please add cutoffs in add.significance.stars()

[landroni]

Please add a cutoffs=c(0.05, 0.01, 0.001) argument to add.significance.stars(). Now the function defaults to the same cut off points as used in base R, c(0.05, 0.01, 0.001), but in other contexts other values may be more suitable (e.g. c(0.10, 0.05, 0.01).

[landroni]

Here's my attempt at addressing this:

add.significance.stars <- function(p, cutoffs=c(0.05, 0.01, 0.001)){
    stopifnot(length(cutoffs)==3)
    if (inherits(p, c("matrix", "data.frame")) && length(dim(p)) == 2) {
        apply(p, c(1, 2), add.significance.stars, cutoffs = cutoffs)
    }
    else {
        if (length(p) > 1) {
            sapply(p, add.significance.stars, cutoffs = cutoffs)
        }
        else {
            ifelse(p > cutoffs[1], "", ifelse(p > cutoffs[2], " *", 
                                              ifelse(p > cutoffs[3], " * *", " * * *")))
        }
    }
}

> xa <- c(2.34067096410392e-14, 0.045492875421418, 1.91932500195526e-40, 3.42330287138604e-15, NA)
> add.significance.stars(xa)  ##R defaults
[1] " * * *" " *"     " * * *" " * * *" NA      
> add.significance.stars(xa, c(0.10, 0.05, 0.01))  ##stargazer defaults
[1] " * * *" " * *"   " * * *" " * * *" NA      

[daroczig]

Thanks for this idea -- please see above commit & let me know how it works.

[landroni]

Looks good! Thanks