TI Technology Manual

to Accompany

Mind on Statistics

Cengage Learning. All rights reserved. No distribution allowed without express authorization.

FIFTH EDITION

Jessica M. Utts
University of California, Irvine
Irvine, CA

Robert F. Heckard
Pennsylvania State University
State College, PA

Prepared by
Melissa M. Sovak
California University of Pennsylvania, California, PA

Australia Brazil Mexico Singapore United Kingdom United States

ISBN-13: 978-1-285-83862-5
ISBN-10: 1-285-83862-9

2015 Cengage Learning

ALL RIGHTS RESERVED. No part of this work covered by the
copyright herein may be reproduced, transmitted, stored, or
used in any form or by any means graphic, electronic, or
mechanical, including but not limited to photocopying,
recording, scanning, digitizing, taping, Web distribution,
information networks, or information storage and retrieval
systems, except as permitted under Section 107 or 108 of the
1976 United States Copyright Act, without the prior written
permission of the publisher except as may be permitted by the
license terms below.

For product information and technology assistance, contact us at

Cengage Learning Customer & Sales Support,
1-800-354-9706.
For permission to use material from this text or product, submit
all requests online at www.cengage.com/permissions
Further permissions questions can be emailed to
[email protected].

Cengage Learning
200 First Stamford Place, 4th Floor
Stamford, CT 06902
USA
Cengage Learning is a leading provider of customized
learning solutions with office locations around the globe,
including Singapore, the United Kingdom, Australia,
Mexico, Brazil, and Japan. Locate your local office at:
www.cengage.com/global.
Cengage Learning products are represented in
Canada by Nelson Education, Ltd.
To learn more about Cengage Learning Solutions,
visit www.cengage.com.
Purchase any of our products at your local college
store or at our preferred online store
www.cengagebrain.com.

NOTE: UNDER NO CIRCUMSTANCES MAY THIS MATERIAL OR ANY PORTION THEREOF BE SOLD, LICENSED, AUCTIONED,
OR OTHERWISE REDISTRIBUTED EXCEPT AS MAY BE PERMITTED BY THE LICENSE TERMS HEREIN.

READ IMPORTANT LICENSE INFORMATION

Dear Professor or Other Supplement Recipient:
Cengage Learning has provided you with this product (the
Supplement) for your review and, to the extent that you adopt
the associated textbook for use in connection with your course
(the Course), you and your students who purchase the
textbook may use the Supplement as described below.
Cengage Learning has established these use limitations in
response to concerns raised by authors, professors, and other
users regarding the pedagogical problems stemming from
unlimited distribution of Supplements.
Cengage Learning hereby grants you a nontransferable license
to use the Supplement in connection with the Course, subject to
the following conditions. The Supplement is for your personal,
noncommercial use only and may not be reproduced, or
distributed, except that portions of the Supplement may be
provided to your students in connection with your instruction of
the Course, so long as such students are advised that they may
not copy or distribute any portion of the Supplement to any third
party. Test banks, and other testing materials may be made
available in the classroom and collected at the end of each class
session, or posted electronically as described herein. Any

TI is a trademark of Texas Instruments.

Printed in the United States of America

1 2 3 4 5 6 7 17 16 15 14 13

material posted electronically must be through a passwordprotected site, with all copy and download functionality disabled,
and accessible solely by your students who have purchased the
associated textbook for the Course. You may not sell, license,
auction, or otherwise redistribute the Supplement in any form. We
ask that you take reasonable steps to protect the Supplement from
unauthorized use, reproduction, or distribution. Your use of the
Supplement indicates your acceptance of the conditions set forth in
this Agreement. If you do not accept these conditions, you must
return the Supplement unused within 30 days of receipt.
All rights (including without limitation, copyrights, patents, and trade
secrets) in the Supplement are and will remain the sole and
exclusive property of Cengage Learning and/or its licensors. The
Supplement is furnished by Cengage Learning on an as is basis
without any warranties, express or implied. This Agreement will be
governed by and construed pursuant to the laws of the State of
New York, without regard to such States conflict of law rules.
Thank you for your assistance in helping to safeguard the integrity
of the content contained in this Supplement. We trust you find the
Supplement a useful teaching tool.

Contents

Chapter 1: Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition ......... 1
Chapter 2: Turning Data into Information ..................................................................................... 2
Chapter 3: Relationships Between Quantitative Variables .......................................................... 36
Chapter 4: Relationships Between Categorical Variables ........................................................... 56
Chapter 5: Sampling: Surveys and How to Ask Questions.......................................................... 67
Chapter 6: Gathering Useful Data For Examining Relationships ................................................ 72
Chapter 7: Probability .................................................................................................................. 77
Chapter 8: Random Variables ...................................................................................................... 82
Chapter 9: Understanding Sampling Distributions: Statistics as Random Variables................... 97
Chapter 10: Estimating Proportions With Confidence................................................................. 108
Chapter 11: Estimating Means With Confidence ........................................................................ 117
Chapter 12: Testing Hypotheses About Proportions.................................................................... 136
Chapter 13: Testing Hypotheses About Means ........................................................................... 150
Chapter 14: Inference About Simple Regression ......................................................................... 167
Chapter 15: More about Categorical Variables............................................................................ 190
Chapter 16: Analysis of Variance ................................................................................................ 200
Appendix: Troubleshooting the TI-83 and TI-84........................................................................ A1

Chapter 1
Introduction to the TI-83 Plus
Silver Edition and the TI-84
Plus Silver Edition
1.1

Getting Started
This chapter represents a brief introduction to the TI-83 Plus Silver Edition (hereafter referred to the TI-83 Plus SE) and the TI-84 Plus Silver Edition (hereafter
referred to the TI-84 Plus SE). Basic commands, techniques and the use of lists are
discussed briefly in this introduction. Detailed descriptions of built in calculator
functions are given in the TI-83 Plus SE and TI-84 Plus SE guidebooks.
After reading this chapter you should be able to:
1. Turn the calculator on and off.
2. Adjust the display contrast.
3. Evaluate an expression.
4. Use last entry to edit an expression and evaluate and expression.
5. Access menu options.
6. Display the mode settings.
7. Graph a function.
8. Enter a list.
9. Plot a statistical data set.
10. Save a list using a descriptive name.
11. Clear lists.

1.2

Features
The keypad on the TI-83 Plus SE and TI-84 Plus SE are virtually identical. The
TI-84 Plus SE, TI-83 Plus SE, and the TI-83 Plus are keystroke-for Keystroke compatible. The keyboard is divided into zones: graphing keys, editing keys, advanced
function keys, and scientific calculator keys. The graphing keys access the interactive graphing features and are located on the first row at the top of the keyboard.
The editing keys allow you to edit expressions and values and are located on the
second and third rows below the graphing keys. The advanced function keys display menus that access the advanced functions: MATH, APPS, PRGM, VARS and
are located on the fourth row below the graphing keys. The scientific calculator
1

Chapter 1

Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition

keys access the capabilities of a standard scientific calculator and are the remaining keys located on rows five through ten.
The TI-83 Plus SE and TI-84 Plus SE uses Flash technology, which lets you upgrade to future software versions without buying a new graphing handheld calculator. As new software becomes available, you can electronically upgrade your
TI-84 Plus from the Internet. The primary differences between the TI-83 Plus SE
and the TI-84 Plus SE occur in
1. The TI-83 Plus SE is preloaded with one application. The TI-84 Plus SE is
preloaded with numerous applications.
2. The TI-83 Plus SE uses a TI-Graph link that is available as an accessory from
TI. The TI-84 Plus SE comes with a USB unit-to-unit cable to connect and
communicate with another TI-84 Plus Silver Edition. With TI Connect software and a USB computer cable, you can also link the TI-84 Plus SE to a
personal computer.

1.3

The Basics
Keystrokes Introduced
1. ON turns on the calculator.
2. 2nd OFF turns the calculator off.
3. 2nd N darkens the screen; 2nd H lightens the screen.
4. 2nd MEM accesss the MEMORY menu.
5. 2nd QUIT returns to the home screen.
6. ENTER may be used to evaluate an expression or execute a menu option.
7. 2nd ENTER recalls the last entry.
8. STAT displays the STAT menu.
9. ALPHA H moves the cursor down one screen at a time.
10. MODE displays the mode settings.
11. Y= displays the Y= editor.
12. WINDOW displays the current window variable values.
13. GRAPH displays the graph of a selected function.
14. ZOOM >ZStandard sets the standard window variables.
To turn on the calculator press the ON key, and the key sequence 2nd OFF
turns the calculator off. There is a battery saving feature on the calculator that will
automatically turn off the TI-83 Plus SE and the TI-84 Plus SE.

1.3

Evaluating Expressions

The 2nd key located on the top left and the up and down cursor movement keys
located on the top right portion of the keypad are used to adjust the screen contrast. The keystrokes 2nd N darken the screen and 2nd H lighten the screen.
This keystroke sequence, when repeated, will continuously darken or lighten the
screen.
You can adjust the display contrast to suit your viewing angle and lighting conditions. As you change the contrast setting, a number from 0 (lightest) to 9 (darkest)
in the top-right corner indicates the current level. You may not be able to see the
number if contrast is too light or too dark. Both the TI-83 Plus SE and the TI-84
Plus SE have 40 contrast settings, so each number 0 through 9 represents four settings. When the batteries are low, a low-battery message is displayed when you
turn on the calculator.
Variables (real or complex number, list, matrix, Y= variable, program, Apps, AppVars, picture, graph database, or string) stored in the calculator may be selectively
deleted. The 2nd MEM keystrokes access the MEMORY menu as shown in
Figure 1.1.

Figure 1.1

Home Screen
The home screen is the primary screen of the TI-83 Plus SE and the TI-84 Plus SE.
The appearance of the cursor indicates what will happen when you press the next
key or select the next menu item to be pasted as a character on the home screen.
On this screen, you may enter instructions to execute and evaluate expressions.
Answers are displayed on this home screen. The blinking rectangular cursor, ,
indicates the calculator is ready to accept commands. To return to the home screen
from any other screen, use 2nd QUIT .

Evaluating Expressions
The order of operations applies to all expressions entered into the calculator. Parentheses should be used to ensure the desired order of operations, with the grey negation key - , being used for negation. The grey negation key is located on the bottom row, column four of the keyboard. After entering an expression, press the
ENTER key to evaluate the expression. Figures 1.2 and 1.3 illustrate several

Chapter 1

Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition

arithmetic calculations.

Figure 1.2

Figure 1.3

Last Entry
When you press ENTER on the home screen to evaluate an expression or execute
an instruction, the expression or instruction is placed in a storage area called ENTRY (last entry). When you turn off the TI-84 Plus, ENTRY is retained in memory.
To recall ENTRY, press 2nd ENTER . The last entry is pasted to the current cursor location, where you can edit and execute it. On the home screen or in an editor,
the current line is cleared and the last entry is pasted to the line.
Example 1.1 A dataset consists of handspan values in centimeters for six females;
the values are 21, 19, 20, 20, 29, and 19. The mean is the numerical average, calculated as the sum of the data values divided by the number of values. (Utts/Heckard,
Statistical Ideas and Methods, p32)
Follow these steps to learn the process of editing an expression.
1. Enter the data and determine the mean.
Enter the data as shown in Figure 1.4. Press ENTER to evaluate the expression.
2. Edit the expression.
An error was found in the data recording. Examination of the data indicates that
the 29 should actually be a 22. Press 2nd ENTRY to display the expression
once again. Use the up arrow key, N , placing the cursor on the 9 of the value
29. Change the 29 to 22. Press ENTER to evaluate the revised expression.
This process is illustrated in Figures 1.4, 1.5, and 1.6.

Figure 1.4

Figure 1.6

Figure 1.6

Menus
You can access the TI-83 Plus SE and TI-84 Plus SE operations using menus. When
4

1.3
you press a key or key combination to display a menu, one or more menu names
appear on the top line of the screen. When you press a key that displays a menu, that
menu temporarily replaces the screen where you are working. For example, when
you press STAT , the STAT menu is displayed as a full screen, as shown in Figure
1.7. The current, or active, menu will be highlighted or darkened. The left and right
arrow keys, J and I , move the cursor to the other menu options. To select a
menu option, press the number of the menu option desired, or move the cursor up
or down with the arrow keys, N and H to highlight the desired selection and
press ENTER . Observe that if the left-most menu option is highlighted, pressing
the left arrow, J , causes the cursor to highlight the right-most menu option. If
more than a screen-full of menu options press ALPHA H to move down one
screen at a time.

Figure 1.7

Figure 1.8

Display Modes
Mode settings control how the TI-83 Plus SE and TI-84 Plus SE displays and interprets numbers and graphs. Mode settings are retained by the Constant Memory
feature when the TI-83 Plus SE and TI-84 Plus SE is turned off. All numbers, including elements of matrices and lists, are displayed according to the current mode
settings. The MODE key, 2nd row, 2nd column, is used to view and/or change
the mode settings. To select a particular setting, move the cursor with the arrow
keys to the desired option and press ENTER to highlight that option. Once you
have selected the desired settings, press 2nd QUIT . Recommended settings are
shown in Figure 1.8.

Graphing
You can store, graph, and analyze up to 10 functions, up to six parametric functions, up to six polar functions, and up to three sequences. You can use DRAW
instructions to annotate graphs. Mode settings must be changed appropriately.
Functions
You can store, graph, and analyze up to 10 functions, up to six parametric functions, up to six polar functions, and up to three sequences. You can use DRAW
instructions to annotate graphs. Mode settings must be changed appropriately.
Example 1.2 Normal random variables are the most common type of continuous

Functions

Chapter 1

Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition

random variables. The bell-shaped normal curve illustrates the distribution of these
normal random variables. (Utts/Heckard, Statistical Ideas and Methods, p268)
Follow these steps to graph a normal probability function y =

1 2
s1 e 2 x .
2

1. Enter the function.

Press Y= , row 1, column 1, to enter the function, as shown in Figure 1.9.
p
Press (1/ (2))e(1/2)(x2 ). The left and right parenthese are located on
row 6. Press 2nd ,  is located on the 5th row, right column above the ^
key. Press 2nd e, e is located on the 8th row, left column above the LN key.
Be sure to use the grey negation key when you enter (1/2).
2. Set the Window viewing variables in order to view the graph.
Press WINDOW , row 1, column 2. Set Xmin to -3, being sure to use the grey
negation key. Set Xmax to 3; Xscl to 1; Ymin to -0.2, again being sure to use
the grey negation key. Set Ymax to 0.5; Yscl to 1; Xres to 1. These settings
are illustrated in Figure 1.10
3. View the graph.
Press GRAPH , row 1, column 5. The graph of the normal curve is shown in
Figure 1.11.

Figure 1.9
Figure 1.10
Figure 1.11
4. Set the graph window to standard viewing and clear the function .
Press ZOOM and select 6: ZStandard to restore the default graph window
settings. Press Y= and press CLEAR to remove the function.

1.4

Statistics
Keystrokes Introduced
1. STAT displays the Stat menu.
2. STAT >CALC displays the STAT >CALC menu.
3. 2nd DISTR displays the distributions menu.
4. VARS displays the VARS menu.
5. DISTR >DISTR displays menu options for calculating values of common
probability distributions.

1.4

Plotting Statistical Data

6. DISTR >DRAW displays menu options for shading areas under a probability
distribution function.
7. 2nd STAT PLOT displays statistical plot options.
8. STO stores values to a list or a single value to a variable.
9. ZOOM >ZoomStat redefines the viewing window so that all statistical data
points are displayed.
10. TRACE may be used to trace a plot of statistical data.
11. ClrList clears from memory the elements of one or more listnames.
12. 2nd A-LOCK sets alpha lock on; ALPHA turns alpha lock off when alpha
lock is on.
13. STAT >SetUpEditor clears the list editor and restores the built in list L!-l6.
The TI-83 Plus SE and TI-84 Plus SE have several functions for analyzing data.
Many of these functions are contained in the STAT >CALC and STAT >TESTS
menu options. The STAT key is located on the 3rd row, 3rd column. These menus
are shown in Figure 1.12 and Figure 1.13. These functions provide summary statistics, regression lines, confidence intervals, hypothesis tests, and analysis of variance.
Other statistical functions are contained in the 2nd DISTR menu, located on the
4th row, 4th column above VARS . . DISTR >DISTR provide menu options
for calculating values of common probability distribution functions, and is shown
in Figure 1.14; DISTR >DRAW provide menu options for shading areas under a
probability distribution function, and is shown in Figure 1.15.

Figure 1.12

Figure 1.13

Figure 1.14

Figure 1.15

Plotting Statistical Data

You can plot statistical data by selecting 2nd STAT PLOT , located directly over
Y= .The 2nd STAT PLOT menu options provides access to statistical plot op7

Chapter 1

Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition

tions and the capability of turning on/off all statistical plots, as shown in Figure
1.16. One, two, or all three statistical plots may be displayed on the screen simultaneously. The TI-83 Plus SE and TI-84 Plus SE can display a scatter plot, xyLine,
histogram, modified box plot, regular box plot, and normal probability plot.

