Bioequivalence and Bioavailability Forum

Dear all,

because bad habits apparently die hard, I am proud to present a fantastic new piece of buggy R code.

It is nothing less than a sensational Silly-o-Meter which takes the temperature of the way a model is specified and gives a diagnosis.

ElMaestros.Silly.o.Meter=function (Some.Model)
{
  mx = model.matrix(Some.Model)
  nRows = dim(mx)[1]
  nCols = dim(mx)[2]
  nSilly = 0
  for (c in 1: nCols)
   {
     n=0
     for (r in 1: nRows) {if (mx[r,c] != 0) n = n+1}
     if (n==0) nSilly = nSilly+1     
   }
  x = 100 * nSilly/nCols
  cat("Congratulations! You specified a model with a model matrix that is", x, "percent meaningless.\n")
}

How to use this magnificent tool:
If you have a model specified like
M = lm(logAUC ~ 0 + Trt + Per + Subj%in%Seq + Seq)

then simply run it with the simple syntax:
ElMaestros.Silly.o.Meter(M)

Good luck to everyone. Am looking forward to hearing some hi-scores.