Significance Tests & P Values

Question 1

Why do we need sig tests?

Answer 1

• diff estimates give diff answers

- want to make sure results not due to chance or over/underestimate of true answer

Question 2

What is null hypothesis?

Answer 2

• starting assumption about pop of interest

Question 3

What form is null hypo usually in?

Answer 3

• there is no real diff/ass

Question 4

What is P value?

Answer 4

• provides way of weighing ev against null hypo in order to reject it/not
• probab of obtaining sample data/diff as extreme/more extreme/large/larger than that observed if null hypo true
• larger diff observed in sample, less likely it could have come from pop where no diff exists

Question 5

What does P value result from?

Answer 5

• results from conducting stat sig test e.g. t-test/chi squared test

Question 6

What happens to P value if larger diff observed

+ vice versa?

Answer 6

• get smaller P value - smaller probab sample could have come from pop where null hypo is true

Question 7

What value is used as a cut-off for determining stat sig?

Answer 7

0.05

Question 8

How do you interpret P<0.05?

Answer 8

• good ev to reject null hypo
• conclude real diff/ass does exist in pop of interest
• stat sig result