公卫人

 找回密码
 立即注册

QQ登录

只需一步,快速开始

不劳无获:如何获取钢镚? 因为论坛,所以相逢。 捐赠百科答题至尊

公卫考研:一起风雨兼程 因为梦想,所以努力。 真题答案政治英语

职称考试:诸君逢考必过 因为热爱,所以执着。 模拟考场技能执医中级

查看: 101|回复: 0
打印 上一主题 下一主题

[分享] 既然有强大的社区命令,为何Stata16依然开发了官方meta套件?

[复制链接]
跳转到指定楼层
1#
alian 发表于 2019-8-25 21:42:38 | 只看该作者 回帖奖励 |倒序浏览 |阅读模式

注册后推荐绑定QQ,之后方才可以使用下方的“用QQ帐号登录”。

您需要 登录 才可以下载或查看,没有帐号?立即注册

x
在Stata Forum里,Suhail Doi提出了一个关于Stata官方新命令套件的疑问:
4 f4 i' ?9 O/ |* i+ J* d
& F# M) T" z1 A& a! Y。 uThe Stata 16 implementation of meta-analysis comes at a time when there are comprehensive user written programs such as admetan and related routines for meta-analysis - not sure why StataCorp chose to go down this path and introduce a weaker suite rather than collaborate with those authors??4 O) W6 O5 y( G* @9 Y$ [
$ R7 Z% B6 W2 e/ j
出处:
5 @$ c" b。 J" h( T( L
5 u; W7 S2 n9 k0 ]会员回复如下:
) ~0 v- d! b/ {( G) A. ~- v5 B8 v% p: W9 u
Suhail Doi There are plenty of powerful user-written packages that are not integrated into factory main modules. For example some commands for building publication quality tables, commands for executing latent and mixed models, not to mention numerous commands for data management and graph written by Nick Cox. The Mplus recommended by Richard Williams may also be regarded as an example although it is another commercial statistical software. I guess StataCorp just don't want to destroy users' custom and increase their burden, 'cause old-users may be used to exercise user-written packages for certain task.
: X+ ]7 _. p9 L5 o4 E! f9 B6 \; [. o" k
6 T# L6 m  V+ z! j: V$ V
#44 The introduction of meta-analysis in official Stata is not competition, but rather signals a firm long-term commitment by the company to the field, which thus includes technical support. The acknowledgments on pp.44-45 of show how this introduction builds on a great deal of community-contributed work over more than 20 years. What those acknowledgments don't spell out, and no blame attaches, is that user-programmer support and maintenance for even the most important community-contributed commands has sometimes been a little uncertain, given personal changes of various kinds, including death (alas), retirement, moving to other fields, being too busy, and so on and so forth. Positively, there is also a long history of programmers handing on the baton to others. Naturally users can continue with whatever commands they want to employ.
* {1 @9 U9 e( ^* M
4 R  t; F) {+ B& IThis is a personal interpretation, and not advertising copy!
; \* i* {, F1 @" [: w8 e) E5 Z$ ?9 P) U( |# g3 D2 q5 z

2 k. i) w2 H1 O/ X+ f#46 Agree with you on that point, my colleague sadly passed away in mid 2017 and development of our MetaXL software () has become considerably challenging. I am also grateful for the user developed commands that currently exist so certainly Stata (all versions) still has robust options in this area.( ~2 y1 C  w! W* ?

0 ?$ L% |5 m! y* e7 y' d
/ m! F& s* R2 BI would like to add to Nick Cox on the question raised by Suhail Doi regarding the new meta commands. Regarding the other commands on meta-analsyis, I would concur with Nick Cox that many of those programmes were maintained by individuals and sometimes it would take ages for them to be updated.
# D* \' t- ^9 R% [* D/ P3 A3 G$ _: M6 s/ i; n
It is great that Stata has produced this suite and now we can be confident that it will be supported and updated going forward. I have been reading the meta documentation in detail, have tested it for a couple of weeks and am currently running an analysis for publication. I can confirm that the new meta suite of commands is very user friendly and contains almost all of the functions that would be needed by someone running a meta-analysis. There are probably only a few things I could do with the flexibility of admetan, metan etc than in meta suite itself; for 95% of people it is easier to use, graphics are splendid and the saved results are detailed. I really like the subgroup analysis implementation. The documentation is also very well written; it is clear and gives a good introduction to meta-analysis and the discusses some recent advances in this area. Good work.
3 s7 t$ Q7 ^' [4 G- U: O+ Z$ Y% Q, `8 C& _+ n+ e$ K& F
There are a couple of things that can improve this suite further, I will be giving my feedback to Stata about these. 3 L; Z1 q$ ]. s. T
( S& M- l8 a$ S9 A

本帖被以下淘专辑推荐:

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

提现|充值|至尊会员||公卫人 ( )

GMT+8, 2019-9-18 11:49 , Processed in 0.137498 second(s), 30 queries , Gzip On.

Powered by X3.4

© 2001-2017

快速回复 返回顶部 返回列表
北京赛车时间表 江苏快3 pk10代理网址 北京赛车论坛 秒速时时彩 秒速时时彩官网 安徽快3 秒速时时彩平台 北京赛车微信二维码进群玩 华盈彩票网