They didn't understand that, errors correlated means that a 1-2% shift changes the whole election. Different than 90% chance!
-
-
Replying to @zeynep @derekwillis
A lot of these complaints don't square w/ fact that 538 had the odds of an Obama W @ >90% in '12. Far greater than 71%
1 reply 0 retweets 1 like -
Replying to @Drew106 @derekwillis
My issue isn't forecasts have error and uncertainty. Of course they do. They need to communicate this better.
1 reply 0 retweets 1 like -
Replying to @zeynep @derekwillis
@NateSilver538 explicitly refuted both the "blue wall" idea and that states move independently of 1 another many times1 reply 0 retweets 0 likes -
I read it! On twitter. In footnotes. Long form. It was correct. Topline continued to not convey this.
2 replies 0 retweets 2 likes -
Risk losing nuances when polling is presented as giant theoretical team scoreboards.
1 reply 0 retweets 0 likes -
538 was better than anyone else; but my issue is conveying of error pattern. That was lacking.
2 replies 0 retweets 6 likes -
Replying to @zeynep @NateSilver538 and
The problem with forecasting is a false certainty which may influence campaign coverage
1 reply 0 retweets 0 likes -
Replying to @Lee__Drake @zeynep and
In a sense, it may encourage journalistic herding on what issues are important
1 reply 0 retweets 0 likes -
Replying to @Lee__Drake @zeynep and
This is not a criticism of models, but rather of their interpretation by journalists
2 replies 0 retweets 1 like
It's on the data journalists to push for better communication. Journos are often statistically blind.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.