Week 11 Flashcards
What is one way to think about an ANCOVA?
A statistical control procedure.
What are the two applications of ANCOVA?
- elimination of experimental confounds
2. clarification of experimental findings
What is the benefit of ANCOVA?
Allows a more sensitive test of treatment effects (as you are have higher control)
What is a good way of adjusting peoples scores after the test to compare what they would be, if they had the same baseline?
ANCOVA
How can we clarify experimental findings of MANOVA?
A step-down ANCOVA analysis, to determine which set of correlated DV’s was (most) responsible for a significant MANOVA.
What is being measured in an ANCOVA?
- main effect of treatment
- ‘main effect’ of covariate
- interaction between the treatment and the CV
Describe what we are testing when we test the main effect of treatment in an ANCOVA?
Tests whether the means of the groups are significantly different after controlling for the effect of the CV on the DV.
If there is no significant main covariate effect in ANCOVA, what does tis tell us?
That there’s essentially is no point controlling for it. In other words, you don’t need to do a ANCOVA.
If the linear regressions lines in ANCOVA are parallel (have the same slope) for each experimental group, then we can do what
Pick the mid point of the covariate and identify the size of the difference between groups on the DV.
What is one assumption regarding linearity in ANCOVA?
The the lines are parallel/slopes are the same.
This is really important.
if the covariate against the DV linear regression lines in ANCOVA are different for different grades, what does this mean?
That the size of the difference between the groups, is going to be not the same at different levels of the covariate.
What do we do if the linear regression lines in ANCOVA are not parallel?
We have to come up with a really good reason as to why we you are comparing the groups at certain values of the covariates and not others.
What is it called when you have parallel linear lines in an ANCOVA?
homogeneity
What is an interaction between the IV and the CV actually telling us in ANCOVA?
Would tell us whether the slope of the DV-CV regression line is the same for each treatment group.
Non significant interaction tells us we have parallel lines.
Does ANCOVA have an interaction analysis term?
No - it ASSUMES that the slope if the same for all groups (homogeneity)
What is the main effect of the IV in ANCOVA telling us?
Are the intercepts for the regression lines for each group the same.
If not, then there is a significant difference between the groups on the DV when the CV = 0.
What does it tell us if we have a significant interaction in and ANCOVA?
Tells us we do not have equal slopes in both groups (BAD)
If the CV X IV interaction is non-significant:
then we have homogeneity of slopes for each group.
If the CV X IV interaction is significant:
then the slope on the DV and CV is not the same for each group.
What are the assumptions for ANOVA and ANCOHET?
- linear relationships between CV and DV
- Normal distribution of scores, homogeneity of variance and independence of residuals
- homogeneity of regression
Homogeneity of regression is assumed in ANCOVA, but is:
Explicitly testes in ANCOHET.
In a perfect world, what relationship do we want between the CV and the IV?
No relationship.
How can we determine the relationship between the CV and the IV?
Perform a one way ANOVA with CV as the DV, and the treatment as the IV. If significant, then the IV and CV are not independent.
What happens if there is a significant relationship between the the IV and the CV?
Then if we were to remove the CV, it may remove some of the variance in the real treatment effect (leave a type 2 error).
How can we use the different DV’s in MANOVA to do a step-down analysis?
We can turn one of the DV’s into a covariate and do an ANCOVA step-down analysis to see if IV is having an effect on one DV, or on both together.
How do we go about doing an ANCOVA in a step-down analysis following a MANOVA?
The DV which is deemed to be the most important (based on theoretical considerations) is tested first in a univariate ANOVA. The remaining DV’s are then successively tested in univariate ANVOCA’s using the first DV as a covariate.
What happens if our ANCOVA step-down analysis is not significant?
We can interpret that the DV does not add anything that has not already been accounted for by the most important DV.
What happens if our ANCOVA step-down analysis is significant?
It means that the IV has an impact on the DV beyond the impact that it had on the most important DV (adds unique variance).
What does ANCOVA do with DV score adjustment?
Adjusts individual scores on the DV to what it would be if everyone had the same score on the covariate.