I am analyzing dataset where I would like to try to predict which of the following measurements scales will best predict scores on likability scales that measure how much participants "like" specific pictures.
Dependent Variable (measurements scales): Scale: Likability of specific picture Factor: Pictures Types: Levels: A, B, C Factor: Pictures Gender: Levels: Female, Male Pictures: Male_A + Female_A + Male_B + Female_B + Male_C + Female_C
Independent Variable: ScaleX + ScaleY + ScaleZ + ScaleW
What I am interested in is to answer how well will: ScaleX + ScaleY + ScaleZ + ScaleW Predict score in: Male_A + Female_A + Male_B + Female_B + Male_C + Female_C
Formula: Pictures(Male_A + Female_A + Male_B + Female_B + Male_C + Female_C) ~ Scales (X,Y,Z,W)
I could theoretically try to run several linear regression but this is really just an desperate attempt and it will not explain which of XYZW scales would be best to predict the Likability score.
My thoughts are that it could be Structural equation modeling with some more simple model without any latent variables, more about measurement. However I have not much experience and I do not wish to invest huge amount of time to learn it and then realize it is useless in this example.
Question is: What approach for analyzing this data should I use? I will welcome even more technical answers with R examples as this will be software of my choice.
Thank you kindly for answers, Cheers! :)
