Week 11 Lecture 11 - non-parametric tests

Question 1

What do NP stats not make assumptions about?

Answer 1

underlying population distributions

“distribution free stats”

Question 2

Do NP stats have more or less power than parametric equivalents?

Answer 2

• less
• higher risk of Type 2 error

Question 3

Should NP stats be used first or as a backup?

Answer 3

backup

Question 4

What can the normality assumption be assessed statistically using?

Answer 4

Shapiro-Wilk test

Question 5

What is the NP equivalent of a b-p t-test?

Answer 5

Mann-Whitney U test

Question 6

What is the NP equivalent of w-p t-test?

Answer 6

Wilcoxon t-test

Question 7

What is the NP equivalent of 1-way independent ANOVA?

Answer 7

Kruskal Wallis test

Question 8

What is the NP equivalent of 1-way RM ANOVA?

Answer 8

Friedman test

Question 9

What is the NP equivalent of factorial ANOVAS?

Answer 9

no NP equivalent

Question 10

What is the NP equivalent of Pearson’s correlation coefficient?

Answer 10

• Spearman’s rho (where N>20)
• Kendall’s Tau (where N<20)

Question 11

What is the NP equivalent of partial correlation?

Answer 11

no NP equivalent

Question 12

What is the NP equivalent of regression

Answer 12

no NP equivalent

Question 13

When is it best to use NP tests for tests of relationship ?

Answer 13

if either variable is measured on an ordinal scale (especially if concerned that the intervals between measures aren’t equal)

Question 14

What are 2 tests for analysis of categorical data?

Answer 14

• 1 variable Chi-square (AKA goodness of fit test)
• Chi-square test of independence (2 variables)

Question 15

Is there a parametric equivalent to the 1 variable chi-square or the chi-square tests of independence?

Answer 15

• no
• DV (outcome variable) is measured on a categorical scale so NP tests are the only option