This video demonstrates how to conduct and interpret a One-Way MANCOVA in SPSS. Assumptions for MANCOVA are tested, including homogeneity of variance-covariance and homogeneity of regression slopes.
I like the idea of setting up the custom MANCOVA so that when going to conduct the second MANCOVA you don't have to make any changes other than the full factorial button. This keeps things simpler. I also paid particular attention to which output to interpret and which test to interpret depending on the outcome of the Box's test. Important things to remember
I have chose a MANCOVA design, possibly a factorial one, this video helps me to see the interaction of the independent variables on the dependent ones and it is very helpful to see the full effect of the data analysis tool is SPSS.
Thank you Dr. Grande. Your videos are valuable. I would like to request a video of Two-way MANCOVA and Cox Regression with testing of assumptions. Thank you in advance.
Dear Dr Grande, Thank you for this tutorial. If after having conducted the initial test for normality, all variables have violated the assumption how do you suggest to continue ? Any help is greatly appreciated
When checking for outliers, shouldn't the examination be conducted within each group of the independent variable (IV) rather than just in the overall distribution of the dependent variables (DVs) and covariates?
It was interesting to see an example of a non-significant study. Although the groups differed in the treatments it wasn't significantly different to matter and to go further with the study. How do we know that the Box's test should be .01? Is that just something to remember? What is the significance between that statistically significant value and the .05 significance cutoff?
Very informative and much to take in, but I would guess that comes with experience. I agree with Ashley AP in that I found the scatterplot interpretation to be rather subjective. I also appreciate the custom MANCOVA. I think I find it most confusing when determining which test to use for significance when there are multiple options presented.
Hi, many thanks for the helpful video. I would like to do repeated measures MANCOVA, I was wondering if anyone have seen a video on how to do that in SPSS? Thanks
I looked for alternative statistics for the MANCOVA and did not find it mentioned in much of the literature. One book mentions the Bonferroni F procedure can be done with a MANCOVA (Huitema, 2011). Are there other alternatives to MANCOVA? Huitema, B. E. (2011). Analysis of single-case reversal designs, in the analysis of covariance and alternatives: Statistical methods for experiments, quasi- experiments, and single-case studies (2nd ed.). Hoboken, NJ: John Wiley & Sons, Inc.
Hello, can you please explain the missing variable for MACOVA? why do we need to test that? and are we allowed to have missing cases? or must we delete via Listwise? thank you, i love your videos
In the preliminary test of assumptions we look at significance in the box tests but how to we know that that will correspond to interpreting Wilks Lambda.??
Kind of urgent question! How do I test for homogeneity of regression slopes in a MANCOVA, when I have 2 IVs and 2 covariates? Just so you have a better idea, the DVs are post-test scores on 2 tests, the IVs are treatment group, propensity layers, and the covariates are pre-test scores on the 2 tests. Do I just do the same thing as you have done, but with many more interaction terms? Thanks!
So do you interpret Pillai's Trace only when Box's Test of Equality of Covariance Matrices is significant? Or would you interpret it if ANY assumption is not met? It seems like in this video you're indicating it would only be if Box's Test is significant. In the MANOVA video I thought you said it would be if any assumption is not met.
+Gregg Elliott I had noted that as well. Mertler and Vanetta (2013) indicates interpreting with Pillai's Trace only when the homogeneity of variance-covariance is not met. However, I have also read that it is prudent to use the more robust Pillai's Trace regardless. I would be interested to hear from Dr. Grande.
+Ivy Byun I would lean toward using Pillai's Trace if any assumption is not met, but one must consider which assumptions are violated and to what degree.
Hi, Dr Grande. I assume the DVs in this MANCOVA video are gain scores. Is that correct? I don't see that often, but I wish to run a similar MANCOVA in which my DVs are gain scores (therapeutic alliance and client motivation, which I predict will be greater in the treatment group vs control group) and plan to treat their pre-test total symptom score as a covariate. Does that sound appropriate?
How to test for covariance when I have only two groups and non normal data ? Log transformation hasn’t lead to a normal distribution. Any non parametric equivalent of MANCOVA in spss?
Hey! I have a question for the first Mancova you did (preanalysis): at 8:04 there is the Levene test - do we not have to interpret it? for homoskedastizity? Thanks for the great video :)