send me a pm

I have 2 questions from my sister's study guide that I cant answer

Answer and explain why and i will paypal you $10

plz o do by tomorrow @ 12

her exam is at one


1.

Your teacher has decided to put 30 subjects in group one and 90 subjects in group 2 for her dissertation. Assuming the standard deviations are approx the same in both groups, what would be the least # of subjects in each group that you could use and get a standard error of the statistic |x1 - |x2 that is equal to your teacher's standard error?

12/5/2006 1:27:09 AM

$40

12/5/2006 1:37:35 AM

tough one. have you tried emailing anyone in the stats department? reich@stat.ncsu.edu . id tell him you have a test coming up and would like some insight into how to solve this problem.


worst hes gonna say is "no".

12/5/2006 6:37:37 PM

you should have started here and then moved on to the other forums

and the answer is 15

[Edited on December 5, 2006 at 7:33 PM. Reason : not really, i have no idea]

12/5/2006 7:33:02 PM

To get the same standard error as the teacher if she did the group 1 average minus the group 2 average?

...

all of them?

12/5/2006 8:22:03 PM

yea what is the smallest sample size(# of subjects) you could put in each group to have the same as your teacher's

12/5/2006 8:29:16 PM

give google a few more years and they will be able to help with these type issues

12/5/2006 9:00:30 PM

i mean i have even a hard time finding a stats messageboard

every google search I get is the statistic for some random messageboard

12/5/2006 9:03:58 PM

I'd love to help, but the way the question is worded, I'm not exactly sure what to do. Is there a reason why this question is so important to answer?

I'd try Yahoo Answers and/or Google Answers though.

12/5/2006 9:21:37 PM

its the one type of question we dont know how to answer

we've worked through most of the rest

12/5/2006 9:26:12 PM

If the question is saying that you should form your own groups of optimal size to get the same standard error when you take the difference, I'll go with what graywolf said in the other thread.

sqrt((s/30+(s/90))=sqrt((s/45)+(s/45))

(s is the standard deviation, but they're all the same and cancel out to one)

45

[Edited on December 5, 2006 at 9:40 PM. Reason : actually it should be s^2 inside the sqrt's, but they still cancel out.]

12/5/2006 9:32:16 PM