As Eric Schnell points out, although ISI impact factors have been criticized before, a recent editorial in the Journal of Cell Biology addresses the validity of the underlying data.
Now might be a good time to re-visit the h-index, which Scopus employs as an alternate method of estimating scientific impact.