Related Posts
More Posts
Carl Nassib 👏🏻👏🏻👏🏻
Accenture India Lately there are so many hr openings with Accenture in almost all portals naukri, indeed, monster, linked in etc but there is not a single response infact I have the cc number wherein status is reflecting as hr will contact you ..but I have not received any revert, tried emailing the hr through a reference got an auto email stating out of office nd shared his contact number to connect, when tried calling hr,he was annoyed but I guess he forgot that he himself has shared the number. Accenture Ind
Age you made manager?
They are so darn cuddly!

Additional Posts in Politics
Trump still claims the tornado hit Alabama? 🤡
NYT must do better

Do as I say, not as I do

Well, it looks like everything's cleared up then

This should go well

New to Fishbowl?
unlock all discussions on Fishbowl.





538 got 2016 wrong
Always love 538. Good to read both. And btw, just a pre-warning: anyone who says that 538 “got 2016 wrong” is an idiot. Those people always find these threads, and they should go back to elementary school to learn how probability works.
Really like 538 but wonder if the methodology they use on poll corrections feels like republicans circadian 2012 saying to “unskew the polls”. And if so, will dems be in for surprise on election night. It’s working in both directions but eyeballing it looks like most of the corrections are in the Dems favor
I don’t see a Dem bias. The Senate numbers have become more GOP-leaning lately.
538’s numbers are good. Their punditry is bad.
Correcting for historical house effect is not “unskewing.” Unskewing is using a factor built into the polls and double-counting it. For instance: “I think Beto will win even though he’s down by 4 now because his team’s ground game is much stronger.”
Beto might win, but the polls will have been ‘wrong’ because of sampling error, not because of ground game. The ground game is built in.
I h. I thought the whole unskewing thing from 2012 was folks saying that the demographic waiting’s weren’t correct so if u re-weighted to what they though the electorate would look like (eg leas black and more in line with historical figures) Romney would have been leading more in polls. From 538 - I know they’re doing corrections but a bit tough to parse everything.
538 and others got 2016 wrong because they use polling data which is aspirational in nature and fails to capture implicit bias. Polling / focus group data is always sub-par in dynamic environments. The aspiration nature helps for some advertising uses but fails for others. On pure political prognostication, U of Trento in Italy did a better job but still not ideal. They nailed BREXIT. The science of prediction (threat detection / extracting implicit bias and true customer intent) has very little relationship to polling and what 538 does. News flash: When we find foreign manipulation events that move large groups it generally has nothing to do with hyper-targeting. You don’t move a crowd by tapping everyone on the shoulder and asking their name