|
Now that we have a new thread, in order to ensure that this thread continues to meet TL standards and follows the proper guidelines, we will be enforcing the rules in the OP more strictly. Be sure to give them a complete and thorough read before posting! NOTE: When providing a source, please provide a very brief summary on what it's about and what purpose it adds to the discussion. The supporting statement should clearly explain why the subject is relevant and needs to be discussed. Please follow this rule especially for tweets.
Your supporting statement should always come BEFORE you provide the source.If you have any questions, comments, concern, or feedback regarding the USPMT, then please use this thread: http://www.teamliquid.net/forum/website-feedback/510156-us-politics-thread |
CNN's ratings continue to slump by the looks of things. Not good at all.
|
On May 08 2019 08:27 Doodsmack wrote:
Compare this fact to Trump's public persona, and you can see that Trump is a fraud through and through. We probably should have a sitting special counsel for the duration of Trump's time in office, monitoring everything he does and says. Otherwise, he can't be trusted.
Beat me to it. Few things.
1. Trump is nothing if not a con-man. His most impressive con has been convincing millions of people that he is some elite, genius businessman. I wont deny that he is a good con-man.
2. Based on what is known I think a legitimate argument could be made that Trump is actually the worst businessman in America. This will likely become more obvious when (not if) his tax returns are torn into. Between NY state going after them and the House's request (Mnuchin is going to get crushed in court), they will be known soon enough.
3. Where was all this shit back in 2016, NYT? So many stories in the last 2 years about someone who is a bad businessman, and worse person, but all too late to stop the trainwreck.
|
On May 08 2019 08:27 ChristianS wrote:Show nested quote +On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:On May 08 2019 03:42 GreenHorizons wrote:On May 07 2019 23:15 ChristianS wrote:On May 07 2019 13:29 GreenHorizons wrote:On May 07 2019 09:47 Mohdoo wrote:On May 07 2019 06:52 Plansix wrote:
My favorite candidate will be the one that knocks Biden out of the race, because my god we need to escape the 1990s Democrats. Next thing you know he will be apologizing for Bill Clinton. Jeb looked favored at one point too. Biden has no charisma. His numbers will drop like a brick after the first debate. I don't think it's even going to take a debate, from what I gather these recent polls showing biden surging (beyond what's typical for an announcement) we're really polling people over 50 which is about the split we'd expect and Biden's had for months. Good chance his polling of people under 50 is already plummeting, particularly when one considers he's been pretty universally shitcanned here. I hesitate to post this because I don’t like horse race coverage of elections, not least because people tend to intentionally or unintentionally use it as a cover for advocacy (e.g. instead of saying “I support Buttigieg” people say “Buttigieg sure is doing well in the polls, let’s talk about it”). So let me explicitly say I have no intention of voting for Biden in the primary. That said, I think you’re misreading the polls on Biden and Bernie somewhat. Nate Silver has written some about this and can defend the point better than I can (although he’s had some takes recently I’m more or less certain you’d hate), but the gist is Biden has consistently outperformed Bernie in the polls and most of the methodological criticisms people have for those polls (e.g. sample size or sampling bias) are off-base. We can certainly expect Biden’s numbers to drop in coming weeks, but not because recent polls were badly done - post-announcement polling bounces usually fade, that’s normal and to be expected. I think a lot of your prediction that Biden will fail is based more on your intuition than on the data; I won’t argue with your intuition there. But as far as the data is concerned, I have trouble seeing how you could look at Biden as anything other than a frontrunner right now. It's worth noting Nate Silver has basically become what he came to fix. A pundit that opines sans the data. The easiest example is when he tiered the candidates. You can put Biden and Bernie together or Biden on top then Bernie, but you can't have Buttigieg and Harris with Sanders, let alone Harris over Sanders. Or at least he didn't provide any data to support his pushing of Harris. As far as Biden's polling he has consistently held a lead, I can't speak to every recent poll but the CNN one showing him "solidifying his frontrunner status" definitely didn't have a significant sample from people under 50. cdn.cnn.comIf other recent ones showing his surge did I haven't seen it yet (could be, I just haven't seen it). Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart. On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm: If(Supports my hypothesis) retweet uncritically Else quibble with methodology has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst. Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is?
Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50.
Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
On May 08 2019 08:28 Gahlo wrote:Show nested quote +On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half.
Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak
|
On May 08 2019 08:51 On_Slaught wrote:Show nested quote +On May 08 2019 08:27 Doodsmack wrote:Compare this fact to Trump's public persona, and you can see that Trump is a fraud through and through. We probably should have a sitting special counsel for the duration of Trump's time in office, monitoring everything he does and says. Otherwise, he can't be trusted. https://twitter.com/maggieNYT/status/1125900816550907905 Beat me to it. Few things. 1. Trump is nothing if not a con-man. His most impressive con has been convincing millions of people that he is some elite, genius businessman. I wont deny that he is a good con-man. 2. Based on what is known I think a legitimate argument could be made that Trump is actually the worst businessman in America. This will likely become more obvious when (not if) his tax returns are torn into. Between NY state going after them and the House's request (Mnuchin is going to get crushed in court), they will be known soon enough. 3. Where was all this shit back in 2016, NYT? So many stories in the last 2 years about someone who is a bad businessman, and worse person, but all too late to stop the trainwreck.
It would be amusing, if they are as bad as perceived due to the situation. He can't deny their contents as being incorrect without saying his returns are knowingly wrong and we'd see him attack the IRS for producing a fake return as the deep state umbrella widens.
|
On May 08 2019 08:52 GreenHorizons wrote:Show nested quote +On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:On May 08 2019 03:42 GreenHorizons wrote:On May 07 2019 23:15 ChristianS wrote:On May 07 2019 13:29 GreenHorizons wrote:On May 07 2019 09:47 Mohdoo wrote:
[quote]
Jeb looked favored at one point too. Biden has no charisma. His numbers will drop like a brick after the first debate. I don't think it's even going to take a debate, from what I gather these recent polls showing biden surging (beyond what's typical for an announcement) we're really polling people over 50 which is about the split we'd expect and Biden's had for months. Good chance his polling of people under 50 is already plummeting, particularly when one considers he's been pretty universally shitcanned here. I hesitate to post this because I don’t like horse race coverage of elections, not least because people tend to intentionally or unintentionally use it as a cover for advocacy (e.g. instead of saying “I support Buttigieg” people say “Buttigieg sure is doing well in the polls, let’s talk about it”). So let me explicitly say I have no intention of voting for Biden in the primary. That said, I think you’re misreading the polls on Biden and Bernie somewhat. Nate Silver has written some about this and can defend the point better than I can (although he’s had some takes recently I’m more or less certain you’d hate), but the gist is Biden has consistently outperformed Bernie in the polls and most of the methodological criticisms people have for those polls (e.g. sample size or sampling bias) are off-base. We can certainly expect Biden’s numbers to drop in coming weeks, but not because recent polls were badly done - post-announcement polling bounces usually fade, that’s normal and to be expected. I think a lot of your prediction that Biden will fail is based more on your intuition than on the data; I won’t argue with your intuition there. But as far as the data is concerned, I have trouble seeing how you could look at Biden as anything other than a frontrunner right now. It's worth noting Nate Silver has basically become what he came to fix. A pundit that opines sans the data. The easiest example is when he tiered the candidates. You can put Biden and Bernie together or Biden on top then Bernie, but you can't have Buttigieg and Harris with Sanders, let alone Harris over Sanders. Or at least he didn't provide any data to support his pushing of Harris. https://twitter.com/NateSilver538/status/1116106327208747008As far as Biden's polling he has consistently held a lead, I can't speak to every recent poll but the CNN one showing him "solidifying his frontrunner status" definitely didn't have a significant sample from people under 50. cdn.cnn.comIf other recent ones showing his surge did I haven't seen it yet (could be, I just haven't seen it). Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart. On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm: If(Supports my hypothesis) retweet uncritically Else quibble with methodology has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst. Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. Show nested quote +On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak
Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this?
|
Mike Pompeo: Climate change isn't happening.
Also Mike Pompeo: Reductions in Arctic sea ice are opening new trade routes.
User was warned for this post
|
On May 08 2019 09:57 ChristianS wrote:Show nested quote +On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:On May 08 2019 03:42 GreenHorizons wrote:On May 07 2019 23:15 ChristianS wrote:On May 07 2019 13:29 GreenHorizons wrote:
[quote]
I don't think it's even going to take a debate, from what I gather these recent polls showing biden surging (beyond what's typical for an announcement) we're really polling people over 50 which is about the split we'd expect and Biden's had for months.
Good chance his polling of people under 50 is already plummeting, particularly when one considers he's been pretty universally shitcanned here. I hesitate to post this because I don’t like horse race coverage of elections, not least because people tend to intentionally or unintentionally use it as a cover for advocacy (e.g. instead of saying “I support Buttigieg” people say “Buttigieg sure is doing well in the polls, let’s talk about it”). So let me explicitly say I have no intention of voting for Biden in the primary. That said, I think you’re misreading the polls on Biden and Bernie somewhat. Nate Silver has written some about this and can defend the point better than I can (although he’s had some takes recently I’m more or less certain you’d hate), but the gist is Biden has consistently outperformed Bernie in the polls and most of the methodological criticisms people have for those polls (e.g. sample size or sampling bias) are off-base. We can certainly expect Biden’s numbers to drop in coming weeks, but not because recent polls were badly done - post-announcement polling bounces usually fade, that’s normal and to be expected. I think a lot of your prediction that Biden will fail is based more on your intuition than on the data; I won’t argue with your intuition there. But as far as the data is concerned, I have trouble seeing how you could look at Biden as anything other than a frontrunner right now. It's worth noting Nate Silver has basically become what he came to fix. A pundit that opines sans the data. The easiest example is when he tiered the candidates. You can put Biden and Bernie together or Biden on top then Bernie, but you can't have Buttigieg and Harris with Sanders, let alone Harris over Sanders. Or at least he didn't provide any data to support his pushing of Harris. https://twitter.com/NateSilver538/status/1116106327208747008As far as Biden's polling he has consistently held a lead, I can't speak to every recent poll but the CNN one showing him "solidifying his frontrunner status" definitely didn't have a significant sample from people under 50. cdn.cnn.comIf other recent ones showing his surge did I haven't seen it yet (could be, I just haven't seen it). Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart. On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm: If(Supports my hypothesis) retweet uncritically Else quibble with methodology has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst. Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this?
