Q1) Write up your results using APA style. Additionally, reoly the following questions about your results:

i) Using overlapping circles, illustrate how SS total was partitioned in ANOVA (to predict Y from A) versus ANCOVA (to predict Y from A, controlling for Xc). Point out what SS values correspond to each slice of circles in these diagrams.

ii) Compare unadjusted SS terms from your ANOVA (for error and effect of treatment) with adjusted SS terms from your ANCOVA. Was effect of including covariate primarily to reduce error variance, mainly to take variance away from treatment variable, or both? Does effect of treatment look stronger or weaker when you control for covariate?

iii) Note pattern of differences across treatment groups on Xc, covariate or pre-test. To what extent were there large differences in aggressiveness before treatment was even administered? Determine the nature of confound: Did most aggressive boys get treatment that was highly effective or ineffective at eliciting aggression?

iv) Look at adjusted and unadjusted deviations which you got when you ran ANCOVA using ANOVA program. Determine and report the within-group slope which relates Y to Xc (for this, you will have to look at your regression printout). Illustrate how one of the adjusted means can be computed.

v) After adjustment, were group means closer together or farther apart? Is there any change in rank order of group means when you compare adjusted versus unadjusted means? Do you believe these adjustments are believable?

vi) We make a number of suppositions to do ANCOVA: that treatment doesn't affect covariate; that there is no treatment by covariate interaction; and that covariate is reliably measured. How well do you believe these suppositions were met in this case? Why would it be problem if these suppositions were not met?

Warner Table