p values/stars in plot when input is not a model

[DominikVogel]

First of all, thanks a lot for this very helpful package!

I am dealing with a quite specific issue. I am working with data of employees nested in teams and therefore want to use clustered standard errors. Hence, I fit the polynomial model on my own and use the resulting coefficients, standard errors, and covariances in RSA.ST() to get the response surface values.

However, when I plot the results using plotRSA(), there is no way to add the SEs and COVs and therefore the displayed response surface values a1-a5 do not contain a star indicating a significant value. Is there a way to modify plotRSA() to display a1-a5 with the stars?

I suspect the problem is that in line 331 of plotRSA.R the functions calls RSA.ST() but without SEs and COVs. I tried to modify plotRSA() to accept SEs and COVs as input and use them in line 331 but it did not work. Did I look at the wrong part of the code?

Thanks a lot!

[DominikVogel]

I made a mistake when modifying plotRSA(). I forgot to provide the degrees of freedom too.

I submitted a pull request (#17) to enable others to benefit from the modification.