I already explained?
Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it.
|
School shooting in Denver. At least 8 injured.
Officials say eight students have been injured in a shooting at a public charter school in Highlands Ranch, Colo., a suburb south of Denver.
Two suspects are currently in custody. Douglas County Sheriff Tony Spurlock said one suspect is an adult male and the other is a juvenile male. Both are students at STEM School. Spurlock said their names are being withheld pending further investigation. The suspects were not previously known to local law enforcement.
"This is a terrible event. This is something that no one wants to have happen in their community," Spurlock said in a press conference late Tuesday afternoon. "Two individuals walked into the STEM school, got deep inside the school and engaged students in two separate locations."
"We do have eight students in area hospitals right now. Several of them are in critical condition," he added.
All of the victims are 15 years and older.
Source
Not much to go on besides a lot of prayers and thoughts being given from the reporting. Will probably include more information over the coming days, but this is pretty standard these days in the US. I wonder how the Democrat candidates will/have responded.
|
On May 08 2019 09:51 Gahlo wrote:Show nested quote +On May 08 2019 08:51 On_Slaught wrote:On May 08 2019 08:27 Doodsmack wrote:Compare this fact to Trump's public persona, and you can see that Trump is a fraud through and through. We probably should have a sitting special counsel for the duration of Trump's time in office, monitoring everything he does and says. Otherwise, he can't be trusted. https://twitter.com/maggieNYT/status/1125900816550907905 Beat me to it. Few things. 1. Trump is nothing if not a con-man. His most impressive con has been convincing millions of people that he is some elite, genius businessman. I wont deny that he is a good con-man. 2. Based on what is known I think a legitimate argument could be made that Trump is actually the worst businessman in America. This will likely become more obvious when (not if) his tax returns are torn into. Between NY state going after them and the House's request (Mnuchin is going to get crushed in court), they will be known soon enough. 3. Where was all this shit back in 2016, NYT? So many stories in the last 2 years about someone who is a bad businessman, and worse person, but all too late to stop the trainwreck. It would be amusing, if they are as bad as perceived due to the situation. He can't deny their contents as being incorrect without saying his returns are knowingly wrong and we'd see him attack the IRS for producing a fake return as the deep state umbrella widens.
His lawyer literally does just this. Says they are inaccurate. Article calls out that BS as Bs.
|
On May 08 2019 10:08 GreenHorizons wrote:Show nested quote +On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:On May 08 2019 03:42 GreenHorizons wrote:On May 07 2019 23:15 ChristianS wrote:
[quote]
I hesitate to post this because I don’t like horse race coverage of elections, not least because people tend to intentionally or unintentionally use it as a cover for advocacy (e.g. instead of saying “I support Buttigieg” people say “Buttigieg sure is doing well in the polls, let’s talk about it”). So let me explicitly say I have no intention of voting for Biden in the primary.
That said, I think you’re misreading the polls on Biden and Bernie somewhat. Nate Silver has written some about this and can defend the point better than I can (although he’s had some takes recently I’m more or less certain you’d hate), but the gist is Biden has consistently outperformed Bernie in the polls and most of the methodological criticisms people have for those polls (e.g. sample size or sampling bias) are off-base. We can certainly expect Biden’s numbers to drop in coming weeks, but not because recent polls were badly done - post-announcement polling bounces usually fade, that’s normal and to be expected.
I think a lot of your prediction that Biden will fail is based more on your intuition than on the data; I won’t argue with your intuition there. But as far as the data is concerned, I have trouble seeing how you could look at Biden as anything other than a frontrunner right now. It's worth noting Nate Silver has basically become what he came to fix. A pundit that opines sans the data. The easiest example is when he tiered the candidates. You can put Biden and Bernie together or Biden on top then Bernie, but you can't have Buttigieg and Harris with Sanders, let alone Harris over Sanders. Or at least he didn't provide any data to support his pushing of Harris. https://twitter.com/NateSilver538/status/1116106327208747008As far as Biden's polling he has consistently held a lead, I can't speak to every recent poll but the CNN one showing him "solidifying his frontrunner status" definitely didn't have a significant sample from people under 50. cdn.cnn.comIf other recent ones showing his surge did I haven't seen it yet (could be, I just haven't seen it). Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart. On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm: If(Supports my hypothesis) retweet uncritically Else quibble with methodology has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst. Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Show nested quote +Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it.
