Process of conducting ANCOVA

Report: Using pairwise comparisons with a Bonferroni adj for error rate, revealed after controlling for effect of Covariate group 1 IV (M= , SE =) reported sig greater/lesser DV than group 2 IV (M = , SE = ), t(df error) = _, p = (tailed), r =__ (S/M/W effect). etc.

Step 1. Mathematical Assumptions

Step 4.FINAL WRITE UP

Parallelism

Step 2. Conduct ANCOVA

Normality

Shapiro Wilks

Homogeneity of Variance

Levene's

Independence

Linearity

Pearson's Correlation; between Cov and DV should be SIG

Check Int effect in ANCOVA box F should be NON-sig

Report: the sig effect of IV on DV after controlling for the Covariate
F(df1,df2)=_, p= (1 / 2 tailed), Eta sq = (strong/mod/weak effect), obs power =
Covariate sig relationship with DV
F(df1,df2)=_, p= (1 / 2 tailed), r =(result of Pearsons), (strong/mod/weak effect), obs power =

Because we obtained a sig main effect for IV and because it contains more than 2 groups, we must conduct pairwise comparisons to see which group means adjusted for the effect of covariate are different -Use Bonferoni or sidak

Step 3.

Examine CONTRASTS to see what is sig

Calculate t-values by Mean Diff/SE

Report: Simple contrasts revealed after controlling for effect of Covariate group 1 IV (M= , SE =) reported sig greater/lesser DV than group 2 IV (M = , SE = ), t(df error) = _, p = (tailed), r =__ (S/M/W effect). etc.

REPORT MEANS AND SE FROM ESTIMATED MARGINAL MEANS
BOX LABELLED 'ESTIMATES'

Examine PAIRWISE to see what is SIG

There was a Sig effect OF the IV ON the DV after controlling for the effect of the covariate, F(df1, df2)= _, p = ( tailed), eta sq=(s/m/w effect), obs power = __
The covariate was also sig related to DV , F(df1, df2)= _, p = ( tailed), eta sq=(s/m/w effect), obs power = __
Simple contrast (or) pairwise revealed that after controlling for effect of covariategroup 1 of IV, (M = , SE = ) reported more/less DV than group 2 of IV (M = , SE = ), t(df within )=, p = (tailed), r= (s/m/w effect) etc...