W3 - Power & Post-hoc

Question 1

A significant Levene’s test doesn’t affect F score as long as

Answer 1

Levene’s F is less than 4

Question 2

Examine normal distribution by

Answer 2

Checking plot

Question 3

What do we check before analysing with ANOVA?

Answer 3

Homogeneity of variance, normal distribution, independence of observations

Question 4

How do we check independence of observations?

Answer 4

We investigate the research design

Question 5

Power is an estimate of

Answer 5

Correctly rejecting the null hypothesis

Question 6

Estimating how many participants are required before a test is called

Answer 6

‘a priori’ power analysis

Question 7

We use Tukey’s test and REGWQ when

Answer 7

Equal variances are assumed (sample size and homogeneity of variance)

Question 8

Why is REGWQ superior to Tukey’s test?

Answer 8

Both have type I error adjustment built in but REGWQ holds superior control over power

Question 9

When equal variances are not assumed, we use

Answer 9

Games-Howell

Question 10

What does it mean if Levene’s test is significant?

Answer 10

Homogeneity of variance is violated

Question 11

Which tests can be used to estimate power?

Answer 11

Cohen’s d and f

Question 12

Estimating power reflects

Answer 12

The non-centrality parameter