I'm struggling with the analysis of data of an experimental video-vignettes study. I haven't been able to find comparable examples in the literature, so I am hoping that any of you can be of help.
I am testing the effect of physician communication on trust. Standardized 'vignettes' (short, enacted, consultations) have been videorecorded, in which 3 elements of physician communication have been manipulated. For each of these three elements we have created 2 conditions (standard vs. enhanced). This results in a 2x2x2 factorial design, with in total 8 varieties of the video*.
Participants first view one randomly selected version of the 8 videos, and indicate their level of trust in the observed physician. After a short distraction they view a second, randomly selected, version of the video, and again assess their trust in the observed physician.
My question is, whether it would be possible to use data of both observations of each participant. My aim would be to test not only the isolated effect of one of the three elements of communication on trust, but also to be able to test what their combined effect is on trust. I can see that there is some kind of 'nestedness' of 2 observations per participants, but I am not sure whether multilevel analysis would be appropriate here. I've been considering whether the combination of videos participants saw should be used as the unit of analysis (e.g. first version 7, then version 2). If so, there would be 28 possible combinations (56 if you take the order into account). Sample size is 345, which means that approximately 12 people (6 people taking order into account) saw the same combination of videos.
Does anyone have any thoughts on how to enable analysis including both observations of each participant? Or is there anyone who see an analogy with other types of research with two observations per participant? All suggestions are more than welcome!
Thanks a lot in advance for your help!
*Elements of communication that were manipulated:
C1 (standard/enhanced), C2 (standard/enhanced) and C3 (standard/enhanced)
8 versions of the video:
C1stand, C2stand, C3stand
C1enhanc, C2stand, C3stand
C1stand, C2enhanc, C3stand
C1stand, C2stand, C3enhanc
C1enhanc, C2enhanc, C3stand
C1stand, C2enhanc, C3enhanc
C1enhanc, C2stand, C3enhanc
C1enhanc, C2enhanc, C3enhanc
You may leave the list at any time by sending the command
to [log in to unmask], leaving the subject line blank.