ANACOVA might be used to compare the mean re-sponses to different treatments when a quanti-tative variable x, called the covariate, is mea-sured on each experimental unit. Can you read more on ANCOVA Kamal Takhdat ?

Thank you Haileleul Zeleke Woldemariam for your replay.

But these subgroups had received the same intervention. I want at the post-test compare their scores to know what subgroup had recorded the most Knowledge gain

I think that ANCOVA is applied for independent groups.

Thank you so much

You can work on gain scores Then compare gain scores of three groups using ANOVA one-way and if F-ratio comes out to be significant, employ t-test to find the pair-wise difference.

thank you so much Mohua Khosla . I have 3 related groups. I think I should not use ANOVA with is applied for independent samples.

https://www.researchgate.net/post/Three_means_comparison_by_t_test_or_ANOVA

thank you Amjed Ahmed

Good question

Follow

Use F-test

But how the groups are related when they have been categorized on the basis of Pre-test scores.

It's not the case of same group undergoing different treatments. Then we use repeated measure ANOVA

You can use F-test for repeated measures (ANOVA).

Since the group received the same intervention as you said, I support to use ANOVA

than you so much for your reply and reflections Mohua Khosla , Tebaa Farouq Frederick Ramos Abdulrahaman Aliu Mahmood A. Al-Shareeda

You can use ANOVA; if you meet the 5 assumptions, if you do not meet the assumptions then I propose Kruskal Wallis

I guess study used an intact class. That is there wasn't any randomization of subjects. If my guess is correct, then ANCOVA (Analysis of Covariance) will be most appropriate. The pretest, in this case, serves as the covariate.

ANCOVA is appropriate because you have two groups and pretest measure will be your covariate.

I support the group for the use of ANOVA as long as what you would compare is the means.

Yes, Analysis of covariance (ANCOVA)

method will be most appropriate.

thank you all for your reply

Sure ANCOVA will be the most appropriate

They can be compared with an ANOVA test of mean differences between independent groups by posing as hypothesis: Ho: Ho: µ1= µ2=µ3

H1: at least one of the socks is different

Available at: http://www.ics-aragon.com/cursos/salud-publica/2014/pdf/M2T08.pdf

thank you Wilmer Lopez

For the comparison of two or more means, we have the analysis of variance, which is a parametric test. It is advisable to check if the data meet the assumptions of this test.

Article Comparing groups for statistical differences: How to choose ...

If you are interested in the 'influence' of the pre-test score on the post-test score, i.e., if you want to know which learners benefited (most) from the intervention, you could also - instead of artificially dividing the group - use a regression with pre-test as predictor and pre-post-gain as criterion variable?

If the intention is to only compare the pretest and post test, it means there was no treatment of subject in the form of an experiment. In such case, a dependent t-test is more appropriate. Remember, you're only exploring the relationship between two scores obtained by same person. I don't seem to think regressing the scores will give meaning to the research.

thank you all for your responses Isaac Ofem Ubi Andreas Wirag .

My dear, it's a simple case if I understood you properly. You are trying to compare the post test means of three sub groups created from a sample of respondents after an intervention. A one way Analysis of Variance would be most ideal to test the hypothesis that there is a/no significant difference in the post test means of the three subgroups (a, b, c) after the intervention (treatment). Best regards

Owan Valentine

thank you Valentine Joseph Owan