significance and power

Question 1

What is H0?

Answer 1

-Null hypothesis
-Have to assume that this is true

Question 2

Define p-value

Answer 2

-Probability of finding a large effect assuming that the null hypothesis is true

Question 3

What are some issues with NHST?

Answer 3

-P values don’t look at chance
-P values aren’t good enough alone

Question 4

Define a type 1 error

Answer 4

-Where you state you have found n effect even though you haven’t

Question 5

Define type 2 error

Answer 5

-Where you state you haven’t found an effect even though you have

Question 6

Define power

Answer 6

-Probability of finding an effect assuming that one exists within the population
-Calculated by 1-B

Question 7

What is B?

Answer 7

-Probability of not finding an effect
-Typically 0.2 (Cohen)

Question 8

What are the 3 things that affect power?

Answer 8

1: Effect size = magnitude of the effect
2: Num. of ppts = if theres a large effect size then you need less ppts to show a ‘real’ effect
3: Alpha = probability of finding a type 1 error

Question 9

What is a family wise experimental error rate?

Answer 9

-Running multiple tests
-Increases likelihood of type 1 error