This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Dr. V.R. Bencivenga
Economics 329 Economic Statistics DESCRIPTIVE STATISTICS 3—INDEX CARD QUESTIONS 2 AND 3
IC Q 2
MBA tuition and starting salaries. Business schools that train MBA’s vary a great deal in the
tuition they charge. As one would expect, starting salaries also vary substantially across MBA
programs. Is there a relationship between the two?
Let’s investigate, using Business Week data for 50 business schools. These data are from an
online article, “Return on Investment,” March 7, 2007. If you are interested, the article is
available at: http://www.businessweek.com/magazine/content/07_12/b4026072.htm. The
data are available on Blackboard as an Excel file: MBA salary relative to tuition BUS WK.xls.
The article ranks 29 private schools and 21 public schools (separately) according to the ratio of
median starting salary to tuition. The relationship we are interested in can be expressed as a regression model: SALARYi 0 1TUITIONi ui
where SALARYi = median starting salary of MBA’s graduating from school i
TUITIONi = annual tuition charged by school i
The results of estimating the model on observations on private schools are:
SALARY = 43610 + 0.207 TUITION
The results of estimating the model on observations on public schools are:
SALARY = 45039 + 0.327 TUITION
The results of estimating the model on the entire sample of 50 schools are:
SALARY = 46481 + 0.118 TUITION a. ˆ
Interpret 1 for public schools. Is this value consistent with the hypothesis that a more
expensive MBA is expected to result in a higher starting salary? Explain. According to
these estimated relationships, does an additional dollar of tuition increase predicted
starting salary more for an MBA at a public school or a private school? (Hint: Resources
to help you with this are the lecture slides and the beginning of Ch 11. Example 11.2 on
pg. 430 is a good example to follow.) b. ˆ
Are you surprised at the value of 1 for the entire sample of 50 schools? Can you think of
ˆ
an explanation for why 1 is smaller for the entire sample than it is for public schools or
private schools separately? c. These samples were chosen by Business Week on the basis of the ratio of starting salary
to tuition (large values of the ratio were chosen). Is there any selection issue here that
ˆ
might bias 1 (compared to if we had data from a random sample of business schools)? IC Q 3
You have data on two variables, X and Y. Consider the following two models: Yi 0 1 X i ui
X i 0 1 Yi v i
I.e., one model regresses Y on X, and the other regresses X on Y. Both are estimated with least
1
ˆ 1 if and only if rXY 1 or rXY 1 (i.e., if and only if the
squares. Show that
ˆ
1
observations on X and Y lie exactly on a straight line)! Note: Answering this question will give
you a lot of intuition about correlation—in particular, it will show you why 1 rXY 1 ! ...
View
Full
Document
This note was uploaded on 02/26/2012 for the course ECONOMICS 329 taught by Professor Bencivenga during the Spring '12 term at University of Texas at Austin.
 Spring '12
 BENCIVENGA
 Economics

Click to edit the document details