My research design is as follows:

I have these Between Subjects IVs:

  • Experiment Condition - 5 levels

  • Facebook User status - 2 levels (yes/no)

My supervisor also wants me to see if there are significant effects of:

  • Gender - 2 levels

  • Relationship Status - 2 levels

  • Relationship Satisfaction - 2 levels

And these within subjects IVs:

  • Mood at Time 1

  • Mood at Time 2

I also have three DVs

  • Attraction level

  • Frequency of Thought

  • Mood State

I am looking at the effect of a certain condition on Facebook users, vs non Facebook users. Within this I want to look at whether there are differences between the genders; whether there is a difference between people in a relationship vs. not in a relationship and the effect of relationship satisfaction. It's a big study, but I'm not sure about how to describe the design.


  • How should I describe the design of this study?

Initial thoughts

I'm thinking it has to be a MANOVA design, however what I am finding confusing are the various IVs. Is it a 2 x 5 Factorial MANOVA? I'm just so confused with all the IVs flying around.

  • $\begingroup$ Which variables were measured at two time points? Just mood? If not all variables, when are the other variables measured? Also, your within subjects iv is "Time" which has two levels 1 and 2. $\endgroup$ – Jeromy Anglim Jun 6 '11 at 4:54
  • $\begingroup$ Just mood. Attraction and Frequency of Thought are both only measured at time 2. $\endgroup$ – emilie rankine Jun 6 '11 at 5:26
  • There is a difference between the design and a statistical test. Your design presumably incorporates random assignment of participants to one of five conditions, actively sampling (perhaps an even number?) of facebook and non-facebook users, and to some extent the study of time.
  • You might describe your design as a 5 by 2 by 2 mixed design (in the case of mood) with condition (5 levels) and facebook status (2 levels) as between subjects factors and time (2 levels) as a within subjects factor; and a 5 by 2 factorial design for the dependent variables only measured at time 2.
  • Any description of your design should make it clear which between subjects factors were achieved through random allocation.
  • In terms of statistical analyses, you may choose to run ANOVAs or MANOVAs and it sounds like you are interested in covariates (i.e., gender, relationship status, relationship satisfaction). I label these covariates because they were not part of either random allocation or the process of sampling participants.
  • $\begingroup$ Thanks Jeremy, this sounds good. Although I am a little concerned about using a MANOVA. What do you think about using a Multivariate Regression instead? $\endgroup$ – emilie rankine Jun 6 '11 at 23:42
  • $\begingroup$ @Emlie You have a lot of variables where you could explore different relationships. I suggest asking another question on this site specifically focused on how to analyse a particular aspect of your data. $\endgroup$ – Jeromy Anglim Jun 7 '11 at 2:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.