So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold.
Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally?
|
On May 08 2019 10:30 ChristianS wrote:Show nested quote +On May 08 2019 10:08 GreenHorizons wrote:On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:On May 08 2019 03:42 GreenHorizons wrote:[quote] It's worth noting Nate Silver has basically become what he came to fix. A pundit that opines sans the data. The easiest example is when he tiered the candidates. You can put Biden and Bernie together or Biden on top then Bernie, but you can't have Buttigieg and Harris with Sanders, let alone Harris over Sanders. Or at least he didn't provide any data to support his pushing of Harris. https://twitter.com/NateSilver538/status/1116106327208747008As far as Biden's polling he has consistently held a lead, I can't speak to every recent poll but the CNN one showing him "solidifying his frontrunner status" definitely didn't have a significant sample from people under 50. cdn.cnn.comIf other recent ones showing his surge did I haven't seen it yet (could be, I just haven't seen it). Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart. On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm: If(Supports my hypothesis) retweet uncritically Else quibble with methodology has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst. Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it. So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold. Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally?
That depending on what the responses were, and how they are weighted, what the poll is reflecting may be more or less accurate of reality. That there is "more noise" would be fair way to put it if one were arguing that it's an accurate reflection of the headlines it got.
Nate Silver is the person who is supposed to bring that to his consumers attention (I don't expect this minutia from typical corporate punditry).
The CNN poll thing is typical and will sort itself out after Biden gets some quick fundraising out of it, Nate Silver is on my shitlist for the far less defensible (than leaving out this information) tiers he put out there in early April for clicks (along with some other stuff that's less relevant atm).
|
On May 08 2019 08:50 iamthedave wrote:
CNN's ratings continue to slump by the looks of things. Not good at all.
Death of CNN is inevitable, their own fault for reporting fake news and pushing fear so hard the past 3 years.People are tired of that.
|
On May 08 2019 10:58 iPlaY.NettleS wrote:Show nested quote +On May 08 2019 08:50 iamthedave wrote:
CNN's ratings continue to slump by the looks of things. Not good at all. Death of CNN is inevitable, their own fault for reporting fake news and pushing fear so hard the past 3 years.People are tired of that.
Projection is a hell of a drug.
|
On May 08 2019 10:49 GreenHorizons wrote:Show nested quote +On May 08 2019 10:30 ChristianS wrote:On May 08 2019 10:08 GreenHorizons wrote:On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:On May 08 2019 05:05 ChristianS wrote:
[quote]
Nate Silver is hit or miss for sure, imo. He has an obnoxious contrarian streak thst makes him say/do stuff seemingly just to get a rise out of people, and he likes to act self-aware about it, but he’s not self-aware enough to realize how that tendency toward “spicy” takes actually makes him more similar to a standard pundit; he thinks it sets him apart.
On the polling, though, part of why I said something is because I think there’s a temptation to dig into methodology any time a poll suggests something you don’t like, and there’s been quite a bit of that among Sanders supporters recently. The algorithm:
If(Supports my hypothesis)
retweet uncritically
Else
quibble with methodology
has an obvious confirmation bias to it when considered in those terms, but it’s very easy to fall into unintentionally. Speaking of Nate Silver, I think avoiding this kind of thinking is still one of his strengths as an analyst.
Like I said, you tend to make pretty strong (that is, high certainty) predictions which seem more based on your intuition than any particular piece of data, and I wouldn’t know how to argue with you on that; if your intuition is that Bernie is the inevitable nominee, I don’t know what to say besides “we’ll see.” Personally I’m a bit scared that we’ll wind up with a contested convention scenario with no clear frontrunner; but I don’t have any particular evidence to support a prediction, so I wouldn’t know how to argue who’s right. I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them). The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability. I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it. So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold. Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally? That depending on what the responses were, and how they are weighted, what the poll is reflecting may be more or less accurate of reality. That there is "more noise" would be fair way to put it if one were arguing in favor that it's an accurate reflection of the headlines it got. Nate Silver is the person who is supposed to bring that to his consumers attention (I don't expect this minutia from typical corporate punditry). The CNN poll thing is typical and will sort itself out after Biden gets some quick fundraising out of it, Nate Silver is on my shitlist for the far less defensible (than leaving out this information) tiers he put out there in early April for clicks (along with some other stuff that's less relevant atm).
