Libertarian's Forum
Libertarian Forum to discuss politics and free market economics.
Libertarian's ForumLibertarian's ForumFreedom Forum › The Replication Crises
Page Index Toggle Pages: [1] 2 3 ... 84 Send TopicPrint
Very Hot Topic (More than 25 Replies) The Replication Crises (Read 6556 times)
SnarkySack
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 3148
Location: Republic of Me
Joined: Sep 11th, 2017
The Replication Crises
Apr 2nd, 2018 at 1:02pm
Print Post  
The Replication Crises is a phenomena happening in the arena of scientific research.  This happens when there is a new and surprising result of some experiment and then, following its publication, other researchers are unable to get the same results by following similar protocols.

There are three main reasons for this:

1)  Bias on the part of the researcher toward "successful" results.

2)  Bias among the professional journals toward publishing surprising results or results that support the philosophy of the journal's editors.

3)  In order for a result to be considered statistically significant, it must have a "p-value" of between one and five percent.  This means that there is between a one percent and a five percent probability that the results are caused by a fluke during the random sorting of participants among the experimental and the control group.  For example if one were testing a new "smart pill" and it showed a twelve percent increase in test scores of the experimental group over the control groups, there is at least a one in a hundred chance the differences is the result of pure chance and not the pill.  If that's the case it would take another fluke to replicate the results.

A one in one hundred chance is not that daunting a probability when you consider the thousands of experiments that look for such results.  Flukes are going to happen and happen often.  But for a study about smart pills with a p-value of 1% to be replicated by another study of smart pills with a p-value of 1% means the probability is one in ten thousand that both would be flukes.  That's the value of replication and it replication is becoming more and more rare.

Why is this important to libertarianism?  I'll let you digest this and explain why in another thread.
  

I used to be burnsred . . .
Back to top
 
IP Logged
 
Don_G
Ex Member


Re: The Replication Crises
Reply #1 - Apr 2nd, 2018 at 1:13pm
Print Post  
SnarkySack wrote on Apr 2nd, 2018 at 1:02pm:
The Replication Crises is a phenomena happening in the arena of scientific research.  This happens when there is a new and surprising result of some experiment and then, following its publication, other researchers are unable to get the same results by following similar protocols.

There are three main reasons for this:

1)  Bias on the part of the researcher toward "successful" results.

2)  Bias among the professional journals toward publishing surprising results or results that support the philosophy of the journal's editors.

3)  In order for a result to be considered statistically significant, it must have a "p-value" of between one and five percent.  This means that there is between a one percent and a five percent probability that the results are caused by a fluke during the random sorting of participants among the experimental and the control group.  For example if one were testing a new "smart pill" and it showed a twelve percent increase in test scores of the experimental group over the control groups, there is at least a one in a hundred chance the differences is the result of pure chance and not the pill.  If that's the case it would take another fluke to replicate the results.

A one in one hundred chance is not that daunting a probability when you consider the thousands of experiments that look for such results.  But for a study about smart pills with a p-value of 1% to be replicated by another study of smart pills with a p-value of 1% means the probability is one in ten thousand that both would be flukes.  That's the value of replication and it is becoming more and more rare.

Why is this important to libertarianism?  I'll let you digest this and explain why in another thread.


Well it certainly does need an explanation of what you are trying to say!
Examples of the two sorts of 'biases' would maybe help?

All I can get out of this is that you are taking a position against science or scientists in general.
  
Back to top
 
IP Logged
 
SnarkySack
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 3148
Location: Republic of Me
Joined: Sep 11th, 2017
Re: The Replication Crises
Reply #2 - Apr 2nd, 2018 at 1:49pm
Print Post  
Quote:
Well it certainly does need an explanation of what you are trying to say!
Examples of the two sorts of 'biases' would maybe help?

All I can get out of this is that you are taking a position against science or scientists in general.


Then I don't think explaining it to you would help.

  

I used to be burnsred . . .
Back to top
 
IP Logged
 
Don_G
Ex Member


Re: The Replication Crises
Reply #3 - Apr 2nd, 2018 at 2:10pm
Print Post  
SnarkySack wrote on Apr 2nd, 2018 at 1:49pm:
Then I don't think explaining it to you would help.



I'll wait until you try to explain it to the others. Don't be too snarky with me. Some of your issues you raise are of little interest to the others and so they don't bother to comment.

Then without me, you're left with Craig Sickler who is likely to tell you fooch or change the topic to slavery.

I fully acknowledge the value of your comments in answer to some of my posts. There's no point in denying each other yet!

And FWIW, you really hit on something with the thread on high cost pharmas, etc.
  
Back to top
 
IP Logged
 
SnarkySack
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 3148
Location: Republic of Me
Joined: Sep 11th, 2017
Re: The Replication Crises
Reply #4 - Apr 2nd, 2018 at 3:45pm
Print Post  
Quote:
I'll wait until you try to explain it to the others. Don't be too snarky with me.


I just have to think that as quickly as you answered, you did not have time to digest it.

I had already said that I would relate the replicaiton crises to libertarianism.

If you don't buy that there is any problem with replication, then you don't share the concerns of the scientific community and so my explanation would be meaningless anyway.
  

I used to be burnsred . . .
Back to top
 
IP Logged
 
The Opposition
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 7224
Joined: Apr 30th, 2014
Re: The Replication Crises
Reply #5 - Apr 2nd, 2018 at 3:48pm
Print Post  
The problem with replicability is a function of the value of scientific conclusions on the free market.

People can use conclusions with the "scientific" label to sell more of their product of their competitor causes cancer, or just push the media agenda to generate political power.

These things have worth.

Demand for the "right" scientific conclusions is high.

