LUBS5134M-Regression
| On Minerva you will find a data file InstructorRatings.dta that contains data on course evaluations, course characteristics, and instructor characteristics for 463 courses at the University of Texas at Austin. A detailed description is given in InstructorRatingsDescription.pdf. One of the characteristics is an index of the instructor’s “beauty” as rated by a panel of six judges. In this exercise you will investigate how course evaluations are related to the instructor’s characteristics and characteristics of the course. The dataset is a modified version of the data used in a paper by Hamermesh and Parker (2005). Hamermesh, Daniel and Amy Parker (2005). “Beauty in the Classroom: Instructors’ Pulchritude and Putative Pedalogical Productivity”, Economics of Education Review 24(4), pp. 369-376. Read the paper by Hamermesh and Parker (2005) before proceeding to the questions below. The assignment consists of three questions. The total mark is 90. |
Assignments should be a maximum of 2000 words in length. The maximum word length for the answer to question 3 is 750 words.
Question 1. [20 Marks]
Run a regression of course_eval on beauty. Answer parts a)-d).
- Derive the normal equations corresponding to this estimation and explain your understanding of them. [7]
- Obtain the expression for the sample mean of course_eval in terms of the estimated model parameters. Comment on this. [5]
- Construct and implement a two-tailed t-test to determine whether the estimated coefficient on beauty is significant at the 5% level. Comment on your finding. [5]
- Calculate the p-value from a one-tailed t-test to determine whether the estimated coefficient on beauty is statistically greater than zero at the 1% level. [3]
Question 2. [35 Marks]
Run a regression of course_eval on beauty and include all available control variables. Answer parts a)-e).
- Describe a suitable measure of the model’s fit to the data. Comment on the fit indicated by your measure. [5]
- What is the nature of the relationship between an instructor’s age and course_eval? Construct a test for the joint significance of age and age2 on course_eval. Implement the test and comment on your finding. [7]
- Professor Jones is a 40 year old black female with average beauty. She teaches a one-credit upper-division course. Use the estimation output to predict Professor Jones’s course evaluation. [5]
- Professor Jones has a makeover and her beauty increases by one standard deviation. What is the expected increase in her evaluation arising from the makeover? Construct the 90% confidence interval for this increase. [8]
- Explain how you could test whether beauty’s effect on course evaluation differs for men and women. Conduct the test and comment on your findings. [10]
Question 3. [35 Marks]
In no more than 750 words critically evaluate estimations when a regression of course_eval on beauty is conducted (Question 1) and when all the available control variables are included (Question 2).
END