T-tests

Question 1

T-tests are used

Answer 1

to test whether the means in two groups are diff from each other

Question 2

Assumptions of T-tests

Answer 2

• data in each group is normally distributed
• SD os constant across groups
• data points are independent of each other

Question 3

what sort of data

Answer 3

continous

inferential

Question 4

unequal variance

Answer 4

a version of t-test which assumes unequal variance, instead of equal
-for example if we collect data and the SD is different for both groups, then we would use this test

Question 5

paired t-test

Answer 5

paired data e.g. blood pressure before and after taking statins for the same people. These data are no longer independent and therefore we use a paired t-test

Question 6

three types of t-test

Answer 6

• unpaired assuming unequal variance
• unpaired data assuming equal variance
• paired t-test

Question 7

t-test example

Answer 7

-we want to know if statins lower bp
1) define population- all those at risk of CV disease
2)study and hypothesis: recruit 500 people at risk of CV event and give half statins, then measure blood pressure. We hfypoheise that those who have taken statins will have a lower BP
3) null hypothesis: statins do not lower bp
4)analyse data:
No statins:174
statins: 154.

Here the difference between means is 20. Our null hypothesis is a difference of 0. When a T-test is applied p=0.02, therefore we have good evidence to show that statins do lower BP