I think that headlines are very rarely, if ever, written by the author of the article. Many years ago my wife had a job writng headlines for articles. After a couple of months she got "unhired" because she was working to hard to put the gist in the headline and not a "grabber" to get people to read it.
This is precisely right. Even in high school, headlines are not written by the author of the article. It's a separate job, and a skill all its own. I was the headline editor in high school, a fun job.
My high school didn't have a paper to edit, but it was in Texas, and thus a participant in UT Austin's UIL events, one of which was Headline Writing. (Each entrant had to read several articles within a time limit and write a headline that best captured each one while staying within a certain width range, complicated by the fact that some letters like capital M ate up more width than, say, small l.)
I think the letter width rules might be dated today, but if we learned nothing else from that event, it was that the article writer and headline writer were usually two different people, with different incentives.
I think you're right, but it's quite possible Vance really does mean "creating a story that the media perforce must address" as opposed to "creating a lie to attract media attention."
He apparently did really get such complaints from his constituents. However I'd say he pretty much blew his ooportunity to just say, "I was repeating things told me by my constituents that no one was paying attention to. Good reporting should discover if those stories are true."
“They are not independent, of course, but the probability of winning all three cannot be greater than the probability of winning Pennsylvania and that is less than .5, so, accepting Silver’s figures, the probability that she will win “through the Blue Wall” must be less, probably much less, than even, ”
I love this piece.
However, your phrase “probably much less” in the above, is probably untrue, and definitely unfounded.
The correlation across winning MI and WI if Harris wins PA is in “fact” pretty darn high (there just aren’t that many scenarios where she wins PA but does not win both MI and WI), and so while I agree 100% with the rest of your logic, including of course that the chance is less than 0.5, it in fact is probably NOT *a lot less*.
Unless you have reason to be fairly certain that it is the case that Harris has a lot less than a 35% chance of winning the presidency, which would put you at odds (pun intended) with all the polling average sites AND all the betting market aggregates.
It isn't her odds of winning that I claim are probably much less than even but her odds of winning through the blue wall. Silver says if she gets those three states she will probably win. I don't know what "probably" means, but it is less than certainty — say 70%. If the three states are perfectly correlated that gets it down to 35%. But they aren't perfectly correlated. They all look like close elections, about equally close, and each will be affected by small effects special to that state. So I think my "probably much less is justified.
In today’s Silver Bulletin post, Silver describes concisely most of the point I was making below verbosely:
“When polls miss low on Trump in one key state, they probably also will in most or all of the others.
In fact, because polling errors are highly correlated between states — and because Trump was ahead in 5 of the 7 swing states anyway — a Trump sweep of the swing states was actually our most common scenario, occurring in 20 percent of simulations. Following the same logic, the second most common outcome, happening 14 percent of the time, was a Harris swing state sweep.”
“Silver says if she gets those three states she will probably win. I don't know what ‘probably’ means, but it is less than certainty — say 70%.”
Sorry, with all due respect, on THIS point you are simply wrong, if by “winning through the Blue Wall” you mean her chances of winning the election if she wins all 3 Blue Wall states.
It’s called “The Blue Wall” for a reason.
If you believe there are only 7 swing states in this election, wins in all 3 Blue Wall states give Kamala 276 Electoral votes, even if she loses all four of the other swing states.
Trump would have to win a “safe” state where Kamala is and has all along been up by more than 5 points, in addition to winning all 4 other swing states plus all *his* “safe” states, in order to get 270+ votes when Kamala wins the 3 Blue Wall states (MI, WI, PA).
This is of course technically possible, but *extremely* unlikely. I’m sure Silver has calculated odds on this (I’m not a paying subscriber, so I couldn’t easily find them), but Trump’s chances in this case are surely less than 10%, and almost certainly less than 5%
Silver - and for that matter his imo lesser cohorts who do this kind of polling average analysis - knows well that each state election is NOT in fact a mostly independent election, but that they are fairly highly correlated nationally and very highly correlated across the Blue Wall states.
Now if you merely meant above what are her chances of winning the other two Blue Wall states if she wins PA, this was *exactly* the point of my prior response where I note you are merely *probably* wrong. If she wins PA it means she very likely wins WI and MI, but I do agree that it’s not as high as 95% probability in that case, exactly for the “small differences” reason you cite.
