Please enable JavaScript.
Coggle requires JavaScript to display documents.
ANCOVA - Coggle Diagram
ANCOVA
Covariate
Variable related to outcome variable
can sometimes be controlled for through alterations in design
Impact
increases bias
increases within-group error variance
How covariate is accounted for
linear regression done between DV and covariate to explain variation due to covariate
Shared variance between DV and IV determined through linear regression, accounting for variance by Covariate
Covariates must be linearly related to DV.
Normality assumptions
DV is normally distributed between each level
Ancova is robust to outliers
ANCOVA assumptions
IV and covariates are not related
must be done before running ANCOVA
regression lines are the same for each group formed by the categorical IV.
Homogeneity assumption
Leven's test