week 3 ANCOVA Flashcards
(5 cards)
1
Q
how do you determine if C is a usefull covariate?
A
- check if there is a relation between X and C (ANOVA, test between subjects)
- Correlation between C and Y?–>yes?–>Adding C will reduce error variance (good thing)
2
Q
what is the function of this graph?
ANCOVA
A
- check the slopes for suspision of parallelism
- In this graph we see suspision of an interaction, because the lines are not completely parallel.
3
Q
using this ANCOVA table. can the null hypothesis of parallel regression lines be rejected?
A
NO–>interaction is non sign–>regression lines between C and Y have the same regression weight.
4
Q
Where can you see if an effect is still sign after correction of a covariant?
ANCOVA
A
note F, df1, df2, p
5
Q
in ANCOVA how do you check which groups have significantly different estimated marginal means?
A
perform an ANCOVA to assess the effect of phone use on reaction time using age as a covariate. In EM Means ask for the estimated marginal means of phone, check the Compare Main Effect option and select Bonferroni for CI adjustment.
