Analysis assumptions

Question 1

Assumption of normality

Answer 1

Assumes that the dependent variable is normally distributed within groups.
- tested w oa Shapiro-Wilk Test
- if violated: use nonparametric tests

Assumption for analysis
- all of them

Question 2

Assumption of independence

Answer 2

Assumes that each result is independent. Eg all participants are independent, the effect of one measure does not effect the other etc. Tested using:
- Multilevel analysis

If violated:
- The N is too high
- The SE is too low ( SE = Sd/VN)
- The t stat is too high, making the p value too low,
Possible shows sign effect where there is none, therefore; most important assumption

Analysis:
- All of them

Question 3

Homoscedasticity/homogeneity/equal variance

Answer 3

Assumes that the variance does not differ between groups.
- tested using Levene’s test (in assumed, if sign. is lower than .05
- If violated: adapted test.

Analysis
- t-test