Please only accept this question if you have the eviews software. question can only be answered via eviews. attached will be an excel file for the data to use with the eviews software.
1.Working in the provost’s office at NU, your job is to use regression to empirically estimate the impacts of different explanatory variables on NU students’ GPA. Use the EViews data file labeled “HW1.wf1� (found in Canvas) to complete the following questions.
a.You are uncertain whether to use colgpa1 or colgpa2 as your dependent variable. Find the descriptive statistics for both variables and paste your results into a word document. Which of the two series should you use and why?
b.Use EViews to create a scatterplot of your preferred GPA measure and the student’s ACT score (act). Paste your graph into a word document. In part c, you will run a regression that looks at how different variables (including ACT score) predict a students’ college GPA Based on your scatterplot, would you expect the estimated coefficient on ACT to be positive or negative?
c.Start by running a regression with the following independent variables: gender (male), student age (age), ACT score (act), average number of lectures missed/week (skipped), and the average number of days/week that the student consumed alcohol (alcohol). Use your preferred GPA measure as the dependent variable. Paste your regression results from EViews.
d.Consider a 21-year-old male student that had an ACT score of 26, never skips class, and consumes alcohol two days per week. Based on your model, what is his predicted GPA? If his actual GPA is 3.75, compute the value of his residual? (Note, male = 1 for male students)
e.You should find that the coefficient on alcohol is positive. Why might you be hesitant to tell your boss that one way to increase student GPA is to offer students free beer and wine (which, presumably, would increase the number of days/week that students consume alcohol)?
f.We could also include a variable for a student’s relationship status (bgfriend). This variable takes on a value of 1 if the student is in a relationship and 0 for single students. What special type of variable is bgfriend? How should this variable impact our regression equation?
g.Rerun the regression with bgfriend added to your independent variables. Paste your new EViews regression results.
h.Based on your result in part g, do students in relationships earn a higher or lower GPA on average? Is this true at any value of the other variables (such as all possible ACT scores)?