In my research (The effects of mentoring EFL teachers on their learners’ writing ability: Does a mentoring software make any difference?) the writing ability (DVs) of three diffrent groups of learners are needed to be compared.
The (IVs) were the two mentoring styles applied on their EFL Teachers. (Traditionla mentoring, Software Mentoring and the control group with no mentoring styles at all). Pre-test and post-test were used before and after the manuplation of the IVs. The EFL teacher In group one was traditionally mentored. The teacher in Group 2 was mentored by using an e-mentoring software and the learners in group 3 were considered as control group and their teacher received no mentoring treatment at all. At the end the the means of the writing ability of these three diffrent groups of learners needed to be compared. What kind of statistical tests should I use?
ANOVA? or ANCOVA? one way? or two ways? or WHAT?
Please exactly tell me which statistical tests and how?
If I understand your question correctly a 3 group 1 way ANOVA for the 3 groups before the intervention and the same after the intervention should tell you what you want to know. I believe you can find more complicated approaches but I believe this answers your question. Best, David Booth
PS I would also do a paired t-test on each group to see how each group progressed on the intervention. BTW I would recommend using statistical software for your computation. Google The name of the statistical test along with the name of your software and you can usually get detailed instructions. Further some basic statistical software is available online at no cost. You might check this out if needed.
Hello Ahmad,
As David Eugene Booth 's suggestion notes, the basic analytic approach that would appear suitable is a one-way anova (3 groups being compared). As well, I would suggest that you use the pretest scores as a covariate, to control for any pre-existing differences in groups on the writing scores prior to treatment. The DV would be the (adjusted) posttest scores. So, one-way ancova probably is likely the better choice.
You didn't say whether groups were randomized (e.g., were individual participants assigned at random to one of the three treatments?). Doing so is considered the best approach, since it gives confidence that results aren't a function of some extraneous ("noise," or "nuisance") variable that is confounded with group construction.
Good luck with your work.
Thank you Mr. David Eugene Booth and Mr. David Morse
With respect, the groups were randomized and a paired t-test on each group had been applied. We have decided to apply One-Way Anova. Now The question is on the SPSS window On Analyze >> Compare Means >> One-Way Anova >> Post hoc window
Which option in the "Equal Variances Assumed" and "Equal Variances Not Assumed" should be ticked?
The (IVs) were the two mentoring styles applied on their EFL Teachers. (Traditionla mentoring, Software Mentoring and the control group with no mentoring styles at all). Pre-test and post-test were used before and after the manuplation of the IVs. The EFL teacher In group one was traditionally mentored. The teacher in Group 2 was mentored by using an e-mentoring software and the learners in group 3 were considered as control group and their teacher received no mentoring treatment at all. At the end the the means of the accuracy, fluency, and complexity of their writing tasks and the total writing ability of these three diffrent groups of learners need to be compared.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA