# Homework #3 Project on Regression Analysis by nha12847

VIEWS: 8 PAGES: 1

• pg 1
```									Eco 320, Fall 2002
Prof. Lahiri

Homework #4: Project on Regression Analysis
Posted on the web on 10/30/2002 (Wednesday). Homework #3 is due on
11/08/2002 (Friday).

Use Microsoft Excel to do the calculations. All the results should be
presented, interpreted and printed out in Microsoft Word format. Excel
output can be easily copied and pasted on the Word file. Nothing should be
submitted handwritten. Presentation should be as tidy as possible.

information on 30 baseball teams in year 2000 on number of Wins (Y), and
Total Payroll (X). You are supposed to perform a simple regression analysis
of the performance (number of Wins) of the major baseball teams on the size
of their team Payrolls during the 2000 season.

Based on the data, conduct at least the following analysis:

1. Draw the scatter diagram of Y against X;
2. Obtain descriptive statistics (Mean, Median and Mode, Variance, Std
Dev., etc.);
3. Obtain a and b of the regression equation, Se, Sb, and the Coefficient
of Determination from the Excel regression output;
4. Compute the above statistics in 3 step by step using XiYi, Xi, Yi,
Xi2, Yi2 from Excel, and compare them with the results in 3.
5. Draw the fitted regression line on the scatter diagram, obtain the
residuals and plot them on the scatter diagram too;
6. Statistically test the null hypothesis that Payroll does not affect Wins
at 1% level of significance against the alternative that Payroll does
matter positively (one sided alternative);
7. Calculate the square of correlation coefficient, and check if it is the
same as the coefficient of determination;
8. Test if the correlation coefficient is 0 against the alternative that it is
not at the 10% level of significance;
9. Write a paragraph or so on any observations you may have on the
data, regression estimates or the regression residuals;
10. Calculate the additional number of games New York Mets would have
won if it had spent additional 20 million dollars on Payroll in 2000.

```
To top