So I was doing a little bit of research into early exit polling data during the election, and I came across an interesting article that went through the data (I'll post a link below). It shows that exit polls underestimated Trump's performance nationwide by about 1.5%, but was also off in some key states. It shows Ohio as a 0.1% lead for Trump (was 8 points by the end), it shows Trump winning Minnesota by a slim margin (clinton won by a slim margin), while Clinton would win NC, Florida, Pennsylvania, and Wisconsin by varying margins. It also shows Michigan as a tie (which was actually pretty close to the final result), and the results show that Trump was underestimated by as much as 11 points (New Jersey), while being overestimated by as much as 5 points (New York). Now it doesn't show the results of some states, but I really don't think any of the states omitted from this chart were ever in doubt.
Link to article:
http://heavy.com/news/2016/11/2016-exit-polls-did-hillaty-clinton-win-presidential-election-voter-fraud-donald-trump-lose-rigged/