19 word tweet is not a "good reply" to a statistical analysis. That is exactly the problem. For too many, a "good reply" is a reply that reinforces preexisting beliefs. A true good reply would tell me where the original analysis got it wrong (which they very well may have)