x*******u
发帖数: 500 | 1 请教meta analysis 高人。
据说 stata 软件 meta 被metan 代替了，以前的同事用meta做的分析， 我用她的code
, 她的数据，用meta 和metan 能得到类似的结果。
但是现有一个data, 有rr, lcl, ucl. 用两个命令，得到完全不同的结果。
meta rr lcl ucl, ci eform
得到 Test for heterogeneity: Q= 184.364 on 135 degrees of freedom (p= 0.003)
metan rr lcl ucl
得到 Heterogeneity chi-squared = 120.16 (d.f. = 135) p = 0.815
请问我有啥地方做错了？ 万分感谢。 | D*********2
发帖数: 535 | 2 never used before, software is like a black box, try read the book.
All the meta-analysis calculations available in metan are based on standard
methods, an overview of which may be found in chapter 15 of Deeks, Altman,
and Bradburn (2001) | g********d
发帖数: 2022 | 3 自己算的话是
Q=sum(w_i*(ES_i-ES_bar)^2)
df=k-1 for k studies.
standard
【在 D*********2 的大作中提到】
: never used before, software is like a black box, try read the book.
: All the meta-analysis calculations available in metan are based on standard
: methods, an overview of which may be found in chapter 15 of Deeks, Altman,
: and Bradburn (2001)
| e******e
发帖数: 410 | 4 yes, i used it in my qualifying proprosal.
however, i didn't use stata. Instead, i used spss. there is a famous book
about meta-analysis "Practical Meta-Analysis ", only around 200 pages, easy
to read, and you can get spss code from this book. I highly recommend it.
see the book link here:
http://www.amazon.com/Practical-Meta-Analysis-Applied-Research-
code
003)
【在 x*******u 的大作中提到】
: 请教meta analysis 高人。
: 据说 stata 软件 meta 被metan 代替了，以前的同事用meta做的分析， 我用她的code
: , 她的数据，用meta 和metan 能得到类似的结果。
: 但是现有一个data, 有rr, lcl, ucl. 用两个命令，得到完全不同的结果。
: meta rr lcl ucl, ci eform
: 得到 Test for heterogeneity: Q= 184.364 on 135 degrees of freedom (p= 0.003)
: metan rr lcl ucl
: 得到 Heterogeneity chi-squared = 120.16 (d.f. = 135) p = 0.815
: 请问我有啥地方做错了？ 万分感谢。
|
|