But those 3 state elections are indeed sufficiently correlated that the chances of her winning both MI and WI if she wins PA are in fact MUCH higher than 50%.
I *suspect* - but of course cannot know for sure - that you are confusing the statement “Kamala wins PA” with the different statement “Kamala barely wins PA by a tiny margin”. Because in that latter case I would agree with you that it is much less certain what the outcomes of MI and WI are (they are each well over 50%, and likely over 60%, even in that case, but definitely nowhere near highly probable for her to win both).
Putting this all together if we take as a given the accuracy of the claim that she has a 46% chance to win PA, then if/when she wins PA she is very likely to win both MI and WI, and that is why I said her chances are NOT “probably much less” in your original piece.
-
Summarizing a la Bryan Caplan (Bet on It) I would *gladly* give you 4 to 1 odds on a contingent bet that *if* she wins the 3 Blue Wall states she wins the election. I’d be willing to do it at 9 to 1.
But if I misinterpreted your latest comment and we are in fact talking solely about her chances to win MI and WI, I’d *gladly* give even money odds on a contingent bet that if she wins PA she also wins both MI and WI, and I’d be willing to do it at 2-1. Because the true odds of that bet are in fact somewhere north of 70%, not south.
"There is no good reason to expect an author writing for USA Today to pay attention to the details of legal rules or one writing for Newsweek to understand conditional probability, no reason to expect either to think clearly or to care much about getting a story right."
Isn't that kind of supposed to be their job?! Isn't that what they tell us they do in all their advertising when they tell us how much we need to listen to/read their output?
Lies on top of lies. Whether through clear intent or malfeasant neglect.
While that's true, those reasons in their ads are the only reason to use their media. A burger may or may not be the best tasting burger in the world as the ads would have us believe, but it's still got some nutrition. When "news" is content-free it has no value or even negative value.
If we don't believe their ads at all, then there is zero or even negative reasons to partake of their media.
Their news isn't content free. In both those cases I was able, from the information they provided, to figure out the relevant facts. The headline in each case was misleading, which is a reason not to rely on the headline alone.
More generally, you can't get information by just believing what one source tells you. You have to be willing to think about it, to recognize obvious bias, to use multiple sources. I have a bunch of substack posts about how to figure out what is true at: http://www.daviddfriedman.com/Sorted_Posts.html#Discovering_Truth_
"When "news" is content-free it has no value or even negative value."
Content-free in what sense? Containing useful information? Most news does not contain useful information for the typical reader. If the news reports that Biden met with the ambassador of France yesterday to discuss a trade deal involving pinot noir grapes and American wheat, that isn't going to affect anyone who isn't in the wheat or grape trade (roughly speaking), and if it's a discussion of something specific, it might affect even fewer people. If there's a news story about how Harris went up 2 points in one state, that may affect a few people in that state - most people are decided on whether they will vote, and for whom, and that story won't change their decision.
Point being, if what I said is true, then one explanation is that most news for most people is only there for entertainment - something to talk about at the water cooler. News doesn't have to be terrifically informative to serve that purpose, however. Indeed, it suffices to be textually true, even if it suggests something that isn't.
Another possibility is that journalists either don't understand or don't care about their subject matter, likely because they are writing to a deadline as a job and the specifics are generally immaterial to them. I had heard it said many times that people who were the subjects of newspaper interviews or stories, or people who knew the details, see how much newspapers get wrong. Having been associated with several stories in the local paper, I agree fully. They just don't seem that interested in being factually accurate.
Add some political bias or other forms of bias, and it seems very likely that not only will details be wrong, but often in a predictable direction (nobody expects the NYT or WaPo to accidentally make a pro-Trump argument through an error, for instance).
“get clicks” is likely for every story - each is aimed at their respective audience and to boost readership. Though you could call this nefarious, it’s really just yellow journalism as it has always been.
I think I am distinguishing in my mind between getting clicks by telling people what they want to believe, which looks a lot like political bias although it doesn't depend on the reporter's bias but on what he thinks his readers' bias is, and getting clicks by saying something startling or interesting, which doesn't depend on anyone's political bias. "Silver says Kamala is going to win" would be a story you would want to read whether you are for or against her.