But this was exactly my point to start with, then. You’re giving Nate shit for presenting this as one of a bunch of polls showing good numbers for Biden post-announcement. You want him to put an asterisk saying they had a low response rate for young demographics that they offset with demographic weighting. But there isn’t actually anything wrong with that methodology, nor is there particular reason to think that poll was an outlier. Putting asterisks about methodology when there’s a result you don’t like is exactly the kind of subtle confirmation bias I was talking about.
It would be less responsible reporting for him to say “CNN’s poll said this, but maybe don’t trust that because they got a pretty low response from young voters” than to just report the result, because nothing you’ve brought up about methodology is actually a reason to distrust the result.
|
On May 08 2019 11:15 ChristianS wrote:Show nested quote +On May 08 2019 10:49 GreenHorizons wrote:On May 08 2019 10:30 ChristianS wrote:On May 08 2019 10:08 GreenHorizons wrote:On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:On May 08 2019 05:28 GreenHorizons wrote:
[quote]
I know you said something but I'm not sure it helps without some sort of data to support it (like I did with the CNN poll that was widely reported showing a surge that could also be explained by not including younger voters (people under 50) in the poll, and it turns out CNN poll did not include them).
The thing about probability predictions, is that in elections, they're useless. It means I can say X has a 99.9% chance of winning, then they lose, but I still get to be right. The least risky position is simply saying you'll wait for the results, next is probability prediction, then a real prediction that can be wrong. I find the first two completely pointless as far as connecting analysis with accountability.
I know a contested convention was Democrats plan since they made their super delegate compromise. They are putting a lot of faith in Biden/Harris/O'Rourke/Warren the other 15+ to get them there. I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:On May 08 2019 07:45 RenSC2 wrote:
The interesting thing about something like the CNN poll is that if it's slanted towards older people, it might actually make it more accurate. Young people don't vote enough. Maybe the "beat Trump" or "Bernie!" waves will finally get them to vote this time, but it didn't happen in 2016 general election or primary. Boring old political insider Hillary crushed Bernie(!) in the primary despite so much of the internet energy being behind Bernie. Then those same internet people start shouting about how it must be rigged. No. The people who vote just quietly preferred Hillary and voted that way. It was the silent majority in action.
It doesn't seem that different this time around either. I'm currently undecided and I'll wait until at least after the first debate to make up my mind. I've seen a few candidates interview with Bill Maher and the only unimpressive one was Julian Castro, the others all impressed in their own ways. I'm not personally excited about Biden and never have been, but he does connect with a lot of people who will quietly vote and just want a return to normal. I'd prefer one of the lower tier candidates to understand how to capture the media spotlight, and do it in a positive way. That's who will beat Trump.
In the end, I'd prefer any of the democratic candidates to Trump and will vote for whoever wins.
However, I would get a little bit of schadenfreude if Bernie got the democratic nomination and then lost to Trump. I don't mind him personally, but his supporters are insufferable. This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. In 17 of the 24 states for which we have both 2008 and 2016 estimates, the percentage of young (ages 17-29) eligible voters who cast a ballot in 2016 was equal to or greater than in 2008. These included the states with the lowest (Nevada, 5%) and highest (New Hampshire, 43%) youth turnout, both of which were the same in ’08 and ’16. In several states, the estimated youth turnout jumped by a substantial amount, increasing by 6+ percentage points in Illinois (18% to 26%), Missouri (21% to 27%), North Carolina (15% to 24%), and Wisconsin (25% to 33%). In Michigan, it nearly doubled: from 14% to 27%. In contrast, in each of the seven states where youth participation was lower in 2016 than in 2008 (Iowa, South Carolina, Georgia, Massachusetts, Tennessee, Texas, and Ohio) the turnout rate dropped by 3 percentage points or less. civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it. So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold. Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally? That depending on what the responses were, and how they are weighted, what the poll is reflecting may be more or less accurate of reality. That there is "more noise" would be fair way to put it if one were arguing in favor that it's an accurate reflection of the headlines it got. Nate Silver is the person who is supposed to bring that to his consumers attention (I don't expect this minutia from typical corporate punditry). The CNN poll thing is typical and will sort itself out after Biden gets some quick fundraising out of it, Nate Silver is on my shitlist for the far less defensible (than leaving out this information) tiers he put out there in early April for clicks (along with some other stuff that's less relevant atm). But this was exactly my point to start with, then. You’re giving Nate shit for presenting this as one of a bunch of polls showing good numbers for Biden post-announcement. You want him to put an asterisk saying they had a low response rate for young demographics that they offset with demographic weighting. But there isn’t actually anything wrong with that methodology, nor is there particular reason to think that poll was an outlier. Putting asterisks about methodology when there’s a result you don’t like is exactly the kind of subtle confirmation bias I was talking about. It would be less responsible reporting for him to say “CNN’s poll said this, but maybe don’t trust that because they got a pretty low response from young voters” than to just report the result, because nothing you’ve brought up about methodology is actually a reason to distrust the result.
