WEEK 10

Question 1

Parametric test assumptions

Answer 1

-underlying probability distributions
-DV measured at interval or ratio level
- no outliers
-homogeneity
-linearity

Question 2

what happens when assumptions are not met?

Answer 2

• data may be skewed
  -data may not be at the required level
  -there may be outliers
  -the sample size may be small
  -there may be groups of unequal sample sizes if using groups

Question 3

Parametric tests guidelines

Answer 3

Independent T test: N>26 per group
Related T test: N>15
Pearson’s correlation: N>28

Question 4

Non- Parametric Tests

Answer 4

Mann-Whitney U: Alternative to the independent samples T test
Wilcoxn: Alternative to the related samples test
Spearman’s: Alternative to the Pearson’s product moment correlation

Question 5

Spearman’s Rho

Answer 5

-non-normal distribution of data
-Small sample size
-Outliers
-Interval, ratio or ordinal data

Question 6

Spearman’s vs Pearson’s

Answer 6

-Pearson’s r assesses the linear relationship
-Spearman’s assesses only a monotonic relationship
-If similar results and near normal use Pearson
-Non-normal use Spearman

Question 7

Mann-Whitney U test

Answer 7

-alternative to the independent samples t test
-when we have two unpaired groups
- use when data is non-normally distributed
-Small sample size
-Interval, ratio or ordinal data

Question 8

effect size for mann whitney U

Answer 8

• convert the z score into effect size, r
  r= z divided by the square root of N
  z is from output
  N is the total number of observations

Question 9

Wilcoxn signed rank test

Answer 9

• non parametric alternative to the paired/ related samples t-test
  -pps perform in both conditions
  -use when non normal data distribution
  -interval,ratio or ordinal data