Discrepancies, Control, and RegressionA psychologist playing the variable-based research game wishes to predict positive affect fromthe discrepancy between the actual and ideal selves with regard to conscientiousness.Specifically, he hypothesizes that large discrepancies between the actual and ideal selves willcorrespond with lower positive affect. This is similar to Carl Rogers’ old idea that greater self-discrepancies will accompany higher depression and other clinical symptoms. The psychologistencounters four opinions with regard to how he should analyze his data. These opinions arebelow.1. One quant advisor suggests that he simply compute the difference score between the ideal andactual selves and then correlate the difference score with the positive affect scores. Do this inSPSS and report your results. The data set is located on the course website(DASBigFiveStudy3.sav). The variables are: PA_1 (positive affect); C2_E2 (Ideal self onconscientious “measure”); C2_E1 (Actual self ratings on conscientious “measure”). Computethe difference score: C2_E2 - C2_E1. Correlate the difference score with PA_1. Create ascatterplot as well to look for anomalies. Summarize the results. 2. A second quant advisor suggests that he should instead run a hierarchical regression toexamine the variance explained by each of the selves. Does the ideal self contribute significantlyto the prediction of PA_1 above and beyond the actual self ratings? Run a 2-step hierarchicalregression analysis. In step 1 regress PA_1 onto C2_E1 (actual self). In step 2 add in the C2_E2(ideal self) variable. Request your R-squared change statistic, and run all the appropriatediagnostics on the final model. Summarize your results. Did the ideal self ratings contribute tothe prediction of PA_1?3. A third quant advisor suggests that he should run a hierarchical regression analysis to examineif the difference score contributes anything unique to the prediction of PA_1. Run a 2-stephierarchical regression. In step 1 regress PA_1 onto C2_E1. In step 2, include the differencescore that you computed for #1 above. Request your R-squared change statistic, and run all theappropriate diagnostics on the final model. Summarize your results. Did the discrepancycontribute to the prediction of PA_1?4. Finally, a fourth advisor agrees with advisors #2 and #3 above. In discrepancy research theidea is to show that the difference is predictive in a way that goes beyond the actual or idealratings alone. He claims, however, that running a simple regression analysis will do the trick:PA_1 = C2_E1, C2_E2. Run this model and diagnose it for anomalies. Also, be sure to examinethe regression weights and whether or not they are significant and opposite in sign, as expected.What are the squared semi-partial correlations for each predictor? Do the results suggest that thediscrepancy matters, or is it only one predictor that counts. Compare your various models, R statistics, regression weights, and conclusions for #1-#4.2Which advisor offers the best advice?As another example of using hierarchical regression, a psychologist wishes to predict InternalReligiosity from the Big Five personality traits. It is well known, however, that positive andnegative affect correlate with introversion/extraversion. Consequently, the researcher wishes tocontrol for the variance explained by affect when predicting religiosity. Do this with the samedata set in SPSS. Here are the relevant variables:1. Internal Religiosity: IntRel_1 (this will be your DV)2. Positive Affect: PA_13. Negative Affect: NA_14. Big Five self ratings: C1_E1 (openness), C2_E1 (conscientiousness), C3_E1 (extraversion), C4_E1 (agreeableness), C5_E1 (neuroticism)You need to run a 2-step hierarchical regression analysis. Be sure to diagnose your final model,as usual and report an anomalies. Summarize your