W7

Question 1

What are some characteristics of non-normal data distributions?

Answer 1

Non-normal data distributions may exhibit features such as bimodal distribution or a decaying distribution.

Question 2

What does the Shapiro-Wilk W metric indicate?

Answer 2

It indicates how “normal” the data is. Higher values of Shapiro-Wilk W indicate data that is closer to a normal distribution.

Question 3

What does the Shapiro-Wilk p-value indicate?

Answer 3

It indicates the probability of any observed difference from normality being significant.

Question 4

What non-parametric test can replace one-sample and paired sample t-tests?

Answer 4

The Wilcoxon Signed Rank Test.

Question 5

What non-parametric test can replace independent samples t-tests?

Answer 5

The Mann-Whitney U Test.

Question 6

What are the issues with normality testing and sample size?

Answer 6

In small samples, even large departures from normality can appear to pass the assumption of normality.

Question 7

When should you run rank-based non-parametric tests?

Answer 7

Rank-based non-parametric tests are valid for both normal and non-normal data sets. They are also valid for ordinal, interval, and ratio data, and are preferred when comparing medians rather than means.

Question 8

Wilcoxon’s W=?

Answer 8

Min(pos_rank, neg_rank)

Question 9

Which non-parametric test should you use for a within subjects design?

Answer 9

Wilcoxon Signed Rank Test.

Question 10

Which non-parametric test should you use for a between subjects design?

Answer 10

Mann-Whitney U Test.