Science
Related: About this forumCan someone help with a statistical problem?
Ok, I understand stats at a gut level but am totally untrained as to performing equations to solve practical problems. I suspect this is easy and I'm asking for a stats DUer to help
The problem:
! have 100 widgets. All of the widgets are either red or blue.
I select 3 widgets randomly. All of the widgets I selected are blue.
What is the implied percentage of blue widgets out of the 100?
Hopefully, I have stated a problem that can be solved with the information presented
DetlefK
(16,423 posts)This number is only true with a certain probability ("confidence" . The actual value lies somewhere in a confidence-interval around those "100%".
As your sample-size increases to encompass more of the whole set, the confidence-interval shrinks until only the true value is left.
EDITED
Is it possible to state a "margin of error" for this scenario?
My example mirrors a real-life situation and so I cannot have more than 3 in my initial sample.
DetlefK
(16,423 posts)That's really difficult.
http://en.wikipedia.org/wiki/Confidence_interval
http://en.wikipedia.org/wiki/Sample_mean_and_sample_covariance
I have a book about statistics lying around somewhere. If I find an easy way to calculate this, I will come back to you later.
sendero
(28,552 posts)... the answer I have is good enough! Thanks again.
StatGirl
(518 posts)The true proportion of the blue ones could be as high as 100%, but the result that you got is not very likely if the proportion is 37% or less.