Figure 1.16
Lists
Lists represent a set of observations. A list may contain up to 999 numerical values
and is the principal way to store data for analysis. Many of the built-in statistical
functions and programs operate on data sorted in a list or lists.The TI-83 Plus SE
and TI-84 Plus SE have six list names in memory: L1, L2, L3, L4, L5, and L6.
The list names L1 through L6 are on the keyboard above the numeric keys 1
through 6 . To paste one of these names to a valid screen, press 2nd , and then
press the appropriate key. L1 through L6 are stored in stat list editor columns 1
through 6 when you reset memory. Lists may also be created with a descriptive
name. The name must be a string of up to 5 characters. The first letter must be
a letter which may be folowed by letters, numbers, or . The number of lists is
limited by available memory. Lists may be created on the home screen, or in the
STAT list editor.
Example 1.3 Here are the weights (in pounds) of 18 men who were on the crew
teams at Oxford and Cambridge universities (The Independent, March 31, 1992),
also Hand, D. J. et al., 1994, p337.): (Utts/Heckard, Statistical Ideas and Methods,
p27)
Cambridge 188.5 183.0 194.5 185.0 214.0
203.5 186.0 178.5 109.0
Oxford

186.0
202.5

184.5
174.0

204.0
183.0

184.5
109.5

195.5

Follow these steps to create two lists.

1. Create list L1 on the home screen.
On the home screen, curly braces ({}) are used to enclose lists. Numbers are
seperated by commas. Enter the weights for Cambridge within curly braces
seperated by commas, as shown in Figure 1.17. Store the list by using the
keystrokes STO ; 2nd L1 ; ENTER , storing the data in list L1. After
pressing ENTER , the contents of the list are displayed on the home screen.
Note that spaces rather than commas seperate values in a displayed list. You
8

1.4
may use the left and right arrow keys, J and I ,to scroll through the list.
2. Create list L2 using the STAT list editor.
Press STAT ENTER to select the STAT list editor. Note that the weights for
Cambridge are displayed in list L1. Place the cursor on list L2 row 1 to make
L2(1) the active list row, as shown in Figure 1.18. Enter the weights for Oxford
pressing ENTER after each entry. The list is partially entered in Figure 1.19.
Press 2nd QUIT to quit the STAT editor.

Figure 1.17
Figure 1.18
Figure 1.19
3. Plot the statistical data by creating modified box plots for the weights of the
crew teams at Oxford and Cambridge universities.
Press 2nd STAT PLOT accessing the StatPlot menu, as shown in Figure 1.20.
Press ENTER , selecting Plot 1. Place the cursor on ON and press ENTER .
Use the down arrow key and the right arrow key to select the first icon in the
second row, the modified box plot. Press ENTER . Use the down arrow key
to select L1 as the list, 2nd L1 , as shown in Figure 1.21.
Use the up arrow key to place the cursor on Plot2. Place the cursor on ON and
press ENTER . Use the down arrow key and the right arrow key to select the
first icon in the second row, the modified box plot. Press ENTER . Use the
down arrow key to select L2 as the list, 2nd L2 , as shown in Figure 1.22.
Press ZOOM , ZoomStat to view the graph, as shown in Figure 1.23.

Figure 1.20

Figure 1.21

Figure 1.22

Figure 1.23
9

Lists

Chapter 1

Introduction to the TI-83 Plus Silver Edition and the TI-84 Plus Silver Edition

4. Identify outliers.
Press the TRACE key and the left arrow key to identify an outlier (109.0) in
list L1. Use the down arrow key and the left arrow key to identify an outlier
(109.5) in list L2.
5. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
6. Save list L1 as CAMBR and list L2 as OXFRD.
Press 2nd L1 STO 2nd A-LOCK and type CAMBR; press ENTER .
Press 2nd L2 STO 2nd A-LOCK and type OXFRD; press ENTER .
These outliers indicate that the last weight given in each list is very different
from the others. In fact, those two men were the coxswains for their teams, while
the other men were the rowers.
Clearing Lists
To clear all of the entries in a list, or lists, press STAT , selecting 4: ClrList. Press
the appropriate key. L1 through L6, as shown in Figure 1.24. To clear all lists,
press STAT , selecting 4: ClrList. Press the appropriate keys, seperating each list
name by a comma, as shown in Figure 1.25.

Figure 1.24

Figure 1.25

Dispaying Lists
The menu option of STAT , SetUpEditor used without any arguments clears the
list editor and restores the built in lists L1-L6. SetUpEditor followed by a sequence
of up to 20 lists replaces the stat list editor with the new sequence of lists.

10

Chapter 2
Turning Data Into Information
2.1

Introduction
In this chapter, you will learn how to create simple summaries and pictures from
various kinds of raw data.
After reading this chapter you should be able to:
1. Change frequencies to a percentage falling into each category.
2. Create a bar chart for a single categorical variable.
3. Create a bar chart displaying two categorical variables.
4. Obtain the five-number summary for quantitative data.
5. Plot statistical data by creating a histogram for a quantitative variable.
6. Create comparative boxplots for quantitative variables.
7. Draw a histogram with s superimposed normal curve.
8. Calculate the variance and standard deviation for a small data set.

2.2

Raw Data
Raw data is a term used for numbers and category labels that have been collected
but have not yet been processed in any way. For example, here is a list of questions
asked in a large statistics class and the raw data given by one of the students:
Question
Raw Data
1. What is your sex (m = male,f = female)?
m
2. How many hours did you sleep last night?
5 hours
3. Randomly pick a letter-S or Q.
S
4. What is your height in inches?
67 inches
5. Randomly pick a number between 1and 10.
3
6. Whats the fastest youve ever driven a car (mph)? 110 mph
7. What is your right handspan In centimeters?
21.5 cm
8. What is your left handspan in centimeters?
21.5 cm

2.3

Types of Variables
Different types of summaries are appropriate for different types of variables. It
makes sense, for example, to calculate the average number of hours of sleep last
night for the members of a group, but it doesnt make sense to calculate the average
sex (male, female) for the group. For gender data, it makes more sense to determine

11

Chapter 2

Turning Data Into Information

the proportion of the group thats male and the proportion thats female.
We learned in a previous section that a variable is a characteristic that differs from
one individual to the next. A variable may be a categorical characteristic, like a
persons sex, or a numerical characteristic, like hours of sleep last night.
Example
Raw data from categorical variables consist of group or
category names that dont necessarily have a logical ordering. eye color
Categorical variables for which the categories have
a logic ordering are called ordinal variables.

highest degree
earned

Raw data from quantitative variables consist of numerical

values taken on each individual
height in inches
TI calculators allow only for numerical values to be used in a statistical analysis.
For example, the text Male or Female can not be used for the Sex variable
in the PennState1 worksheet. Neither can we use the letters M or F since
these letters are replaced by the value stored in memory for the M and F variables in the calculator.
The solution to the problem is to assign a unique numerical code for each value of
the variable. In this case, you might code Male = 0 and Female = 1 on the TI
calculator.
Values of the other categorical categorical variables (SQpick and Form) in the
PennState1 worksheet could also be coded. For example, you might code S = 0
and Q = 1. Other numerical values could also be used.
The quantitative variables in the PennState1 worksheet: (Hours of sleep the previous night, Choice of either S or Q, Reported height, inches , Random pick of
a number between 1 and 10, Fastest speed ever driven, mph, Measured stretched
right handspan, cm, Measured stretched left handspan, cm ) can be handled by the
TI calculator without coding.

2.4

Summarizing One or Two Categorical Variables

Numerical Summaries
To summarize a categorical variable, the first step is to count how many individuals
fall into each possible category. Percents usually are more informative than counts
so the second step is to calculate the percent in each category. These two easy steps
can also be used to summarize a combination of two categorical variables.
Keystrokes Introduced
1. 2nd LIST >MATH>sum(

returns the sum of the elements within a list.

2. ZOOM >ZStat redefines the viewing window so that all statistical data points
12

2.4

Numerical Summaries

are displayed.
3. 2nd STAT PLOT accesses the StatPlot menu.
4. 2nd DRAW > Text(

draws text on a graph screen.

5. STAT >CALC> 1: 1-VarStats analyzes data for one quantitative variable.

6. 2nd LIST I , OPS. Select 1: SortA(
order.

sorts elements of a list in ascending

Example 2.1 Seatbelt Use by 12thGraders

How often do you wear a seatbelt when driving a car? This is one of many questions asked in a biennial nationwide survey of American high school students. The
survey, conducted as part of a federal program called the Youth Risk Behavior Surveillance System (YRBSS), is sponsored and organized by the U.S. Centers for
Disease Control (CDC). Survey questions concern potentially risky behaviors such
as cigarette smoking, alcohol use, and so on. For the question about seatbelt use
when driving, possible answers were Always, Most times, Sometimes, Rarely, and
Never. An additional choice allowed respondents to say they dont drive, which
often was the case because many survey participants were under the minimum
legal driving age. Table 2.1 summarizes responses in the 2003 survey given by
12thgrade students who said they drive.
Response
Count
Always
1686
Most times
578
Sometimes
414
Rarely
249
Never
115
Table 2.1
Follow these steps to determine the percentage of students falling into each category.
1. Clear any data from lists L1 and L2.
Press STAT ENTER to select the STAT list editor. Place the cursor at the
top of list L1. Press CLEAR followed by the down arrow key,H to clear any
data from list L1. Place the cursor at the top of list L2. Press CLEAR followed
by the down arrow key,H to clear any data from list L2.
2. Enter the data.
Place the cursor on list L1 row 1 to make L1(1) the active list row, as shown
in Figure 2.1. Enter the counts for the responses pressing ENTER after each
entry. The list is entered in Figure 2.2.
3. Enter an expression to determine the percentage of students falling into each
13

Chapter 2

Turning Data Into Information

category.
Move the cursor to the top of list L2. With the cursor at the top of list L2
type 2nd L1 2nd LIST J , selecting sum( , and press ENTER . Type
2nd L1 and a ). These steps are reflected in Figures 2.3 and 2.4. Press ENTER
to evaluate the expression.

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4
Figure 2.5
Notice that a majority, 1686/3042= .554 or 55.4%, said they always wear a seatbelt when driving, while just 115/3042= .032 or 3.2% said they never wear a
seatbelt. Because 55.4% said they always wear a seatbelt, we can calculate the
percent who dont always wear a seatbelt as 100%-55.4% =44.6% . Alternatively, the percent saying they dont always wear a seatbelt could be detennined
as 19.0% + 13.6% + 8.2% + 3.8%, the sum of the percents for all categories
other than Always.

Frequency and Relative Frequency

Frequency is a synonym for the count of how many observations fall into a category. The proportion or percent in a category is a type of a relative frequency,
the count in a category relative to the total count over all categories. A frequency
distribution for a categorical variable is a listing of all categories along with their
frequencies (counts). A relative frequency distribution is a listing of all categories
along with their relative frequencies (given as proportions or percents, for example). It is commonplace to give the frequency and relative frequency distributions
together, as was done in Table 2.1

Visual Summaries for Categorical Variables

There are two simple visual summaries used for categorical data:
a. Pie charts are useful for summarizing a single categorical variable if there
are not too many categories. Unfortunately, pie charts are not built-in to
the TI-83 Plus SE nor the TI-84 Plus SE.

14

2.4

Visual Summaries for Categorical Variables

b. Bar graphs are useful for summarizing one or two categorical variables
and are particularly useful for making comparisons when there are two
categorical variables.
Both of these simple graphical displays are easy to construct and interpret, as
the examples in the text demonstrate.
Example 2.3 Random Numbers Question 5 in the class survey described in Section 2.1 asked students to Randomly pick a number between 1and 10. The pie
chart shown in Figure 2.1 of the text illustrates that the results are not even close
to being evenly distributed across the numbers. Notice that almost 30%of the students chose 7 while only just over 1% chose the number l. The data is displayed
as an ungrouped frequency distribution in Table 2.2.
Random Number 1 2
3
4
5
6
7
8
9 10
Percent 1 4.7 11.6 11.0 9.5 12.1 29.5 10 7.4 3.2
Frequency 2 9
22
21
18 23
56
19 14 6
Table 2.2
Follow these steps to create a bar chart for the categorical variable random number.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6, as shown in Figure 2.6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the data for the categorical variable random number in list L1.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the data: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 pressing ENTER after each entry.
Place the cursor on list L12 row 1 to make L2(1) the active list row. Enter
the frequencies: 2, 9, 22, 21, 18, 23, 56, 19, 14, 6 in L2 pressing ENTER

15

Chapter 2

Turning Data Into Information

after each entry, as shown in Figure 2.7.

Figure 2.6

Figure 2.7

3. Plot the statistical data by creating a bar chart for the categorical variables random number.
Press 2nd STAT PLOT accessing the StatPlot menu.
Press ENTER , selecting Plot 1. Place the cursor on ON and press ENTER .
Use the down arrow key and the right arrow key to select the third icon in the
first row, the histogram (bar chart). Press ENTER . Use the down arrow key
to select list L1 as the list, 2nd L1 . Use the down arrow key to enter list L2
as the Freq: 2nd L2 . The settings for Plot 1 are shown in Figure 2.8.
4. View the graph.
Press ZOOM , 9: ZoomStat to view the graph, as shown in Figure 2.9.

Figure 2.8

Figure 2.9

Example 2.4 Myopia A survey of 479 children found that those who had slept
with a nightlight or in a fully lit room before the age of 2 had a higher incidence
of nearsightness (myopia) later in childhood (Sacramento Bee, May 13, 1999,
pp. A1, A18). The raw data for each child consisted of two categorical variables, each with three categories. Table 2.2 gives the categories and the number
of children falling into each combination of them.
The patten in Table 2.2 is striking. As the amount of sleeptime light inceases,
the incidence of myopia also increases. However this study does not prove that
sleeping with light actually caused myopia in children. There are other possible
explanations. For example, myopia has a genetic component, so those children
whose parents have myopia are more likely to suffer from it themselves. Maybe
nearsighted parents are more likely to proviode light while their children are

16

2.4
sleeping.
Slept with:
Darkness
Nightlight
Full Light
Total

No Myopia
155 (90%)
153 (66%)
34 (45%)
342 (71%)

Myopia
15 (9%)
72 (31%)
36 (48%)
123 (26%)
Table 2.2

Visual Summaries for Categorical Variables

High Myopia
2 (1%)
7 (3%)
5 (7%)
14 (3%)

Total
172
232
75
479

Follow these steps to create a bar chart for the categorical variables. You will
create a clustered bar chart displayed in percentages of the row totals for each
of the categorical variables.
5. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6, as shown in Figure 2.10. Press ENTER to execute the command.
6. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter codes for categorical variable Slept with in odd.numbered lists.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
1, 2, 3 pressing ENTER after each entry.
Place the cursor on list L3 row 1 to make L3(1) the active list row. Enter
5, 6, 7 pressing ENTER after each entry.
Place the cursor on list L5 row 1 to make L5(1) the active list row. Enter
9, 10, 11 pressing ENTER after each entry.
b. Enter the percentages, as whole numbers, for the categorical variable Myopia in even numbered lists.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter
the percentages: 90, 9, 1 pressing ENTER after each entry.
Place the cursor on list L4 row 1 to make L4(1) the active list row. Enter
the percentages: 66, 31, 3 pressing ENTER after each entry.
Place the cursor on list L6 row 1 to make L6(1) the active list row. Enter
17

Chapter 2

Turning Data Into Information

the percentages: 45, 48, 7 pressing ENTER after each entry.

The results of the data entry process are shown in Figures 2.11 and 2.12.

Figure 2.10

Figure 2.11

Figure 2.12

7. Plot the statistical data by creating a clusted bar chart for the categorical variables Slept with and Myopia.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the third icon in the first row, the histogram. Press ENTER . Use the
down arrow key to select L1 as the Xlist, 2nd L1 . Use the down
arrow key to select L2 as the Freq:, 2nd L2 . The settings for Plot
1 are shown in Figure 2.13.
(ii) Use the up arrow key to place the cursor on Plot2. Place the cursor on ON and press ENTER . Use the down arrow key and the
right arrow key to select the third icon in the first row, the histogram.
Press ENTER . Use the down arrow key to select L3 as the Xlist,
2nd L3 . Use the down arrow key to select L4 as the Freq:, 2nd L4 .
The settings for Plot 2 are shown in Figure 2.14.
(iii) Use the up arrow key to place the cursor on Plot3. Place the cursor on ON and press ENTER . Use the down arrow key and the
right arrow key to select the third icon in the first row, the histogram.
Press ENTER . Use the down arrow key to select L5 as the Xlist,
2nd L5 . Use the down arrow key to select L6 as the Freq:, 2nd L6 .
The settings for Plot 3 are shown in Figure 2.15.

Figure 2.13

Figure 2.14

Figure 2.15

8. Set the Window viewing variables in order to view the graph.

Press WINDOW , row 1, column 2. Set Xmin to 1. Set Xmax to 12; Xscl to
1; Ymin to -10, being sure to use the grey negation key. Set Ymax to 105; Yscl

18

2.5

Interesting Features of Quantitative Data

to 10; Xres to 1. These settings are illustrated in Figure 2.16

9. View the graph.
Press GRAPH to view the graph, as shown in Figure 2.16.

Figure 2.16
Figure 2.17
10. Optional: Add text to the histogram (bar chart).
Press 2nd DRAW , selecting 0: Text from the DRAW menu, as shown in Figure 2.18. Use the arrow keys to position the cursor. Press 2nd A-LOCK
to type the labels. You may have to select 2nd DRAW , selecting 1: ClrDraw from the DRAW menu and GRAPH to attempt once again to position
the labels to your satisfaction. The finished graph is displayed in Figure 2.19.

Figure 2.18
Figure 2.19
The first cluster on the left of the clustered bar chart displays the category Darkness of the Slept with variable. The heights of the bars indicate relative
frequencies of 90%, 9%, and 1%. The middle cluster of the clustered bar chart
displays the category Nightlight of the Slept with variable. The heights of
the bars indicate relative frequencies of 66%, 31%, and 3%. The third cluster
from the left of the clustered bar chart displays the category Full light of the
Slept with variable. The heights of the bars indicate relative frequencies of
45%, 48%, and 7%.
11. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
These outliers indicate that the last weight given in each list is very different
from the others. In fact, those two men were the coxswains for their teams, while
the other men were the rowers.

19

Chapter 2

Turning Data Into Information

2.4 - (continued) Interesting Features of Quantitative Data

Looking at a long, disorganized list of data values is about the same as looking at a
scrambled set of letters. To begin finding the information in quantitative data, we
have to organize it using visual displays and numerical summaries. In this section
we focus on interpreting the main features of quantitative variables. More specific
details will be given in the following sections.
Example 2.5 Right Handspans. Table 2.3 displays the raw data for the right
handspan measurements (in centimeters) made in the student survey described in
Section 2.1 of the text. The measurements are listed separately for males and females, but are not organized in any other way. Imagine that you know a female
whose stretched right handspan is 20.5 em. Can you see how she compares to the
other females in Table 2.3? That probably will be hard because the list of data values is disorganized.
We will organize the handspan data in Table 2.3 using a five-number summary,
which consists of the median, the quartiles (roughly, the medians of the lower and
upper halves of the data), and the extremes (high, low).

Males (87 students)

21.5, 22.5, 23.5, 23.0, 24.5, 23.0, 26.0, 23.0, 21.5, 21.5,
24.5, 23.5, 22.0, 23.5, 22.0, 22.0, 24.5, 23.0, 22.5, 19.5,
22.5, 22.0, 23.0, 22.5, 20.5, 21.5, 23.0, 22.5, 21.5, 25.0,
24.0, 21.5, 21.5, 18.0, 20.0, 22.0, 24.0, 22.0, 23.0, 22.0,
22.0, 23.0, 22.5, 25.5, 24.0, 23.5, 21.0, 25.5, 23.0, 22.5,
24.0, 21.5, 22.0, 22.5, 23.0, 18.5, 21.0, 24.0, 23.5, 24.5,
23.0, 22.0, 23.0, 23.0, 24.0, 24.5, 20.5, 24.0, 22.0, 23.0,
21.0, 22.5, 21.5, 24.5, 22.0, 22.0, 21.0, 23.0, 22.5, 24.0,
22.5, 23.0, 23.0, 23.0, 21.5, 19.0, 21.5
Females (103 students)
20.00, 19.00, 20.50, 20.50, 20.25, 20.00, 18.00, 20.50, 22.00,
20.00, 21.50, 17.00, 16.00, 22.00, 22.00, 20.00, 20.00, 20.00,
20.00, 21.70, 22.00, 20.00, 21.00, 21.00, 19.00, 21.00, 20.25,
21.00, 22.00, 18.00, 20.00, 21.00, 19.00, 22.50, 21.00, 20.00,
19.00, 21.00, 20.50, 21.00, 22.00, 20.00, 20.00, 18.00, 21.00,
22.50, 22.50, 19.00, 19.00, 19.00, 22.50, 20.00, 13.00, 20.00,
22.50, 19.50, 18.50, 19.00, 17.50, 18.00, 21.00, 19.50, 20.00,
19.00, 21.50, 18.00, 19.00, 19.50, 20.00, 22.50, 21.00, 18.00,
22.00, 18.50, 19.00, 22.00, 12.50, 18.00, 20.50, 19.00, 20.00,
21.00, 19.00, 19.00, 21.00, 18.50, 19.00, 21.50, 21.50, 23.00,
23.25, 20.00, 18.80, 21.00, 17.00, 21.00, 20.00, 20.50, 20.00,
19.50, 21.00, 21.00, 20.00
Table 2.3

20

2.5

Interesting Features of Quantitative Data

Follow these steps to obtain the five-number summaries for females and males.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the Stretched Right Handspans (cm) of the 190 College students
in lists L1 and L2.
Enter the Stretched Right Handspans (cm) for the Males (87 students)
in list L1. Place the cursor on list L1 row 1 to make L1(1) the active list
row. Enter 21.5, 22.5, 23.5, etc. pressing ENTER after each entry.
Enter the Stretched Right Handspans (cm) for the Females (103 students) in list L2. Place the cursor on list L2 row 1 to make L2(1) the active
list row. Enter 20, 19, 20.5, etc. pressing ENTER after each entry.
The data in lists L1 and L2 are displayed in Figure 2.20
3. Obtain the five-number summaries for females and males.
Press STAT I to obtain the STAT CALC menu, as shown in Figure 2.21.
a. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the
Stretched Right Handspans (cm) of the males. Press ENTER . Use
the down arrow key, H , five times. The output from the TI calculator is
displayed in Figure 2.22.
b. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the
Stretched Right Handspans (cm) of the males. Press ENTER . Use
the down arrow key, H , five times. The output from the TI calculator is
displayed in Figure 2.23.
4. Save list L1 as MRSN1 and list L2 as FRSN1.
Press 2nd L1 STO 2nd A-LOCK and type MRSN ALPHA 1; press

21

Chapter 2

Turning Data Into Information

ENTER . Press 2nd L2 STO 2nd A-LOCK and type FRSN ALPHA 1;
press ENTER .

Figure 2.20

Figure 2.21

Figure 2.22

Figure 2.23

Remember that the five-number summary approximately divides the dataset

into quarters. For example, about 25% of the female handspan measurements are
between 12.5 and 19.0 centimeters, about 25% are between 19 and 20 em, about
25% are between 20 and 21 em, and about 25 % are between 21 and 23.25 em. The
five-number summary gives us a good idea of where our imagined female with the
20.5 centimeter handspan fits into the distribution of handspans for females. Shes
in the third quarter of the data, slightly above the median (the middle value).

2.6

Pictures for Quantitative Data

There are three similar types of pictures that are used to represent quantitative
variables, all of which are valuable for assessing center, spread, shape, and outliers. Histograms are similar to bar graphs and can be used for any number of
data values, although they are not particularly informative when the sample size is
small. Stem-and-Leaf plots and dotplots present all individual values, so for very
large datasets they are more cumbersome than histograms. A fourth kind of picture, called a boxplot or box-andwhisker plot, displays the information given in a
five-number summary. It is especially useful for comparing two or more groups
and for identifying outliers. The TI-83 Plus SE and the TI-84 Plus SE are well
suited for displaying histograms, Stem-and-Leaf plots and boxplots. The TI-83
Plus SE and the TI-84 Plus SE do not have build in features for creating dotplots.
We will begin by creating a histogram of womens right handspans.
Example 2.5 Right Handspans. Table 2.3 displays the right handspan measurements (in centimeters) made in the student survey described in Section 2.1 of the
text. The measurements are listed separately for males and females. Recall that
the right handspan measurements for the females are stored in the list FRSN1.
Follow these steps to obtain the histogram of right handspans for females.
22

2.6

Pictures for Quantitative Data

1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the Stretched Right Handspans (cm) of the females in lists L1.
Place the cursor at the top of list L1. Press 2nd LIST , selecting the list
FRSN1, as shown in Figure 2.24. Press ENTER to drive the data into the
working list L1. The data from the list FRSN1 is displayed in list L1, as
shown in Figure 2.25.

Figure 2.24

Figure 2.25

3. Plot the statistical data by creating a histogram of the right handspan measurements for the females.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the third icon in the first row, the histogram. Press ENTER . Use
the down arrow key to select L1 as the list, 2nd L1 . Use the down
arrow key to enter 1 as the Freq:. The settings for Plot 1 are shown in
Figure 2.26.
4. View the graph.
Press ZOOM , 9: ZoomStat to view the graph, as shown in Figure 2.27.

23

Chapter 2

Turning Data Into Information

Figure 2.26
Figure 2.27
5. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
The histogram shows the distribution of the data, the pattern of how often the
various measurements occurred. The histogram is useful for assessing the location,
spread, and shape of a distribution and may be useful for detecting outliers. Notice
that the values are centered around 20 em, which is the median value. There are
two possible outlier values that are low compared to the bulk of the data that are
evident in the histogram. Except for those values, the handspans have a range of
about 7 em, extending from about 16 to 23 em. They tend to be clumped around
20 and taper off toward 16 and 23.
Example 2.5 Continued. Right Handspans. Table 2.3 displays the right handspan
measurements (in centimeters) made in the student survey described in Section 2.1
of the text. The measurements are listed separately for males and females. Recall
that the right handspan measurements for the females are stored in the list FRSN1
and the right handspan measurements for the males are stored in the list MRSN1.
Follow these steps to obtain the boxplot of right handspans for females and males.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.

24

2.6

Pictures for Quantitative Data

2. Enter data using the STAT list editor.

Press STAT ENTER to select the STAT list editor.
a. Enter the Stretched Right Handspans (cm) of the females in lists L1 and
the Stretched Right Handspans (cm) of the males in lists L2.
Place the cursor at the top of list L1. Press 2nd LIST , selecting the list
FRSN1, as shown in Figure 2.28. Press ENTER to drive the data into the
working list L1. Place the cursor at the top of list L2. Press 2nd LIST ,
selecting the list MRSN1, as shown in Figure 2.29. Press ENTER to
drive the data into the working list L1. The data from the list FRSN1 and
MRSN1 is displayed in list L1 and list L2, as shown in Figure 2.30.

Figure 2.29

Figure 2.29

Figure 2.30
3. Plot the statistical data by creating comparative boxplots of the right handspan
measurements for the females and males.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the first icon in the second row, the modified boxplot. Press ENTER .
Use the down arrow key to select L1 as the list, 2nd L1 . Use the
down arrow key to enter 1 as the Freq:. The settings for Plot 1 are
shown in Figure 2.31.
(ii) Use the up arrow key to place the cursor on Plot2. Place the cursor
on ON and press ENTER . Use the down arrow key and the right
arrow key to select the first icon in the second row, the modified box
plot. Press ENTER . Use the down arrow key to select L2 as the
25

Chapter 2

Turning Data Into Information

list, 2nd L2 . Use the down arrow key to enter 1 as the Freq:. The
settings for Plot 2 are shown in Figure 2.32.

Figure 2.31

Figure 2.32

4. View the graph.

Press ZOOM , 9: ZoomStat to view the graph, as shown in Figure 2.33.

Figure 2.33
5. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
The comparative boxplots compare the spans of the right hands of males and
females. For each group, the box covers the middle 50% of the data, and the line
within a box marks the median value. With the exception of possible outliers,
the lines extending from a box reach to the minimum and maximum data values.
Possible outliers are marked with an square.

2.7

Numerical Summaries of Quantitative Variables

We discussed the interesting features of a quantitative dataset in Section 2.4 of the
text, and in Section 2.5 of the text we learned how to look for them using use visual
displays of the data. In this section we learn how to compute numerical summaries
of these features for quantitative data.

Quartiles and Five-Number Summaries

A simple way to find the quartiles is to split the ordered values into the half that
is below the median and the half that is above the median. The lower quartile (Ql)
is the median of the data values ,that are below the median. The upper quartile
(Q3) is the median of the data values that are above the median. These values are

26

2.7

Quartiles and Five-Number Summaries

called quartiles because, along with the median and the extremes, they approximately divide the ordered data into quarters.We will begin by creating a histogram
of womens right handspans.

Example 2.13 Fastest Speeds. In Case Study 1.1 we summarized responses to

the question Whats the fastest youve ever driven a car? Table 2.4 displays the
response of the 87 males surveyed.
110,109,90,140,105,150,120,110,110,90,115,95,145,
140,110,105,85,95,100,115,124,95,100,125,140,85,
120,115,105,125,102,85,120,110,120,115,94,125,80,
85,140,120,92,130,125,110,90,110,110,95,95,110,
105,80,100,110,130,105,105,120,90,100,105,100,120,
100,100,80,100,120,105,60,125,120,100,115,95,110,
101,80,112,120,110,115,125,55,90
Table 2.4
Follow these steps to obtain the five-number summary for the 87 speeds.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the Fastest Speeds in lists L1.
Enter the Fastest Speeds for the 87 students in list L1. Place the cursor
on list L1 row 1 to make L1(1) the active list row. Enter 110, 109, 90, etc.
pressing ENTER after each entry. After entering all of the data, select

27

Chapter 2

Turning Data Into Information

2nd QUIT .

Figure 2.34

Figure 2.35

Figure 2.36

b. Save list L1 as MFST1.

Press 2nd L1 STO 2nd A-LOCK and type MFST ALPHA1; press
ENTER , as shown in Figure 2.35.
c. Sort the data in ascending order.
Press 2nd LIST I , OPS. Select 1: SortA(, pressing ENTER . Press
2nd L1 ) and press ENTER , as shown in Figure 2.36.
d. Examine the data set.
Press STAT ENTER to select the STAT list editor. The sorted data now
appears in list 1, as shown in Figure 2.37.
Use the down arrow key, H , to locate the 44th value in the list, as shown
in Figure 2.38.
The median is the middle value in an ordered list, so for 87 values, the
median is the (87 + 1)/2 =88/2 =44th value in the list. The 44th value is
110, and this value is shown in bold in the data list, as shown in Figure
2.38.

Figure 2.37

Figure 2.38

Aside from the middle value of 110, there were 43 values at or below 110,
and another 43 values at or above 110. Notice that there are many responses
of 110, which is why we are careful to say that 43 of the values are at or
above the median.
There are 43 values on either side of the median. To find the quartiles, simply find the median of each of those sets of 43 values. The lower quartile
is the (43 + 1)/2 = 22nd value from the bottom of the data.
Use the up arrow key, N , to locate the 22nd value in the list, as shown in
Figure 2.39. The value of Q1 is 95.
Use the down arrow key, H , to locate the 22nd value from the top, as
shown in Figure 2.40. The upper quartile is the 22nd value from the top;

28

2.8

Features of Bell-Shaped Distributions

the value of Q3 is 120.

Figure 2.39

Figure 2.40

The median and quartiles divide the data into equal numbers of values but
do not necessarily divide the data into equally wide intervals. For example,
the lowest 1/4 of the males had responses ranging over the 40-mph interval
from 55 mph to 95 mph, while the next 1/4 had responses ranging over only
a l5-mph interval, from 95 to 110. Similarly, the third quarter had responses
in only a 10-mph interval (110 to 120), while the top 1/4 had responses in
a 30-mph interval (120 to 150). It is common to see the majority of values
clumped in the middle and the remainder tapering off into a wider range.
e. Find the Summary Measures (mean, median, quartiles, low and high values, range and interquartile range).
Press STAT > CALC, selecting 1: 1Var Stats. Press ENTER . Press
2nd L1 and ENTER , as shown in Figure 2.41. The results are shown
in Figure 2.42. Use the down arrow key, H , five times to obtain the summary measure shown in Figure 2.43.

Figure 2.41
Figure 2.42
Figure 2.43
The calculator output indicates the mean = 107.4, miniumum =55, Q1 =
95, median = 110, Q3 = 120 and maximum = 150. The range is maximum
- minimum= 150-55 = 95 and the interquartile range is Q3-Q1 = 120 - 95
= 25.

2.8

Features of Bell-Shaped Distributions

Nature seems to follow a predictable pattern for many kinds of measurements.
Most individuals are clumped around the center, and the greater the distance a
value is from the center, the fewer individuals have that value. Except for the
two outliers at the lower end. that pattern is evident in the females right handspan
measurements, as shown in Example 2.5, Figure 2.27. If we were to draw a smooth
curve onnecting the tops of the bars on a histogram with this shape, the smooth
curve would resemble the shape of a bell.
Numerical variables that follow this pattern are said to follow a bell-shaped curve,
or to be bell-shaped. A special case of this distribution of measurements is so

29

Chapter 2

Turning Data Into Information

common it is also called a normal disbibution or normal curve.

Example 2.5 - Revisted - Womens Right Hand Spans. Table 2.3 displays the
raw data for the right handspan measurements (in centimeters) made in the student
survey described in Section 2.1 of the text. The measurements are listed separately
for males and females, but are not organized in any other way. In Example 2.5, you
have saved the data for the males in list MRSN1 and the data for the females in list
FRSN1.
We will draw a histogram of the womens right handspans, with a superimposed
normal curve.
Follow these steps to draw the histogram of the womens right handspans, with a
superimposed normal curve.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the Stretched Right Handspans (cm) of the females in list L1.
Place the cursor at the top of list L1. Press 2nd LIST , selecting the list
FRSN1, as shown in Figure 2.44. Press ENTER to drive the data into the
working list L1. The data from the list FRSN1 is displayed in list L1, as
shown in Figure 2.45.

Figure 2.44
Figure 2.45
Figure 2.46
b. Obtain the numerical summaries of the womens right handspans.
30

2.8

Features of Bell-Shaped Distributions

Press STAT I to obtain the STAT CALC menu.

c. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the
Stretched Right Handspans (cm) of the females. The output from the
TI calculator is displayed in Figure 2.46.
Observe that the mean of the womens right handspans is 20.017 and the
standard deviation is 1.764.
3. Set up the plot for the histogram of the right handspan measurements for the
females.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the third icon in the first row, the histogram. Press ENTER . Use
the down arrow key to select L1 as the list, 2nd L1 . Use the down
arrow key to enter 1 as the Freq:. The settings for Plot 1 are shown in
Figure 2.47.
4. Enter the function to superimpose the normal curve on the histogram.
Press Y= , row 1, column 1, to enter the function, as shown in Figure 1.9. Press
p
(18/1.764 (2))e((1/2)(x  20.017)2 /1.7642 ). Observe that the mean
of the womens right handspans, 20.017 and the standard deviation, 1.764, are
entered into the function to determine the y-values of the graph. The 18 is a
scaling factor designed to make the plot of the histogram and the normal curve
coincide. Other scaling factors can be explored. The left and right parentheses
are located on row 6. Press 2nd ,  is located on the 5th row, right column
above the ^ key. Press 2nd e, e is located on the 8th row, left column above
the LN key. Be sure to use the grey negation key when you enter (1/2).
The function is shown in Figure 2.48.
5. Set the Window viewing variables in order to view the graph.
Press WINDOW , row 1, column 2. Set Xmin to 11, Xmax to 27; Xscl to 1;
Ymin to -5, being sure to use the grey negation key. Set Ymax to 31; Yscl to
1; Xres to 1. These settings are illustrated in Figure 2.49
6. View the graph.
Press GRAPH , to view the graph, as shown in Figure 2.50.

31

Chapter 2

Turning Data Into Information

Figure 2.47

Figure 2.48

Figure 2.49
Figure 2.50
7. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
8. Clear the function.
Press Y= , and Press Y= and press CLEAR to remove all functions For
each line that is not blank, place the cursor on the function and press CLEAR
Press 2nd QUIT .

The Concept of Standard Deviation

Because normal curves are so common in nature, a whole set of descriptive features has been developed that apply mostly to variables with that shape. In fact,
two summary features uniquely determine a normal curve, so that if you know
those two summary numbers, you can draw the curve precisely. The first summary
number is the mean, and the bell shape is centered on that number. The second
summary number is called the standard deviation, and it is a measure of the spread
of the values.
You can think of the standard deviation as roughly the average distance values fall
from the mean. Put another way, it measures variability by summarizing how far
individual data values are from the mean.
The formula for calculating the standard deviation is a bit more involved than the
conceptual interpretation just discussed. This is the first instance of a summary
measure that differs based on whether the data represent a sample or an entire population. The version given here is appropriate when the dataset is considered to
represent a sample from a larger population. The value of s2 , the squared standard
deviation is called the (sample) variance. The formula for the (sample) variance
is
P
(x  x)2
s2 =
n1
In practice, statistical software like Minitab, a spreadsheet program like Excel, or

32

2.8

The Concept of Standard Deviation

a TI calculator typically is used to find the standard deviation lor a dataset. For
situations where you have to calculate the standard deviation by hand, here is a
step-bystep guide to the steps involved:
Step 1: Calculate x, the sample mean.
Step 2: For each observation, calculate the difference between the data
value and the mean.
Step 3: Square each difference calculated In step 2.
Step 4: Sum the squared differences calculated in step 3, and then divide
this sum by n  1.The answer for this step Is called the varIance.
Step 5: Take the square root of the variance calculated in step 4.
The answer for this step is called the standard deviation.

Example 2.18 Calculating a Standard Deviation.

deviation of the four pulse rates 62, 68, 74, 76.

You will calculate the standard

Follow these steps to calculate the standard deviation.

1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the four pulse rates 62, 68, 74, 76 in list L1.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
62, 68, 74, 76 pressing ENTER after each entry. The data is displayed in

33

Chapter 2

Turning Data Into Information

list L1, as shown in Figure 2.45. Press 2nd QUIT to exit the Stat Editor.

Figure 2.51
b. Obtain the mean, variance, and standard deviation of the pulse rates using
the definitions.
Press 2nd LIST I I to obtain the LIST MATH menu.
(i) Obtain the mean of the pulse rates.
Select 3: mean( and press ENTER . Press 2nd L1 to select the
four pulse rates. Press ENTER . The output from the TI calculator
is displayed in Figure 2.53, indicating the mean, 70.
(ii) Obtain the sum of the squared differences.
On the homescreen, press 2nd L1 - 2nd LIST I I , selecting
3: mean(. Press 2nd L1 and ) . Press STO 2nd L2 ENTER ,
storing the differences between the data value and the mean in list L2.
Press 2nd L2 , x2 . Press STO 2nd L3 ENTER , storing the
squared differences in list L3. The output from the TI calculator is displayed in Figure 2.52, indicating the sum of the squared difference is
120.
(iii) Obtain the variance.
Press 2nd LIST I I to obtain the LIST MATH menu. Select
5: sum( and press ENTER . Press 2nd L3 and ) to obtain the
sum of the squared differences in list L3. To obtain the variance, divide this sum by n  1. Do this by pressing 2nd LIST I I to
obtain the LIST MATH menu. Select 5: sum( and press ENTER .
Press 2nd L3 and ) . Press , and 2nd LIST I to obtain
the LIST OPS menu. Select 3: dim( and press ENTER . Press
2nd L1 and ) . Press ENTER . The output from the TI calculator
is displayed in Figure 2.53, indicating the variance, 40.
(iv) Obtain the standard deviation.
Press 2nd

40 ) to obtain the standard deviation, 6.32, as shown

34

2.8

The Concept of Standard Deviation

in Figure 2.54.

Figure 2.52
Figure 2.53
Figure 2.54
Observe that the mean of the pulse rates is 70, the variance is 40, and
the standard deviation is 6.32.
3. Obtain the mean, variance, and standard deviation of the pulse rates using the
STAT CALC menu..
Press STAT I to obtain the STAT CALC menu.
a. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the four
pulse rates. Press ENTER . The output from the TI calculator is displayed
in Figure 2.55.

Figure 2.55
Observe that the mean of the pulse rates is 70 and the standard deviation
is 6.32.

35

Chapter 3
Relationships Between
Quantitative Variables
Introduction
In this chapter, we will learn how to describe the relationship between two quantitative variables. Remember (from Chapter 2) that the terms quantitative variable
and measurement variable are synonyms for data that can be recorded as numerical values and then ordered according to those values. The relationship between
weight and height is an example of a relationship between two quantitative variables.
The questions we ask about the relationship between two variables often concern
specific numerical features of the association. For example, we may want to know
how much weight will increase on average for each 1-inch increase in height. Or,
we may want to estimate what the college grade point average will be for a student
whose high school grade point average was 3.5.In this chapter, you will learn how
to create simple summaries and pictures from various kinds of raw data.
After reading this chapter you should be able to:
1. Display a scatterplot of two quantitative variables.
2. Display subgroups of two quantitative variables on a scatterplot.
3. Display a scatterplot with the regression equation superimposed upon the scatterplot.
4. Make predictions using a regression equation.
5. Obtain the residuals.
6. Find the correlation coefficient and the coefficient of determination for two
quantitative variables.
7. Obtain the regression output, identifying the slope, intercept, r2 , SSTO, and
SSE for two quantitative variables.
Keystrokes Introduced
1. 2nd STAT PLOT > scatterplot displays a scatterplot of two quantitative
variables.
2. STAT CALC> 8: LinReg (a + bx) calculates a regression equation for two
quantitative variables.
3. 2nd CATALOG >DiagnosticOn displays r, the correlation coefficient, and
r2 , the coefficient of determination when a linear regression equation is ob-

36

3.1

Looking for Patterns With Scatterplots

tained.
4. VARS> 5: Statistics I I accesses the regression equation storage registers.
5. STAT >CALC> 1: 1-VarStats analyzes data for one quantitative variable.
6. 2nd LIST >MATH>sum( returns the sum of the elements within a list.

3.1

Looking for Patterns With Scatterplots

A scatterplot is a two-dimensional graph of the measurements for two numerical
variables. A point on the graph represents the combination of measurements for
an individual observation. The vertical axis, which is called the y axis, is used to
locate the value of one of the variables. The horizontal axis, called the x axis, is
used to locate the value of the other variable.
Questions to AskAbout a Scatterplot
What is the average pattern? Does it look like a straight line or is it curved?
What is the direction of the pattern?
How much do individual points vary from the average pattern?
Are there any unusual data points?
Example 3.1 Height and Handspan
Tables 3.1a and 3.1b display the observations of a dataset that includes the heights
(in inches) and fully stretched hands pans (in centimeters) of 167 college students.
The data values for all 167 students are the raw data for studying the connection
between height and handspan. Imagine how difficult it is to see the pattern in the
data from all 167 observations were shown in Table 3.1. Even when we just look
at the data for the first12 students, it takes a while to confirm that there does seem
to be a tendency for taller people to have larger handspans.
Follow these steps to display a scatterplot of handspan and height measurements
for all 167 students.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,

37

Chapter 3

Ss
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

Relationships Between Quantitative Variables

L5, L6, as shown in Figure 3.1. Press ENTER

Height Hand
Ss Height Hand
68
21.5
31 67
20.0
71
23.5
32 67
20.0
73
22.5
33 66
19.0
64
18.0
34 62
17.0
68
23.5
35 72
22.0
59
20.0
36 71
22.0
73
23.0
37 61
17.5
75
24.5
38 63
19.0
65
21.0
39 66
19.0
69
20.5
40 71
22.0
69
20.5
41 71
22.0
64
18.5
42 66
18.5
67
21.0
43 70
20.0
67
19.5
44 67
20.5
69
22.0
45 69
21.0
73
22.0
46 67
19.5
62
20.0
47 68
20.0
69
22.5
48 67
21.5
64
18.5
49 68
22.5
74
21.5
50 71
20.0
73
24.5
51 70
22.5
66
20.5
52 74
24.5
74
24.5
53 60
18.5
73
21.0
54 65
20.0
69
21.0
55 72
24.0
64
18.5
56 76
23.5
67
18.0
57 66
21.0
60
19.5
58 64.5
19.5
75
20.5
59 71
20.0
64
21.0
60 69
22.5
Table 3.1a

to execute the command.

Ss Height Hand
61 64
20.0
62 65
20.0
63 74
24.0
64 68
21.0
65 68
21.5
66 69
18.5
67 68
23.0
68 67
23.0
69 61.5
20.5
70 63
16.5
71 67
19.5
72 71
23.0
73 73
22.5
74 63
18.5
75 61
18.5
76 67
21.5
77 72
20.5
78 72
20.5
79 68
20.0
80 66
21
81 67
21.5
82 67
20.5
83 72
20.5
84 67.5
21.0
85 63.75
21.5
86 72
21.5
87 69
22.5
88 68
21.0
89 71
21.0
90 71
22.0

2. Enter data using the STAT list editor.

Press STAT ENTER to select the STAT list editor.
a. Enter the data for the quantitative variables height and handspan in
lists L1 and L2.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the height data: 68, 71, 73, ... pressing ENTER after each entry.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter the

38

3.1

Looking for Patterns With Scatterplots

hand data: 21.5, 23.5, 22.5, ... in L2 pressing ENTER after each entry,
as shown in Figure 5.2.

Figure 3.1
Ss
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116

Height
63
70
68
67.5
75
75
71
71
64
71
69
65
69
63
70
71
64
63
65
66
66
65
67.5
57
72
64

Hand
19.0
23.0
20.5
20.5
21.0
24.0
22.0
21.0
19.5
21.0
19.5
19.0
23.0
20.5
24.0
22.0
20.0
21.5
19.0
19.0
20.0
19.5
20.0
16.0
22.5
17.5

Ss
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142

Figure 3.2
Height Hand
67
19.5
68
22.5
63
20.0
67
21.5
66
20.5
72
23.5
74
22.0
69
18.0
68
19.0
65
19.5
64
19.0
67
20.0
74
23.5
73
24.0
64
18.5
76
24.5
68
20.0
76
23.0
64.25
22.0
69
22.5
75
24.5
61.5
17.0
69
22.0
67
22.0
74
24.5
74
24.0
Table 3.1b

Ss
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167

Height
71
71
63
67
65
68
67
73
78
62
70
64
64
72
74
70
70
62
64
66
60
73
66
68
73

Hand
18.5
21.5
21.0
22.0
20.5
19.0
20.5
23.0
25.5
18.5
19.0
19.0
20.0
20.5
24.0
22.0
23.5
17.0
18.5
20.0
17.0
23.0
18.5
21.0
21.0

3. Plot the statistical data by creating a scatterplot of handspan and height measurements for all 167 students.
Press 2nd STAT PLOT accessing the stat plot menu.
Press ENTER , selecting Plot 1. Place the cursor on ON and press ENTER .
Use the down arrow key and the right arrow key to select the first icon in the
first row, the scatterplot. Press ENTER . Use the down arrow key to select list

39

Chapter 3

Relationships Between Quantitative Variables

L1 as the list, 2nd L1 . Use the down arrow key to enter list L2 as the Ylist:
2nd L2 . Use the down arrow key to select the second icon for the mark. The
settings for Plot 1 are shown in Figure 3.3.
4. View the graph.
Press ZOOM , 9: ZoomStat to view the graph, as shown in Figure 3.4.

Figure 3.3
5. Save list L1 as HGHT and list L2 as HAND.

Figure 3.4

Press 2nd L1 STO 2nd A-LOCK and type HGHT; press ENTER .
Press 2nd L2 STO 2nd A-LOCK and type HAND; press ENTER .
Figure 3.4 is a scatterplot that displays the handspan and height measurements
for all 167 students. The hands pan measurements are plotted along the vertical
axis (y), and the height measurements are plotted along the horizontal axis (x).
Each point represents the two measurements for an individual.
We see that taller people tend to have greater handspan measurements than
shorter people do. When two variables tend to increase together, as they do
in Figure 3.4, we say that they have a positive association. Another noteworthy characteristic of the graph is that we can describe the general pattern of
this relationship with a straight line. In other words, the hands pan and height
measurements may have a linear relationship.

Indicating Groups Within the Data on Scatterplots

When we examined the connection between height and hands pan in Example 3.1,
you may have wondered whether we should be concerned about student gender.
Both height and hands pan tend to be greater for men than for women, so we should
consider the possibility that gender differences might be completely responsible for
the observed relationship.
Its easy to indicate subgroups on a scatterplot. We just use different symbols or
different colors to represent the different groups.

40

3.1

Indicating Groups Within the Data on Scatterplots

Example 3.1 Height and Handspan Continued The data for

females is displayed in Table 3.2.The data for males is displayed inTable 3.3.
Height and HandSpans for 89 Females
Ss Height Hand
Ss Height Hand
Ss Height Hand
1
68
21.5
31
64
20.0
61
57
16.0
2
64
18.0
32
65
20.0
62
64
17.5
3
59
20.0
33
68
21.5
63
67
19.5
4
65
21.0
34
69
18.5
64
68
22.5
5
69
20.5
35
68
23.0
65
63
20.0
6
64
18.5
36
61.5
20.5
66
66
20.5
7
67
21.0
37
63
16.5
67
68
19.0
8
67
19.5
38
67
19.5
68
65
19.5
9
62
20.0
39
63
18.5
69
64
19.0
10
64
18.5
40
61
18.5
70
67
20.0
11
66
20.5
41
68
20.0
71
64
18.5
12
64
18.5
42
66
21.0
72
68
20.0
13
67
18.0
43
67
20.5
73
64.25
22.0
14
60
19.5
44
63.75
21.5
74
61.5
17.0
15
64
21.0
45
72
21.5
75
71
18.5
16
67
20.0
46
68
21.0
76
63
21.0
17
67
20.0
47
63
19.0
77
65
20.5
18
66
19.0
48
68
20.5
78
68
19.0
19
62
17.0
49
67.5
20.5
79
67
20.5
20
61
17.5
50
64
19.5
80
62
18.5
21
63
19.0
51
69
19.5
81
70
19.0
22
66
18.5
52
65
19.0
82
64
19.0
23
67
20.5
53
63
20.5
83
64
20.0
24
67
19.5
54
64
20.0
84
62
17.0
25
68
20.0
55
63
21.5
85
64
18.5
26
71
20.0
56
65
19.0
86
66
20.0
27
60
18.5
57
66
19.0
87
60
17.0
28
65
20.0
58
66
20.0
88
66
18.5
29
66
21.0
59
65
19.5
89
68
21.0
30
64.5
19.5
60
67
20.0
Table 3.2
Follow these steps to display a scatterplot of handspan and height measurements
for the 89 female students and the 78 male students.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .

41

Chapter 3

Relationships Between Quantitative Variables

b. Clear all lists in the Stat editor.

Ss
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26

Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6, as shown in Figure 3.5. Press ENTER to execute the command.
Height and HandSpans for 78 Males
Height Hand
Ss Height Hand
Ss Height Hand
71
23.5
27
72
24.0
53
71
22.0
73
22.5
28
76
23.5
54
72
22.5
68
23.5
29
71
20.0
55
67
21.5
73
23.0
30
69
22.5
56
72
23.5
75
24.5
31
74
24.0
57
74
22.0
69
20.5
32
68
21.0
58
69
18.0
69
22.0
33
67
23.0
59
74
23.5
73
22.0
34
71
23.0
60
73
24.0
69
22.5
35
73
22.5
61
76
24.5
74
21.5
36
67
21.5
62
76
23.0
73
24.5
37
72
20.5
63
69
22.5
74
24.5
38
72
20.5
64
75
24.5
73
21.0
39
67
21.5
65
69
22.0
69
21.0
40
72
20.5
66
67
22.0
75
20.5
41
67.5
21.0
67
74
24.5
72
22.0
42
69
22.5
68
74
24.0
71
22.0
43
71
21.0
69
71
21.5
66
19.0
44
71
22.0
70
67
22.0
71
22.0
45
70
23.0
71
73
23.0
71
22.0
46
75
21.0
72
78
25.5
70
20.0
47
74
24.0
73
72
20.5
69
21.0
48
71
22.0
74
74
24.0
67
21.5
49
71
21.0
75
70
22.0
68
22.5
50
71
21.0
76
70
23.5
70
22.5
51
69
23.0
77
73
23.0
74
24.5
52
70
24.0
78
73
21.0
Table 3.3

2. Enter data using the STAT list editor.

Press STAT ENTER to select the STAT list editor.
a. Enter the data for the quantitative variables height and handspan for
females in lists L1 and L2.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the height data for females: 68, 64, 59, ... pressing ENTER after each
entry.

42

3.1 Indicating Groups Within the Data on Scatterplots

Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter
the hand data for females: 21.5, 18.0, 20.0, ... in L2 pressing ENTER
after each entry, as shown in Figure 3.6.
b. Enter the data for the quantitative variables height and handspan for
males in lists L3 and L4.
Place the cursor on list L3 row 1 to make L3(1) the active list row. Enter
the height data for males: 71, 73, 68, ... pressing ENTER after each entry.
Place the cursor on list L4 row 1 to make L4(1) the active list row. Enter
the hand data for males: 23.5, 22.5, 23.5, ... in L2 pressing ENTER after
each entry, as shown in Figure 5.7.

Figure 3.5

Figure 3.6

Figure 3.7

3. Plot the statistical data by creating a scatterplot indicating groups within the
data.
Press 2nd STAT PLOT accessing the StatPlot menu.
Create a scatterplot of female heights and handspans with heights on the horizontal axis and handspan on the vertical axis.
Press ENTER , selecting Plot 1. Place the cursor on ON and press ENTER .
Use the down arrow key and the right arrow key to select the first icon in the
first row, the scatterplot. Press ENTER . Use the down arrow key to select list
L1 as the Xlist, 2nd L1 . Use the down arrow key to enter list L2 as the Ylist:
2nd L2 . Use the down arrow key to select the second icon for the mark. The
settings for Plot 1 are shown in Figure 3.8.
Create a scatterplot of male heights and handspans with heights on the horizontal axis and handspan on the vertical axis.
Use the up arrow key, to select Plot 2. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select the first
icon in the first row, the scatterplot. Press ENTER . Use the down arrow key
to select list L3 as the Xlist, 2nd L3 . Use the down arrow key to enter list
L4 as the Ylist: 2nd L2 . Use the down arrow key to select the third icon for
the mark. The settings for Plot 2 are shown in Figure 3.9.
4. Set the Window viewing variables in order to view the graph.

53

Chapter 3

Relationships Between Quantitative Variables

Press WINDOW , row 1, column 2. Set Xmin t o 5 5 . S e t X m a x t o

8 0 ; X s c l to 1; Ymin to 15. Set Ymax to 26; Yscl to 10; Xres to 1. These
settings areillustrated in Figure 3.10
5. View the graph.
Press GRAPH to view the graph, as shown in Figure 3.11.

Figure 3.8

Figure 3.9

Figure 3.10
Figure 3.11
6. Save list L1 as HGHTF and list L2 as HANDF.
Press 2nd L1 STO 2nd A-LOCK and type HGHTF; press ENTER .
Press 2nd L2 STO 2nd A-LOCK and type HANDF; press ENTER .
7. Save list L3 as HGHTM and list L4 as HANDM.
Press 2nd L1 STO 2nd A-LOCK and type HGHTM; press ENTER .
Press 2nd L2 STO 2nd A-LOCK and type HANDM; press ENTER .
Notice that the positive association between hands pan and height appears to
hold within each sex. For both men and women, hands pan tends to increase
as height increases.

3.2

Describing Linear Patterns With a Regression Line

Scatter plots show us a lot about a relationship, but we often want more specific
numerical descriptions of how the response and explanatory variables are related.
Imagine, for example, that we are examining the weights and heights of a sample
of college women. We might want to know what the increase in average weight
is for each I-inch increase in height. Or, we might want to estimate the average
weight for women with a specific height, like 510.
Regression analysis is the area of statistics used to examine the relationship between a quantitative response variable and one or more explanatory variables. A
key element of regression analysis is the estimation of a regression equation that
describes how, on average, the response variable is related to the explanatory vari-

44

3.2

Describing Linear Patterns With a Regression Line

ables. This regression equation can be used to answer the types of questions that
we just asked about the weights and heights of college women.
A regression equation can atso-be used to predict values of a response variable
using known values of an explanatory variable. For instance, it might be useful
for colleges to have an equation for the connection between verbal SAT score and
college grade point average (GPA). They could use that equation to predict the potential GPAs of future students, based on their verbal SAT scores. Some colleges
actually do this kind of prediction to decide whom to admit, but they use a collection of variables to predict GPA.
There are many types of relationships and many types of regression equations. The
simplest kind of relationship between two variables is a straight line, and thats the
only type we will discuss here. Straight-line relationships occur frequently in practice, so this is a useful and important type of regression equation. Before we use
a straight-line regression model, however, we should always examine a scatterplot
to verify that the pattern actually is linear.
Example 3.2 Driver Age and the Maximum Legibility Distance of Highway
Signs In a study of the legibility and visibility of highway signs, a Pennsylvania
research firm determined the maximum distance at which each of 30 drivers could
read a newly designed sign. The 30 participants in the study ranged in age from 18
to 82 years old. The government agency that funded the research hoped to improve
highway safety for older drivers and wanted to examine the relationship between
age and the sign legibility distance.
Table 5.4 lists the data. We will use the TI calculator to display a scatterplot to show
that the relationship between maximum distance and age has a straight line
pattern and to find the best line for this set of measurements. We will display a
line that describes the average relationship between the two variables.
Age Distance
Age Distance
Age Distance
18
510
37
420
68
300
20
590
41
460
70
390
22
560
46
450
71
320
23
510
49
380
72
370
23
460
53
460
73
280
25
490
55
420
74
420
27
560
63
350
75
460
28
510
65
420
77
360
29
460
66
300
79
310
32
410
67
410
82
360
Table 3.4
Follow these steps to display a scatterplot with the regression equation superimposed upon the scatterplot.
1. Preparations:

45

Chapter 3 Relationships Between Quantitative Variables

a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
c. Turn Diagnostics On to display r, the correlation coefficient, and r2 , the
coefficient of determination.
Press 2nd CATALOG , located on the bottom row, 2nd column from the
left above the 0. Press ALPHA D, and use the down arrow key to locate
DiagnosticOn, as shown in Figure 5.12. Press ENTER to select the command and press ENTER once again to execute the command.

Figure 3.12

2. Enter data using the STAT list editor.

Press STAT ENTER to select the STAT list editor.
a. Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the Age data: 18, 20, 22, ... pressing ENTER after each entry.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter
the Distance data: 510, 590, 560, ... in L2 pressing ENTER after each
entry, as shown in Figure 5.13.

Figure 3.13
3. Obtain the regression equation.
Press STAT I to obtain the STAT CALC menu.

46

3.2 Describing Linear Patterns With a Regression Line

a. Use the down arrow key, H , seven times and press ENTER, or just press 8
to select 8: LinReg (a+bx), as shown in Figure 3.14. Press 2ndL1 to select
2ndL2 to select the Distance data, as shown
the Age data. Press

in Figure 5.15. Press ENTER to execute the command. The output from
the TI calculator is displayed in Figure 5.16.

Figure 3.14
Figure 3.15
Figure 3.16
The regression line y=577 - 3x describes how the maximum sign legibility
distance (the y variable) is related to driver age (the x variable).
4. Obtain data points to plot the regression equation.
Press VARS, row 4, column 4. Select 5: Statistics. Use the right arrow, I ,
twice, selecting 2: a. Press +, VARS, 5: Statistics, and the right arrow, I ,
twice, selecting 3: b. Press , 2ndL1. Press STO 2nd L3 . Your screen
should look like Figure 5.17. These data points represent the predicted values
of Distance from the Age variable stored in list L1. These predicted values
of Distance are stored in L3.

Figure 3.17
5. Display a scatterplot with the regression equation superimposed upon the scatterplot.
Press 2nd STAT PLOT accessing the StatPlot menu.
Press ENTER , selecting Plot 1. Place the cursor on ON and press ENTER .
Use the down arrow key and the right arrow key to select the first icon in the
first row, the scatterplot. Press ENTER . Use the down arrow key to select list
L1 as the Xlist, 2nd L1 . Use the down arrow key to enter list L2 as the Ylist:
2nd L2 . Use the down arrow key to select the second icon for the mark. The
settings for Plot 1 are shown in Figure 3.18.
Use the up arrow key to place the cursor on Plot2. Place the cursor on ON and
press ENTER . Use the down arrow key and the right arrow key to select the
second icon in the first row, the xyLine. Press ENTER . Use the down arrow
key to select L1 as the Xlis t , 2nd L1 . Use the down arrow key to select
L3as the Ylist:, 2nd L3 . The settings for Plot 2 are shown in Figure 3.19.
47

Chapter 3

Relationships Between Quantitative Variables

6. View the graph.

Press ZOOM , 9: ZoomStat to view the graph, as shown in Figure 3.20.

Figure 3.18
Figure 3.19
Figure 3.20
Earlier, we asked these two questions about distance and age:
(i) How much does the distance decrease when age is increased?
(ii) For drivers of any specific age, what is the average distance at which
the sign can be read?
The slope of the equation can be used to answer the first question. Remember
that the slope is the number that multiplies the x varia,ble and the sign of the slope
indicates the direction of the association. Here, the slope tells us that, on average,
the legibility distance decreases 3 feet when age increases by one year. This information can be used to estimate the average change in distance for any difference
in ages, For an age increase of 30 years, the estimated decrease in legibility distance is 90 feet because the slope is -3 feet per year.
The question about estimating the average legibility distances for a specific age is
answered by using the specific age as the x value in the regression equation. To
emphasize this use of the regression line, we write it as
Average distance = 577  3 Age
1. Make predictions for specific ages, 20, 50, and 80, finding the average distance
at which the sign can be read.
Press VARS, row 4, column 4. Select 5: Statistics. Use the right arrow, I ,
twice, selecting 2: a. Press + , VARS, 5: Statistics, and the right arrow, I ,
twice, selecting 3: b. Press , 20. Press ENTER.
Press VARS, row 4, column 4. Select 5: Statistics. Use the right arrow, I ,
twice, selecting 2: a. Press + , VARS, 5: Statistics, and the right arrow, I ,
twice, selecting 3: b. Press , 50. Press ENTER.
Press VARS, row 4, column 4. Select 5: Statistics. Use the right arrow, I ,
twice, selecting 2: a. Press + , VARS, 5: Statistics, and the right arrow, I ,
twice, selecting 3: b. Press , 80. Press ENTER. The results of these three

48

3.3

Measuring Strength and Direction with Correlation

calculations are shown in Figure 3.21.

Figure 3.21
For any given line, we can calculate the predicted value yb for each point in
the observed data. To do this for any particular point, we use the observed x
value in the regression equation. The prediction error for an observation is the
difference between the observed y value and the predicted value yb; the formula
is error = (y  yb). The terminology error is somewhat misleading, since
the amount by which an individual differs from the line is usually due to natural
variation rather than errors in the measurements. A more neutral term for the
difference (y  yb) is that it is the residual for that individual.
2. Obtain the residuals.
Recall that the predicted values of Distance, based upon the Age variable
are stored in L3 and the observed values of Distance are in L2.
Press STAT ENTER to select the STAT list editor.
a. Place the cursor at the top of list L4. Press 2nd L2 , - 2nd L3 , as shown
in Figure 3.22, pressing ENTER to obtain the residuals.
b. Place the cursor at the top of list L5. Press 2ndLIST, selecting the list
RESID, as shown in Figure 3.23. Press ENTER to drive the residuals,
that are automatically generated on the TI calculator with that list name,
into L5. Observe that the residuals displayed in lists L4 and L5 are identical, as shown in Figure 3.24.

Figure 3.22

Figure 3.23

Figure 3.24

3. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.

49

Chapter 3

3.3

Relationships Between Quantitative Variables

Measuring Strength and Direction with Correlation

The linear pattern is so common that a statistic was created to characterize this type
of relationship. The statistical correlation between two quantitative variables is a
number that indicates the strength and the direction of a straight-line relationship.
(i) The strength of the relationship is determined by the closeness of the
points to a straight line.
(ii) The direction is determined by whether one variable generally increases or generally decreases when the other variable increases.
As used in statistics, the meaning of the word correlation is much more specific
than it is in everyday life. A statistical correlation only describes linear relationships. Whenever a correlation is calculated, a straight line is used as the frame of
reference for evaluating the relationship. When
Example 3.2 Driver Age and the Maximum Legibility Distance of Highway
Signs Revisted In a study of the legibility and visibility of highway signs, a
Pennsylvania research firm determined the maximum distance at which each of 30
drivers could read a newly designed sign. The 30 participants in the study ranged
in age from 18 to 82 years old. The government agency that funded the research
hoped to improve highway safety for older drivers and wanted to examine the relationship between age and the sign legibility distance.
Table 3.3 lists the data. We will use the TI calculator to determine the correlation
coefficient between maximum distance and age.
Follow these steps to find the correlation coefficient and the coefficient of determination.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor: Caution: If the Age and Distance data
are within L1 and L2, Do NOT execute this step .
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
c. Turn Diagnostics On to display r, the correlation coefficient, and r2 , the
coefficient of determination.
Press 2nd CATALOG , located on the bottom row, 2nd column from the
left above the 0. Press ALPHA D , and use the down arrow key to

50

3.3

Calculating the Sum of Squared Errors

locate DiagnosticOn, as shown in Figure 3.12. Press ENTER to select

the command and press ENTER once again to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the Age data: 18, 20, 22, ... pressing ENTER after each entry.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter
the Distance data: 510, 590, 560, ... in L2 pressing ENTER after each
entry, as shown in Figure 3.13.
3. Obtain the regression equation, correlation coefficient and the coefficient of
determination.
Press STAT I to obtain the STAT CALC menu.
a. Use the down arrow key, H , seven times and press ENTER , or just press
8 to select 8: LinReg (a+bx), as shown in Figure 3.14. Press 2ndL1 to
2nd L2 to select the Distance data,
select the Age data. Press

as shown in Figure 3.15. Press ENTER to execute the command. The

output from the TI calculator is displayed in Figure 3.16 and Figure 3.25.

Figure 3.25
For the data shown in Figure 3.20 relating driver age and sign legibility
distance, the correlation is r = 0.80. This value indicates a somewhat
strong negative association between the variables.

Calculating the Sum of Squared Errors

A least squares line has the property that the sum of squared differences between
the observed vaIues of y and the predicted values is smaller for that line than it
is for any other line. Put more simply, the least squares line minimizes the sum
of squared prediction errors for the observed data set. The notation SSE, which
stands for sum of squared errors, is used to represent the sum of squared prediction
errors. The least squares line (the regression line) has a smaller SSE than any other
regression line that might be used to predict the response variable.
51

Chapter 3

Relationships Between Quantitative Variables

Example - Exam Scores Suppose that x = score on exam 1 in a course and y=

score on exam 2, and that the first two rows in Table 5.5 (shown below) give x
values and y values for n = 6 students. We will use the TI calculator to obtain the
regression output, identifying the slope, intercept, r 2 , SSTO, and SSE for this set
of measurements.
x = Exam 1 score 70 75 80 80 85 90
y = Exam 2 score 75 82 80 86 90 91
Table 3.5
Follow these steps to obtain the regression output, identifying the slope, intercept,
r2 , SSTO, and SSE for this set of measurements.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
c. Turn Diagnostics On to display r, the correlation coefficient, and r2 , the
coefficient of determination.
Press 2nd CATALOG , located on the bottom row, 2nd column from the
left above the 0 . Press ALPHA D , and use the down arrow key to
locate DiagnosticOn, as shown in Figure 5.12. Press ENTER to select
the command and press ENTER once again to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
the x = Exam 1 score data: 70, 75, 80, ... pressing ENTER after each
entry.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Enter
the y =Exam 2 score data: 75, 82, 80, ... in L2 pressing ENTER after

52

3.3

Calculating the Sum of Squared Errors

each entry, as shown in Figure 3.13.

Figure 3.26
3. Obtain the regression equation.
Press STAT I to obtain the STAT CALC menu.
a. Use the down arrow key, H , seven times and press ENTER , or just press
8 to select 8: LinReg (a+bx), as shown in Figure 3.27. Press 2nd L1
to select the x = Exam 1 score data. Press
2nd L2 to select the y

=Exam 2 score data, as shown in Figure 3.28. Press ENTER to execute

the command. The output from the TI calculator is displayed in Figure
.29.

Figure 3.27
Figure 3.28
Figure 3.29
The regression equation is y = 20 + 0.8x; the y-intercept is 20 and the
slope is 0.8. The correlation coefficient, r = .918 describes a moderately
strong positive association. The squared correlation is r = ( .918) 2 =
0.842. The x = Exam 1 score explains 84.2% of the variation among
the y =Exam 2 score data.
4. Obtain the sum of square errors.
Press VARS , row 4, column 4. Select 5: Statistics. Use the right arrow, I ,
twice, selecting 2: a. Press +, VARS , 5: Statistics, and the right arrow, I ,
twice, selecting 3: b. Press , 2nd L1 . Press STO 2nd L3 . Your
screen should look like Figure 5.30. These data points represent the predicted
values, yb, from x =Exam 1 score variable stored in list L1. These predicted
values, yb, of Exam 2 score are stored in L3.
Press STAT ENTER to select the STAT list editor.
a. Place the cursor at the top of list L4. Press 2nd L2 , - 2nd L3 , as
shown in Figure 3.31, pressing ENTER to obtain the residuals. The resid-

53

Chapter 3

Relationships Between Quantitative Variables

uals are shown in Figure 3.32.

Figure 3.30

Figure 3.31

b. Press 2nd QUIT . Press 2nd LIST I I , selecting 5: sum(. Press

2nd L4 x2 ) . Press
Figure 3.33.

ENTER . Your screen should look like

Figure 3.32
The sum of squared errors is SSE = 30.

Figure 3.33

5. Obtain the total sum of squares, SST O = sum ((y-y)2 ).

To obtain the mean of the predicted y values, press STAT > CALC, selecting
1: 1Var Stats. Press ENTER . Press 2nd L3 and ENTER , as shown in
Figure 5.34. The results are shown in Figure 5.35.

Figure 3.34
The mean of the predicted y values is 84.

Figure 3.35

Press STAT ENTER to select the STAT list editor.

a. Place the cursor at the top of list L5. Press 2nd L2 , - 8 4 , as shown in
Figure 3.36, pressing ENTER . The results are shown in Figure 3.37.
b. Press 2nd QUIT . Press 2nd LIST I I , selecting 5: sum(. Press
2nd L5 x2 ) . Press ENTER . Your screen should look like Figure

54

3.3

Calculating the Sum of Squared Errors

3.38.

Figure 3.36
Figure 3.37
Figure 3.38
2
The total sum of quares, S S T O = sum ((y-y) ) is 190.
The coefficient of determination, r2 =

SST OSSE
SST O

55

19030
190

= 0. 84211.

Chapter 4
Relationships Between
Categorical Variables
This chapter is about the analysis of the relationship between two categorical variables, so lets begin by recalling the meaning of the term categorical variable.The
raw data from categorical variables consist of group or category names that dont
necessarily have any ordering. Eye color and hair color, for instance, are categorical variables.
We can also use the methods of this chapter to examine ordinal variables. Ordinal
variables can be thought of as categorical variables for which the categories have
a natural ordering. For example, a researcher might define categories for quantitative variables, like age, income, or years of education.
Although there are many questions that we can and will ask about two categorical
variables, in most cases the principal question that we ask is: Is there a relationship between the two variables, so that the category into which individuals fall for
one variable seems to depend on the category they are in for the other variable?
After reading this chapter you should be able to:
1. Construct a table of Frequency Counts from raw data including row and column
percents.
2. Conduct a chi-square test, including finding observed counts, computing a chisquare statistic and find the p-value.
3. Find a p-value given the chi-square value and degrees of freedom.
Keystrokes Introduced
1. 2nd LIST I OPS>3: dim(listname ) returns the dimension (number of elements) of listname.
2. 2nd LIST I I MATH > 5: sum(list [,start,end ]) returns the sum of the
elements of list from start to end.
3. 2nd MATRIX I I , selecting the MATRIX EDIT menu. This command
enables you to edit a matrix element value including the dimensions and the
elements of the matrix.
4. 2nd MATRIX >NAMES selecting a matrix, pressing ENTER to view the
elements of the matrix.
5. STAT I I , selecting the STAT TEST menu. You will select C: 2 Test,
performing a 2 test, where the observations have been entered into matrix A.
6. 2nd DISTR , using the down arrow key, H , several times to select 7: 2 cdf(.
The arguments in 2 cdf are(lowerbound,upperbound,df ). The command com-

56

4.1

Displaying Relationships

putes the distribution probability between lowerbound and upperbound for the
specified degrees of freedom df.

4.1

Displaying Relationships
We have already encountered several examples of the type of problem we will study
in this chapter. In Chapter 2, for instance, we described a study of 479 children
that found that children who slept either with a nightlight or in a fullylit room
before the age of two had a higher incidence of myopia (nearsightedness) later
in childhood. Data from the PennState1 worksheet will be used toillustrate
how relationships betweencategorical variables may be presented.
Example
In an experiment done in a statistics class, 92 college students were given a form
read Randomly choose one of the letters S or Q. Another 98 students were given
a form with the order ofthe letters reversed, to read Randomly choose one of the
letters Q or S. The purpose was to determine if the order of listing the letters might
influence the choice of letters. The possible influence of the order of listing items is
a concern in elections. Many election analysts feel a candidate gains an advantage
if he or she is the first candidate listed on the ballot. The data is contained in the
PennState1 worksheet and is displayed in Table 4.1 and Table 4.2.
SSSSSSSSSSSSSSSSSSS
SSSSSSSSSSSSSSSSSSS
SSSSSSSSSSSSSSSSSSS
SSSSQQQQQQQQQQQQQQQ
QQQQQQQQQQQQQQQQ
Table 4.1 Randomly pick a letter-S or Q
SSSSSSSSSSSSSSSSSSS
SSSSSSSSSSSSSSSSSSS
SSSSSSSQQQQQQQQQQQQ
QQQQQQQQQQQQQQQQQQQ
QQQQQQQQQQQQQQQQQQQ
QQQ
Table 4.2 Randomly pick a letter-Q or S
TI calculators allow only for numerical values to be used in a statistical analysis.
We can not use the letters S or Q since these letters are replaced by the value
stored in memory for the S and Q variables in the calculator.
The solution to the problem is to assign a unique numerical code for each value
of the variable. In this case, you might code S = 0 and Q = 1 on the TI
calculator.
Follow these steps to construct a table of frequency counts from the raw data,
including row and column percents.

57

Chapter 4 Relationships Between Categorical Variables

1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the data found in Table 4.1 Randomly pick a letter-S or Q.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Code
S = 0 and Q = 1 entering 0 for S and 1 for Q pressing ENTER
after each entry.
b. Enter the data found in Table 4.2 Randomly pick a letter-Q or S.
Place the cursor on list L2 row 1 to make L2(1) the active list row. Code
S = 0 and Q = 1 entering 0 for S and 1 for Q
pressing
ENTERafter each entry, as shown in Figure 4.1.

Figure 4.1

3. Count the number of observations entered into list L1 and list L2.
On the homescreen press 2nd

LIST I to select the LIST OPS menu.

Select 3: dim(. Press ENTER . Press 2nd L1 ) . Press ENTER to execute

the command.We know that there are 92 entries in list L1, corresponding to the
92 students, as shown in Figure 4.2.
On the homescreen press 2nd LIST I to select the LIST OPS menu.
Select 3: dim(. Press ENTER . Press 2nd L2 ) . Press ENTER to execute
the command.We know that there are 98 entries in list L2, corresponding to the

58

4.1

Displaying Relationships

98 students, as shown in Figure 4.2.

Figure 4.2
4. Count the number of 0s or the number of occurrences of the letter Q.
On the homescreen press 2nd LIST I I to select the LIST MATH menu.
Select 5: sum(. Press ENTER . Press 2nd L1 ) . Press ENTER to execute the command. Since the letter Q was coded as a 1, we know that Q
occurred 31 times out of the 92 responses, as shown in Figure 4.3.
On the homescreen press 2nd LIST I I to select the LIST MATH menu.
Select 5: sum(. Press ENTER . Press 2nd L2 ) . Press ENTER to execute the command. Since the letter Q was coded as a 1, we know that Q
occurred 53 times out of the 98 responses, as shown in Figure 4.3.

Figure 4.3
As a result of counting the number of occurrences of the letter Q, you are
now able to construct a frequency table of the number of occurrences of the
letter Q and the letter S. This information is shown in Table 4.3.
Letters picked
Form
S
Q
Total
S first
61
31
92
Q first

45

53

98

Total
106
84
190
Table 4.3 Occurrences of the letters S and Q

Follow these steps to construct a table of row percents based upon Table 4.3.
1. Enter the data found in Table 4.3 using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
59

Chapter 4 Relationships Between Categorical Variables

a. Place the cursor on list L3 row 1 to make L3(1) the active list row. Focus
on the row labeled S first. Enter the count for S,61 and the count for
Q, 31, pressing ENTER after each entry.
b. Place the cursor on list L5 row 1 to make L5(1) the active list row. Focus
on the row labeled Q first. Enter the count for S,45 and the count for
Q, 53, pressing ENTER after each entry.
2. Calculate the row percents for S first.
On the homescreen, press 2nd L3 2nd LIST I I selecting the LIST
MATH menu. Select 5: sum( and press 2nd L3 ) 100 STO 2nd L4 ,
as shown in Figure 4.4.
3. Calculate the row percents for Q first.
On the homescreen, press 2nd L5 2nd LIST I I selecting the LIST
MATH menu. Select 5: sum( and press 2nd L5 ) 100 STO 2nd L6 ,
as shown in Figure 4.5.

Figure 4.4

Figure 4.5

4. Select STAT ENTER to select the STAT list editor.

a. Place the cursor on list L4 row 1 to make L4(1) the active list row. View
the row percents for S first displayed in list L4, as shown in Figure 4.6.
b. Place the cursor on list L6 row 1 to make L6(1) the active list row. View
the row percents for Q first displayed in list L6, as shown in Figure 4.6.

Figure 4.6

Figure 4.7

As a result of the calculation of the row percents, you are now able to add
the row percent to the table of the number of occurrences of the letter Q
and the letter S. This information is shown in Table 4.4. The % of Row

60

4.1
for Total were calculated seperately.
Letters picked
Form
S
Q
S first
61
31
% of Row
66.3
33.7
Q first
% of Row

45
45.9

53
54.1%

Displaying Relationships

Total
92
100.0
98
100.0

Total

106
84
190
55.8
44.2
100.0
Table 4.4 Occurrences of the letters S and Q
We can use row percents to compare the rates of the letters picked by those
who received the form S first and Q first. The first row of the table
gives the data for those who received the form S first. Among the 92
individuals who received the form S first, 66.3% picked the letter S, and
33.7% picked the letter Q. The second row of the table gives the data for
those who received the form Q first. Among the 98 individuals who
received the form Q first, 45.9% picked the letter S, and 54.1% picked
the letter Q. The difference between the two sets of row percents appears
to indicate a relationship. There is a relationship between two categorical variables forming a two-way table if two or more rows have different
distributions of row percents.
Follow these steps to construct a table of column percents based upon Table
4.3.
1. Enter the data found in Table 4.3 using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Place the cursor on list L3 row 1 to make L3(1) the active list row. Focus
on the column labeled S. Enter the count for S first,61 and the count
for Q first, 45, pressing ENTER after each entry.
b. Place the cursor on list L5 row 1 to make L5(1) the active list row. Focus
on the column labeled Q. Enter the count for S first,31 and the count
for Q first, 53, pressing ENTER after each entry.
2. Calculate the column percents for S.
On the homescreen, press 2nd L3 2nd LIST I I selecting the LIST
MATH menu. Select 5: sum( and press 2nd L3 ) 100 STO 2nd L4 ,
as shown in Figure 4.4.
3. Calculate the column percents for Q.

61

Chapter 4

Relationships Between Categorical Variables

On the homescreen, press 2nd L5 2nd LIST I I selecting the LIST

MATH menu. Select 5: sum( and press 2nd L5 ) 100 STO 2nd L6 ,
as shown in Figure 4.5.

Figure 4.8

Figure 4.9

4. Select STAT ENTER to select the STAT list editor.

a. Place the cursor on list L4 row 1 to make L4(1) the active list row. View
the column percents for S displayed in list L4, as shown in Figure 4.10.
b. Place the cursor on list L6 row 1 to make L6(1) the active list row. View
the column percents for Q displayed in list L6, as shown in Figure 4.11.

Figure 4.10

Figure 4.11

As a result of the calculation of the column percents, you are now able to
add the column percents to the table of the number of occurrences of the
letter Q and the letter S. This information is shown in Table 4.5. The
% of Column for Total were calculated seperately.
Letters picked
Form
S
Q
Total
S first
61
31
92
% of Column
57.5
36.9
48.4
Q first
% of Column

45
42.5

53
63.1

Total

98
51.6

106
84
190
100.0
100
100.0
Table 4.5 Occurrences of the letters S and Q
We can use column percents to compare the rates of the form received by
those who picked the letter S and those who picked the letterQ. The
first column of the table gives the data for those who picked the letter S.
Among the 106 individuals who picked the letter S, 57.5% received the
form S first, and 42.5% received the form Q first. The second column
of the table gives the data for those who picked the letter Q. Among
62

4.4

Assessing the Statistical Significance of a 22 T a b l e

the 84 individuals who picked the letter Q, 36.9% received the form S
first, and 63.1% received the form Q first. The difference between the
two sets of column percents appears to indicate a relationship. There is a
relationship between two categorical variables forming a two-way table if
two or more columns have different distributions of column percents.

4.4

Assessing the Statistical Significance of a 22 T a b l e

Example - Continued
Question 3 in the class survey described in Section 2.1 asked 92 college students
to Randomly pick a letter-S or Q. Another 98 college students were asked to
Randomly pick a letter-Q or S. The data is contained in the PennState1 worksheet and is displayed in Table 4.1 and Table 4.2. Table 4.3 contains the frequency
table of the number of occurrences of the letter Q and the letter S.
The steps involved in computing a chi-square-test on the TI-83 S.E. and TI-84 S.E.
require:
entering the observed counts in a matrix. Enter that matrix variable name at the
Observed: prompt in the 2 .Test editor with the default being matrix A: [A]. At
the Expected: prompt, enter the matrix variable name to which you want the computed expected counts to be stored with the default being matrix B: [B].
Calculating the 2 test statistic.
Examining the matrix of Expected counts obtained by calculating the 2 test statistic.
Follow these steps to conduct a chi-square test, including observed counts, computing a chi-square statistic and finding the p-value.
1. Enter the observed counts in matrix A.
a. Press 2nd MATRIX , located on row 4, left hand column. and press
I I selecting the MATRIX EDIT menu, as shown in Figure 4.12.
Select matrix [A]. Press ENTER .
b. Enter the dimensions of matrix A. Press 2 (rows), ENTER , press 2 (columns)
ENTER . Refer to Table 4.3 to obtain the observed counts. Enter the element in the first row, first column: 61; press ENTER . Enter the element
in the first row, second column: 31; press ENTER . Enter the element in
the second row, first column: 45; press ENTER . Enter the element in the
second row, second column: 53; press ENTER . The resulting matrix is

63

Chapter 4

Relationships Between Categorical Variables

shown in Figure 4.13.

Figure 4.12

Figure 4.13

Press 2nd QUIT .

2. Compute the chi-square statistic.
Press STAT I I selecting the STAT TEST menu. Select C: 2 Test, as
shown in Figure 4.14. Press ENTER . If matrix A is not listed in Observed:
and matrix B is not listed in Expected,as shown in Figure4.15 then follow
these instructions:
a. Place matrix A in Observed: by selecting 2nd MATRIX >NAMES 1:
[A] and pressing ENTER , as shown in Figure 4.16.
b. Place matrix B in Expected: by selecting 2nd MATRIX >NAMES 2:
[B] and pressing ENTER , as shown in Figure 4.16.

Figure 4.14

Figure 4.15

Figure 4.16

c. Calculate the chi-squared statistic by highlighting Calculate and press ENTER ,

as shown in Figure 4.17. The results are shown in Figure 4.18.

Figure 4.17
Figure 4.18
Using the TI calculator, the p-value is found to be 0.005. The p-value tells
us that the chance is only 0.005 (which is really 5 in 1000) that we would
get a chi-square statistic as large as 7.995 (or larger) if there really is no
relationship between the order of the letters on the form and the letter that
would be picked by people in this population. In the context of this problem, this means that there is a statistically significant relationship between
the form of the question (S first or Q first) and the letter picked in the
population.
d. View the expected counts for the two-way table..

64

4.4

Finding a P-value

Press 2nd MATRIX >NAMES 2: [B] and pressing ENTER to view

the expect counts shown in Figure 4.19.

Figure 4.19
Use the right arrow key, I , to scroll through the expected values. As a
result of viewing the expected values, you are now able to add the expected
values to the table of the number of occurrences of the letter Q and the
letter S. This information is shown in Table 4.6. The % of Column for
Total were calculated seperately.
Letters picked
Form
S
Q
Total
S first
61
31
92
Expected Value
51.33 40.67
Q first
Expected Value

45
54.67

53
43.33

98

Total
106
84
190
Table 4.6 Occurrences of the letters S and Q

The Family of 2 -Distributions

A 2 -distribution is used to find the p-value for an 2 -test of the null hypothesis that there is no association between the two variables. The family of 2 distributions is a family of skewed distributions, each with a minimum value of 0.
A specific 2 distribution is indicated by the parameter called degreesof freedom
In 2 -test,the degrees of freedom is df = k  1 (number of groups - 1).

Finding a P-value
Finding a p-value given the 2 value and degrees of freedom. In Example 4.10,
Figure 4.18, the p-value is reported as part of the output. The TI distribution function 2 cdf (lowerbound, upperbound, df ) computes the 2 distribution probability
between the lowerbound and upperbound for the specified df.

65

Chapter 4

Relationships Between Categorical Variables

Follow these steps to obtain the p-value obtained in Example 4.10, Figure 4.18.
1. Find a p-value for an  2 -distribution:
Press 2nd DISTR , using the down arrow key, H , several times to select
7: 2 cdf(, as shown in Figure 4.20 and press ENTER. Type the lowerbound,
7.995 , upperbound,1E99, df,1 ) . The upperbound,1E99, is translated as 1
1099 . Press ENTER to execute the command. The results are shown in Figure
4.21.

Figure 4.20
Figure 4.21
The area to t he right of 2 = 7.995 under the 2 - distribution is the same as
the p-value, 0.005.

66

Chapter 5
Sampling: Surveys and How
to Ask Questions
There are two major categories of statistical techniques that can be applied to data.
The first is descriptive statistics, in which we use numerical and graphical summaries to characterize a dataset. We partially covered descriptive statistics in Chapter 2, and we introduced additional descriptive techniques in Chapters 3 and 4.The
second important category of statistical techniques is inferential statistics, i n
which we use sample data to make conclusions about a broader range of individuals than just those who are observed. For example, in Case Study 1.6 about aspirin
use and the risk of heart disease, the data from a sample of 22,071 physicians was
used to infer that taking aspirin helps prevent heart attacks for all men similar to
the participants.
In Chapters 5 and 6, you will learn how to collect representative data. In these
chapters you willleam that the data collection method used affects the extent to
which you can use sample data to make inferences about a larger population. Descriptive summaries such as the mean and standard deviation, as well as graphical
techniques, can be used whether the data are from a sample or from an entire population, but inferential methods can be used only when the data in hand are from
a representative sample for the question being asked a about a larger population.
When you use inferential methods, a key concept is that you have to think about
both the source of the data and the question(s) of interest. A dataset may contain
representative infonnation for some questions but not for others.
The Fundamental Rule for Using Data for Inference is that available data can
be used to make inferences about a much larger group if the data can be considered ,to be representative with regard to the question(s) of interest.
After reading this chapter you should be able to:
1. Select a simple random sample.
Keystrokes Introduced
1. 2nd LIST I , selecting 5: seq( from the OPS menu. The arguments in seq(
are seq(expression,variable,begin,end [,increment ]). You will use
seq(expression,variable,begin,end [,increment ]) to create a column of ID lables.
2. MATH I I I selecting 5: randInt( from the PRB menu. The arguments
in randInt( are randInt(lower,upper [,numtrials ]). You will use randInt( are
randInt(lower,upper [,numtrials ]) to randomly select students.

67

5.1

5.1

Simple Random Samples

Populations, Samples, and Simple Random Samples

In most statistical studies, the objective is to use a small group of units to make
an inference about a larger group. The larger group of units about which inferences are to be made is called the population. The smaller group of units actually
measured is called the sample. Sometimes measurements are taken on the whole
group of interest, in which case these measurements comprise a census of the whole
population. Occasionally you will see someone make the mistake of trying to use
census data to make inferences to some hypothetical larger group when there
isnt one.

Simple Random Samples

Rememberthe fundamental rule for making valid inferences about the group represented by the sample for which the data were measured: The data must be representative of the larger group with respect to the question of interest.The principal
way to guarantee that sample data represents a larger population is to use a simple
random sample from the population.
With a simple random sample, every conceivable group of units of the required
size from the population has the same chance to be the selected sample.
An ideal data collection method is to obtain a simple random sample of the population of interest, or to collect sample data using one of the more complex random
sampling methods described later in this chapter. In some research studies, however, random sampling is not possible for both practical and ethical reasons. For
instance, suppose researchers want to study the effect of using marijuana to reduce
pain in cancer patients. It would be neither practical nor ethical to select a random
sample of all cancer patients to participate. Instead, the researchers would use
volunteers who want to take part, and hope these volunteers represent the larger
population of all cancer patients. The use of volunteers will be discussed more
fully in Chapter 6, when we cover randomized experiments.
Simple random samples and related sampling methods are typically used for one
type of statistical study: sample surveys or polls. Remember from Chapter 1 that in
a sample survey the investigators gather opinions or other information from each
individual included in the sample. Because this gathering of information is usually
not time consuming or invasive, it is often both practical and ethical to contact a
large random sample from the population of interest. Throughout this chapter we
will learn more about how to select simple random samples and how to conduct
sample surveys.

68

Chapter 5

Sampling: Surveys and How to Ask Questions

Example: Finding a Simple Random Sample Using UCDavis1.

Students in a liberal arts course in statistical literacy were given a survey that
inc1uded questions on how many hours per week they watched television. The
responses are shown in Table 5.1 and are contained in the UCDavis1 data file.
13
2 20 15
8
3
2
4
8
1
8 28
4
11 10
1 10 10
1
4
2 40 16 10 30 10
2 10 15
4
6 100
6 15
1
2
4 10
1
1
4
1
6
2
4 10 18 20
4 20
5
0
11
0
1
2
8
1
3
6 10 15 15 12
2
4 15 21
4
4
8
2
4 10
2
9
7 14
2
4
0 10 10
25
6 14
0 21 14 11
8
2
2 14
2
6
20 14
1 14 10 15
2 10
6 20 20 35 15
5 14 35
1
4
0 14
5
5
5
1
1
9
15
5
8
1 10
2
7 14
1
1
2
1
4
3
8
1
3 12 30 15
1
9 25
2
3
1
4 30 20
3
2 15 16
5
8 10
2
8 10
10
6
4
8
3
1
5
8
2
9
1
5
Table 5.1
Follow these steps to find a simple random sample of 10 students weekly television watching amounts (variable is TV).
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.
Press STAT ENTER to select the STAT list editor.
a. Enter the weekly television watching amounts (variable is TV) in list L3.
Place the cursor on list L3 row 1 to make L3(1) the active list row. Enter the
weekly television watching amounts from Table 5.1 row by row. type 13,
2, 20, ...9, 1, 5 pressing ENTER after each entry to enter all 173 weekly
television watching amounts, as shown in Figure 5.1.

69

5.1

Simple Random Samples

b. Create a column of ID lables in list L1.

Press 2nd QUIT . On the homescreen, we will create a list of ID labels
from 1 to 173 in list L1. Press 2nd LIST I , selecting 5: seq(from the
OPS menu, as shown in Figure 3..
2. Press ENTER , placing seq( on the
homescreen.
The arguments in seq( are seq(expression,variable,begin,end [,increment ]).
Enter the expression by pressing X,T ,,n found on row three, column
two; press , .
Enter the variable by pressing X,T ,,n , , .
Enter the values for begin, end, and increment. Type 1 , 173 , 1 followed
by ) .
Store the results in list L1 by pressing STO 2nd L1 . Press ENTER to
execute the comand. The homescreen is shown in Figure 5.3.

Figure 5.1

Figure 5.2

Figure 5.3

3. Randomly select 10 students weekly television watching amounts, storing the

results in list L2.
On the homescreen, press MATH I I I selecting 5: randInt( from the
PRB menu, as shown in Figure 5.4. Press ENTER placing randInt( on the
homescreen.
The arguments in randInt( are randInt(lower,upper [,numtrials ]).
Enter the lower, upper and numtrials by typing 1 , 173 , 10 followed by STO 2nd L2 .
.Press ENTER to execute the comand. The results are shown in Figure 5.5.
4. View the randomly selected 10 students weekly television watching amounts.
Press STAT ENTER to select the STAT list editor.
The ID lables are displayed in list L1, the 10 randomly selected 1D labels are
displayed in list L2, and the weekly television watching amounts are displayed
in list L3, as shown in Figure 5.6.

Figure 5.4

Figure 5.5
70

Figure 5.6

Chapter 5

Sampling: Surveys and How to Ask Questions

The TI output, as shown in Figure 5.6, indicates the follwing 10 randomly

selected ID lables act as pointers to 10 students weekly television watching
amounts: 25 30, 173 5, 98 14, 114 4, ..., 15$ 10.
Your results most certainly would be different since these are randomly selected
labels.

71

Chapter 6
Gathering Useful Data
For Examining Relationships
In this chapter, we learn about ways to collect data in order to examine relationships
between variables. We have already seen several examples that involved possible
links between variables. In Chapter 2, Example 2.1 was about the connection between gender and seat belt use for 12th grade students. Example 2.2 was about a
possible connection between the use of nightlights in infancy and nearsightedness.
Case Study 1.6 described a study that demonstrated a link between taking an aspirin a day and a decreased risk of heart attacks for men.
In studies like these, we want to know if a cause-and-effect relationship exists.
That is, we want to know if changing the value of one variable cause ehanges in
another variable. We will learn in this chapter that the way a study is conducted
affects our ability to infer that a cause-and-effect relationship exists.

6.1

Speaking the Language of Research Studies

Although there are a number of different strategies for collecting meaningful data,
there is common terminology used in most of them. Statisticians tend to borrow
words from common usage and apply a slightly different meaning, so be sure you
are familiar with the special usage of a word in a statistical context.
Types of ResearchStudies
There are two basic types of statistical research studies conducted to detect relationships between variables:
observational studies
experiments
In an observational study, the researchers simply observe or question the participants about opinions, behaviors, or outcomes. Participants are not asked to do
anything differently. For example, Case Study 1.5 described an observational study
in which blood pressure and frequency of certain types of religious activity (like
prayer and church attendance) were measured. The goal was to see if people with
higher frequency of religious activity had lower blood pressure. Researchers simply measured blood pressure and frequency of religious activity. They did not
ask participants to change how often they prayed or went to religious services, or
change any other aspect of their lives.
In an experiment, researchers manipulate something and measure the effect of the

72

Chapter 6

Gathering Useful Data For Examining Relationships

manipulation on some outcome of interest. Randomized experiments are experiments in which the participants are randomly assigned to participate in one condition or another. The different conditions are called treatments.
A major theme of this chapter will be that a randomized experiment provides
stronger evidence of a cause-and-effect relationship than an observational study.

6.2

Designing a Good Experiment

An experiment measures the effect of manipulating the environment of the participants in some way. With human participants, the manipulation may include
receiving a drug or medical treatment, going through a training program, agreeing
to a special diet, and so on. Most experiments on humans use volunteers because
you cant very well force someone to accept a manipulation. Experiments are also
done on other kinds of experimental units, such as when different growing conditions are compared for their effect on plant yield, or different paints are applied on
highways to see which ones last longer. The idea is to measure the effect of the
feature being manipulated, the explanatory variable, on the response variable.
In a randomized experiment, participants usually are randomly assigned to either
receive a specific treatment or to take part in a control group. The purpose of the
random assignment is to make the groups approximately equal in all respects except for the explanatory variable, which is purposely manipulated. Differences in
the response variable between the groups, if large enough to rule out natural chance
variability, can then be attributed to the manipulation of the explanatory variable.
After reading this chapter you should be able to:
1. Use simulation to obtain a random sample.
Keystrokes Introduced
1. MATH I I I the PRB (probability) menu, select 5: randInt(lower,upper
[,numtrials ]). The command generates and displays a random integer within a
range specified by lower and upper integer bounds for a specified number of
trials numtrials.
2. 2nd LIST I OPS>1: SortA(listname ) sorts elements of listname in ascending order.
3. DEL used to delete an entry.

73

6.3

6.3

Simple Random Sampling and Randomization

Simple Random Sampling and Randomization

We have already encountered several examples of the type of problem we will study
in this chapter. In Chapter 2, for instance, we described a study of 479 children
that found that children who slept either with a nightlight or in a fullylit room
before the age of two had a higher incidence of myopia (nearsightedness) later
in childhood. Example 6.10 - Finding Gifted ESP Participants will be used to
illustrate how relationships between categorical variables may be presented.
Example 6.3 - Assigning Children to Lift Weights
In Case Study 6.2, 43 children were randomly assigned to one of three treatment
groups. Children in group 1 performed weight-lifting repetitions with a heavy load,
group 2 performedmore repetitions but with a moderate load, and group 3 was a
control group that did not lift weights. There were 15 children assign to group 1,
16 to group 2, and 12 to group 3.
Suppose we are asked to randomly assign children to treatment goups. How could
we carry out this randomization?
One way would be to use a random integer generator, like the one on the
T I - 8 3 P l u s S.E. or TI-84S.E. First , we think of the children as being labeled
with integers from 01 to 43. Then you will choose a simple random sample of size
15.
Follow these steps to choose 15 children to assign to Group 1 and the 12 children
to assign to the Control Group.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Use the random integer function to select 15 children to assign to Group 1.
a. On the homescreen, press MATH , located on the fourth row, left column.
Press I I I to select the PRB (probability) menu. Select 5: randInt(
to generate and store a random integer.Type 1,43,15 to select 15 integers
from 1 to 43. Press ) STO$ 2nd L1 to store the random integers in
list L1, as shown in Figure 6.1.

74

Chapter 6

Gathering Useful Data For Examining Relationships

b. Check for duplicates.

On the homescreen, press 2nd LIST , located on the third row, column
3. Press 2: SortA( and 2nd L1 ) ENTER to place the list in ascending
order, as shown in Figure 6.2. Choose STAT >EDIT to view the list in the
STAT list editor. If duplicates exist, repeat all of Step 2, until no duplicates are present by pressing 2nd QUIT , exiting the STAT editor.Press
2nd ENTER 2nd ENTER to execute the commands once again.

Figure 6.1

Figure 6.2

3. Use the random integer function to select 12 children to assign to the Control
Group.
a. On the homescreen, press MATH , located on the fourth row, left column.
Press I I I to select the PRB (probability) menu. Select 5: randInt(
to generate and store a random integer.Type 1,43,30 to select 30 integers
from 1 to 43. Press ) STO$ 2nd L2 to store the random integers in
list L2, as shown in Figure 6.3.
b. Check for duplicates.
On the homescreen, press 2nd LIST , located on the third row, column 3.
Press 2: SortA( and 2nd L2 ) ENTER to place the list in ascending order, as shown in Figure 4.4. Choose STAT >EDIT to view the lists in the
STAT list editor. View list L1 and list L2 side by side. If duplicates exist
in L2, press DEL to delete a duplicate entry in L2. If an entry appear in
list L1 and also in list L2, press DEL to delete the entry in list L2. Do this
until only 12 entries appear in list L2. You may wish to repeat all of Step
3, until no duplicates are present in list L2 by pressing 2nd QUIT , exiting the STAT editor.Press 2nd ENTER 2nd ENTER to execute the
commands once again. The results are shown in Figure 6.5. Your results
are most likely to be different since these are random numbers.