Your point is one I considered from the beginning. Which is why I pointed out that's not really why he's on my shitlist, while I disagree with your particular take on the polling he did enough to cover that it's probably going to go away very soon so really the only impact was a bunch of misleading headlines (with less aware analysis than nate provided) and probably a good fundraising week, but it's not the basis of my distrust of Nate.
Also that I hold Nate to a higher standard than to simply say that the math checks out when the polls all seem to reflect the same thing because of the same problem rather than despite it, for which he gets credit for pointing out it's other likely inflationary factors.
|
On May 08 2019 11:21 GreenHorizons wrote:Show nested quote +On May 08 2019 11:15 ChristianS wrote:On May 08 2019 10:49 GreenHorizons wrote:On May 08 2019 10:30 ChristianS wrote:On May 08 2019 10:08 GreenHorizons wrote:On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:On May 08 2019 06:31 ChristianS wrote:[quote] I mean, I can google some polls for you, or I can link the 538 article that goes through the analysis showing Biden has a polling bounce from his announcement, but it doesn’t really matter that much anyway, since it was already apparent that a) Biden’s ahead, and b) people usually surge temporarily in polls after they announce, so it’s not surprising that he’s up. My point was only that you should be cautious doing methodological quibbling with specific polls that don’t show what you want. I don’t think it’s that important to your prediction anyway, although I’m unclear if it was supposed to have a more sinister undertone (i.e. does “CNN’s poll showing a Biden bounce sampled too many old people” have the subtext “CNN is rigging polls in Biden’s favor”?) Stochastic predictions are the only honest way to predict stochastic events. Does it make judging the quality of a prediction more statistically complex? Sure. Such are the struggles of living in a statistically complex world. Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked. CNN’s poll found Biden at 39 percent — up 11 points from 28 percent in their previous poll in March — and well ahead of Bernie Sanders, who was at 15 percent. then notes Biden had 46 percent support from Democrats age 50-64 in CNN’s poll and 50 percent support from those 65 and older. But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:On May 08 2019 07:58 GreenHorizons wrote:[quote] This is a little revisionist if you ask me. Namely that Hillary "crushed" Bernie rather than "held on", which in no small part can be attributed to a deliberate effort at coronation by the DNC and corporate media. Obama was an anomaly of sorts but youth turnout was in fact higher than usual for the primaries. [quote] civicyouth.org "Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history. Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it. So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold. Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally? That depending on what the responses were, and how they are weighted, what the poll is reflecting may be more or less accurate of reality. That there is "more noise" would be fair way to put it if one were arguing in favor that it's an accurate reflection of the headlines it got. Nate Silver is the person who is supposed to bring that to his consumers attention (I don't expect this minutia from typical corporate punditry). The CNN poll thing is typical and will sort itself out after Biden gets some quick fundraising out of it, Nate Silver is on my shitlist for the far less defensible (than leaving out this information) tiers he put out there in early April for clicks (along with some other stuff that's less relevant atm). But this was exactly my point to start with, then. You’re giving Nate shit for presenting this as one of a bunch of polls showing good numbers for Biden post-announcement. You want him to put an asterisk saying they had a low response rate for young demographics that they offset with demographic weighting. But there isn’t actually anything wrong with that methodology, nor is there particular reason to think that poll was an outlier. Putting asterisks about methodology when there’s a result you don’t like is exactly the kind of subtle confirmation bias I was talking about. It would be less responsible reporting for him to say “CNN’s poll said this, but maybe don’t trust that because they got a pretty low response from young voters” than to just report the result, because nothing you’ve brought up about methodology is actually a reason to distrust the result. Your point is one I considered from the beginning. Which is why I pointed out that's not really why he's on my shitlist, while I disagree with your particular take on the polling he did enough to cover that it's probably going to go away very soon so really the only impact was a bunch of misleading headlines (with less aware analysis than nate provided) and probably a good fundraising week, but it's not the basis of my distrust of Nate. Also that I hold Nate to a higher standard than to simply say that the math checks out when the polls all seem to reflect the same thing because of the same problem rather than despite it, for which he gets credit for pointing out it's other likely inflationary factors.
Then you considered my point from the beginning, but still thought it was good analysis to dismiss Biden’s lead by pointing out obscure methodological details in one of the many polls showing that lead. If you were gonna make an argument like “I think the polls are underestimating Bernie’s support because many polling methods still only call people with landlines and Bernie’s supporters are less likely than other members of their demographics to not have a landline,” then there’d be something to talk about. It’d be silly to expect pundits to put an asterisk next to every poll result saying “maybe don’t trust this though, landline polling is flawed,” but it’s a valid criticism at least. This is where we get to the point that polling isn’t a perfect way of knowing public opinion, just less bad than any other method I know of.
|
On May 08 2019 10:58 iPlaY.NettleS wrote:Show nested quote +On May 08 2019 08:50 iamthedave wrote:
CNN's ratings continue to slump by the looks of things. Not good at all. Death of CNN is inevitable, their own fault for reporting fake news and pushing fear so hard the past 3 years.People are tired of that.
