近日，一篇已发表的文献[1]也专门研究了这一情况，聚焦的则是顶尖的期刊：NEJM、柳叶刀和JAMA，纳入的是近20年来发表在这三个顶级期刊上的研究。主要挑选那些结果的P值接近0.05的文章：（1）P＜0.05（0.036到0.050之间）；（2）P＞0.05（0.050到0.064之间）

 单独来看一下P＞0.05的情况，在97篇文章里，可以看到有40篇文章在面对略高于0.05的P值时，存在暗示有统计学差异的英文表述。

 接下来让我们一起欣赏下顶级期刊的论文的英文写法（注：仅挑选一部分，为了原汁原味的展示作者们的措辞，以下英文描述不再翻译）。

 期刊来源：JAMA

1、There were 5 stent thromboses (0.7%) in the sirolimus-eluting stent group vs 13 stent thromboses (1.9%) in the paclitaxel-eluting stent group, a difference that nearly reached statistical significance (P = 0.06 by Fisher exact test)

2、After controlling for … the inverse relationship between amount of moderate to vigorous exercise and sudden cardiac death risk was attenuated; however, a trend persisted (P = 0.06 for trend)

3、Surprisingly, lower rates were observed for severe haemorrhage (44 vs 28; P = 0.06)

4、Increased income did not attenuate the association between being uninsured and using fewer services for diabetes, although increased income trended towards significantly increasing the likelihood of uninsured compared with insured diabetics receiving foot examinations (P for interaction = 0.06)

5、The age of energy from protein and carbohydrates marginally increased in the intervention schools compared with the controls (P = 0.07 and P = 0.06, respectively)

6、Sodium consumption, adjusted for energy intake, also marginally increased among intervention students (P = 0.06)

7、In addition, there was a borderline effect in which members of the comparison group demonstrated a somewhat higher hazard rate for diabetes mellitus (P = 0.06)