1
$\begingroup$

The pretest scores of a study I'm analysing is significant: the experimental group outperformed the control group in one of the measured dependent variables (academic achievement) while the control group did better in another measured dependent variable (retention) in pretest.

Meanwhile, the experiment (treatment) has been carried out and post test administered.

Since significant difference was observed in both pretest and post test scores of the groups, is there a way of reporting that the treatment produced an effect?

$\endgroup$
2
$\begingroup$

You have baseline differences on the pretest. You can attempt a regression adjustment. Use both pretest and group assignment to predict posttest in a regression model.

The treatment effect in this model is the average difference on the posttest between those in the treatment and control group for those who had the same pretest score.

But if the groups differed on the pretest, both groups may differ in other ways not captured by the pretest. So you cannot interpret the group difference from your regression coefficient causally. It is simply an observed group difference for those with similar scores on the pretest. Be sure to check the data to see that there are enough individuals in both groups with substantial overlap on the pretest.

This is a starting point from which you can explore further.

$\endgroup$
  • $\begingroup$ Thank you, Jim. But, I did not get your explanation in your second paragraph $\endgroup$ – Adaobiagu Nov 3 '18 at 12:21
  • $\begingroup$ @Adaobiagu that's just the standard interpretation of a regression coefficient, or the group difference in ANCOVA. Do you have any familiarity with regression models? $\endgroup$ – Heteroskedastic Jim Nov 3 '18 at 12:31
  • $\begingroup$ @Adaobiagu you're welcome. Since you're new, maybe I should add that if the answer satisfies your question, you can check accept it as the correct answer. $\endgroup$ – Heteroskedastic Jim Nov 4 '18 at 14:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.