Poll: Read the OP First
Biased
Scientist
[Show Results]
 
An Experiment: Requires A Friend To Help!
Post Reply
 
Thread Rating:
  • 0 Votes - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
09-07-2013, 09:01 PM
Question An Experiment: Requires A Friend To Help!
So here's the experiment: your friend will check the "spoiler" below. Then you'll be given 3 numbers that follow a certain rule. To figure out what that rule is, give your friend a series of 3 numbers back and he or she will tell you if they follow that rule with a simple "yes" or "no". Do this as many times as you like. Once you think that you know the rule, state it for your friend and he or she will tell you if you're right. Keep giving your friend sets of 3 numbers and/or guesses at the rule until you guess it correctly.

The example series of numbers to say out loud is 2, 4, 6. The rule for numbers (that you won't tell your friend until they guess it right) is that they are any three numbers in ascending order -- that is, they go from smallest to largest.

After doing the test with a friend:

This is a test for confirmation bias. Confirmation bias is a cognitive bias that drives you to seek only evidence that confirms what you already know or believe. When you guessed the rule for the first time, was it after getting a no from the friend who knew the rule or did you only seek yeses? Before being told "no", the rule could have been "any set of 3 numbers" for all you knew. The scientific method rules out confirmation bias and specifically seeks disconfirming evidence. A biased person doesn't look for disconfirming evidence, but a scientist does... so if you guessed the rule before you were told "no" at least once, you should answer "biased" above. Otherwise, chalk it up as "scientist". You'll note that I already answered for myself -- I failed this test the first time I took it, too Wink




My girlfriend is mad at me. Perhaps I shouldn't have tried cooking a stick in her non-stick pan.
Visit this user's website Find all posts by this user
Like Post Quote this message in a reply
09-07-2013, 09:51 PM
RE: An Experiment: Requires A Friend To Help!
what if I guessed the rule at the first attempt and so I never get a no for an answer? Am I still biased?

[Image: sigvacachica.png]
Find all posts by this user
Like Post Quote this message in a reply
10-07-2013, 05:31 AM
RE: An Experiment: Requires A Friend To Help!
(09-07-2013 09:51 PM)nach_in Wrote:  what if I guessed the rule at the first attempt and so I never get a no for an answer? Am I still biased?

Yeah. As said above, there's no way to know the rule isn't simply "any 3 numbers" until you at least narrow it down with a single disconfirming guess.

My girlfriend is mad at me. Perhaps I shouldn't have tried cooking a stick in her non-stick pan.
Visit this user's website Find all posts by this user
Like Post Quote this message in a reply
10-07-2013, 10:51 AM
RE: An Experiment: Requires A Friend To Help!
What if "no" is the first answer? Then there is no chance to not seek disconfirming evidence. But if "yes" is the first answer, then it has the problem that nach_in mentioned. Am I missing something? I'm not seeing how this works.

"Religion has caused more misery to all of mankind in every stage of human history than any other single idea." --Madalyn Murray O'Hair
Find all posts by this user
Like Post Quote this message in a reply
10-07-2013, 12:05 PM
RE: An Experiment: Requires A Friend To Help!
Oh I get it, if I guess the rule correctly the first time, then I should guess numbers against my rule to make sure it works by failing Tongue

[Image: sigvacachica.png]
Find all posts by this user
Like Post Quote this message in a reply
10-07-2013, 11:20 PM
RE: An Experiment: Requires A Friend To Help!
(10-07-2013 10:51 AM)Impulse Wrote:  What if "no" is the first answer? Then there is no chance to not seek disconfirming evidence. But if "yes" is the first answer, then it has the problem that nach_in mentioned. Am I missing something? I'm not seeing how this works.

This experiment has already been performed before elsewhere. Typically someone gets a theory about the rule, such as "they are consecutive even numbers" and then produces a confirming set of numbers such as "4, 6, 8" or "112, 114, 116". After getting confirmation that they fit (a "yes"), the subject will typically guess the rule. After being told they got it wrong, they'll often for a new theory (the number are consecutive sets of 2) and guess another confirming set like "3, 5, 7" before guessing the rule yet again. It is unusual for someone who is trying to guess the rule to mess up so badly as to get a "no" before at least one "yes", because that seems to reflect a lack of pattern recognition.

