Register | Sign In


Understanding through Discussion


EvC Forum active members: 65 (9162 total)
8 online now:
Newest Member: popoi
Post Volume: Total: 915,815 Year: 3,072/9,624 Month: 917/1,588 Week: 100/223 Day: 11/17 Hour: 7/1


Thread  Details

Email This Thread
Newer Topic | Older Topic
  
Author Topic:   More Bunk Science
Wounded King
Member
Posts: 4149
From: Cincinnati, Ohio, USA
Joined: 04-09-2003


Message 61 of 64 (630581)
08-26-2011 11:03 AM
Reply to: Message 60 by Bolder-dash
08-26-2011 10:17 AM


Re: The search for the great bin of useless studies
I would disagree with you that this study is large enough to consider the small differences they noted as being that significant
Well duh! You have said so several times, you just don't seem to have any actual rationale for saying it, beyond your own prejudices.
but actually I was not referring to the p value when I said the two groups were statistically almost identical, I was referring to the calculations for FA. it seems to me the two groups could well overlap with those margins for error.
Clearly your ability to understand statistics is non-existent. The whole point is that the p-value shows us that despite the overlap in the raw measurements, in terms of standard deviation, the populations are still significantly different in their means at the P<0.05 level. What margins of error are you talking about if not the standard deviations from table 1? Why do you choose to ignore the fact that the standard error is considerably smaller than that.
Is playing a game where you have to choose which is the best way to make points really a test of co-operation? You seem to have completely skipped this fact.
Because this fact is incidental to whether or not this paper is bunk science as you claimed. If they have showed a significant association but you disagree with the interpretation then fine. You want to put it down to intelligence and claim it doesn't really reflect cooperativeness, I'm not much bothered. These elements are interpretation and discussion of the results, they aren't the results themselves. There is even research out there that would support your view since FA has also been shown to negatively correlate with some measures of intelligence (Burlow et al 1997). I haven't really found any equivalent studies looking at whether IQ correlates with defect/cooperate behaviour in the prisoner's dilemma. The only one I have come across looked at groups playing the game over a repeated series of games rather than individuals in a one off situation, in that case they found that groups of higher intelligence seemed to cooperate more (Jones, 2008).
That is a fictitious scenario created only in the minds of the authors.
Except that it wasn't, it was created more than 50 years ago by game theorists and is one of the most studied game theory scenarios.
And by the way, I thought scientists had the utmost integrity, so why would there be a bias towards only publishing studies that support their assertions? Naughty naughty.
It may shock you to learn this but scientists don't get to dictate when and where their research gets published. They have to get it through peer review and past journal editors. There are a variety of factors which have lead to what most people would agree is an invidious situation, a large part of which is that scientific publishing is by and large a commercial enterprise.
I want to read those 20 other studies they did that showed no correlation whatsoever between FA and parlor games.
And I want you to stop making things up, but I guess you can't always get what you want.
TTFN,
WK
Edited by Wounded King, : No reason given.

This message is a reply to:
 Message 60 by Bolder-dash, posted 08-26-2011 10:17 AM Bolder-dash has replied

Replies to this message:
 Message 62 by Bolder-dash, posted 08-26-2011 12:09 PM Wounded King has replied

  
Bolder-dash
Member (Idle past 3630 days)
Posts: 983
From: China
Joined: 11-14-2009


Message 62 of 64 (630582)
08-26-2011 12:09 PM
Reply to: Message 61 by Wounded King
08-26-2011 11:03 AM


Re: The search for the great bin of useless studies
Wow, talk about pedantic. Of course I was calling into question the validity of the research conclusions, not simply how they got their numbers. What makes it bunk science is not what decimal point they used-this is a study of evolutionary psychology, not the weight of a bag of beans.
The study is designed to show a correlation between how you act in public, and the size of your ears. Its does no such thing in my opinion. And all this study really is about is opinion. The authors opinion that this factor played a more important role than any other of one thousand factors.
And regarding those numbers that you love so much that they block your view of the forest, there is no way that you could tell me that if they played this game with 10,000 participants instead of 100 that the numbers couldn't just as easily flip and go the other direction. There is no such significant trend here no matter what p value you use. If they would like to play a game where they predict what someone will decide beforehand, based on the shape of their two ears, I will take them on in a betting game anytime those authors wish.

This message is a reply to:
 Message 61 by Wounded King, posted 08-26-2011 11:03 AM Wounded King has replied

Replies to this message:
 Message 63 by Wounded King, posted 08-26-2011 12:52 PM Bolder-dash has not replied
 Message 64 by Taq, posted 08-26-2011 1:02 PM Bolder-dash has not replied

  
Wounded King
Member
Posts: 4149
From: Cincinnati, Ohio, USA
Joined: 04-09-2003


Message 63 of 64 (630586)
08-26-2011 12:52 PM
Reply to: Message 62 by Bolder-dash
08-26-2011 12:09 PM


Re: The search for the great bin of useless studies
The study is designed to show a correlation between how you act in public, and the size of your ears.
No it isn't.
there is no way that you could tell me that if they played this game with 10,000 participants instead of 100 that the numbers couldn't just as easily flip and go the other direction.
Yes there is, because that is what the data strongly suggests. The chances are pretty darn high that you would see significant differences in the means in the same way from a larger population. There is a small chance that the trend might disappear or reverse, but the whole point of calculating the statistics and specifically the p-value, is to give us a measure of our confidence that it won't. You don't seem to understand what statistics is for.
TTFN,
WK
Edited by Wounded King, : No reason given.

This message is a reply to:
 Message 62 by Bolder-dash, posted 08-26-2011 12:09 PM Bolder-dash has not replied

  
Taq
Member
Posts: 9972
Joined: 03-06-2009
Member Rating: 5.5


Message 64 of 64 (630588)
08-26-2011 1:02 PM
Reply to: Message 62 by Bolder-dash
08-26-2011 12:09 PM


Re: The search for the great bin of useless studies
And regarding those numbers that you love so much that they block your view of the forest, there is no way that you could tell me that if they played this game with 10,000 participants instead of 100 that the numbers couldn't just as easily flip and go the other direction.
Do you even understand how statistical analyses work? The 0.015 p-value tells us that there is a 1.5% chance that the null hypothesis is correct. There is a 98.5% chance that the hypothesis is correct. That seems pretty good to me. Why do you have a problem with this?
There is no such significant trend here no matter what p value you use.
So what statistical tests should be used to determine a statistically signficant trend? What analyses should they have used instead of a t-test? Should they have used a Chi-squared test?

This message is a reply to:
 Message 62 by Bolder-dash, posted 08-26-2011 12:09 PM Bolder-dash has not replied

  
Newer Topic | Older Topic
Jump to:


Copyright 2001-2023 by EvC Forum, All Rights Reserved

™ Version 4.2
Innovative software from Qwixotic © 2024