证据中国

 找回密码
 立即注册

QQ登录

只需一步,快速开始

搜索
热搜: 视频教程
查看: 1367|回复: 1

假如你的p>0.05怎么办?这里告诉你一些方法。实用经验,包括生存曲线

[复制链接]

17

主题

30

帖子

96

积分

注册会员

Rank: 2

积分
96
发表于 2017-10-30 01:04:39 | 显示全部楼层 |阅读模式
本帖最后由 277051445 于 2017-11-1 00:31 编辑

很多时候实验结论并不能一蹴而就,并且实验是做出来了,趋势很明显,但是P>0.05,应该怎么办???我该如何解释我这样的P值呢?
答案分为以下几点。

1、你的P值确实大太多:比如0.3或者0.8等等,那么你就说他没有意义就行了,很多时候有可能你需要重新做实验,确定你的阴性结果,多少大牛的文章,NCS的文章就是因为总是做不出阳性结果,从阴性结果出发,而发现新世界所以你的P值较大,没意义,也未尝不是好事。
2、P接近于0.05但是仍然稍微大一点:这种情况可能是最经常遇见的,并且国内文章总体来说,这个其中的猫腻,大家都懂,后面处理他的方法也很多,包括用种种手段使其最终小于0.05,虽然我说的直接,但是这确实是真理,是事实,这事情都能写一个社论发表NATURE了。他们绝对愿意看这个。最近看见一个帖子,里面总结了一些常用的英语,来忽悠审稿人解释P微弱大于0.05的句子,好几百句,随便学两句吧,以后能用来自己实践。
尝试了各种合适的统计学方法,P值依然稳稳地略高于0.05,你应该得出结论“无统计学意义(non-significant)”,并这样写到文章里。然而,对于很多作者来说,这并不是他们想寻找的答案:因为感觉发布“阴性结果”的文章要比“阳性结果”的文章难呀。
很多人的解决方案就是采用由来已久的迂回策略,将无统计学意义的结果装扮地更有趣。下面这个列表,来自于同行评审期刊中的文章,这些文章的共同点是:(1)作者设定P<0.05为具有统计学意义;(2)结果未能达到P<0.05;(3)文章在描述的时候让P值看起来更有意思了。为了原汁原味的展示作者们的有趣措辞,以下英文描述不再翻译,
  • hint of significance (p>0.05)/only a little short of significance (p>0.05)/a strong tendency towards statistical significance (p=0.051)/significantly better overall (p=0.051)/slightly missed the margin of significance (p=0.051)/narrowly avoided significance (p=0.052)
  • not exactly significant (p=0.052)/a possible trend toward significance (p=0.052)/(barely) not statistically significant (p=0.052)
  • borderline level of statistical significance (p=0.053)
  • just lacked significance (p=0.053)
  • at the very edge of significance (p=0.053)
  • on the cusp of conventional statistical significance (p=0.054)
  • at the limit of significance (p=0.054)
  • close to being statistically significant (p=0.055)
  • not unequivocally significant (p=0.055)
  • noticeably significant (p=0.055)
  • on the boundary of significance (p=0.055)
  • very closely approaches the conventional significance level (p=0.055)
  • weakly statistically significant (p=0.0557)
  • better trends of improvement (p=0.056)
  • near to statistical significance (p=0.056)
  • just marginally significant (p=0.0562)
  • at the verge of significance (p=0.058)
  • just borderline significant (p=0.058)
  • hovered at nearly a significant level (p=0.058)
  • in the edge of significance (p=0.059)
  • on the threshold of significance (p=0.059)
  • virtually significant (p=0.059)
  • on the very borderline of significance (0.05<p<0.06)
  • an observed trend (p=0.06)
  • approached near significance (p=0.06)
  • at the brink of significance (p=0.06)
  • barely outside the range of significance (p=0.06)
  • close to being significant (p=0.06)
  • equivocal significance (p=0.06)
  • just outside the traditional bounds of significance (p=0.06)
  • leaning towards statistical significance (p=0.06)
  • not as significant (p=0.06)
  • not currently significant (p=0.06)
  • slight non-significance (p=0.06)
  • slightly exceeded significance level (p=0.06)
  • tended to be significant (p=0.06)
  • almost significant tendency (p=0.06)
  • an inverse trend toward significance (p=0.06)
  • slightly failed to reach statistical significance (p=0.061)
  • slightly missed the conventional level of significance (p=0.061)
  • fell narrowly short of significance (p=0.0623)
  • near nominal significance (p=0.064)
  • not significant in the normally accepted statistical sense (p=0.064)
  • very nearly significant (p=0.0656)
  • closely approaches the statistical significance (p=0.0669)
  • barely failed to attain statistical significance (p=0.067)
  • slightly below the level of significance (p=0.068)
  • a considerable trend toward significance (p=0.069)
  • approaching clinical significance (p=0.07)
  • difference was apparent (p=0.07)
  • fell marginally short of significance (p=0.07)
  • just fails to reach conventional levels of statistical significance (p=0.07)
  • nearly reached a significant level (p=0.07)
  • slightly insignificant (p=0.07)
  • tendency toward statistical significance (p=0.07)
  • close to borderline significance (p=0.072)
  • a barely detectable statistically significant difference (p=0.073)
  • slightly outside the significance level (p=0.077)
  • did not quite reach conventional levels of statistical significance (p=0.079)
  • a trend close to significance (p=0.08)
  • an expected trend (p=0.08)
  • approaching, although not reaching, significance (p=0.08)
  • not significant in the formal statistical sense (p=0.08)
  • not very definitely significant from the statistical point of view (p=0.08)
  • almost non-significant (p=0.083)
  • not fully significant (p=0.085)
  • approaching a level of significance (p=0.089)
  • not highly significant (p=0.089)
  • approaching, but not reaching, significance (p<0.09),a numerical trend (p=0.09)
  • a significant trend (p=0.09)
  • a small trend (p=0.09)
  • highly significant (p=0.09)
  • tended to approach significance (p=0.09)
  • does not reach the conventional significance level (p=0.098)
  • very slightly significant (p<0.1)
  • almost clinically significant (p<0.10)
  • an adverse trend (p=0.10)
  • essentially significant (p=0.10)
  • not entirely significant (p=0.10)
  • possibly statistically significant (p=0.10)
  • very slightly non-significant (p=0.10)
  • tantalisingly close to significance (p=0.104)
  • approaching formal significance (p=0.1052)
  • not decisively significant (p=0.106)
  • tentatively significant (p=0.107)
  • verged on being significant (p=0.11)
  • possibly marginally significant (p=0.116)
  • fell just short of statistical significance (p=0.12)
  • fell somewhat short of significance (p=0.138)
  • approaches but fails to achieve a customary level of statistical significance (p=0.154)
  • close to the limit of significance (p=0.17)
  • a very slight trend toward significance (p=0.20)