I would have to agree with all of that. I also find the focus on subtle political lies that have been coming from liberals to be an interesting choice, one week before what the author hopes will be the presidential re-election of the most mendacious individual anyone can think of, the politician who single handedly caused all the efforts to police political disinformation on social media and elsewhere. When a hundred million fools have bought into the idea that the last election was stolen, and when a billionaire who controls the most important platform makes sure that lies cannot be identified as such for the news consumer, then we'll focus on a subtle lie about the billionaire. It's like getting heavily into trying to cure your athlete's foot while you're dying of cancer.
Trump's political lies are rarely subtle. If you can point me at pro-Trump news stories that are subtle lies I will add them to my next collection — I expect to do this again in a few months when I have accumulated enough material.
[added later]
A lot more of mass media are biased towards the Democrats than towards the Republicans — my impression is that reporters are overwhelmingly left of center. Do you disagree? If so, finding news stories that are distorted by left wing bias is likely to be easier than finding ones distorted by right wing bias, which fits my experience.
I don't know if you read my two posts on Vance. I think they make it reasonably clear that I am not an admirer of the Republican ticket. Of course, I'm not an admirer of their opponents either.
Today my daughter confessed to me that she voted Libertarian! I am so glad she voted her beliefs, even though I know a single vote will not make a difference.
Her mother started as a Trotskyist who grew into the real world, and I am a Social Democrat from way back who now voted for Trump. Eventually I became the left winger in the family, and I guess I still am. :-)
I think that headlines are very rarely, if ever, written by the author of the article. Many years ago my wife had a job writng headlines for articles. After a couple of months she got "unhired" because she was working to hard to put the gist in the headline and not a "grabber" to get people to read it.
This is precisely right. Even in high school, headlines are not written by the author of the article. It's a separate job, and a skill all its own. I was the headline editor in high school, a fun job.
My high school didn't have a paper to edit, but it was in Texas, and thus a participant in UT Austin's UIL events, one of which was Headline Writing. (Each entrant had to read several articles within a time limit and write a headline that best captured each one while staying within a certain width range, complicated by the fact that some letters like capital M ate up more width than, say, small l.)
I think the letter width rules might be dated today, but if we learned nothing else from that event, it was that the article writer and headline writer were usually two different people, with different incentives.
My wife liked it and worked hard until they told her she was doing it wrong. LOL
It really upset her sense of propriety.
I guess one could expect a better understanding of probabilities by those who are writing about them.
When it comes to J.D. Vance statement:
“I say that we’re creating a story, meaning we’re creating the American media focusing on it.”
It sounds like a verbal contortion to hide the plain meaning of ‘creating a story’.
I think you're right, but it's quite possible Vance really does mean "creating a story that the media perforce must address" as opposed to "creating a lie to attract media attention."
He apparently did really get such complaints from his constituents. However I'd say he pretty much blew his ooportunity to just say, "I was repeating things told me by my constituents that no one was paying attention to. Good reporting should discover if those stories are true."
"Creating" was an unforced error.
“They are not independent, of course, but the probability of winning all three cannot be greater than the probability of winning Pennsylvania and that is less than .5, so, accepting Silver’s figures, the probability that she will win “through the Blue Wall” must be less, probably much less, than even, ”
I love this piece.
However, your phrase “probably much less” in the above, is probably untrue, and definitely unfounded.
The correlation across winning MI and WI if Harris wins PA is in “fact” pretty darn high (there just aren’t that many scenarios where she wins PA but does not win both MI and WI), and so while I agree 100% with the rest of your logic, including of course that the chance is less than 0.5, it in fact is probably NOT *a lot less*.
Unless you have reason to be fairly certain that it is the case that Harris has a lot less than a 35% chance of winning the presidency, which would put you at odds (pun intended) with all the polling average sites AND all the betting market aggregates.
It isn't her odds of winning that I claim are probably much less than even but her odds of winning through the blue wall. Silver says if she gets those three states she will probably win. I don't know what "probably" means, but it is less than certainty — say 70%. If the three states are perfectly correlated that gets it down to 35%. But they aren't perfectly correlated. They all look like close elections, about equally close, and each will be affected by small effects special to that state. So I think my "probably much less is justified.
