The Student Room Group

T-Test Statistics Question - Help Please?

Hey Guys;

Well, I have an exam on Wednesday and kind of find myself stuck on this type of question which MIGHT or MIGHT NOT come up, so, if someone wouldn't mind answering question 2, I would be grateful. =)

Is a Statistics Year 2 Module for Biomedical Science.

The question is the entire 2nd page and contains about 10 marks.

Thanks. =)
(edited 7 years ago)
Original post by Scienceisgood
Hey Guys;

Well, I have an exam on Wednesday and kind of find myself stuck on this type of question which MIGHT or MIGHT NOT come up, so, if someone wouldn't mind answering question 2, I would be grateful. =)

Is a Statistics Year 2 Module for Biomedical Science.

The question is the entire 2nd page and contains about 10 marks.

Thanks. =)


Well, tell us what your thoughts are about the answers. To get you started, notice the phrasing of the question: "Nine samples of the cereal were used. Each was divided into two parts" - what feature of the test is this crying out for?
Original post by Gregorius
Well, tell us what your thoughts are about the answers. To get you started, notice the phrasing of the question: "Nine samples of the cereal were used. Each was divided into two parts" - what feature of the test is this crying out for?


Tempted to say paired T-Test because there are two possible methods and the P Values and CI have changed?
Am I right?

EDIT:
Put two sample by mistake...
(edited 7 years ago)
Original post by Scienceisgood
Tempted to say paired T-Test because there are two possible methods and the P Values and CI have changed?
Am I right?



It is a paired t-test - because the samples are naturally paired.
Original post by Gregorius
It is a paired t-test - because the samples are naturally paired.


So, would the answer for B (assumptions of the data) be things like, assume a normal distribution at a 95% Significance Level?
(edited 7 years ago)
Original post by Scienceisgood
So, would the answer for B (assumptions of the data) be things like, assume a normal distribution at a 95% Significance Level?


Normal distribution is one of the assumptions. Significance level isn't - that's imposed upon the test itself, not the data. The other big assumptions are that the samples have to be independent of each other, and have to be random samples from the underlying population.

Quick Reply