MANCOVA and MANOVA Discussion > What stat tool should I use?
Thanks for the great question, Kenney. From what I read here, I think an ANCOVA is what you are looking for. An ANCOVA is similar to a MANCOVA, except an ANCOVA (ANCOVA= analysis of covariance) only has one dependent variable (DV). By comparison, MANCOVA has multiple dependent variables (MANCOVA=multivariate analysis of covariance).
In your example, your post assessment could be your DV, your treatment group variable your Independent variable (IV), and your pre assessment, grade level, age, and ethnicity, would be covariates. Your results would indicate whether your treatment group had higher post assessment scores than the non-treatment group, holding all covariates equal (including your pre data).
Keep in mind that when you include ethnicity, you'll need to dummy code. However, if you are using SPSS, you can use the "contrast" function in the General Linear Model analysis (GLM) and it will do the dummy coding automatically for you. I hope this is helpful and please let me know if you have other questions. Happy Holidays!
On a side note: be careful how many predictors you include in the model, because with only 20 people in each group you may have problems with insufficient power to detect effects.
Jeremy Taylor
Thnak you so much. You have saved me!
Kenney Moore
I'm glad I could help, Kenney!
Jeremy Taylor
I am working on a college stat problem. I am intested in if students' partcipation in using a computer-based reading program improves their reading score from a pre and post assessment that isn't part of the computer program. I have 20 students in the treatment group (computer-based program) and 20 students that will not reveive the intervention. I am thinking a MANCOVA would be useful becuase it was account for the pre and post scores , the treatment, as well as covariates like students' grade level, age, ethnicity...Would MANCOVA suffice?