[HN Gopher] A Model for Journalistic Copypasta
___________________________________________________________________
A Model for Journalistic Copypasta
Author : Gadiguibou
Score : 43 points
Date : 2022-08-11 17:27 UTC (5 hours ago)
(HTM) web link (dynomight.net)
(TXT) w3m dump (dynomight.net)
| ZeroGravitas wrote:
| Is there a browser add-on/site that'll figure out the original
| paper when reading a science news story?
|
| Maybe it can use the fact that multiple outlets will copypasta at
| the same time to help
| james-redwood wrote:
| This is perhaps one of the best pieces of writing I've ever seen
| on HN. Very, very pertinent.
| WFHRenaissance wrote:
| > Now, there's no special brilliance needed to come up with this
| kind of test. Frankly, it's pretty obvious and I'm sure the
| Wirecutter is aware they could do it. They just don't.
|
| They are giving others way too much credit here. If Wirecutter is
| aware they could do it they probably also think that their test
| and this test are equivalent.
| mherdeg wrote:
| I see that subsection 5 mentions the Gell-Mann Amnesia effect so
| it is again time for me to again scream into the void about this
| ( https://news.ycombinator.com/item?id=18005236 ):
|
| In the speech where Crichton proposes the Gell-Mann amnesia
| effect he argues against almost all forms of attempts to predict
| the future. Crichton opposes most or all "speculation"; I think
| he would find probabilistic reasoning distasteful at best. His
| approach is a remarkably nihilistic response to the normal human
| situation of reasoning under uncertainty.
|
| Read it for yourself. He proposed this 20 years ago (
| http://web.archive.org/web/20070714204136/http://www.michael...
| ).
|
| And to his great credit, he gave a specific example of what he
| was talking about!
|
| He claimed that it was "useless" to write or read a March 2002
| newspaper article quoting experts predicting the impact of the
| 2002 United States steel tariffs. Specialists should not be
| quoted, he said, because "Nobody knows the future."
|
| Can we check this out?
|
| Let's look at the 2002 United States steel tariffs. Crichton is
| dismayed by the following predictions in a newspaper article:
|
| (1) Mr. Bush's action "is likely to send the price of steel up
| sharply, perhaps as much as ten percent.."
|
| (2) American consumers "will ultimately bear" higher prices.
|
| (3) America's allies "would almost certainly challenge" the
| decision. Their legal case "could take years to litigate in
| Geneva, is likely to hinge" on thus and such.
|
| (4) In addition, there is a further vague and overarching
| speculation. The Allies' challenge would be "setting the stage
| for a major trade fight with many of the same countries Mr. Bush
| is trying to hold together in the fractious coalition against
| terrorism." In other words, the story speculates that tariffs may
| rebound against the fight against terrorism.
|
| He _hates_ that someone wrote this. He thinks it is the biggest
| waste of anyone 's time. None of it should have been printed, he
| says.
|
| So - let's check. Were these predictions useless? Were they
| correlated in some way to reality?
|
| I argue that these predictions do two valuable things:
|
| First, they may help a contemporary reader know what's coming.
|
| Second, they may help future readers judge whether the quoted
| experts were capable in the past of predicting things, which
| might be useful to know.
|
| So:
|
| (1) The price of steel did not go up ten percent as predicted.
| Instead, the price of some steel products rose 60-80% from
| January 2002 to July 2002 according to this random PDF I found
| from a group that publishes studies about trade (page 6 :
| http://www.tradepartnership.com/pdf_files/2002jobstudy.pdf )
| whose authors appear to have international-trade credentials. The
| authors say that the tariffs contributed to the price increase
| along with other factors.
|
| (2) Whether US consumers bore higher prices is unclear. I do see
| at
| https://en.wikipedia.org/wiki/2002_United_States_steel_tarif...
| that a study determined 'the impact of the tariffs on the U.S.
| welfare ranged between a gain of $65.6 million (0.0006% of GDP)
| to a loss of $110.0 million (0.0011% of GDP), "with a central
| estimate of a welfare loss of $41.6 million."'
|
| (3) Whether allies challenged the decision is an easy one! Yes,
| the decision was challenged, and an overwhelming, strong
| international trade war occurred. The US backed down. See
| https://en.wikipedia.org/wiki/2002_United_States_steel_tarif...
| ($2B of WTO sanctions threatened, threat of retaliatory tariffs
| from the EU).
|
| (4) I could not find any evidence that the steel tariffs made it
| harder to enlist other countries in the Iraq War.
|
| Side note -- contra Crichton, I think I am glad that people
| publicly predicted what would happen with the Iraq War and that
| we are able to compare their predictions versus reality (
| https://www.csmonitor.com/World/Backchannels/2011/1222/Iraq-...
| ).
|
| If defense officials had instead merely said "well, we're
| launching a war, but we will not speculate on how long it will
| take, what it will cost, or how many people will die, because no
| one can know the future", well gosh, that would really have been
| something.
|
| Overall, I think 3 of the 4 pieces of expert speculation about
| the 2002 United States steel tariffs which Crichton cited as "a
| complete waste of time" were interesting. They presented an
| expert's testable hypothesis about the future result of actions;
| and they helped others judge the credibility of those experts in
| the future.
