How Pure Is This Cup of Joe? • Hacker News Board

Re: How Pure Is This Cup of Joe?

#21

Post by xapata » Tue, Nov 07, 2017, 12:43 PM UTC

gwern wrote: post #15

xapata wrote: post #11

Earlier quoted context omitted.

Reverse isn't the right word here, either. Better to restate the definition: the p-value is the probability of observing the data (as represented by its aggregate statistics), under the condition of the null hypothesis.

I say 'reverse' because that's what it is formally: confusing P(data|null) with P(null|data). It reverses the conditional probability.

Reversing the arguments of the conditional probability, yes, but just saying "it's the reverse" is a bit ambiguous.

> If it is less than 5% probable that the null hypothesis is correct given our data, we will reject the null hypothesis and consider our alternate hypothesis (HA) to be correct.

You could be reversing the 5% to 95%, could be reversing null hypothesis to alternate hypothesis, could be reversing reject to accept, etc. Too many moving parts. P-values are so often misinterpreted, it's best to err on the side of specificity.

Re: How Pure Is This Cup of Joe?

#22

Post by FiatLuxDave » Tue, Nov 07, 2017, 2:57 PM UTC

BugsJustFindMe wrote: post #6

justesjc wrote: post #5

Earlier quoted context omitted.

Why stomp all over somebody's fun? I enjoyed reading it.

Because awareness of the gap between measurement and conclusion might just be the most important thing in the world, and people and governments fail spectacularly at it all the time with significant lasting effects on society.

For anyone reading this who doubts what BugsJustFindMe said here, as an example I invite you to examine how the government measures the loudness of broadcast commercials compared to the programs that surround them. We had seventy years of blasting people with commercials, all because the way of measuring the loudness does not match the actual "loudness".

Re: How Pure Is This Cup of Joe?

#23

Post by lordlimecat » Tue, Nov 07, 2017, 6:44 PM UTC

justesjc wrote: post #5

BugsJustFindMe wrote: post #3

[edit] Here's a photo of the back of a bag of his "reference" Trung Nguyen S blend. It explicitly says that it has flavorings in it, including "coffee flavor". Omfg. https://imgur.com/a/KeSYq --- I guess the guy had fun, but the conclusion is a few giant leaps away from the actual results. Many of his decisions and assumptions were entirely arbitrary, not well substantiated, and lacking just about anything approachin…

Why stomp all over somebody's fun? I enjoyed reading it.

Because that somebody is pretending they have useful quantified results, and they don't.

And "man plays chemist with iodine for a few hours" probably wouldnt make the front page.

Re: How Pure Is This Cup of Joe?

#24

Post by bewo001 » Wed, Nov 08, 2017, 10:13 AM UTC

BugsJustFindMe wrote: post #3

[edit] Here's a photo of the back of a bag of his "reference" Trung Nguyen S blend. It explicitly says that it has flavorings in it, including "coffee flavor". Omfg. https://imgur.com/a/KeSYq --- I guess the guy had fun, but the conclusion is a few giant leaps away from the actual results. Many of his decisions and assumptions were entirely arbitrary, not well substantiated, and lacking just about anything approachin…

Catimor is a variety of arabica, not a different species like robusta. Excelsa seems to refer to the Colombian coffee grade which may include different arabica varieties.