In today’s Silver Bulletin post, Silver describes concisely most of the point I was making below verbosely:
“When polls miss low on Trump in one key state, they probably also will in most or all of the others.
In fact, because polling errors are highly correlated between states — and because Trump was ahead in 5 of the 7 swing states anyway — a Trump sweep of the swing states was actually our most common scenario, occurring in 20 percent of simulations. Following the same logic, the second most common outcome, happening 14 percent of the time, was a Harris swing state sweep.”
https://www.natesilver.net/p/the-model-exactly-predicted-the-most#footnote-4-151345713
“Silver says if she gets those three states she will probably win. I don't know what ‘probably’ means, but it is less than certainty — say 70%.”
Sorry, with all due respect, on THIS point you are simply wrong, if by “winning through the Blue Wall” you mean her chances of winning the election if she wins all 3 Blue Wall states.
It’s called “The Blue Wall” for a reason.
If you believe there are only 7 swing states in this election, wins in all 3 Blue Wall states give Kamala 276 Electoral votes, even if she loses all four of the other swing states.
Trump would have to win a “safe” state where Kamala is and has all along been up by more than 5 points, in addition to winning all 4 other swing states plus all *his* “safe” states, in order to get 270+ votes when Kamala wins the 3 Blue Wall states (MI, WI, PA).
This is of course technically possible, but *extremely* unlikely. I’m sure Silver has calculated odds on this (I’m not a paying subscriber, so I couldn’t easily find them), but Trump’s chances in this case are surely less than 10%, and almost certainly less than 5%
https://www.270towin.com/
Silver - and for that matter his imo lesser cohorts who do this kind of polling average analysis - knows well that each state election is NOT in fact a mostly independent election, but that they are fairly highly correlated nationally and very highly correlated across the Blue Wall states.
Now if you merely meant above what are her chances of winning the other two Blue Wall states if she wins PA, this was *exactly* the point of my prior response where I note you are merely *probably* wrong. If she wins PA it means she very likely wins WI and MI, but I do agree that it’s not as high as 95% probability in that case, exactly for the “small differences” reason you cite.
But those 3 state elections are indeed sufficiently correlated that the chances of her winning both MI and WI if she wins PA are in fact MUCH higher than 50%.
I *suspect* - but of course cannot know for sure - that you are confusing the statement “Kamala wins PA” with the different statement “Kamala barely wins PA by a tiny margin”. Because in that latter case I would agree with you that it is much less certain what the outcomes of MI and WI are (they are each well over 50%, and likely over 60%, even in that case, but definitely nowhere near highly probable for her to win both).
Putting this all together if we take as a given the accuracy of the claim that she has a 46% chance to win PA, then if/when she wins PA she is very likely to win both MI and WI, and that is why I said her chances are NOT “probably much less” in your original piece.
-
Summarizing a la Bryan Caplan (Bet on It) I would *gladly* give you 4 to 1 odds on a contingent bet that *if* she wins the 3 Blue Wall states she wins the election. I’d be willing to do it at 9 to 1.
But if I misinterpreted your latest comment and we are in fact talking solely about her chances to win MI and WI, I’d *gladly* give even money odds on a contingent bet that if she wins PA she also wins both MI and WI, and I’d be willing to do it at 2-1. Because the true odds of that bet are in fact somewhere north of 70%, not south.
"There is no good reason to expect an author writing for USA Today to pay attention to the details of legal rules or one writing for Newsweek to understand conditional probability, no reason to expect either to think clearly or to care much about getting a story right."
Isn't that kind of supposed to be their job?! Isn't that what they tell us they do in all their advertising when they tell us how much we need to listen to/read their output?
Lies on top of lies. Whether through clear intent or malfeasant neglect.
The fact that someone selling a product tells you something in his advertising is not a good reason to believe it.
While that's true, those reasons in their ads are the only reason to use their media. A burger may or may not be the best tasting burger in the world as the ads would have us believe, but it's still got some nutrition. When "news" is content-free it has no value or even negative value.
