关于小世界网络文章reviewer的返回意见,严老师和臧老师能否帮我分析一下,万分感谢

Submitted by zhangyue_one on
严老师,臧老师:
       您们好!我前段时间投稿一个病人(PHN疼痛病)小世界研究的文章后,reviewer返回了这个意见,我查阅了很多资料,不太清楚,应该选取哪种合适的统计方法来回答reviewer的问题,因为local efficiency 和global efficiency 是两个不同的测量量,PHN(一种疼痛)对local efficiency 的影响怎么和对global efficiency 的影响比较起来?非常感谢两位老师!
                                                                                                                               yue zhang

- "Certainly the data presented does not allow the conclusion that PHN was associated with more impairment to short range connections than the long-range connections. There are at least two reasons for this:
o A. Significant effect on local efficiency combined with non-significant effect on global efficiency does not mean that effect on the local efficiency would be significantly different from the effect of global efficiency. However, it is easy to verify whether the effects are significantly different and this is what the paper should report (in addition to the statistical tests presently reported).
o B. The test above still concerns global efficiency and local efficiency with only indirect linkage to short and long range connections. However, also the effects of PHN pain to short and long range connections could be tested by correlating the connection length and the strength of the connection (correlation coefficient in the correlation matrix that is thresholded to create a network) in HCs versus PHNs. I am not an expert enough on PHN to recommend if such a test if useful or not, but to claim something along these lines, a proper test is required. "

The point in A was not optional, it is mandatory. I am sorry for not communicating this clearly enough. A significant effect on local efficiency and a not significant effect on global efficiency does not permit to make any claims about the difference between the two types of efficiency but these differences should be tested with a separate (paired) test. Of course, when dealing with different quantities (global and local efficiency), a specific care must be taken that this test is valid (i.e. report the standard deviations also).

Also, it should be clearly stated in the discussion what conclusions are made based on the data (local and global efficiency) and what are speculations based on those results (long and short range connections).

YAN Chao-Gan

Mon, 07/01/2013 - 15:59

1. 你把Eloc和Eglob分别做检验的?他质疑的是你的推论“effect on the local efficiency would be significantly different from the effect of global efficiency”,这个推论有点大。要么更改这个推论,要么做一些interaction的检验?心理学上有不少统计方法,但这个关于两个不同量交互效应的检验,我不太熟悉。
2. 我同意Eloc和Eglob不能解释为长距和短距。长距和短距可以直接检验的,比如说用欧氏距离。
Forums