|
From the Randomwaffle thread:
|
# ? Jul 16, 2023 21:38 |
|
|
# ? May 28, 2024 18:10 |
|
Powered Descent posted:The one that confuses me is "owns a flashlight". I'd have figured just about everybody would have at least a cheap one rolling around in the junk drawer; it never struck me as a particularly white thing to have. It's for when we go downstairs to investigate what that noise is instead of turning all the lights on, getting in the car, and driving thirty-five miles the gently caress away
|
# ? Jul 17, 2023 22:22 |
|
Based on the highest numbers being clustered around the rate of white people in the U.S., I don't think this chart is listing anything like "predictive capability" whatever that means. I think it is just listing the percentage of each ownership group that are white i.e. If someone owns a pet in the US in 2016 there is a 63.4% chance they are white. This is of course a terrible way to display the data for the point they are trying to make. Whether these numbers make for strong indicators doesn't depend on the raw numbers listed; it depends on how far those numbers are above or below the (unlisted) proportion of whites in the general population.
|
# ? Jul 18, 2023 23:41 |
|
Agnostalgia posted:Based on the highest numbers being clustered around the rate of white people in the U.S., I don't think this chart is listing anything like "predictive capability" whatever that means. I think it is just listing the percentage of each ownership group that are white i.e. If someone owns a pet in the US in 2016 there is a 63.4% chance they are white. That’s basically what predictive capability is. If someone has a pet, there is a 63.4% chance that they are white. Owning a pet is 63.4% predictive of whiteness. (I wonder if this will convince my wife to let me get a kitten.)
|
# ? Jul 19, 2023 02:36 |
|
if the predictive capability is basically the same as the demographics, doesnt that also mean that if you have a random black person and a random white person, they are equally likely to own a pet?
|
# ? Jul 19, 2023 08:35 |
|
Agnostalgia posted:Based on the highest numbers being clustered around the rate of white people in the U.S., I don't think this chart is listing anything like "predictive capability" whatever that means. I think it is just listing the percentage of each ownership group that are white i.e. If someone owns a pet in the US in 2016 there is a 63.4% chance they are white. Predictive capability is relevant twice. First in the measure. There's no need for prediction if you have a true census on a matter. By census, I mean you measure the item of interest on every target of the survey. Example, how many people at a hypothetical high school are taller than 5 ft by measuring the height of each person. Checking the washing machine status of every white American is a labor intensive task probably not worth the immense cost and probably isn't feasible. Hiring enough people to check every squarefoot of America for both white people and washing machines would be costly, and there are plenty of places where you may not have permission to check. So you take a survey by selecting what is hopefully a random representative sample and then using that to predict the value for everyone. So if I somehow achieve a true random and unbiased sample of people, the "expected value" of that sample is equal to the total proportion of people. But it's only a prediction. To demonstrate, imagine if you only asked one white person whether they have a cat. Now imagine the answer is yes. If we extrapolate that data to a ratio of all people, then we can conclude that 100% of white people have cats. Probably not true though, so you need to achieve a sufficient sample. This is where the weird stats backwards speak. What's a sufficient sample? Well you want to show that a null hypotheses is extremely unlikely. This is calculated by determining how many samples you need for the probability that 'this is just a fluke' falls beneath your acceptable significance level (α). But at the end this is just a prediction about the true proportion. So in statistics it's decided how many white people own a cat by determining how likely that some measure of white people with cats was an exceptional circumstance, then drilling the exceptionality of that circumstance to the extreme. Interpreting data then is also making predictions. How does knowing that 40% of white people own cats help me solve a problem? If I want to know how many cat owners are in a room of 100 white people, I can use that statistic to make a prediction. If I actually know about those specific white people, then why do I care about the stat? The value of the stat is to use prediction so I don't have to measure or for when I can't measure. There may be some unusual edge cases where this isn't the case, but the vast majority of interpretations are just complicated and obfuscated* versions of this. *much like my posting.
|
# ? Jul 19, 2023 10:49 |
|
Subjunctive posted:(I wonder if this will convince my wife to let me get a kitten.) Do you really think "this will make me more like a white person" is a great argument for anything?
|
# ? Jul 19, 2023 12:00 |
|
BonHair posted:Do you really think "this will make me more like a white person" is a great argument for anything? "Rescuing a cat isn't the worst reason to ascribe to genetic determinism," is a technically true statement
|
# ? Jul 19, 2023 12:05 |
|
Can someone explain how predictive capability is calculated?
|
# ? Jul 19, 2023 14:43 |
|
if it's not bayesian surprise/information entropy calculated in shannon bits, i'm not interested
|
# ? Jul 19, 2023 15:33 |
|
|
# ? Jul 19, 2023 17:54 |
|
Unironic support for frequentism. None of the people using bayesian analysis are producing anything useful for daily thinking.
