Maybe I'm just too skeptical of these types of stories coming from generic polling, but I will say that I don't have a high degree of confidence in polling done by random phone calls (the CNN/ORC poll listed says that on page 14 [1]). My wife frequently answers the political poll questions when people call, in part because she feels sorry for the poller. Answering those calls can take 15-30 minutes. So you're depending on the results of people who have 15-30 minutes to answer 35 questions about the government. I feel like there's a selection bias at work there already, and you're assuming that people will give thoughtful answers to the questions, that the order of questions doesn't affect responses, etc, etc, etc. You're depending on 1018 people who actually may not be statistically relevant, since America is a big, big country, to drive your thinking on that country.

Plus, would the responses to the surveys be different if the word 'disappointed' was used instead of angry? And would the perception of the responses be different?

Secondarily, and this is far more anecdotal, but I am friends with a fairly wide swath of people via social media, from conservatives to libertarians to liberals to progressives. The key takeaway that I have from watching all of their posting and all of their interactions is that people are angry at others. And people have always been easily angered by others, it's just that the modern American landscape allows for a) easier access to others and b) easier access to well-written, rational condemnation of others than can be easily shared. But reading through my feed,

Maybe Americans aren't actually angry, but since it's easier to construct understandable stories using language like anger, Americans feel like they should be angry?

[1]http://edition.cnn.com/2015/12/29/politics/cnn-orc-poll-full...