While ratings are down, revenue is up, their market isn't just television anymore. I also highly doubt a $3.2 Billion company is going to just "die".
|
The shift in denial of climate change is that man-made climate change isn't happening so we don't have to do anything.
On May 08 2019 10:58 iPlaY.NettleS wrote:Show nested quote +On May 08 2019 08:50 iamthedave wrote:
CNN's ratings continue to slump by the looks of things. Not good at all. Death of CNN is inevitable, their own fault for reporting fake news and pushing fear so hard the past 3 years.People are tired of that.
I think the death of cable news is inevitable it's a format i don't think works for younger generations, i know very few who aren't 50+ that watch cable news regularly regardless of the channel. It's aging itself out of existence.
|
On May 08 2019 11:32 ChristianS wrote:Show nested quote +On May 08 2019 11:21 GreenHorizons wrote:On May 08 2019 11:15 ChristianS wrote:On May 08 2019 10:49 GreenHorizons wrote:On May 08 2019 10:30 ChristianS wrote:On May 08 2019 10:08 GreenHorizons wrote:On May 08 2019 09:57 ChristianS wrote:On May 08 2019 08:52 GreenHorizons wrote:On May 08 2019 08:27 ChristianS wrote:On May 08 2019 06:49 GreenHorizons wrote:
[quote]
Only one of those polls shows anything about the ages (that I saw) but it's still unclear how much of their sample was under 50. It does also show that the surge is with people over 50 though. Whether it's intentional or not is impossible to tell from here but when someone like Nate Silver opens in the article you linked.
[quote]
then notes
[quote]
But doesn't mention the poll is missing the ages where Bernie polls the strongest, he's not being the contrarian numbers guy, he's being the sellout pundit imo. Sorry, after googling for a moment I’m realizing your methodological complaint about the CNN poll might not even be sound. Can you explain what exactly you think was wrong with that CNN poll, and what you think its significance is? Nothing is "wrong" with the poll, it just didn't have a statistically significant sample of the ages where Bernie performs much better than he does with voters over 50. Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally. For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on. On May 08 2019 08:28 Gahlo wrote:On May 08 2019 08:22 hunts wrote:
[quote]
"Held on" isn't exactly accurate. Given the difference in votes when bernie surrendered, I think "crushed" is quite apt and not revisionist history.
Things look a lot closer when you take superdelegates out of the picture. It basically cuts her margin of victory in half. Ron Paul got crushed, Jeb Bush got crushed, Hillary hung on to what should have been an easy win losing several states in the process. At least that's my perspective. Notably, her political career is over and Bernie's is reaching a peak Okay, so they didn’t sample enough young people to draw statistically significant conclusions worth reporting about the subgroup. What is it you are concluding from that fact? And why do you believe Nate Silver was irresponsible in his analysis for not mentioning this? I already explained? Most polls don't tell us anything about the electorate models they are using either which can make such coincidences even more pronounced in the final tally.
For example if you use a model based on 2012 you get a much more significant spike for Biden than if you use 2008 or 2016 model. Are they using primary electorate data or "Democrat" turnout data? So on and so on.
He knows this and how important it is to the value of the conclusions of the poll and fails to mention it. So let’s start from the beginning here. CNN does a tracking poll, just like every other time they do a tracking poll, and calls random respondents asking a bunch of questions. When they’re done, they tabulate all the results and report them. They don’t, however, report results for subgroups if they didn’t happen to get enough respondents from that subgroup. In this case, “18-35” or w/e didn’t clear that threshold. Am I supposed to think the numbers are off because of this? That they’re overestimating Biden’s support in this poll? That the noise is higher than usual? What is becoming more pronounced in the final tally? That depending on what the responses were, and how they are weighted, what the poll is reflecting may be more or less accurate of reality. That there is "more noise" would be fair way to put it if one were arguing in favor that it's an accurate reflection of the headlines it got. Nate Silver is the person who is supposed to bring that to his consumers attention (I don't expect this minutia from typical corporate punditry). The CNN poll thing is typical and will sort itself out after Biden gets some quick fundraising out of it, Nate Silver is on my shitlist for the far less defensible (than leaving out this information) tiers he put out there in early April for clicks (along with some other stuff that's less relevant atm). But this was exactly my point to start with, then. You’re giving Nate shit for presenting this as one of a bunch of polls showing good numbers for Biden post-announcement. You want him to put an asterisk saying they had a low response rate for young demographics that they offset with demographic weighting. But there isn’t actually anything wrong with that methodology, nor is there particular reason to think that poll was an outlier. Putting asterisks about methodology when there’s a result you don’t like is exactly the kind of subtle confirmation bias I was talking about. It would be less responsible reporting for him to say “CNN’s poll said this, but maybe don’t trust that because they got a pretty low response from young voters” than to just report the result, because nothing you’ve brought up about methodology is actually a reason to distrust the result. Your point is one I considered from the beginning. Which is why I pointed out that's not really why he's on my shitlist, while I disagree with your particular take on the polling he did enough to cover that it's probably going to go away very soon so really the only impact was a bunch of misleading headlines (with less aware analysis than nate provided) and probably a good fundraising week, but it's not the basis of my distrust of Nate. Also that I hold Nate to a higher standard than to simply say that the math checks out when the polls all seem to reflect the same thing because of the same problem rather than despite it, for which he gets credit for pointing out it's other likely inflationary factors. Then you considered my point from the beginning, but still thought it was good analysis to dismiss Biden’s lead by pointing out obscure methodological details in one of the many polls showing that lead. If you were gonna make an argument like “I think the polls are underestimating Bernie’s support because many polling methods still only call people with landlines and Bernie’s supporters are less likely than other members of their demographics to not have a landline,” then there’d be something to talk about. It’d be silly to expect pundits to put an asterisk next to every poll result saying “maybe don’t trust this though, landline polling is flawed,” but it’s a valid criticism at least. This is where we get to the point that polling isn’t a perfect way of knowing public opinion, just less bad than any other method I know of.