|
# ? Jul 19, 2023 18:02 |
|
What was the original version of this "stop doing ______" thing? e: by which I mean, what was the first one?
|
# ? Jul 19, 2023 18:05 |
|
credburn posted:What was the original version of this "stop doing ______" thing? math was the first one
|
# ? Jul 19, 2023 18:06 |
|
Tree Goat posted:math was the first one But how would you know that, without math
|
# ? Jul 19, 2023 18:07 |
|
Captain Hygiene posted:But how would you know that, without math Fingers
|
# ? Jul 19, 2023 18:12 |
|
Captain Hygiene posted:But how would you know that, without math "YEARS OF COUNTING" is great because they are really underselling how long humans have been counting. But to provide a larger and more accurate number, they would have to go beyond ten fingers.
|
# ? Jul 20, 2023 14:55 |
|
jjack229 posted:"YEARS OF COUNTING" is great because they are really underselling how long humans have been counting. But to provide a larger and more accurate number, they would have to go beyond ten fingers. Years can mean anywhere from 2 years up to infinity years. That's not really underselling
|
# ? Jul 20, 2023 15:03 |
|
CainFortea posted:Years can mean anywhere from 2 years up to infinity years. That's not really underselling It is because saying "30,000 years" would be more impactful than a vague "years", which is usually used if the number is small, otherwise "decades" or "centuries", etc would be used.
|
# ? Jul 20, 2023 15:39 |
|
jjack229 posted:It is because saying "30,000 years" would be more impactful than a vague "years", which is usually used if the number is small, otherwise "decades" or "centuries", etc would be used. If I didn't know better, I'd say there's humor intended somewhere.
|
# ? Jul 20, 2023 16:08 |
|
CainFortea posted:Years can mean anywhere from 2 years up to infinity years. That's not really underselling loving inferior English language that only has one plural form instead of like 5
|
# ? Jul 20, 2023 19:18 |
|
Powered Descent posted:If I didn't know better, I'd say there's humor intended somewhere. Tbf there are flat earthers out there so the "math isnt real" guy very well could exist
|
# ? Jul 24, 2023 00:08 |
|
jjack229 posted:It is because saying "30,000 years" would be more impactful than a vague "years", which is usually used if the number is small, otherwise "decades" or "centuries", etc would be used. you're who the picture is warning us about
|
# ? Jul 24, 2023 00:52 |
|
mandatory lesbian posted:Tbf there are flat earthers out there so the "math isnt real" guy very well could exist His name is Terrence Howard
|
# ? Jul 24, 2023 04:11 |
|
mandatory lesbian posted:Tbf there are flat earthers out there so the "math isnt real" guy very well could exist I will never not take the opportunity to point out that one of these guys was prime minister of Australia Malcolm Turnbull posted:The laws of mathematics are very commendable, but the only law that applies in Australia is the law of Australia
|
# ? Jul 24, 2023 07:24 |
|
Keep arguing with a meme, you're winning.
|
# ? Jul 24, 2023 09:24 |
|
credburn posted:What was the original version of this "stop doing ______" thing? stop doing linear time
|
# ? Jul 24, 2023 11:06 |
|
CainFortea posted:Years can mean anywhere from 2 years up to infinity years. That's not really underselling
|
# ? Jul 24, 2023 17:43 |
|
Scarodactyl posted:What, do you say 'one and a half year'? A year and a half.
|
# ? Jul 24, 2023 17:48 |
|
I've lived here for 0.000000042 years
|
# ? Jul 24, 2023 17:54 |
|
‘Over a year’
|
# ? Jul 24, 2023 18:03 |
|
goblin week posted:stop doing linear time Start doing linear crime
|
# ? Jul 24, 2023 19:28 |
|
don't do the linear crime if you can't do the linear time
|
# ? Jul 24, 2023 20:22 |
|
OwlFancier posted:don't do the linear crime if you can't do the linear time So that's why the Bajoran Prophets exist outside of time -- they're just avoiding prosecution. Wormhole? More like loophole
|
# ? Jul 24, 2023 21:19 |
|
|
# ? Jul 31, 2023 03:16 |
|
That's certainly all true.
|
# ? Jul 31, 2023 03:25 |
|
Blue whale vs Vatican City
|
# ? Aug 1, 2023 23:59 |
|
Blue whale vs Vatican City?!
|
# ? Aug 2, 2023 00:03 |
|
I get that's the size of the impact crater, but how does it scale with height?
|
# ? Aug 2, 2023 00:07 |
|
|
# ? May 28, 2024 18:10 |
|
Whale-sized Vatican city vs Vatican City-sized whale
|
# ? Aug 2, 2023 00:56 |