And the free market simply provides.
  

Making Sci-Fi great again since 2063.

Not taking Jeff seriously until he admits this is animal abuse (which he says should be illegal): http://www.youtube.com/watch?v=WE-IT7_CaE4
Back to top
 
IP Logged
 
SnarkySack
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 3148
Location: Republic of Me
Joined: Sep 11th, 2017
Re: The Replication Crises
Reply #6 - Apr 2nd, 2018 at 4:12pm
Print Post  
The Opposition wrote on Apr 2nd, 2018 at 3:48pm:
The problem with replicability is a function of the value of scientific conclusions on the free market.

People can use conclusions with the "scientific" label to sell more of their product of their competitor causes cancer, or just push the media agenda to generate political power.

These things have worth.

Demand for the "right" scientific conclusions is high.

And the free market simply provides.


Well put.  One thing about the free market is that there is a high demand for products that are first of their kind.  So why wait for replication when you've got these awesome new findings to sell?

Note for Don:

This isn't at all about Darwinism or Global Warming.  It couldn't be, because those things are not subject to the kind of experimentation I'm talking about.  They both rely on models that make predictions and then can be changed if the predictions don't pan out.

  

I used to be burnsred . . .
Back to top
 
IP Logged
 
Jeff
Libertarian Freedom Member
*****
Online

Libertarian's Forum

Posts: 32535
Location: USA
Joined: Feb 26th, 2014
Re: The Replication Crises
Reply #7 - Apr 2nd, 2018 at 5:59pm
Print Post  
SnarkySack wrote on Apr 2nd, 2018 at 1:02pm:
The Replication Crises is a phenomena happening in the arena of scientific research.  This happens when there is a new and surprising result of some experiment and then, following its publication, other researchers are unable to get the same results by following similar protocols.

There are three main reasons for this:

1)  Bias on the part of the researcher toward "successful" results.

2)  Bias among the professional journals toward publishing surprising results or results that support the philosophy of the journal's editors.

3)  In order for a result to be considered statistically significant, it must have a "p-value" of between one and five percent.  This means that there is between a one percent and a five percent probability that the results are caused by a fluke during the random sorting of participants among the experimental and the control group.  For example if one were testing a new "smart pill" and it showed a twelve percent increase in test scores of the experimental group over the control groups, there is at least a one in a hundred chance the differences is the result of pure chance and not the pill.  If that's the case it would take another fluke to replicate the results.

A one in one hundred chance is not that daunting a probability when you consider the thousands of experiments that look for such results.  Flukes are going to happen and happen often.  But for a study about smart pills with a p-value of 1% to be replicated by another study of smart pills with a p-value of 1% means the probability is one in ten thousand that both would be flukes.  That's the value of replication and it replication is becoming more and more rare.

Why is this important to libertarianism?  I'll let you digest this and explain why in another thread.
It's important to science, but perhaps even more important to political economy, where results that have not been honestly peer reviewed are presented as "scientific fact" and governments base policies on the un-reviewed or poorly reviewed "discoveries".

Any honest competent scientist will replicate their own experiments before they publish, and maybe ask colleagues to look over their hypothesis and procedures to see if they did something obviously stupid.

Unless of course, as you say, the government granted them lots of money and told them what results were required...

That's fraud and should be prosecuted.
  
Back to top
 
IP Logged
 
Don_G
Ex Member


Re: The Replication Crises
Reply #8 - Apr 2nd, 2018 at 7:43pm
Print Post  
SnarkySack wrote on Apr 2nd, 2018 at 4:12pm:
Note for Don:

This isn't at all about Darwinism or Global Warming.  It couldn't be, because those things are not subject to the kind of experimentation I'm talking about.  They both rely on models that make predictions and then can be changed if the predictions don't pan out.



Your criticism of science, scientists, Darwinism, and AGW are of little interest to we who haven't gone over to the far side burnsred.

You others are sort of, shall we say, a pretty small minority.

I'm sorry, and I know that sometimes the small minority are ahead of the curve and sometimes get things right. It's just that I don't see you bunch as innovators or discoverers on the level of Newton, Darwin, Einstein or some others. I see you more as flat earthers.

You could be right but it's not right for me to put any hope into it with you.

Oppo being the exception as he may discover a new way of winning at video gaming and become famous in his circles!
  
Back to top
 
IP Logged
 
SnarkySack
Libertarian Freedom Member
*****
Offline

Libertarian's Forum

Posts: 3148
Location: Republic of Me
Joined: Sep 11th, 2017
Re: The Replication Crises
Reply #9 - Apr 2nd, 2018 at 9:07pm
Print Post  
Jeff wrote on Apr 2nd, 2018 at 5:59pm:
It's important to science, but perhaps even more important to political economy, where results that have not been honestly peer reviewed are presented as "scientific fact" and governments base policies on the un-reviewed or poorly reviewed "discoveries".

Any honest competent scientist will replicate their own experiments before they publish, and maybe ask colleagues to look over their hypothesis and procedures to see if they did something obviously stupid.


Yes, but many scientists and researchers are not doing that.  Especially in the social sciences.  The "Replication Crises" isn't my phrase.  It is a problem of concern to those who believe science should be scientific.

Quote:
Unless of course, as you say, the government granted them lots of money and told them what results were required...

That's fraud and should be prosecuted.


Yup.

  

I used to be burnsred . . .
Back to top
 
IP Logged
 
Page Index Toggle Pages: [1] 2 3 ... 84
Send TopicPrint
 
Libertarian's ForumLibertarian's ForumFreedom Forum › The Replication Crises
Libertarian's Forum

Libertarian's Forum Information Rules, Agreement and Privacy Policy