S2 lecture 6 - within-subject designs

Question 1

What are the two significant between-participant inferential tests we have studied?

Answer 1

Independent t-test (parametric) and Mann-Whitney U test (nonparametric).

Question 2

Non-parametric Mann-Whitney U tests are less powerful than parametric t-tests so there is a greater chance for…

Answer 2

type II error

Question 3

How might samples differ in between-subject designs compared to within-subject designs?

Answer 3

Within-subject design samples can be smaller because each participant can give you more than one data point, and these designs are more powerful for their sample size because you are more likely to detect an effect if there is one.

Question 4

What are some of the issues that may arise in within-subject designs?

Answer 4

order effects, practice effects, and fatigue effects.

Question 5

What are the two significant within-participant inferential tests we have studied?

Answer 5

paired sample t-tests(parametric) and Wilcoxon signed-rank test(nonparametric).

Question 6

What are the assumptions for paired sample t-tests?

Answer 6

interval/ratio DV, independent data within conditions, normal sampling distribution in each condition. (Note that unlike independent t-tests, paired sample and one sample t-tests do not require homogeneity of variance.)

Question 7

What is type II error?

Answer 7

Missing an effect that is actually there.