Kavli's 2004 dissertation, "Embedded Paralinguistics in Textual Computer-Mediated Communication:
Is a Picture Really Worth a Thousand Words?" used the gold standard of research, an experimental design. This results in the highest possible external validity in determining cause and effect in behavioral research. As there are two DVs, which are scale variables, and two IVs (or factors with different categorical levels), MANOVA is the appropriate multivariate choice if the assumptions of the model can be met.
Discuss how Kavli examined and treated the statistical assumptions of MANOVA and the issues she encountered. Evaluate the findings and the validity of the approach.