Unsurprisingly, those turned out to be fabrications. Regardless, these models aggregate dozens to hundreds of polling organizations. If Monmouth tried to 'create a narrative,' as described in the purported memo it would just end up looking out of step and would lose credibility.

> The "leaked memo" is a small test of readers' credulity. Written in apocalyptic terms, with strange capitalization ("the Liberal base is demoralized — but will become more enthusiastic the more Hillary is seen as an iconoclast who will overturn the last foundations of Western Culture"), it was uploaded to Scribd by a pro-Trump website with the confidence-inspiring name of RealTrueNews. For a short while, it was credited to Nate Silver; his name was blacked out, rendering it slightly less odd that someone who did not work for Monmouth was credited on a Monmouth polling memo.

https://www.washingtonpost.com/news/post-politics/wp/2016/09...

source? id like to read this

That would explain why campaign intern polling is always such a sacred ground.