That's not what I did.
I don't think it's even going to take a debate, from what I gather these recent polls showing biden surging (beyond what's typical for an announcement) we're really polling people over 50 which is about the split we'd expect and Biden's had for months.
Good chance his polling of people under 50 is already plummeting, particularly when one considers he's been pretty universally shitcanned here.
As I've alluded to before though, the real use of these early polls is to tier candidates which where Nate is concerned are coming with increasingly direct "this is bs" disclaimers knowing the people clicking will ignore the "this is bs" part.
EDIT: The real important thing I get from Biden's polling so far is that he'll probably hold 10-15% (as in unshakable, he might get this in states after he dropped out if he did) as long as he's in the race which practically eliminates pretty much everyone consistently polling under 5%.
I recognize Biden still has a real lead that will take work to dwindle but those same polls also showed Sanders is the most popular second choice besides Biden so it's their race to lose to whoever can take the third slot out of super Tuesday (if they make it that far).
|
But this was exactly my point to start with, then. You’re giving Nate shit for presenting this as one of a bunch of polls showing good numbers for Biden post-announcement. You want him to put an asterisk saying they had a low response rate for young demographics that they offset with demographic weighting. But there isn’t actually anything wrong with that methodology, nor is there particular reason to think that poll was an outlier. Putting asterisks about methodology when there’s a result you don’t like is exactly the kind of subtle confirmation bias I was talking about.
It would be less responsible reporting for him to say “CNN’s poll said this, but maybe don’t trust that because they got a pretty low response from young voters” than to just report the result, because nothing you’ve brought up about methodology is actually a reason to distrust the result.
When you look into the methodology of that poll it becomes clear that literally no one under the age of 50 was included, despite the fact that many of those polled in the age ranges of 18 - 35, and 35 - 50 were successfully contacted. I believe the poll sampled 1000 respondents, but only included responses from groups that had over 150 respondents. An actual number of those in the 18 - 50 age range wasn't provided, but it's quite likely that at least some in the range of 18 - 35, and a more substantial number in the 35 - 50 range, were contacted. Its not overreaching to say that CNN has a heavy bias towards the establishment, and it isn't absurd to think that they would purposely weight a poll to ONLY include responses from people who are demographically more likely to support the candidate that they support. The fact that the poll's release so closely coincided with Biden's campaign announcement also casts their impartiality into question. Early polling is unlikely to have a huge affect on the primary, but its abundantly clear, at least to me, that the pseudo left wing corporate media has decided to throw their lot in with Biden only because Booker, Kamela, and Buttigieg have all failed to gain traction, even with overwhelmingly positive reporting for almost 3 months. Where I believe Nate Silver betrays his bias is in failing to mention that this poll is effectively only focusing on Biden's strongest demographics, and in failing to note that the numbers in the poll are identical to earlier polls conducted, but only in showing Biden's strength with older voters, and fail to accurately reflect what a substantial voting block in the Democratic Party believes.
In short - CNN was at best irresponsible, and was unwilling to publish a more credible poll that includes a wide range of Democrat voters, something that could have easily been accomplished, or CNN was willing to publish what they knew to be an inaccurate poll in a desperate attempt to hamper the more progressive, anti corporate candidate in the race. I tend to think of it as the latter rather than the former. I would condemn Bernie Sanders for disingenuously releasing a poll where only 18 - 35 year olds were contacted if he tried to pass it off as reflecting the viewpoint of the general public, its dishonest and it hurts the credibility of the mainstream news in a time when their credibility is already at an all time low.
|
|
|
|
|
|
EPT
