Caesura109 posted:

Any thoughts on Pew Research Center's reliability? Their surveying methods, their ideology etc.

With polling, there are two issues to look at: the technical aspects of how they survey and the questions they ask. On the technical side, they way pollsters solicit respondents (robocall, human call, online, face to face, etc.) can influence how people answer and what demographics they actually reach. There are a variety of ways to try to correct for these selection issues by reweighting the sample, and those options give pollsters more flexibility to manipulate the outcome. You can see the results of this sort of manipulation in election polling--in the US, republican and democratic pollsters will often be biased by 1-3% towards their preferred candidates.

However, these technical issues often aren't too important if you're looking at non-electoral issues--parties have an incentive to re-position themselves when they become uncompetitive, so elections are typically very close, and a 2% bias makes a huge difference in terms of predicting the winner or winning coalition--a candidate getting 51% of the vote vs 49% is huge. However, if you're asking how many people are in unions in the US, finding 9% instead of 11% probably isn't going to change your assessment of the situation.

Much more important is how questions are phrased, and this is what you should be looking at when evaluating the credibility of surveys. Here's a simple, well known example:

Each question is asking the same thing, but the results are dramatically different--the differences are far greater than what plausible methodological choices could generate. You should always look at the actual wording of survey questions (not the gloss in a headline) and think about what other ways the question could have been phrased, and why the surveyor chose the wording they did. That's the easiest way to suss out the slant of the survey, and you don't need to be a statistician to do it.

cars posted:

Haha

cars posted:

Caesura109 posted:

Trying to find a way to deprogram econ undergrad friend from thinking it's a legitimate field of study

we got a regular sisyphus over here

According to a leaked January 2 memo from the ‘Leading Group of Internet Financial Risks Remediation’ — the country’s internet finance regulator which initiated the clampdown on bitcoin — bitcoin miners should make an “orderly exit” from China because they have consumed “huge amounts of resources and stoked speculation of ‘virtual currencies.”

this exodus of bit-tribes seem to be heading towards canada. a reuters report says the 'miners' find manitoba "attractive", quite an unexpected adjective.
all this was mildly amusing but then the circus came to town. i go to pop music to avoid geopolitics but someone found a way to make bitcoin funnier.

slipdisco posted:

a reuters reportsays the 'miners' find manitoba "attractive",

someone found etc.

hey posted:

Like many other financial instruments, cryptocurrencies provides absolutely nothing of value in this world. But finance makes the attempt to hide all the bullshit under layers of convoluted terminology. cryptocurrencies doesn't even bother with that. It's just straight up gambling. Sure you can buy drugs online and an hour of board game time at some nerd cafe with bitcoin, but for all the new up and coming crytocurrencies, they don't do anything. Nothing. Pure speculation. We've finally gotten to the point where people, fully aware, are willing to invest in nothing but speculation. It's fucking bizarre. At least at the casino you don't have to worry about hackers running off with your money.

uh, excusse me but i think you'll find that actually, a new block of transactions is created every ten minutes or so, with 12.5 bitcoins — BTC — reward — plus any fees on the transactions — attached. Bitcoin miners apply as much brute-force computing power as they can to take the prize in this block’s cryptographic lottery. Unprocessed transactions are broadcast across the Bitcoin network. A miner collects together a block of transactions and the hash of the last known block. They add a random ‘nonce’ value, then calculate the hash of the resulting block. If that number is lower than the current mining difficulty, they have mined a block! This successful block is then broadcast to the network, who can quickly check the block is valid. They get 12.5 BTC plus the transaction fees. If they failed, they pick another nonce value and try again. Since it’s extremely hard to pick what data will have a particular hash, guessing what value will give a valid block takes many calculations — as of October 2016 the Bitcoin network was running 1,900,000,000,000,000,000 — 1.9×1018, or 1.9 quintillion — hashes per second, or 1.14 x 1021 — 1.14 sextillion — per ten minutes. The difficulty is adjusted every two weeks — 2016 blocks — to keep the rate of solved blocks around one every ten minutes. The 1.14 sextillion surplus calculations are then thrown away, actually

roseweird posted:

i wish gpus weren't still twice their usual price

there should be a brief window after crypto wipes out that we can acquire the pc gaming engines of our darkest fantasies