#1

ok ive got a quite simple question,

I'm doing my biology extended essay (woop!) and I'm conducting an unequal variance, unequal sample t-test.

Ive got two variables from each site (surface area and polyp density) and two sites.

I've done an F-test on the samples and returned p-values of

0.24 and 0.09

are these enough to back up my decision to do an unequal variance t-test rather than a Student's t-test.

Thanks guys

I'm doing my biology extended essay (woop!) and I'm conducting an unequal variance, unequal sample t-test.

Ive got two variables from each site (surface area and polyp density) and two sites.

I've done an F-test on the samples and returned p-values of

0.24 and 0.09

are these enough to back up my decision to do an unequal variance t-test rather than a Student's t-test.

Thanks guys

#2

You're not getting much help ts...I think you expect a little much from the pit...

#3

You're not getting much help ts...I think you expect a little much from the pit...

This man is right, youve come here with something needing above a high school education

#4

aaw come on I've asked herder maths questions than this before....

#5

They sound different enough.

however whats a student t-test? or a t-test?

I do further maths and i can;t understand this jargon

however whats a student t-test? or a t-test?

I do further maths and i can;t understand this jargon

#6

You're not getting much help ts...I think you expect a little much from the pit...

You'd be surprised, a lot of UGers can probably come up with the answer.

Me personally, well no.

Is this more of a statistics question? Cause I never heard of this in calc.

#7

the answer is:

**maybe**
#8

erm im doing higher maths for IB, i think the A-level stats module is next year? actually I think \m/[-_-]\m/ is right. I'll probably write down the results of the F-test and say that the probability of observing the results with the null hypothesis is blah blah...

EDIT: student's t-test is the run of the mill t-test, equal variance, but its unreliable for populations with underlying different variances.

The unequal sample, unequal variance t-test is called Welch's t-test I think and its slightly more complicated, but quite unknown in reseach communities, hence why I'm using it !

help is still needed for interpreting th p-value people!

EDIT: student's t-test is the run of the mill t-test, equal variance, but its unreliable for populations with underlying different variances.

The unequal sample, unequal variance t-test is called Welch's t-test I think and its slightly more complicated, but quite unknown in reseach communities, hence why I'm using it !

help is still needed for interpreting th p-value people!

*Last edited by YngwieJnr at Aug 17, 2009,*

#9

I say pick B. when in doubt always pick B.

#10

i've done A level biology but don't have a clue about this problem.

Don't you need more values than that to prove that whatever it is unequal?

Don't you need more values than that to prove that whatever it is unequal?

#11

i've done A level biology but don't have a clue about this problem.

Don't you need more values than that to prove that whatever it is unequal?

well technically yes, but the t-test is all about probabilities. So the p-value (in a t-test) gives the probability that the observed values are observed differently due to chance given the null hypothesis is correct. TYhe nulll hypothesis states both populations are correct. if the p-alue is small enough then you can technically disregard the null hypothesis because the probability of it being correct is so small.

If your sample is very small and the difference between the populations isnt very much, your p-value will show this.

The main issue is with the F-test, as I don't know the level of significance i need.

*Last edited by YngwieJnr at Aug 17, 2009,*