The reason that seeking a "no" answer is scientific is because the set of numbers that doesn't fit the rule is a control group, and confirmation bias drives us to avoid using a control group in our tests. We naturally want to feel "right", and we don't want anyone to tell us that we're wrong, but we have to be wrong from time to time in order to even know what right is. You can tell what group you're in by how you approach the test... it's not really about whether you get yeses or nos but rather if you're even trying to get both before you assume that you understand the pattern.

My girlfriend is mad at me. Perhaps I shouldn't have tried cooking a stick in her non-stick pan.
Visit this user's website Find all posts by this user
Like Post Quote this message in a reply
10-07-2013, 11:27 PM
RE: An Experiment: Requires A Friend To Help!
I'd say that there's nothing better to satisfy our confirmation bias than to have a solid control group, that way you'll be even more sure that you got it right Big Grin

[Image: sigvacachica.png]
Find all posts by this user
Like Post Quote this message in a reply
[+] 1 user Likes nach_in's post
10-07-2013, 11:27 PM
RE: An Experiment: Requires A Friend To Help!
The thing is that prayer is often ALWAYS self confirming. If you pray and get your way then god had heard you. If you do not get your way then it is simply god's will. So then, if I fail, I have not really failed, I have simply been redirected by god. Which is itself just as confirming as a win.

"Your mind is twice a valuable as your body. And your ears are twice as valuable as your mouth. People will pay you based on which you use." - A very smart old lawyer
Find all posts by this user
Like Post Quote this message in a reply
[+] 1 user Likes PatThePoltergeist's post
11-07-2013, 05:42 AM
RE: An Experiment: Requires A Friend To Help!
(10-07-2013 11:27 PM)PatThePoltergeist Wrote:  The thing is that prayer is often ALWAYS self confirming. If you pray and get your way then god had heard you. If you do not get your way then it is simply god's will. So then, if I fail, I have not really failed, I have simply been redirected by god. Which is itself just as confirming as a win.

Yeah. In fact, I can say that I "not prayed" for a car this week, and God answered my lack of prayer with a lack of a car. I also "not prayed" for money today, and God answered my lack of prayer by giving me money through my regular income. It's amazing how God answers my "not prayers" by either giving me what I didn't ask for or not giving it to me, and this "seek results and then fit them to an assumed pattern" method works with the same evidence that Christians give for prayer.

My girlfriend is mad at me. Perhaps I shouldn't have tried cooking a stick in her non-stick pan.
Visit this user's website Find all posts by this user
Like Post Quote this message in a reply
[+] 1 user Likes Starcrash's post
11-07-2013, 09:18 AM
RE: An Experiment: Requires A Friend To Help!
(10-07-2013 11:20 PM)Starcrash Wrote:  
(10-07-2013 10:51 AM)Impulse Wrote:  What if "no" is the first answer? Then there is no chance to not seek disconfirming evidence. But if "yes" is the first answer, then it has the problem that nach_in mentioned. Am I missing something? I'm not seeing how this works.

This experiment has already been performed before elsewhere. Typically someone gets a theory about the rule, such as "they are consecutive even numbers" and then produces a confirming set of numbers such as "4, 6, 8" or "112, 114, 116". After getting confirmation that they fit (a "yes"), the subject will typically guess the rule. After being told they got it wrong, they'll often for a new theory (the number are consecutive sets of 2) and guess another confirming set like "3, 5, 7" before guessing the rule yet again. It is unusual for someone who is trying to guess the rule to mess up so badly as to get a "no" before at least one "yes", because that seems to reflect a lack of pattern recognition.

The reason that seeking a "no" answer is scientific is because the set of numbers that doesn't fit the rule is a control group, and confirmation bias drives us to avoid using a control group in our tests. We naturally want to feel "right", and we don't want anyone to tell us that we're wrong, but we have to be wrong from time to time in order to even know what right is. You can tell what group you're in by how you approach the test... it's not really about whether you get yeses or nos but rather if you're even trying to get both before you assume that you understand the pattern.
Oh duh, apparently my brain wasn't working very well yesterday. I was thinking the person was guessing the rule every time (so "yes" would mean the rule was guessed). I see now they are submitting 3 more numbers where "yes" means the numbers follow the rule. I have no clue why I missed that yesterday since you wrote it quite clearly (but I'll blame it on old age creeping up Big Grin ).

"Religion has caused more misery to all of mankind in every stage of human history than any other single idea." --Madalyn Murray O'Hair
Find all posts by this user
Like Post Quote this message in a reply
Post Reply
Forum Jump: