Chapter 16, pages 285-286: 1-4, 6-8, 10

Chapter 5, pages 93-96: 1-4, 8, 9, 11

Chapter 17, pages 304-306: 1, 2, 4, 7-9, 12

Chapter 18, pages 327-329: 2, 4-7, 12-14

2) Worksheet

Rick Weiss and Ka Leo O Hawaii, "Class status linked to determining of IQ",Available from LexisNexis here.University Wire,Sept. 10, 2003.

**Note: You may experience a browser crash with Safari
on a Mac. Try Firefox if you have problems.
**

Use the following to search:

Text Box: Class status linked to determining of IQ

Source: All News (English, Full Text)

Date: Previous 10 years

*Remark*: A common IQ test produces scores that
follow a bell (Normal) curve
with average 100 and SD 15.

*FYI*:
Here is an online IQ test.

- The article discussed, briefly, the controversial
book
__The Bell Curve__. This book, in particular, stated that blacks have lower IQs, on average, than whites. It went on further to say that this difference will eventually cause racial stratification and a resurgence of racism. The book authors' argument relied on IQ being genetically determined. How would you use this article to discount__The Bell Curve__'s conclusion. - Given that children from very low income homes score low on IQ tests and that improving their environments will increase their IQ scores, the average IQ of the entire population will increase. What will happen to the SD?
- The article states that
University of Minnesota behavoiral geneticist Irving
Gottesman "noted that [IQ] remains the best predictor today of
social and economic success in U.S. society."
Argue that the correct statement could be
*social and economic success in U.S. society is the best predictor today of IQ*.

**Preliminary Writeup**

Before you do any computations with Excel, what percent of teams
do you expect to have below-average payrolls? And what percent of teams
do you expect to have payrolls below the median? Do you expect the data
to be normally distributed?

Now, suppose the data is normally distributed. What percent of the teams
would have salaries below the 84^{th} percentile? What value would
the 84^{th} percentile be for a normally distributed dataset with
average $82.63 million and an SD of $33.35 million?

**On the Computer**

Copy the 2007 baseball payroll data below into a spreadsheet program:

2007
Baseball payroll (salaries-only list): Move these into the spreadsheet.

- Create a histogram for the payroll data using class intervals of length $20,000,000, up to $200,000,000.
- Compute the value of the average, median, mode, 84
^{th}percentile and standard deviation for the salaries. (The Excel function for what we call SD is "stdevp" [the "p" stands for population]. The function "stdev" gives what we will later in the course call "SD^{+}", the "sample standard deviation".) Which team would you say lies in the 84^{th}percentile? - Determine the percentage of teams with below average payroll.
- Compare the actual data with your predictions. Is payroll data for baseball players approximately normally distributed? Would you say that the payroll data is symmetric, left-sided, or right-sided. Does the rule, the average follows the tail seem to be valid in this situation?
- Find the correlation coefficient between payroll and number of wins. How would you describe the correlation?
- Create a scatterplot for payroll versus number of wins. Let payroll be on the x-axis. Is the use of the correlation coefficient valid, based on your scatterplot? While you're at it, insert the regression line in the scatterplot.
- What would you predict for the payroll of a team with 88 wins? There are 3 teams with 88 wins. Which teams comes closest to your prediction?
- There seems to be one outlier in the scatterplot. Remove this data point and recalculate the correlation coefficient. What do you get? Did the correlation coefficient change in the way (both in size and direction) you expected? Explain.

Last revised February, 2008. Mail to
arobertson@mail.colgate.edu

Copyright 2008 © Colgate University. All rights reserved.