Present the table with results here the trend coefficient
Interpret estimated coefficients. Interpret R-square and Anova part of the regression table.
You believe that unemployment affects number of accidents with a one-year lag. Write down the equation of the model you will use to estimate such effect. You don’t need to estimate this model, just specify your dependent variable, and independent variables using the equation.
Use joint hypothesis testing to verify if the seasonality is present (Hint: you will be testing if the seasonal coefficients are jointly equal to zero). Please make sure to present all the tables with estimated coefficients used to answer this question here.
Question 4 (25 points) A well known consulting company is interested in comparing the performance of the consultants with MBA degrees and consultants with non-MBA degrees (Ph.D.s, or M.S. degrees)You collect the data on 130 consultants. For each person, you get three pieces of information:
Billing = β0 + β1Experience + β2MBA + β3MBA * Experience + u
You estimate the following regression with the available data and get the following results:
Coefficient estimate | S.E. | T-stat | p-value | |
---|---|---|---|---|
Intercept (β0) | 44.13 | 12.4 | ||
Experience (β1) | 9.07 | 0.62 | ||
MBA (β2) | 68.43 | 20.53 | ||
Experience*MBA (β3) | -1.43 | 0.51 | ||
R-squared (adj) | 0.86 | |||
Number of obs | 130 |
The sample used in estimation consisted of consultants who have been at the company for up to five years. Suppose you are asked to use your results to predict what the difference in billing will be for those with 10 years of experience. What would the estimated regression predict? Use your judgement to discuss the validity of this prediction.
Question 5. (25 points)
Exam 1 is the student’s grade on Exam 1 (out of possible 25 points)
Quiz is the student’s average quiz grade (out of possible 5 points).
I have noticed that there is high positive correlation between quiz and homework scores. What kind of problem does it presents for the regression estimation? How would you deal with this problem?
I am also thinking that the homework score affects male and female students differently. If so, which regression would I be estimating? (you just need to write down the regression equation, you do not need to estimate it.) Explain.