I have a student who is looking at the effects of attractiveness on impression formation. The subjects viewed six different pictures that varied on the dimensions of attractiveness and gender. For each picture the subject rates the person in the picture on a variety of dimensions. The student is interested in whether the ratings on these dimensions are related to one another and what the underlying factor structure might look like. Therefore, he did a factor analysis treating the six pictures as independent ratings. I don't think this is the right approach because they are not independent events. I told him to do separate factor analyses for each of the pictures. Is this right? Is there anyway to analyze the data from all six pictures together? He would like to be able to get factor scores that could be used in subsequent analyses. However, if the 6 pictures are analyzed separately they will not load on the factors in the same way. Therefore, it doesn't seem possible to get comparable factor scores. TIA for any help you can provide. Richard Platt St. Mary's College of Maryland
