圣教，心中充满喜乐), 信区: Biology
标 题: multiple comparison方法的选择。
发信站: BBS 未名空间站 (Sun Feb 21 22:41:55 2010, 美东)
有什么根据吗？在什么情况下选择tukey，什么时候用bonferroni，什么时候用scheffe
？谢谢。 |
|
s*r
发帖数: 2757 | 22 i guess tukey and sheffe are for pairwise comparisons between multiple
levels of two factors
scheffe |
|
a*********r
发帖数: 108 | 23 先把b取个倒数，仍然记作b，然后求b的UMVUE。
这时E(X)=a/b,Var(X)=a/b^2
记y＝sum(x1,...,xn)
则y~Gamma(na,b),a已知
1/y~Inverse Gamma(na,b)
E(1/y)=b/(na-1),Var(1/y)=b^2/((na-1)^2*(na-2))
于是就令估计量bb=(na-1)/Y, Y是y代入x的那些样本值。
Var(bb)=b^2/(na-2)
而算一下可以知道C-R下限是b^2/(na)，就差一点
另一方面算算便可知a已知的情况下，minimal sufficient and complete statistics
就是y＝sum(x)了
bb已经是y的函数，condition on y无法进一步缩小方差，CRLB不可达到。
Rao-Blackwell + Lehmann-Scheffe定理，bb是UMVUE
仅供参考，不知道搞错了没有。 |
|
n*****n
发帖数: 3123 | 24 completeness应该没有sufficent, minimal这样的可以理解的意义。rao-blackwell 说
明如何找better estimator. if T' is an unbiased estimator of theta, and T is
a sufficient statistic, E(T'|T) is a better estimator than T'. While if T is
a complete sufficient statistic, E(T'|T) is the UMVUE by Lehmann-scheff. |
|
l*******d
发帖数: 101 | 25 I also don't find completeness very intuitive, although it is quite useful.
Sufficiency + completeness => minimal sufficiency.
Two major theorems relying on completeness:
1. Basu's theorem: Complete sufficient statistics are independent of
ancillary statistics.
E.g. can be used to show that sample mean and sample variance of a Normal (\
mu, \sigma^2) are independent of each other.
2. Lehmann-Scheffe theorem (see ningyan's reply): Unbiased estimators based
on complete sufficient statistics are u |
|
f*******m
发帖数: 94 | 26 非常感谢 ningyan 和 leftfield 的解释，现在清楚多了。
不过Rao-blackwell 和 Lehmann—scheffe 两个定理很容易搞混。虽然前者是建立在
sufficient 的基础上，后者是建立在 complete sufficient基础上的，但是都是用来
找UMVUE的。 |
|
A*******s
发帖数: 3942 | 27 what is ur purpose?
using Bonferroni means u test a family of hypotheses simultaneously. Say if
you would like to see which pair of groups are more/less different, then do
pairwise comparison one by one and rank the P-values. If you would like to
test a hypothesis that all/some pairs are equal, then do simultaneous test
with Bonf/Tukey/Scheffe.
From what u said before, i think u just want to find out the significantly
different pairs. u don't need Bonferroni.
Just my 2 cents. |
|
a***r
发帖数: 420 | 28 正在复习Casella那本书上的data reduction和point estimation，相当的有点纠结啊
尤其关于Lehmann-Scheffe Theorem:
Unbiased estimators based on complete sufficient statistics are unique.
书上主要做了如下讨论：
Theorem 7.3.23: Let T be a complete sufficient statistic for a parameter
theta, and phi(T)be any estimator based only on T. Then phi(T) is the unique
best unbiased estimator of its expected value q(theta).
这个phi(T)，我的理解是, Rao-Blackwell里的,比如W(X) unbiased for q(theta), 那么
phi(T)=E(W(X)|T(X))就是一个UMVUE.
那么这个是不是可以理解为，对于一个分布p(x;theta)... 阅读全帖 |
|