|
| Crichton seems to be saying instead that there is no point in
| publishing anything about how air purifiers perform, because who
| can know the future? Maybe they will change how they manufacture
| their air purifiers and consumers will start buying bad ones
| instead of the tested good ones. Maybe your house will be totally
| different from the test house used in the dynomight tests. Maybe
| forest fires will make your air quality so bad that you enter a
| regime no one tested in advance.
|
| Now, in retrospect there is a bit of irony in Crichton choosing
| to "predict" a certain future prediction as being a bullshit one
| -- and in being able to see after the fact that it was pretty
| spot-on.
|
| But the deeper irony to me is that Crichton does not ever give
| any evidence for his claim -- he does not publish any percentage
| of news reporting which is bullshit!
|
| Imagine if he said "I have analyzed predictions made in news
| articles for the past X years, and judged the accuracy of N
| predictions -- Y% of them were accurate. Frustratingly none of
| them expressed any degree of confidence in their predictions so I
| treated all equally for this analysis. This percentage is [no
| worse than guessing | worse than guessing, so you should expect
| the opposite of what is predicted with weak/strong confidence |
| better than guessing, so you should expect what is predicted with
| weak/strong confidence]. Here are my data so you can see for
| yourself."
|
| Did Crichton do his homework? It feels like, when he defined the
| Gell-Mann amnesia effect, he just give up in dismay and cherry-
| picked his best examples of failed predictions. I find it hurt to
| trust his anecdotes.
|
| With respect to opinion columnists (not the same as newswriters),
| this work has been done at an undergraduate level -- see
| https://www.hamilton.edu/news/story/pundits-as-accurate-as-c... ,
| https://www.hamilton.edu/documents/an-analysis-of-the-accura....
| Crichton says that pundits are worthless, but this one study
| found that certain specific opinion writers tend to make
| predictions which are accurate, and certain others tend to make
| inaccurate predictions.
| karaterobot wrote:
| Okay, I read the speech.
|
| He's not formulating the Gell-Mann Amnesia effect in that
| speech, he's referencing an existing idea he had some time
| before, and which he's not even the first person to observe.
|
| And that idea is not _about_ speculation per se, even though
| the article is. He 's saying "as a way to understand why
| speculation is dangerous, consider this phenomenon I'll call
| Gell-Mann Amnesia" and then building from there. I don't think
| the meaning or validity of the Amnesia effect is related to
| that specific example about the tariffs.
|
| I didn't need him to give more examples to understand the
| meaning of the observation: isn't the point that we've _all_
| experienced this when reading outsiders write about our own
| domains of expertise?
|
| In general, I think the author of the original, linked article
| uses the Gell-Mann Amnesia effect correctly, and that the
| description given by Crichton in that speech more or less
| accords with what most people understand it to be.
| mherdeg wrote:
| This is thought-provoking and may change my views - thanks
| for taking the time to share.
| FatalLogic wrote:
| You've written a lot there, so on the first line of your
| comment, please write one sentence that summarizes what you
| want to say.
|
| Because without that context, it's very hard to process the
| rest of it, and I guess many people won't bother.
| notahacker wrote:
| His post is quite well summarised by his second sentence.
| "Gell-Mann Amnesia" gets quoted a lot as "why trust
| journalists when you know they misunderstand stuff you
| understand", but Crichton's actual article wasn't just
| targeting the media but expertise in general and the idea
| that there was any point to making any predictions at all
| about the future, without any evidence base and with a
| layman's drive by attack on climate science, economic
| forecasting and medical testing ethics. Ironically
| considering the contexts the Gell-Mann Amnesia meme gets
| used, Crichton's article that coined it was a pretty broad
| ranging rant including about academics doing science!
|
| And I say this as someone who's cited the Gell Mann Amnesia
| effect meme before without having come across Crichton's
| original article.
| 2hh9d8ue wrote:
| mikevin wrote:
| Does you know of a good resource to learn this kind of
| research/fact checking. I don't have an academic background but
| recently got interested in improving my ability to identify
| "fake news" and assess the validity of some paper.
|
| Where does one even start? Even finding the original paper
| seems a challenge, news articles mention one and then just to
| another news article. If I search the web for the title and
| author it feels like most websites are just rehosting the
| content, like with manuals/datasheets. Not sure how to get
| started here but your comment makes me think you might know
| some resources.
| chmod775 wrote:
| I particularly love how the chronological order of events ends up
| being wrong in a lot of articles.
|
| It usually starts with an article being written in not-
| chronological order, jumping ahead at first, rewinding, maybe
| sprinkling in a few interludes, and then some journalist hastily
| copying it assumes it's in chronological order(?) and gets
| everything completely wrong. A common theme with accident reports
| involving something the journalist in question clearly doesn't
| understand.
| bombcar wrote:
| The solution is to follow _curators_ who comb through the
| independents _for you_ to find what is good.
|
| HN itself is one of these curators, and it does require effort.
| You can find others.
| thatmarkdykeman wrote:
___________________________________________________________________
(page generated 2022-08-11 23:01 UTC)