- Within: ACROSS
- Between: DOWN
A One Way ANOVA can compare ____ or more independent groups
- 3
- One IV
- All means should be equal for each Tx group starting out (X1, X2, Control)
What is the F test?
- how much observed variability is explained by treatment means and how much is attributed to unexplained differences among subjects
- Also known as sum of square
- The bigger the F, the smaller thr p-value! This is what we want!
We do not like ____ in subjects for an ANOVA
variability
Total variability attributed the to the treatment effect (between groups) and the unexplained sources of ____ among subjects (within groups).
- variance (error variance)
F tests are ____ tailed
- 1
- Only tell us the difference between at least two of the means
If the F is found to be large, what must be performed?
A post-hoc test to determine where the difference lies
Is this significant?
(F(2,12)=8.67,p-value<0.05)
(F(df, df = F=statistic, p-value)
Yes!
F equal or less than 1 is NOT significant
When looking at the SPSS value of ETA we want an effect size of ____ to know it is significant
Greater than 0.14
Two Way ANOVA
- Involves** two** independent variables (each variable is a factor and is not repeated).
Ex: - Effect of prolonged vs. quick stretch vs. control
- At the same time we are examining knee position (flexed vs. extended).
This is an example of…
A 2 x 3 ANOVA
Main effect
- Collapse factors (pooling) to get marginal means
- Look at one factor at a time
Example
* Main effect for type of stretch (prolonged vs. quick vs. control) irrespective of knee position.
* Main effect for knee position (flexed vs. extended) irrespective of type of stretch.
Interaction Effects
- Look at combinations of levels of each independent variable
- Must plot!
- If parallel, NO interaction
How would you interpret this graph?
No interaction, possible main effect
How would you interpret this graph?
- Likely an interaction effect
An interaction is a significant difference between ____
- simple effects
- Examination of combinations of Tx’s
when no interaction, discuss ____
If interaction, no significant main effects, discuss ____
Harder to interpret with interaction present and significant main effects!!! Look at ____ and ____
- main effects
- interaction
- Disordinal interaction
- Ordinal interaction
What is this considered?
- Disordinal interaction
- Cannot draw any conclusions about the main effect of knee position. Ignore interpretation
What is this considered?
- Ordinal interaction
- Relative ranking of level of stretch does not change with knee position.
- The relative ranking of factor A (Type of Stretch) does not change with factor B (Knee Position). Prolonged stretching better.
If a two way ANOVA is significant, what must be run?
post-hoc test
