ANCOVA

Question 1

ANCOVA: PURPOSE

Answer 1

ANCOVA is used to compare two or more groups, after statistically removing (adjusting, controlling) the effect of extraneous variable (s) on the dependent variable.

The main question for ANCOVA is the same as for ANOVA.

Question 2

ANOVA vs ANCOVA

Answer 2

ANOVA: Partition the total sum of squares for the dependent variable (SST) into sum of square attributable to the independent variable (SSB) and sum of squares within (SSW).

ANCOVA: Sum of squares are adjusted for the covariance’s relationship with the dependent variable. The degree of adjustment is related to correlation of the Covariate and the dependent variable.

Question 3

Covariate

Answer 3

A variable that is statically controlled (held constant) in analysis of covariance

The covariate is typically an extraneous, confounding influence on the dependent variable or a pretest measure of the dependent variable.

Question 4

THE PURPOSES OF ANCOVA

Answer 4

1. Maximize the equivalence of groups prior to testing the effect of the IV on DV.
2. Increase the precision of the experiment by reducing the error term (confounder) and therefore an increase in the power of the analysis.
3. Remove the potential sources of bias in the experiment by adjusting group means.

Question 5

ASSUMPTIONS: ANCOVA

Answer 5

1. The independent variables are made up of mutually exclusive groups.
2. A normally distributed dependent variable.
3. Homogeneity of variance among the groups.
4. The covariate is a continuous variable.
5. The relationship between the DV and the covariate is linear. (Use scatter plot and Pearson r to evaluate this assumption.)
6. Homogeneity of regression slops across groups. The covariate should have the same relationship with the DV in every group being compared. (Refer to Figure 11-2 of Polit’s book.)

Question 6

Steps in Performing ANCOVA

Answer 6

(Optional) - Perform ANOVA on only DV.
1. Examine the means of DV and Covariate (Cov.)
2. Examine the correlation between DV and Cov.
3. Examine the assumption of Homogeneity of Regression slopes by checking the interaction between the Cov. and grouping variable.
4. Perform ANCOVA.
(Optional) - Compare ANOVA with ANCOVA.
5. Compare the means of DV (Unadj with Adj).