This is nonsense. Flesch-Kincaid has been applied to presidential debates or speeches before, it was nonsense then and it's nonsense now.
Fleisch-Kincaid is for written documents. It's dubious whether it's really of value even in that limited circumstance, but it makes no sense for transcriptions of written speech, where punctuation is basically up to whoever's transcribing it. The metric is strongly influenced by sentence length, so you can see why it would cause problems.
Language Log has written about this and why it's ridiculous in detail before.