3、再生存曲线最终结果中没有意义

生存曲线是一个有意思的东西,他最终结果没有意义,但是不代表其一定没有办法描述。
举个例子
f829e85499c3293f687cab7a292c361c (1)_副本.jpg
上图2处是分开的,1处开始交叉,3处开始完全分不开.也因此描述可以这么写,
1、在生存周期中前20来个月时候,高度恶性、中毒恶性、弥漫性大无裂三个分组之间有明显的统计学意义,且中度恶性的人死亡率较高度恶性低10%,较弥漫性大无裂低15%(注释:图是瞎找的,数据是瞎说的,自己可以量啊,比例尺会不会?纵坐标是生存率啊
2、在(1处)第90个月开始,三者生存率无统计学差异。
上面这种描述总是出现在乳腺癌的研究中,尤其是乳腺癌的随机对照实验中,特别是有生存分析的临床实验中。乳腺科的人懂我说了啥,但是这个也能借鉴到其他的论文中。
回复

使用道具 举报

17

主题

30

帖子

96

积分

注册会员

Rank: 2

积分
96
 楼主| 发表于 2017-10-30 01:10:35 | 显示全部楼层
http://www.evidencechina.com/for ... hread&tid=18&extra=
帖子还没完,下面是其他的
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则

Archiver|手机版|小黑屋|Evidencechina Inc.

GMT+8, 2018-12-17 22:20 , Processed in 0.180221 second(s), 23 queries .

Powered by Discuz! X3.4

© 2001-2017 Comsenz Inc.

快速回复 返回顶部 返回列表