Real Clear Science ranks best and worst news sites for scien

They couldve used points... its so fucking hard to read.

>>8728844
So basically
>Best
Those outlets that exist almost solely to report science and technology.
>Not bad
Those outlets that cater to a "general audience"
>Worst
Those outlets that are glorified opinion generators

Who'd have thought it. Also
>New Scientist/BBC/The Guardian
>More compelling content than Physics World

Fucking brainlets.

>>8728844
Looks like fake news

>>8728844
What troubles me about it is one part I agree with -- the abysmally low rankings of the major outlets where most people get most of their news.

Science reporting in the main media outlets sis fucking terrible.

>>8728844
>MIT Technology Review
>not compelling enough to read

>>8728844
Why the fuck is the economist rated that good?

>>8728844
no thanks, i get my news from /pol/

>>8729027
This

/pol/ shitposts are the most reliable source of news in the world

>>8729011
Their special features are generally very in-depth.

FAKE NEWS

>>8728969
>Fucking brainlets.
Exactly. If it is not interesting to lay-people it'll rank lower here. But that's OK since not everything has to cater to the masses.

>>8728844
>VOX
loafsodlfoasdlfoasdgjkserhgliuserhbvldbrvjhlsaebrvsaleryvbslearvgbsaelrygalerghalrbvsleyrhbgjserhbvalweruglqwiuerawiuefgaslrhgvbalvhbalfgbaldfhvbalekrgbvaelrgblewnbslejktvbajrhgb askjwfgjwrua gajrgajrg aj gajd gbsdfjgbwljdfbvayvfcn xc fvbsdfbxdv xc

>Th eguardian

REEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

>>8728844
ifuckinglovescience is off the screen to the upper right-- compelling but worthless. I'd rate infowars much more compelling than they do, very readable and entertaining, just total bullshit.

>>8728844

> IEEE spectrum
> Stratfor
> infowars/drudge

If you want real red pills you need to pick one

>>8729716
whats wrong with IEEE spectrum?

>>8728844
>not usually
>538
Tip top shitposting, 538 and Real Clear are so rude to each other because of some imagined rivalry

I only get my information from natural news and I won't be changing you fucking GMO shilling cunts.

I hope the CEO of mon$anto knows his money is going to waste.

File: 547674764643.jpg (15KB, 216x209px) Image search: [Google]

15KB, 216x209px

>>8728844
>arstechnica
AHAHAHAHHAHAHAHAHHAHAHAHAHHAHAHAHAHHAHAHAHAHHAHAHHAHA

>>8728844
national geographic, who once reported that the japanese are a hive mind that don't mourn their dead

>>8728844
Let's make this clear though. This is brainlets ranking science websites for other brainlets. What do you think they were gonna come up with?

>National Geographic
https://www.youtube.com/watch?v=YQba3ENhlKA

>>8730405
can you tell us yours

>>8728844
Should I get a New Scientist subscription? Love their articles but it's so fucking expensive.

File: www.jpg (4KB, 189x53px) Image search: [Google]

4KB, 189x53px

where's the one i use?

File: nate_silver_election.jpg (799KB, 690x6602px) Image search: [Google]

799KB, 690x6602px

>>8730331
Nate Silver on suicide watch

>>8731082
I don't get people critising Nate Silver. He gave Trump the best chance of winning out of almost all the poll aggregators (something a bit above 30%, which sounds pretty damn reasonable in retrospect to me). Not to mention the election was damn close - if it had swung the other way a half a percent or so suddenly those opinion pieces don't look so silly.

Classic case of hindsight bias

>>8731166
/pol/tards don't understand statistics

>>8731166

>push a false narrative for over a year
>rig polls to back up your "facts"
>get btfo on live television

What's not to love?

