我的论文三修了,但有一个问题一直没解决,现在主编下了最后通牒,说解决这个问题是论文接收的先决条件。 急躁ing...
我分别作了雄树和雌树的生物量方程。(胸径自变量,生物量为因变量)
通过AIC和RMSE选出的最优模型方程的形式是一样的(y=a*x^b)
雄树:y=a1*x^b1
雌树:y=a2*x^b2
然后我说a1不等于阿,b1不等于b2,所以性别间有差异。
但审稿人认为这样没有统计学意义。 要求我用统计学的方法来说明这两个方程参数有差异或者方程有差异。
现在我自己有2个方案:
方案一:将方程转换成线性方程后,用chow 检验(但对数转换的方程曾经遭到审稿人反对)
方案二:将x在取样胸径范围内随机取10000个值,同T检验比较两个方程的y
我不知道response letter的时候说上这两种方案,然后文章中只用其中一种。 还是说回复的时候必须特别坚定地选一个?
或者可以给提出这个问题的责任编委Barry Gardiner发邮件问询吗?
有哪位大神干过这事儿没,在修改之前跟审稿的提出问题的编辑联系?
他的意见如下:
My suggestion is to take your best model for each of the biomass components and perform something like Hotelling's t-test (multivariate t-test). You need to do this to test whether the parameters of the models for male and hermaphrodite trees are statistically different. If you need to I would suggest asking help from a statistician.One option would be to fit the model for eachtree and get the different values of parameters (p0-p3) for the males and thehermaphrodites and then do a t-test to see if the population of p0, p1, p2 andp3 are significantly different. So for example the model for flower using onlydbh as a variable has 2 parameters p0 and p1. If you get p0 and p1 for all themale trees and all the hermaphrodite trees do the populations of p0 and p1 varystatistically between the male and hermaphrodite trees?