Page 1 of 1

2110: "Error Bars"

Posted: Mon Feb 11, 2019 9:03 pm UTC
by svenman
Image

Title text: "...an effect size of 1.68 (95% CI: 1.56 (95% CI: 1.52 (95% CI: 1.504 (95% CI: 1.494 (95% CI: 1.488 (95% CI: 1.485 (95% CI: 1.482 (95% CI: 1.481 (95% CI: 1.4799 (95% CI: 1.4791 (95% CI: 1.4784..."

I wonder whether you could meaningfully formulate a distinction between situations where the outer bounds of these error intervals converge or diverge. And if so, whether you could somehow arrive at an alternative derivation of the Heisenberg uncertainty principle that way.

Re: 2110: "Error Bars"

Posted: Mon Feb 11, 2019 9:17 pm UTC
by moody7277
So many open parentheses...

*feels discomfited*

Re: 2110: "Error Bars"

Posted: Mon Feb 11, 2019 9:19 pm UTC
by Heimhenge
The first thing that comes to mind is that this is a type of fractal, though I don't think Randall implied the error-error would be iterated to infinity. But yeah, it could be convergent either way. Do a search for convergent fractal and you'll find plenty of examples.

EDIT: Just noticed the ellipsis at the end of the title text. So yep, this is a fractal. Or a convergent series, if you want to view it numerically.

Re: 2110: "Error Bars"

Posted: Mon Feb 11, 2019 9:47 pm UTC
by fibonacci
Heimhenge wrote:The first thing that comes to mind is that this is a type of fractal, though I don't think Randall implied the error-error would be iterated to infinity. But yeah, it could be convergent either way. Do a search for convergent fractal and you'll find plenty of examples.

EDIT: Just noticed the ellipsis at the end of the title text. So yep, this is a fractal. Or a convergent series, if you want to view it numerically.


My first impulse was to try to figure out the L-system generator for it.

Re: 2110: "Error Bars"

Posted: Mon Feb 11, 2019 10:36 pm UTC
by dash
Good one!

Re: 2110: "Error Bars"

Posted: Tue Feb 12, 2019 1:16 am UTC
by Mikeski
This comic explains a lot.

Well, maybe not a lot, but at least it explains political polls.

Re: 2110: "Error Bars"

Posted: Tue Feb 12, 2019 3:02 am UTC
by Soupspoon
An error bar is a place you go to to get drunk and spout baseless opinions at others, Shirley…

Re: 2110: "Error Bars"

Posted: Tue Feb 12, 2019 12:04 pm UTC
by cellocgw
Soupspoon wrote:An error bar is a place you go to to get drunk and spout baseless opinions at others, Shirley…


I was in an error bar once when a mean sigma walked in...

Re: 2110: "Error Bars"

Posted: Tue Feb 12, 2019 5:07 pm UTC
by ucim
cellocgw wrote:I was in an error bar once when a mean sigma walked in...
Xe got brought before the Grand Canonical Ensemble, but they didn't have the proper confidence level.

Jose

Re: 2110: "Error Bars"

Posted: Tue Feb 12, 2019 7:49 pm UTC
by ksvanhorn
This is what happens when you use frequentist methods. Bayesians have no trouble propagating errors.

Re: 2110: "Error Bars"

Posted: Wed Feb 13, 2019 11:46 am UTC
by Aardwizz
47.0 ± 5.6 ± 1.2

Re: 2110: "Error Bars"

Posted: Wed Feb 13, 2019 2:43 pm UTC
by colonel_hack
ksvanhorn wrote:Bayesians have no trouble propagating errors.

True.

Re: 2110: "Error Bars"

Posted: Wed Feb 13, 2019 3:23 pm UTC
by DanD
Soupspoon wrote:An error bar is a place you go to to get drunk and spout baseless opinions at others, Shirley…


I thought that was the internet?

Re: 2110: "Error Bars"

Posted: Thu Feb 14, 2019 8:22 am UTC
by FOARP
"Yo Dawg, I heard you like error bars"


Should I be proud that I am the first one on this thread to mention this meme, or sad that this probably indicates how out of date my cultural references are?

Re: 2110: "Error Bars"

Posted: Thu Feb 14, 2019 12:44 pm UTC
by Soupspoon
♬The Error Bar Kid is strong and tough,
Only the best is good enough,
Meaner modes, a taller bar,
The graph based on an error-bar…♬

"The Error Mars are on me!"


(Nichier historicalier meme. :P )
Decided against ♬We are the Error-barries…♬ or "You can have an color [bar], so long as it's black.". Tbis one will be confusing enough.