Lecture 8

Question 1

What parametric test is used for unpaired data?

Answer 1

Independent Samples t-Test

Question 2

What parametric test is used for paired data?

Answer 2

Paired-Samples t-Test

Question 3

What non-parametric test is used for unpaired data?

Answer 3

Mann-Whitney U test (or Wilcoxon rank-sum test)

Question 4

What non-parametric test is used for paired data?

Answer 4

Wilcoxon Signed Ranks test (also Sign test)

Question 5

What are advantages of non-parametric methods?

Answer 5

1. Can be applied to a wide variety of situations because they do not make rigid assumptions about populations
2. Can be applied to non numerical data (ie. ranks or other data at the ordinal level of measurement)
3. Usually involve simpler computations than the corresponding parametric methods and are easier to understand

Question 6

What are disadvantages of non-parametric methods?

Answer 6

1. Tend to waste information because exact numerical data are often reduced to a qualitative form (ie. ranks)
2. Not as efficient as parametric tests, so with non-parametric tests we generally need stronger evidence before we can reject a null hypothesis