File: nateshilverbtfobynntaleb.png (73KB, 635x443px) Image search: [Google]

73KB, 635x443px

>>8731179
Hilarious because I'm actually a statistician

File: nntaleb1.png (146KB, 592x689px) Image search: [Google]

146KB, 592x689px

>>8731166

File: nntaleb2.png (120KB, 591x598px) Image search: [Google]

120KB, 591x598px

>>8731166
.

>>8731750
>>8731752
>>8731753
>I'm a statistician
>So instead of explaining something with my own words about a field I am supposedly proficient in, I'm going to use Twitter screenshots to make a point

>>8731824
My point is there are experts far more knowledgeable in the field who have statistical rebuttals to 538 and the validity of their prediction models. I was the one who was attacked first in this thread as a /pol/tard despite the fact that I spend most of my time here on /sci/ and /g/

>>8728844
info wars in the wrong corner t b h

>>8731166
He did a good job and wasn't as biased as other people but him being so moderate made him extremely easy to pick on, as he didn't fall back on the retarded muh feelings arguement of many hard lefties, but rather used statistics and just happened to be very unlucky.

>>8731839
Please explain Taleb's criticism in your own words.

>>8731867
Infowars engages in investigative journalism and does not allow people to pay them to run a story. CNN does neither of these things.

The chart is just garbage desu.

>>8731935
Are you joking?

>arstechnica
>theAtlantic
>that high
>slate/vox/theguardian/washingtonpost
>not in the ideologically driven category

Which marxist made this?

>>8731944
You know this is specifically about science coverage right?

>>8731935
You forgot to mention that Infowars doesn't drop people from interviews the moment they disagree with Jeff Bewkes narrative.

>>8731942
Have you been living under a rock, or are you just not American? CNN is National Enquirer tier.

>>8728844
I'm not sure that any of those belong anywhere past the middle square, much less the top left.

>>8728844
>vox is evidenced based and not ideologically driven

that is such garbage.

vox frequently puts out articles with which have the pretension of objectivity and factual content but are really laughably biased and flimsy interpetations of facts.

For example their video-article which said something like "how the highway system is racist" because highways were built so that cars did not need to stop off in the middle of town where inner city areas were which is wheree black people often lived, which is racist.

It's a total fucking joke putting the daily mail in the red column but not vox.

someone needs to show the authors of that picture this
https://www.youtube.com/watch?v=odF4GSX1y3c

and ask how they justify putting vox in the "evidence based reporting" column ahead of the telegraph.

what fucking garbage

Also fundamentally there is a problem that evidence-based and ideologically drivenare not opposites of each other.

the dailymail's articles are always based on evidence but they're also ideologically driven (or rather, delivering to an audience that is)

>>8731934
What he was explaining is there is too much random noise in their model to accurately assign a valid figure on the probability. And due to this variance that cannot be measured, it essentially becomes a coin toss.

File: 8a5[1].jpg (22KB, 600x315px) Image search: [Google]

22KB, 600x315px

>>8729249

>>8731953
>>>/pol/

BBC is not ideologically driven? Good joke.

>>8732115
Protip: saying "pol" every time you disagree with someone doesn't win the argument.

>>8730886
Absolute garbage. Their writers often fabricate details in stories when they don't feel like doing research into the topic or want to make it sound flashier. They sensationalize everything and provide almost no detail.

>>8732127
Ignore SJWtards. They don't belong here.

>>8732123
This is about scientific coverage specifically

>>8732073
No that's not actually what he's saying. The issue isn't that there is too much random noise in the model, the issue is that the model lacks random noise. You aren't a statistician, you liar.

i.e. the model is based on the historical predictive power of polls and the current results of those polls, without considering the effect of unforeseeable events which have a high effect on the election between now and then. There is nothing inherently wrong with the model used by Nate Silver, it's the choice of model that Taleb is criticizing.

>>8732019
>the dailymail's articles are always based on evidence but they're also ideologically driven (or rather, delivering to an audience that is)
t. Delusional /pol/tard

>>8732127
That would imply I am trying to argue. Why would I argue with a delusional person? Don't bother answering, just cry more about how mean I'm being.

