Misc. Stats Topics Discussion > Study Design Help
Overall, I would say this is in fact a between-subjects design, because you only have one data collection for each participant. If you collected data several times over a period of time, you could then conduct a within subject design to examine the degree to which individuals within schools change over time.
To answer your second question, just purely by using randomization does not completely eliminate any concerns about validity. You still must be concerned about the measurement of your variables… among other things. Reliability and validity is a very complex topic.
Jeremy Taylor
so do I have 3 Iv's here? treatment 1( cbt) treatment 2( cbt + social learning) and control?
or is this just 1 IV here with my one one DV being the self-report scores ? i'm completely confused here as to if i'm using an anova or manova
Daniel
From what you've described, it sounds like you have 1 IV (with three levels/groups for that IV) and one DV (self-report scores). It sounds like an ANOVA is more appropriate than MANOVA, although you may need ANCOVA, if you want to include covariates.
Jeremy Taylor
I'm trying to design a study incorporating learning theory and a school wide bully prevention program. I want to have three groups 1) Treatment 1 - using cognitive behavioral learning 2) treatment 2 - using social learning techniques 3) control waitlist
My goal is to randomize my schools to receive either one of these groups over one semester. I would incorporate a self-report of victimization/aggression as my pre-post test. Schools selected for treatment would incorporate this only in k-8 classrooms but it would be the whole school not just certain classrooms.
This project would take place over one semester. Is this a between subject design? My reasoning is that it only can be a between subject design because i have different schools allocated to different treatments. Is there a way I could make this a within subject design? If i'm using randomization are all threats to internal validity covered? Any thoughts to set this up would be great