The purpose of this project is to review and to force you to work through everything weâ€™ve learned in chapters 7 and 8 regarding Multiple Regression and Path Analysis. I will grade this as if itâ€™s been submitted as part of your dissertation in an effort to give you feedback as to what weâ€™re looking for with regards to presenting this material. You may use one of the data sets provided by the text, but try to choose either â€˜aâ€™ or â€˜bâ€™ so that the data is a bit more raw, which results in more prescreening on your part (the latter data sets have already prescreened and transformed the majority of their variables). Please address the following topics:
The first portion of the project will require you to build a best fit model (equation of a line) using multiple regression that contains at least 4 input variables in an attempt to best predict one output variable. Choose one variable (DV) as the one that youâ€™d like to predict by the other variables (IVs). The second portion will have you build a path diagram that explains the causal relationships between the variables. The same variables may be used for both portions of the project.
1) Data Screening:
â€¢ Are there any missing data points?
â€¢ Are there any outliers? (compute the Mahalanobis distance for all variables, independent and depen- dent, and look up the cutoff for removal from the table on page 357).
â€¢ Create a scatterplot matrix of all IVs and the DV. Any â€™boomerangâ€™ shaped plots should be transformed with a natural log transformation.
â€¢ Run normality plots with tests and interpret Kolmogorov-Smirnov.
â€¢ Create and display a residual plot, and conclude whether there is linearity, normality, and homoscedas-
ticity.
2) Multiple Regression:
â€¢ You will complete this twice, the difference being the method of input (everything else will be identical).
â€¢ Run multiple regression using the â€™enterâ€™ method. Every variable will be included in your model. Display the best fit line as an equation, and identify which variables are significant (although you will still include all variables in your model).
â€¢ Run multiple regression again using the â€™forwardâ€™ method. You may use the default cutoffs for telling SPSS when to stop adding variables to the model.
â€¢ Your output should not display several models (each one adding an additional variable to the equation). Display the best fit line from the final model, and identify which variables were excluded as insignificant.
3) Path Analysis:
â€¢ You may use the same independent variables and dependent variable as you did in the previous step.
â€¢ Build a lattice (hubs for your path diagram) that you believe accurately displays the flow of the causal relationships between the variables (all of this may be hand drawn).
â€¢ You may input arrows in your path diagram that you feel accurately represent the flow, but you will have to revise them if your model does not accurately predict the observed correlations.
â€¢ Compute a correlation coefficient matrix that you will use as your observed correlations.
â€¢ Run multiple regression on each endogenous variable (every variable with an arrow pointing into it),
and transfer path coefficients to your path diagram.
â€¢ Create a path decomposition table that accounts for every legal path in your diagram.
â€¢ Calculate the reproduced correlations based on your path decomposition table and the coefficients in your diagram.
â€¢ Identify which reproduced correlations differ from the observed correlations by more than .05.
â€¢ If necessary, revise your diagram and repeat the process. Determine which paths are significant and
need to be added to your diagram, and which are insignificant and should be deleted.
â€¢ If necessary, complete the process again, creating a new path decomposition table that accounts for every path in your updated diagram. Calculate the new reproduced correlations, and identify which of them differ from the observed correlations by more than .05.
â€¢ Display your results (observed correlations vs reproduced correlations) in a table similar to that on page 218 and 240.
4) Summary:
Finish the project by summarizing what you have found. You are expected to report your findings in a format similar to the Results section of a research journal, as you will do in your disseration. The textbook provides examples of how to write up your results at the end of each chapter under â€™Presentation of Resultsâ€™. Discuss the model that you have built, and how it accurately predicts (in the case of multiple regression) and explains (in the case of path analysis) the value of the dependent variable from values of the independent variables.

The purpose of this project is to review and to force you to work t...