How do you conduct a t- test for a set of matching pairs to find the t statistic, p value, and significance level? The matching pairs part is what's throwing me off.

Question

amistre64 · Answer

oy this sounds familiar ....

amistre64 · Answer

a set of matching pairs, what is that?

anonymous · Answer

For this example they're comparing two different drugs and the results.

amistre64 · Answer

http://stattrek.com/statistics/dictionary.aspx?definition=matched-pairs_t-test im reading this to refresh anyting i might have lost thru the years

amistre64 · Answer

whats our x1 and x2 ?

anonymous · Answer

this might be easier.

amistre64 · Answer

you lied!!  that is definantly NOT easier :)

anonymous · Answer

haha, I meant easier to see it for me to describe...lol I have the values i L1 and L2 in the calculator, but not sure where to go from there.

amistre64 · Answer

what are the numbers in the reference and generic columns relate to?

anonymous · Answer

Those are the results from the different drugs. So I'm guessing those would be the x1 and x2?

amistre64 · Answer

so each group received the generic and the reference drugs; but at different times and such that yada yada ....

anonymous · Answer

exactly

amistre64 · Answer

but what do the numbers themselves mean?

anonymous · Answer

The absorption of the drug in the blood

amistre64 · Answer

absoprtion extent = 4567  ....

amistre64 · Answer

so they are differing absorption thingys

anonymous · Answer

right.

amistre64 · Answer

and youve created 2 separate lists on your ti83 or similar contraption so far

anonymous · Answer

so far yes. I feel like I need to subtract one from the other though. I'm just not sure on that.

amistre64 · Answer

yes, L1-L2 will give us the required x1-x2 parts to distribute into a curve

amistre64 · Answer

L1-L2 = d

then the mean of d and its associated sds and variances and curvilinear contraptions can be assessed

anonymous · Answer

alright. do I need to subtract them separately? or could I do the total of L1-total of L2?

amistre64 · Answer

you want to so:

sub1 : L1 - L2 = d1
sub2 : L1 - L2 = d2
etc .....

this way you have a running tabulation of effects for each random variable presented

amistre64 · Answer

so separatley is what ive got in mind

anonymous · Answer

okay. Let me do that real quick.

anonymous · Answer

whew. okay, got that done.

amistre64 · Answer

now with the new stuff in "d" as im calling it; we can construct a distribution and ints graph to perform calculations on

anonymous · Answer

ok, so put the new list in the calc. then go from there?

amistre64 · Answer

yes, then we can test hypothesises and the normal day to day unpaired situations

anonymous · Answer

Okay, I have it in. So I guess I need to find the  μ ?

amistre64 · Answer

yes

amistre64 · Answer

http://stattrek.com/hypothesis-test/paired-means.aspx this page actually has a good working example that you can follow that parallels yours i believe

anonymous · Answer

Great. Thank you!

amistre64 · Answer

yw, and good luck with it :)

anonymous · Answer

Lol I think I got this one. We'll see about the rest of the course. :p