The study seeks to examine if the lesson plan in the treatment group is a more efficacious way of teaching the Toulmin model of argumentation (DV1; measured via a quiz on knowledge and application of the Toulmin model) compared to the control. Additionally, we wanted to measure students' need for cognition (DV2; though we don't expect this to change from pre to post), self-assessment of critical thinking (DV3), and argumentativeness (DV4).
The research design is as follows:
Both the treatment and control groups took a pretest with measures of the aforementioned DVs. The control group was then taught the standard Toulmin lesson plan; while the experimental group was taught a different lesson specifically designed for the study. The pre- and posttests were paired, with the same participants completing both tests in each group.
I know I can Select Cases for the control, for example, then do a paired samples t-test to see if there was a statistically significant difference between the pre and post test. However, I'm not sure to compare the control and experimental groups to see if there was a statistically significant difference between the pre- and posttest across the two groups.
Thank you and please let me know if you need any more information.
Hi
1. The ANOVA for correlated groups has to be developed.
2. Yoy have not discussed the sample sizes of the groups. It is one of the characteristics of parameteric or non-parameteric test.
3. For more details, please refer to,
. Landau, S. and Everitt, B. S.(2004). A Handbook of Statistical analyses using SPSS.
.Howitt, D. and Cramer, D.(2008). Introduction to SPSS.
Regards,
Zuhair
Dear Nate,
I'm not sure about the advice the two gentlemen have given above. If you are looking for differences between two groups with multiple DVs then the appropriate test is MANOVA, provided that its assumptions or their robust exceptions are satisfied.
Dear Nate,
I think if your sample size is small and you have choose the paradigm of pre and post where the sample is same, then you have to go with paired t-test because this is more suitable and authentic test for analysis and apart from this you may be multivariate analysis.
I agree with Peter Samuels and Sushma Rathee together. You want a Repeated Measures Manova because you have multiple DVs and a pre- Post- design. You may want to look at the most modern ways to handle these in the following notes from Prof. Marie Davidian.
https://www.stat.ncsu.edu/people/davidian/courses/st790/notes.html
Best wishes, David Booth
PS Nate, You have a local expert on this stuff at ISU, retired Prof. Ken Berk at [email protected]. Ken is retired from the Math Dept. Mention my name. best wishes, Dave
My understanding is that a MANOVA ahould be used if the DVs are related to each other. I dont know enough about these measures to say if they are. MANOVA can be used for any two or more that are related. Standard repeated measures t tests will work For DVs not related to each other. All of this is assuming the measures are interval or ratio scales and that the data meet he assumptions for the tests.
To greatly simplify the problem, calculate the differences (Post - Pre) and then you can compare the two groups per individual DV (independent samples t-test or Mann-Whitney rank test). If you also want to look at the relationships between the four DVs, you might consider a multivariate General Linear Model. Another possibility which may add to your insight, and which also involves analysing the four DVs separately, is to plot scatter plots with Post and Pre on the axes and different symbols for the Treatment and Control groups. If it looks like straight line relationships, you may do a regression analysis with Post as the DV and Pre, Group and Pre X Group as independent variables. If the regression coefficient of the interaction term Pre X Group is significant, it means that the difference between the Groups with respect to Pre and Post depends on the Pre outcome. If the interaction term is not significant, the regression coefficient of Group will show whether or not the teaching methods led to significantly different outcomes.
Nate, are you asking that, e.g., after you perform the paired t-test on the four DVs, you want to know if there is a significant difference among the four average differences tested in the t-test?
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS