Should I handle this as a within-subjects, between-subjects or a mixed design?

#1
My design has two dependent variables (likeability and persuasive power). The independent variable is advertisement version, with 2 levels (individualistic and collectivistic).

The participants of the questionnaire each saw 4 different advertisements. These 4 advertisements were either collectivistic or individualistic. Not one participant saw both the collectivistic as well as the individualistic version of the same advertisement. However, the questionnaire tool decided randomly what condition to place each participant in. This way, some participants saw the same versions of advertisements as other participants, but different ones than others.

My question is now, should I handle this as a within, or inbetween-subject condition? Some people saw the same versions obviously (wihtin-subjects), but saw different ones than others (inbetween-subjects). I also don't think this is a mixed design, as it's not the case that 50% of my participants saw version A and 50% of the participants saw version B.

I'm really at a loss here. I could really appreciate your help! Let me know if there's anything unclear to my explanation of the question.
 

Karabiner

TS Contributor
#2
I'm afraid that I have some difficulties to understand your design.
These 4 advertisements were either collectivistic or individualistic. Not one participant saw both the collectivistic as well as the individualistic version of the same advertisement.
What does this actually mean? Did you have 2 advertisments (A and B),
and each in a collectivistic (c) or individualistic (i) version, so there's
Ac, Ai, Bc, Bi ?
the questionnaire tool decided randomly what condition to place each participant in
What exactely do you mean by "condition", what were the
stimuli presented in each condition?

With kind regards

K.
 
#3
If I understand correctly, you created 4 pairs of ads (eight total), where there was an individualistic and a collectivistic version for each of the four pairs (Pair1Individual, Pair1Collective, Pair2Individual, Pair2Collective, etc.) You then randomly selected which item within each pair a participant would see?

I would imagine that you would want to restructure your data into longform, with four observations for participant. You could then run a mixed model, where you would include a random intercept for participant (to control for the fact that each participant is likely to have systematic biases in how likeable and persuasive they find ads in general, separate from any influence of individualism/collectivism), and simply model the effect of individualistic / collectivist on your DV's while controlling for pair number...