 |
 |
Dear all, I am fully aware that these questions are asked on a nearly daily basis, but I got a question of how to conduct the ANOVA for my study. The study's design is pretty simple, 4 subject groups, that conducted a 4-condition-paradigm each (1 control, 3 experimental condition). Resulting in a groupXcondition design, with each factor having 4 levels. So far, I assumed a flexible factorial design, with the factors 'subject' and 'group'. For the GLM, 'condition' was included as the main effect. Subjects of group one were assigned to condition 1-4, Subjects of group 2 to condition 5-8, and so on. Now I stumbled upon the paper of Jan Gläscher "Contrast weights in flexible factorial design with multiple groups of subjects". According to this, 3 factors should be included: subjects, group and condition. So here are my questions: 1) Which approach is the "right" one to mainly focus on the interaction between group and condition (group comparison): Should I include the factors 'subject' & 'condition', as I did, or stick to Gläscher's approach? Or is there an entirely different approach which fits the study best? 2) Which main effect & interaction should I model in the design? What is the meaning of including a main effect and/or in the interaction in the GLM? Thank you very much in advance! Kind Regards Ruben Scholle, M.Sc. Scientific researcher / PhD candidate Department of Psychiatry, Psychotherapy and Psychosomatics [UKA_banner] Phone: +49 (0)241 80-85027 Fax: +49 (0)241 80-82401 [log in to unmask]<mailto:[log in to unmask]> Universitätsklinikum Aachen Anstalt des öffentlichen Rechts (AöR) Pauwelsstraße 30 52074 Aachen, Germany Phone Call- & Service-Center: 0241 80-84444 [log in to unmask] www.ukaachen.de Aufsichtsratsvorsitzender: Dr. Robert G. Gossink Vorstandsvorsitzender: Professor Dr. Thomas H. Ittel Kaufmännischer Direktor: Dipl.-Kfm. Peter Asché Sitz Aachen Sales tax identification nmumber: DE 813100566