A simple explanation for that is that facts exist and impact the coverage presented by media outlets that make some effort to do what their overt role is.
I mean, alternatively, you could suspect,as you imply, that all non-fringe media that aren't state run propaganda arms of US rivals are, in fact, US state propaganda outlets pushing CIA/NSA scripted narratives.
Well, except when Paul Manafort isn't planting propaganda for Russian client regimes in them, at any rate.
"Hillary is going to win the election!" Was certainly propaganda.
I don't have any theory, but you can see they all somehow sing to the same tune, including UK and EU. They want to give their approved opinions instead of reporting reality. With the exception of the ocasional independent journalist.
I wasn't necessarily meaning that TASS or Reuters were scripted or censored by CIA/Kremlin. However we can pragmatically assert both agencies have biases (cultural, political, financial, etc).
For example, according to Russia, the British government released almost no data about the toxic nerve agent case, no specific proof was made public. Other governments (France, US) said they support the UK.
Now that did lead to more sanctions on Russia, and the expulsion of Russian diplomats, but from an external observer point of view, it isn't rigorous at all. IMHO, many more reports and documents should be made public, especially about the methods used to conclude that the Russian government is the source of the attack.
Would the general public arrive to the same conclusions than the governments?
I'm 27 years old, I didn't live during the ex-URSS or the Cold War. All I see is 2 camps accusing each other and escalating without serious proof, and people are supposed to say nothing about it? To me, the days when we systematically accused the Russian government of all the shady stuff are supposed to be gone, and we are supposed to use rigorous tribunal procedure before sanctioning an entity, especially if it means deteriorating international relations
An opinion that one of two candidates would win the election was "certainly propaganda?"
Those making the 95% prediction were using statistical models established for that purpose before the identity of the candidates or what those models would show were known. They quite arguably are bad models as they assumed state level deviations from polling results are independent where history suggests that, in fact, they are strongly correlated (a fact pointed out by Nate Silver prior to the election, in explaining why 538 had a much lower projection of Clinton's probability of victory.) But choosing a bad model isn't propaganda.
A poll, your favorite. Let's hope they used a good model.