490/506 Psy Advanced Statistical Design First Project Assignment: Create an EFA using as variables the scales listed as independent variables in data_dict.xls. Post to Portal: 1. Complete .sps, ending with final model 2. Narrative describing steps to final model: 1. First model and justification 2. Critique of results and justification for second attempt 3. Brief description of intermediate attempts 4. Full description of final attempt Objective: Develop a measurement model that you will be able to put into a hybrid SEM. Your variables are scales listed as independent variables in data_dict.xls. Note that they are themselves scales and their Cronbach alphas are listed. You have some freedom of choice in how to proceed. To make the SEM go, you will want at least two latent traits with at least three observed variables, although it is possible to have only two or even one indicator variables. Possible Strategy 1: Our theoretical base is in the controversy over how many psychopathies there are and how they differ. Our dependent variable will be antisocial behaviors, so we are coming down on the side of Cooke and Michie and not Newman, Hare & Newman or Walters (the former argue that we need to identify the personality traits associated psychopathy with and antisocial personality whereas the latter argue that antisocial behaviors are essential to the definition of antisocial personality). To this end you will look for two latent traits, one that looks like essential psychopathy (PCL Factor 1, PPI's Fearless dominance) and one that looks like antisocial personality (PCL Parasitic lifestyle, PPI's Impulsive Antisociality). HINT: you may end up with other latent traits in addition or instead of these and that is OK. Alternative Models. You may choose to formulate latent traits of your own devising, still predicting to antisocial behaviors, or even choose another latent trait (e. g., intimacy) as a dependent variable. If you wish to change DVs, you must create an EF...