Formulating any analysis of the Postal Service on 100 letters is like formulating an analysis of the internet based on 1000 packets. Something longitudinal with a vaguely statistically significant sample set would be more meaningful
It seems representative enough to warrant a follow-up investigation, instead of dismissing it as a statistical fluke entirely. All local media AND individuals can try and reproduce this, from anywhere, to anywhere.