The trickiness of statistics ...

Started by Baruch, February 20, 2018, 06:12:37 AM

Previous topic - Next topic

Baruch

In medical research in particular ....

https://www.nytimes.com/2018/02/16/books/review/science-inference-data.html

Researchers don't produce good reproducibility, because they don't apply the technique of statistical experiment design correctly.

I still have a problem with the subjective aspect of Baysian probability, even though it goes back to Pascal's wager ... over games of chance.
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

Hakurei Reimu

You're always going to have a subjective aspect to statistics, from selecting the model to deciding how many tests you do before you give up and accept what the data is telling you. Bayesian statistics forces you to confront this, whereas frequentist statistics allows you to hide behind false "objectivity."
Warning: Don't Tease The Miko!
(she bites!)
Spinny Miko Avatar shamelessly ripped off from Iosys' Neko Miko Reimu

Baruch

Still being a gambler like Pascal and Fermat, oui?
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

SGOS

Statistics have flaws, mostly in interpretation.  Down with statistics, damn them!  And therefore, down with the whole scientific method.  Down with reality!   It's time to open the floodgates of total subjectivity so we can create a better reality, one that is more pleasing to the masses, where clerics can simply pontificate without testing.  I long for the Dark Ages, when things were simpler.

Hakurei Reimu

Quote from: Baruch on February 20, 2018, 07:05:17 PM
Still being a gambler like Pascal and Fermat, oui?
We all gamble. Only some people don't realize they're doing it.
Warning: Don't Tease The Miko!
(she bites!)
Spinny Miko Avatar shamelessly ripped off from Iosys' Neko Miko Reimu

Mermaid

Quote from: Baruch on February 20, 2018, 06:12:37 AM

Researchers don't produce good reproducibility, because they don't apply the technique of statistical experiment design correctly.

That is not true. There are certain experiments that have not held up, certainly, but those come to light if not designed properly. Your statement here makes it sound like nothing is ever done right.
A cynical habit of thought and speech, a readiness to criticise work which the critic himself never tries to perform, an intellectual aloofness which will not accept contact with life’s realities â€" all these are marks, not as the possessor would fain to think, of superiority but of weakness. -TR

Baruch

Quote from: SGOS on February 21, 2018, 07:44:20 AM
Statistics have flaws, mostly in interpretation.  Down with statistics, damn them!  And therefore, down with the whole scientific method.  Down with reality!   It's time to open the floodgates of total subjectivity so we can create a better reality, one that is more pleasing to the masses, where clerics can simply pontificate without testing.  I long for the Dark Ages, when things were simpler.

There you go again! - Ronald Reagan

So, use statistics ... incompetently or odiously if you wish.  Just don't expect me to respect you in the morning.  I didn't say they couldn't be used ... just that ape men are bad at it.
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

Baruch

Quote from: Mermaid on February 21, 2018, 02:08:51 PM
That is not true. There are certain experiments that have not held up, certainly, but those come to light if not designed properly. Your statement here makes it sound like nothing is ever done right.

The article said, and I agree ... they often do it wrong.  You can disagree with the article.  And no, I didn't say that no research (involving statistical analysis) is ever done competently.  Neither the article nor I said that.  I said in the title .. it is tricky, not impossible.  Though perhaps in political campaign analysis (before or after the election) bias may make it impossible.  There is a lot of wishful thinking in elections, and other human activities.
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

Baruch

#8
Quote from: Hakurei Reimu on February 21, 2018, 12:50:50 PM
We all gamble. Only some people don't realize they're doing it.

Do go jump off a high building then ;-)  Statistically, there is a small chance you won't kill yourself.  Or ... think about the odds, and avoid long shots.  And yes, people mostly don't realize anything they are doing.  They aren't Socrates.  The odds of an Athenian knowing what they were doing, was about 100,000 to 1.
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

Hakurei Reimu

Quote from: Baruch on February 21, 2018, 04:44:20 PM
Do go jump off a high building then ;-)  Statistically, there is a small chance you won't kill yourself.  Or ... think about the odds, and avoid long shots.
No shit. I never had any plans of jumping off any high buildings, because there are smart gambles and there are stupid gambles, and jumping off of a building is one of the latter. Obviously.

Was there even a point to the above quote?

Quote from: Baruch on February 21, 2018, 04:44:20 PM
And yes, people mostly don't realize anything they are doing.  They aren't Socrates.  The odds of an Athenian knowing what they were doing, was about 100,000 to 1.
Depends on what the Athenian was doing, doesn't it? An Athenian olive oil pressure pressing olives is right that he's going to get olive oil out of it with odds significantly better than 100,000 to 1.
Warning: Don't Tease The Miko!
(she bites!)
Spinny Miko Avatar shamelessly ripped off from Iosys' Neko Miko Reimu

Baruch

The point being, the Athenians, except for Socrates, led unexamined lives, not worth living.  Almost all people do that, they are wasted oxygen consumers.
Ha’át’íísh baa naniná?
Azee’ Å,a’ish nanídį́į́h?
Táadoo ánít’iní.
What are you doing?
Are you taking any medications?
Don't do that.

Cavebear

Quote from: Hakurei Reimu on February 21, 2018, 07:35:11 PM
No shit. I never had any plans of jumping off any high buildings, because there are smart gambles and there are stupid gambles, and jumping off of a building is one of the latter. Obviously.

Was there even a point to the above quote?
Depends on what the Athenian was doing, doesn't it? An Athenian olive oil pressure pressing olives is right that he's going to get olive oil out of it with odds significantly better than 100,000 to 1.

I recall my Mom demanding (when I asked to do something I considered reasonable) "Well if all your friends jumped off a bridge would you jump too)?.  And thinking "What"?  She was often weird like that.
Atheist born, atheist bred.  And when I die, atheist dead!