Hacker News Reader

Comment by king_panic | Hacker News Reader

We are one country that's been divided by a news media telling us what to believe. Look no further than the dead wrong "polling data" -- again -- to see there's an agenda playing out in our hands and living rooms. If you still believe what you see on MSM, I've got a bridge to sell you.

We don't talk to others anymore. We tweet, we Facebook, we Instagram, we believe what we read on screens and we've made companies feeding us what we want to hear rich.

There is a game being played, you've just missed the players.

View on HN

> Look no further than the dead wrong "polling data" -- again -- to see there's an agenda playing out in our hands and living rooms.

I think this is a tired, lazy argument. First, while the polls were skewed they actually weren't that wrong, and many of the differences are magnified by the fact that the electoral college forces us to chop up one large poll (the national results) into lots of small ones (the states) where statistically you're likely to see more variance no matter what. And as for the skew, the generally accepted explanation that it's easier to poll urban, more highly educated people than rural, lower educated people makes much more sense than there being some coordinated conspiracy.

Don't kid yourself. The polls were dead wrong, they predicted Joe Biden would win an overwhelming popular vote. Believing otherwise is naive.

There is no "generally accepted explanation" for the polling skew, let alone one as obvious as "it's easier to poll highly educated, urban voters than lesser educated rural voter". Believing pollsters are dumb enough to not consider this when conducting polls is lazy thinking.

> The polls were dead wrong, they predicted Joe Biden would win an overwhelming popular vote.

But Biden did win an overwhelming popular vote.

The fact that's your lone point of dispute is reassuring. Thank you.

He's up over 4 million votes. Glad you are reassured.

"lone"? That was the only point you made about how the polls were right or wrong! They disputed all of your one point(s).

> The polls were dead wrong, they predicted Joe Biden would win an overwhelming popular vote.

Major poll-based forecasts (which not only aggregate polls but incorporate modelling of non-sampling error in polls, which only report sampling error in the margin of error) were not “dead wrong” (naively looking at polls as predictions might be dead wrong, but the fact that polls don't work that way is why poll-based forecasts like 538 exist.) The low end of the 80% confidence interval (the 95% CI is what is usually used to state a “margin of error” when a CI isn't explicitly identified and would be much wider) for the 538 forecast for Biden’s vote share was ~50.9%. His actual vote share at current count is ~50.7%.

Results a hair outside the 80% CI doesn't indicate that a forecast is “dead wrong”.

EDIT: While I think 538 has made some progress in simplifying it's reporting to help deal with this, aside from deliberate misinformation (which is a substantial factor in the “criticism”), I think the main problem remains that people really, really don't grasp uncertainty even when it's shoved in their face, and are prone to criticize something for not being a precise oracle even when it's expected error range is prominently presented.

He's up over 4 million votes, though?

In an election with (currently) 146,592,751 votes. I'm sorry but 4 million is _not_ "overwhelming". It's a difference of less than 3% of total voters.

That very much depends on how much of those original 146 odd million votes are 'mobile', and with the way people vote straight 'R' or 'D' for life the margins are a lot thinner than you might expect. > 40% of both parties voters would never in a lifetime vote for the other party no matter how terrible the candidate for 'their' party.

Point taken.

Reply on news.ycombinator.com