[R-sig-eco] Quantile regression

Allan Edelsparre aedelspa at uoguelph.ca
Tue Nov 18 16:39:17 CET 2008


I have a problem with quantile regression when I want to analyse quantiles above the 90th quantile and below the 5 quantile. When I check for significance I use the following commands:

>results<- rq(Total~Pmove, tau = 0.9)
>summary(results, se="nid")

Call: rq(formula = Total ~ Pmove, tau = 0.9)

tau: [1] 0.9

Coefficients:
            Value     Std. Error t value   Pr(>|t|) 
(Intercept)  24.10000   2.13566   11.28456   0.00000
Pmove       -15.00000   3.28563   -4.56533   0.00003


### which is fine, but when I want to look at the 95th quantile I get the following output:

> results<- rq(Total~Pmove, tau = 0.95)
> summary(results, se="nid")

Error in summary.rq(results, se = "nid") : 
  tau + h > 1:  error in summary.rq

### Does anyone know what I am doing wrong here?



More information about the R-sig-ecology mailing list