>>8732160
you are retarded and nobody takes you seriously lmao. this is the fate of all your fellow SJWtards in this universe

>>8732160
I'm not even the person you were talking to, all I was saying is that calling someone "pol" doesn't mean your argument is automatically right

>>8731867
>info wars in the wrong corner t b h

Yeahm they should be on the upper right.

Guys who made the chart erred in not seperating engaging and interesting from true or not, and so got the diagonal line the chart shows.

Boring uninteresting nonsense is no threat to anything, engaging and convincing nonsense gets read and spread.

>>8732149
That doesn't mean the model specifically lacks noise, that's exactly what an unforeseeable event is. You're just mincing words now and resorting to name calling like everyone else in this thread. I'm not an expert on all the statistical lingo of every single model, my specialty is data algorithms and database management.

>>8731166
He insists that his polling has predictive value. It does not, nor does anybody else's polling. Polling data, when it has value at all, is a snapshot of the present moment, it cannot predict how things will change in the future.

Silver is smart enough to know this, but pushes the false story anyway.

>>8732206
So?

>>8732019
>vox frequently puts out articles with which have the pretension of objectivity and factual content but are really laughably biased and flimsy interpetations of facts.
>For example their video-article which said something like "how the highway system is racist" because highways were built so that cars did not need to stop off in the middle of town where inner city areas were which is wheree black people often lived, which is racist.


This chart is based on their science reporting, not their reporting in general.

>>8732229
This. A 1,000x this. Nate's problem is he assumes that the data he is feeding into his model is valid, but that's simply an assumption. It was purely an assumption that the "Blue Wall" was going to hold and follow a similar pattern to 2008/12. I did not think that was going to happen. I kept telling everyone it was going to be more like 2010/14. I predicted 49 out of 50 states correctly (I only got NH wrong). Nate only got 44 out of 50. He was certainly more level-headed than the 98.1% certainty HuffPo "stats experts" who ran 10 million simulations with Clinton winning 9.8 million of those. I bet they wish they were living in one of those simulations where she won.

>>8732220
It doesn't say engaging/interesting, it says "is the science content compelling." Sites which exaggerate trivial results into big things should be in the bottom right.

>>8732149
>the model is based on the historical predictive power of polls

Then the model is nonsense. Polls have essentially zero predictive power beyond a few days or a week, and the electorate can transform within that time frame given the right events.

Politicians who rely on polling to tell them what is going to happen are being fooled by dishonest pollsters.

Polling has value, as a chance to see what the ground truth today looks like, and how what happened yesterday impacted today. It tells you fuck-all about tomorrow.

>>8732259
People also fail to understand that polls have a confidence interval for a reason. +/- 3% can be a huge difference.

>>8732228
No, Taleb is saying the model should have more variance so that the prediction "converges" to 50% until the election. Taleb's criticism is utterly silly and anyone who claims he BTFO Silver doesn't understand statistics. Just another of Taleb's esoteric, nonsensical contrarian rants that only impresses laymen.

>>8732229
>his polling
You have no idea what you're talking about.

>>8728844
Because /sci/ isn't a science website, it's an IQ circlejerk mixed in with bait from right wing retards that came from /pol/

This place is trash compared to any website of repute

>>8732229
Polling, historically, has plenty of predictive value. Nate Silver would have no model at all if they didn't.

>>8732256
>It doesn't say engaging/interesting, it says "is the science content compelling."

But in the article they define that as "very interesting, so-so, or weak?" and talk about crisp, exciting writing as opposed to dry writing as they discuss that metric. They also mention appeal to the layman or nonspecialist.

Defining something not being "compelling" as meaning the science is exaggerated or slanted seems to be using the quality of the science as both axes, which is what I think they have done.