If we don't believe their ads at all, then there is zero or even negative reasons to partake of their media.
Their news isn't content free. In both those cases I was able, from the information they provided, to figure out the relevant facts. The headline in each case was misleading, which is a reason not to rely on the headline alone.
More generally, you can't get information by just believing what one source tells you. You have to be willing to think about it, to recognize obvious bias, to use multiple sources. I have a bunch of substack posts about how to figure out what is true at: http://www.daviddfriedman.com/Sorted_Posts.html#Discovering_Truth_
One you might find useful is https://daviddfriedman.substack.com/p/how-to-learn-what-is-true
"When "news" is content-free it has no value or even negative value."
Content-free in what sense? Containing useful information? Most news does not contain useful information for the typical reader. If the news reports that Biden met with the ambassador of France yesterday to discuss a trade deal involving pinot noir grapes and American wheat, that isn't going to affect anyone who isn't in the wheat or grape trade (roughly speaking), and if it's a discussion of something specific, it might affect even fewer people. If there's a news story about how Harris went up 2 points in one state, that may affect a few people in that state - most people are decided on whether they will vote, and for whom, and that story won't change their decision.
Point being, if what I said is true, then one explanation is that most news for most people is only there for entertainment - something to talk about at the water cooler. News doesn't have to be terrifically informative to serve that purpose, however. Indeed, it suffices to be textually true, even if it suggests something that isn't.
It's a plot to give everyone involved plausible deniability
I feel like these subtle lies go 80/20 in one direction.
Only 80-20? 🤣🤣🤣🤣🤣
Another possibility is that journalists either don't understand or don't care about their subject matter, likely because they are writing to a deadline as a job and the specifics are generally immaterial to them. I had heard it said many times that people who were the subjects of newspaper interviews or stories, or people who knew the details, see how much newspapers get wrong. Having been associated with several stories in the local paper, I agree fully. They just don't seem that interested in being factually accurate.
Add some political bias or other forms of bias, and it seems very likely that not only will details be wrong, but often in a predictable direction (nobody expects the NYT or WaPo to accidentally make a pro-Trump argument through an error, for instance).
Wow, what an eye-opener this post was! I think you're exactly right!
“get clicks” is likely for every story - each is aimed at their respective audience and to boost readership. Though you could call this nefarious, it’s really just yellow journalism as it has always been.
I think I am distinguishing in my mind between getting clicks by telling people what they want to believe, which looks a lot like political bias although it doesn't depend on the reporter's bias but on what he thinks his readers' bias is, and getting clicks by saying something startling or interesting, which doesn't depend on anyone's political bias. "Silver says Kamala is going to win" would be a story you would want to read whether you are for or against her.
I would have to agree with all of that. I also find the focus on subtle political lies that have been coming from liberals to be an interesting choice, one week before what the author hopes will be the presidential re-election of the most mendacious individual anyone can think of, the politician who single handedly caused all the efforts to police political disinformation on social media and elsewhere. When a hundred million fools have bought into the idea that the last election was stolen, and when a billionaire who controls the most important platform makes sure that lies cannot be identified as such for the news consumer, then we'll focus on a subtle lie about the billionaire. It's like getting heavily into trying to cure your athlete's foot while you're dying of cancer.
Trump's political lies are rarely subtle. If you can point me at pro-Trump news stories that are subtle lies I will add them to my next collection — I expect to do this again in a few months when I have accumulated enough material.
[added later]
A lot more of mass media are biased towards the Democrats than towards the Republicans — my impression is that reporters are overwhelmingly left of center. Do you disagree? If so, finding news stories that are distorted by left wing bias is likely to be easier than finding ones distorted by right wing bias, which fits my experience.
I don't know if you read my two posts on Vance. I think they make it reasonably clear that I am not an admirer of the Republican ticket. Of course, I'm not an admirer of their opponents either.
Ah, dear David Friedman,
Today my daughter confessed to me that she voted Libertarian! I am so glad she voted her beliefs, even though I know a single vote will not make a difference.
Her mother started as a Trotskyist who grew into the real world, and I am a Social Democrat from way back who now voted for Trump. Eventually I became the left winger in the family, and I guess I still am. :-)