Behavioral factors show strong correlation with each other,
making multicollinearity an issue for the power and stability of the
regression model. We used FA (Factor Analysis) [29] to remove
multicollinearity of the variables, and identified latent variables
that are not captured by direct behavioral questions. Factor Analysis
reduces the number of variables by replacing them by potentially
lower number of linear combinations of original variables (called
factors) while conserving as much information as possible. Because
these factors are linear combinations of the original variables, they
are easy to interpret and can be assigned physical significance. In
short, Factor Analysis identifies the set of k latent factors