>>8732282
>IQ
>right wing retards from pol
Your low IQ is nobodies fault you stinking SJWtard

>>8732266
I understand that, and it is not relevant to why Silver is a huckster.

>>8732276
Misspoke, apologies, about grade 2.5 micro-sorries.

Substitute "the claims he makes about what polling data reveals to him through magic glitter dust."

>>8732287
>Polling, historically, has plenty of predictive value.

You are wrong.

>Nate Silver would have no model at all if they didn't.

You are right.

>>8732259
Nope, the polls have had predictive power. That's how models like this are created. The first thing you have to do is determine the predictive power of a poll via past data. Just because Trump won doesn't mean you get to revise history and declare polls have no predictive power.

>>8732296
Go away Greek faggot.

>>8732314
>SJWtards hate greeks as well
You trannie pedophiles are really running out of friends.

>>8732331
Do you just shit up the thread and throw an autistic everytime someone says go back to >>>/pol/

>>8732290
No, the quality of the science is an independent variable from the accuracy of the reporting on that science. The fact that inaccurate sites also exaggerate trivial results makes sense, but you could also have sites that only misrepresent big science. The reason you don't is because its harder to misrepresent what everyone else isaccurately reporting on.

>>8728969
tldr
>superiors should only have voting power

>>8732338
Learn english before you try to shitpost you stinking SJWtard. Fucking hell you're dumb.

>>8732364
*English
*shitpost, you
>SJWtard
That's not a real word, honey.

>>8732364
>autistic tantrum continues
>>>/pol/

>>8732374
/pol/tard BTFO

>>8732374
>>8732377
>>8732381
>If I act more retarded maybe people will think I'm trolling.
ssshhh...keep sucking SJWtard...this is why you're here for.

Remember when Scientific American used to be like a light scientific journal instead of the rag that it is now?


>>8732338
>throw an autistic
kek

>>8732241

So if you have no argument, then just say "I give up." Be an adult.

Good day, sir. I SAID GOOD DAY.

>>8732392
>>>/pol/
You really gotta go back

>>8732439
>SJWtard trying to hijack /sci/ with the same shitpost
not possible with your 2 digit IQ.

>>8732457
>>>/pol/
Leave

>>8732492
Sorry this is a science board. You have to go back to your cesspool >>>/t/umblr

>>8732499
>>>/pol/

>>8728844
every single one of those is fake news.

>>8728969
Author must be a complete retard to compare Nature and FOX/CNN. They are completely different and cater to completely different audiences.

>>8732561
All of them report on science (some exclusively, others not). This is about news sites targeted at the general public.

Whats wrong with physics world? Its pretty good.

>>8732580
not compelling enough to read for people with lower IQs

>>8732311
Sometimes, nothing happens to change the outcome of an election, and a candidate with the most support early on goes onto win. Other times, things change (this is the purpose of holding a campaign before an election.) And sometimes a candidate with less support in polling is doing a much better job in find 'em,vote 'em, count 'em GOTV efforts. None of that is predictable by polling, but it is decisive in determining who wins an election..

>>8732019
>the dailymail's articles are always based on evidence
There's no way you typed this with a straight face.

>>8732600
And none of that shows that polls don't have predictive power. I'll wait.

>>8732621
but anon, if they are not right 100 % of the time, how can they have predictive power?

>>8732437
There is no argument in the first place. A delusional individual said something stupid and I mocked him. I have no obligation to refute every nutjob that posts.

>>8732629

They don't have any predictive power when the people doing the polling are literal shills. +5 for Hillary in Arizona? Really?

>>8728844
>natural news

I'll be the first to admit that there are a lot of problems at that website, but at least they encourage healthy living through eliminating vitamin and mineral deficiencies, getting calories through whole foods and not through processed foods, adequate exercise, and consistent quality sleep.

For a group such as yourselves who are so smart, it's amazing how much you literally do not care about your own health. You eat shit and your bodies look like shit. Do you really not want to live long and healthy lives?