75

6.3

Simple Random Sampling and Randomization

Figure 6.3
Figure 6.4
Figure 6.5
The TI calculator output, as shown in Figure 6.5, indicates 15 students labeled 3, 8, 9, ... are in Group 1, while 12 students labeled 5, 6, 7 are in the
Control Group. The remaining 16 children will constitute Group 2.
Other methods of making random assignments would also work. For instance, Minitab or Excell could also be used to create random assignments.

76

Chapter 7
Probability
Statistical methods are used to evaluate information in uncertain situations and
probability plays a key role in that process. Remember our definition of statistics
from Chapter 1: Statistics is a collection of procedures and principles for gathering
data and analyzing information in order to help people make decisions when faced
with uncertainty. Decisions like whether to buy a lottery ticket, whether to buy an
extended warranty on a computer,or which of two courses to take are examples of
decisions that you may have to make that involve uncertainty and the evaluation
of probabilities.
Probability calculations also are a key element of statistical inference. In Chapter
6 we introduced p-values, which are probabilities used to determine if the results
of a study are statistically significant. As a reminder of how p-values are used,
consider Case Study 1.6 in which 22,071 physicians were randomly assigned to
take either aspirin or a placebo. There were 189 heart attacks in the placebo group
but only 104 in the aspirin group. Could this have happened just by the luck of
how the physicians were randomized to the treatment groups?
Suppose that regardless of which group they were in, 104 + 189 = 293 of the men
would have had heart attacks anyway. What is the probability that, just by the luck
of random assignment, the numbers of heart attacks in the two groups would have
been so different? In other words, if aspirin and placebo are equally effective (or
ineffective), what is the probability that we would see such a large discrepancy in
the proportion of heart attacks in the two groups? The answer is the p-value, which
is less than .00001. This is strong evidence that these results did not just occur by
chance. From this, we conclude that aspirin really did reduce the number of heart
attacks in the group that took it.
After reading this chapter you should be able to:
1. Use simulation to estimate probabilities.
Keystrokes Introduced
1. 2nd LIST I OPS>5: seq(expression, variable, begin, end[,increment] )
returns a list.
2. MATH I I -PRB>7: randBin(numtrials,prob,[,numsimulations] ) generates and displays a random real number from a specified Binomial distribution.
3. MATH I I -PRB>6: randInt(lower,upper,numtrials ) generates and displays a random integer within a range specified by lower and upper for a specified number of trials.

77

Chapter 7

7.1

Probability

Using Simulation to Estimate Probabilities

Some probabilities are so difficult or time-consuming to calculate that it is easier
to simulate the situtation repeatedly using a computer or calculator and observe the
relative frequency of the event of interest. If you simulate the random circumstance
n times and the outcome of interest occurs in x out of those n times, then the
estimated probability for the outcome of interest is nx . This is an estimate of the
long-run relative frequency with which the outcome would occur in real life.
Example 7.30 - Finding Gifted ESP Participants
An ESP test is conducted by randomly selecting one of five video clips and playing it in one building, while a participant in another building tries to describe what
is playing. Later, the participant is shown the five video clips and is asked to determine which one best matches the description he or she had given. By chance, the
participant would get this correct with probability 1/5. Individual participants are
each tested eight times, with five new video clips each time. They are identified as
gifted If they guess correctly at least five times out of the eight tries. Suppose
people actually do have some ESPand can guess correctly with probability .30 (instead of the .20 expected by chance). What is the probability that a participant will
be identified as gifted?
In Chapter 8 you will learn how to solve this kind of problem, but we can simulate the answer using a TI calculator to produce the digits 0, 1,2, .. .,9 with equal
likelihood. A random number table is available in the text. Many calculators and
computers will simulate these digits. Here are the steps needed for one repetition:
Each guess is simulated with a digit, equally likely to be 0 to 9.
For each participant, we simulate eight guesses resulting in a string of eight digits.
If a digit is 7, 8, or 9, we count that guess as correct so P(correct) =
3/10 = .3, as required in the problem. If the digit is 0 to 6, the guess Is incorrect.
(There is nothing special about 7, 8,9; we could have used any
three digits.)
If there are five or more correct guesses (digits 7, 8, 9), we count that as
gifted.
The entire process is repeated many times, and the proportion of times the result
is a gifted participant is an estimate of the desired probability.
Follow these steps to simulate this experiment for one participant, exploring the
step-by-step process.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
78

7.1

Using Simulation to Estimate Probabilities

2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Place a sequence of random real numbers from a integer distribution into list
L1.
On the homescreen press 2nd LIST I to select the LIST OPS menu.
a. Obtain a sequence of 8 random digits from {0,1,2,3,4,5,6,7,8,9}.
Press MATH I I I to the PRB menu. Select 5: randInt(. Type
0 , 9 ) x,T,, n , 1 , 8 , 1 ) . The x,T,, n key is located on the 3rd row,
end column. Press ENTER to execute the command, as shown in Figure
7.1.

Figure 7.1

Figure 7.2

3. If a digit is 7, 8, or 9, we count that guess as correct. Code each digit in list

L1 as success (1), if the digit is a 7, 8, or 9. Code each digit in list L1 as failure
(0) if the digit is less than 7. You will store the results in list L2.
Press 2nd TEST displaying the TEST menu.
a. Select 4: > and press ENTER . Press 7 to compare to the smallest success number. Observe that 0 is the code for failure or False (the digit
was < 7). Observe that 1 is the code for success or true (the digit was
> 7). The output from the TI calculator is displayed in Figure 7.2.
4. Count the number of successes.
On the homescreen press 2nd LIST I I to select the LIST MATH menu.
(i) Select 5: sum(. Press ENTER . Press 2nd ANS ) . The ANS
key is located above the gray negation key on the bottom row, column
four. The number of successes are displayed in Figure 7.3.

79

Chapter 7

Probability

Figure 7.3
Notice that this participant got 3 guesses of digits that were a 7, 8,
or 9.
Follow these steps to simulate this experiment for a 100 participants.
1. Place a sequence of random real numbers from a integer distribution into list
L1.
On the homescreen press 2nd LIST I to select the LIST OPS menu.
a. Obtain a sequence of 100 random digits {0,1,2,3,4,5,6,7,8,9}storing the
sequence in list L1.
Press MATH I I I to the PRB menu. Select 7: randBin(. Type
8 , 0.3 ) x,T,, n , 1 , 100 , 1 ) STO$ 2nd L1 . The x,T,, n key
is located on the 3rd row, end column. Press ENTER to execute the command, as shown in Figure 7.4.

Figure 7.4

Figure 7.5

2. If there are five or more correct guesses (digits 7, 8, 9), we count that as
gifted. Code each participant in list L1 who had five or more correct
guesses as success (1). Code each participant in list L1 who had fewer than
five correct guesses as as failure (0). You will store the results in list L2.
Press 2nd TEST displaying the TEST menu.
a. Select 4: > and press ENTER . Press 5 to compare to the smallest success number. Press STO$ 2nd L2 , storing the 1s and 0s in list
L2. Observe that 0 is the code for failure or False (the digit was < 7).
Observe that 1 is the code for success or true (the digit was > 7). The
80

7.1

Using Simulation to Estimate Probabilities

output from the TI calculator is displayed in Figure 7.5.

3. Count the number of successes.
On the homescreen press 2nd LIST I I to select the LIST MATH menu.
(i) Select 5: sum(. Press ENTER . Press 2nd L2 ) , selecting list L2.
The number of successes are displayed in Figure 7.3.

Figure 7.6
Notice that there were 2 participants who got five or more correct, so
the probability of finding a gifted participant in this simulation is
2
= 0.02. In other words, if everyone is equally talented, and
about 100
each guess is correct with probability 0.3, there there will be five or
more correct guesses out of eight tries with probability 0.02, or about
2% of the time.

81

Chapter 8
Random Variables
The numerical outcome of a random circumstance is called a random variable. In
this chapter, well learn how to characterize the pattern of the distribution of the
values that a random variable may have, and well learn how to use the pattern to
find probabilities. Patterns make life easier to understand and decisions easier to
make. For instance, dogs come in a variety of breeds, sizes, and temperaments, but
all dogs fit certain patterns that veterinarians can rely upon when treating nearly
any type of dog. If a veterinarian had to learn a different pattern for treating every
different breed, it might be nearly impossible for any individual to learn enough to
be able to treat dogs in general.
After reading this chapter you should be able to:
1. List the probabilities for a binomial experiment.
2. Find exact and cumulative probabilities for a specific value of x, given n and
p.
3. Find probabilities for a uniform distribution.
4. Find standard normal probabilities.
5. Find probabilities for any normal distribution.
6. Find percentiles for a normal distribution.
Keystrokes Introduced
1. 2nd DISTR , selecting 0: binomialpdf(. The arguments are
binomialpdf(numtrials,p[,x]). The function computes a probability at x for the
discrete binomial distribution wiht the specified numtrials and probability p of
success on each trial.
2. 2nd DISTR , selecting A: binomialcdf( . The arguments are
binomialpdf(numtrials,p[,x]). The function computes a cumulative probability at x for the discrete binomial distribution wiht the specified numtrials and
probability p of success on each trial.
3. 2nd DISTR , selecting 2: normalcdf( . The arguments are
normalcdf(lowerbound,upperbound [,,]). The function computes the normal distribution probability between lowerbound and upperbound for the specified and .
4. 2nd DISTR , selecting 3: invNorm( .The arguments are invNorm(area [,,]).
The function computes the inverse cumulative normal distribution function for
a given area under the normal distribution specified by and .

82

8.1

8.1

Binomial Random Variables

Binomial Random Variables

In this section, we consider an important family of discrete random variables called
binomial random variables. Certain conditions must be met for a variable to fall
into this family, but the basic idea is that a binomial random variable is a count of
how many times an event occurs (or does not occur) in a particular number independent observations or trials that make up a random circumstance.
Binomial Experiments and Binomial Random Variables
The number of heads in three tosses of a fair coin, the number of girls in six independent births, and the number of men who are six feet tall or taller in a random
sample of ten adult men from a large population are all examples of binomial random variables. A binomial random variable is defined as X = number of successes
in the n trials of a binomial experiment.
A binomial experiment is defined by the following conditions:
1. There are n trialswhere n is specified in advance and is not a random value.
2. There are two possible outcomes on each trial, called success and failure
and denoted Sand F.
3. The outcomes are independent from one trial to the next.
4. The probability of a success remains the same from one trial to the next, and
this probability is denoted by p. The probability of a failure is 1p for every
trial.
Example List the Probabilities for a Binomial Experiment
As an example of listing the probabilites for a binomial experiment, let us use
n = 10 and p = 0.25 as an example.
Follow these steps to find the probabilities for a binomial experiment where n =
10 and p = 0.25.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Enter data using the STAT list editor.

83

Chapter 8

Random Variables

Press STAT ENTER to select the STAT list editor.

a. Method 1: List the probability distribution in the STAT editor..
Enter the data for the binomial random variable in list L1.
Place the cursor on list L1 row 1 to make L1(1) the active list row. Enter
values from 0 to 10: 0, 1, 2, ...,10 pressing ENTER after each entry.
Place the cursor at the top of list L2, on the label L2. Press 2nd DISTR ,
selecting 0: binomialpdf( , and Press ENTER . Type the number of trials,
10, a , , the probability of success, 0.25, and ) as shown in Figure 8.2.
Press ENTER to execute the command. The results are shown in Figure
8.3

Figure 8.1

Figure 8.2

Figure 8.3

The TI output, as shown in Figure 8.3, indicates the P (x = 0) = 0.05631,

P (x = 1) = 0.18771, P (x = 2) = 0.28157, etc.
3. Method 2: Listing the probability distribution on the homescreen..
Press 2nd QUIT returning to the homescreen.
Press 2nd DISTR , selecting 0: binomialpdf( , type the number of trials, 10,
a , , the probability of success, 0.25, and ) as shown in Figure 8.4. Press
ENTER to execute the command. The results are shown in Figure 8.5. Use
the right arrow key, I , to view the individual probabilities.

Figure 8.4

Figure 8.5

The TI output, as shown in Figure 8.5, indicates the P (x = 0) = 0.05631,

P (x = 1) = 0.18771, P (x = 2) = 0.28157, etc.
4. Method 3: Producing individual probabilities..

84

8.1

Binomial Random Variables

Press 2nd DISTR , selecting 0: binomialpdf( , type the number of trials, 10,
a , , the probability of success, 0.25, the number of successes, 0, and ) as
shown in Figure 8.6. Press ENTER to execute the command. The results are
shown in Figure 8.7.

Figure 8.6

Figure 8.7

The TI output, as shown in Figure 8.7, indicates the P (x = 0) = 0.05631.

Repeat the above process for another value of x.
Press 2nd DISTR , selecting 0: binomialpdf( , type the number of trials, 10,
a , , the probability of success, 0.25, the number of successes, 1, and ) as
shown in Figure 8.8. Press ENTER to execute the command. The results are
shown in Figure 8.9.

Figure 8.8

Figure 8.9

The TI output, as shown in Figure 8.9, indicates the P (x = 1) = 0.18771.

Example 8.16 Calculations for Number of Girls in Ten Births
Let X = number of girls in ten births, and assume that p = 0.488 is the probability
that any birth is a girl. This value of p is based on birth records in the United States.
Follow these steps to find the probability of exactly 7 girls in ten births.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.

85

Chapter 8

Random Variables

Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
c. Press 2nd DISTR , selecting 0: binomialpdf( , type the number of trials,
10, a , , the probability of success, 0.488, the number of successes, 7, and
) as shown in Figure 8.10. Press ENTER to execute the command. The
results are shown in Figure 8.11.

Figure 8.10

Figure 8.11

The TI output, as shown in Figure 8.11, indicates the P (x = 7) = 0.106.

2. Find cumulative probabilities.
a. Find the probability of having at most 7 girls out of 10 births. An equilavent
statement is to find the probability of having 7 or fewer girls out of 10
births.
Press 2nd DISTR , selecting A: binomialcdf( , type the number of trials,
10, a , , the probability of success, 0.488, the number of successes, 7, and
) as shown in Figure 8.12. Press ENTER to execute the command. The
results are shown in Figure 8.13.

Figure 8.12

Figure 8.13

The TI output, as shown in Figure 8.13, indicates the P (x  7) = 0.9533.

b. Find the probability of having at least 7 girls out of 10 births. An equilavent
statement is to find the probability of having 7 or more girls out of 10 births.
What we will do is to subtract the probability of having at most 6 girls out
of 10 births from the sum of the probabilities, 1.
Press 1 - 2nd DISTR , selecting A: binomialcdf( , type the number of
trials, 10, a , , the probability of success, 0.488, the number of successes,
6, and ) as shown in Figure 8.14. Press ENTER to execute the com-

86

8.2

Continuous Random Variables

mand. The results are shown in Figure 8.15.

Figure 8.14

Figure 8.15

The TI output, as shown in Figure 8.13, indicates the P (x  7) = 0.1529.

8.2

Continuous Random Variables

We learned in Section 8.1 that a .continuous random variable is one for which
the outcome can be any value in .an interval of collection of intervals. In practice, all measurements are rounded to a specified number of decimal places, so we
may not be able to accurately observe all possible outcomes of a continuous variable. Forexample, the limitations of weighing scales keep us from observing that
a weight may actually be 128.3671345993pounds. Generally, however, we call a
random variable a continuous random variable ifthere are a large number of observable outcomes covering an interval or set of intervals.
For a discrete random variable,we can find the probability that the variable X exactly equals a specified value. We cant do this for a continuous random variable.
For a continuous random variable, we are only able to find the probability that X
falls between two values. In other words, unlike discrete random variables, continuous random variables do not have probability distribution functions specifying
the exact probabilities of specified values. Instead, they have probability density
functions, which are used to find probabilities that the random variable falls into a
specified interval of values.
Example 8.19 Time Spent Waiting for a Bus I
A bus arrives at a bus stop every 10 minutes. If a person arrives at the bus stop at
a random time, how long will he or she have to wait for the next bus? Define the
random variable X = waiting time until the next bus arrives. The value of X could
be any value between 0 and 10 minutes, and X is a continuous random variable.
(In practice, the limitations of watches would force us to round off the exact time.)
Figure 8.16 shows the probability density function for the waiting time. Possible
waiting times are along the horizontal axis, and the vertical axis is a density scale.
The height of the curve is .1 for all X between 0 and 1, so the total area between
0 and 10 minutes is (10)(.1) =1.
The density function shown in Figure 8.16 is a flat line that covers the interval of
possible values for X. There is a uniformity to this density curve in that every

87

Chapter 8

Random Variables

interval with the same width has the same probability. A random variable with
this property is called a uniform random variable and is the simplest example of a
continuous random variable.

Density

0.10

0.05

0.00

4
5
6
Waiting Time (min)

10

Figure 8.16
Suppose we want to find the probability that the waiting timeX was in the interval
from 5 to 7 minutes. The general principle for any continuous random variable is
that the probability P (a  x  b) is the area under the curve over the interval
from a to b. In this example, the area under the curve is the area of a rectangle
that has width = 7 - 5 = 2 minutes and height = .1. This area is (2)(.1) = .2, which
is the probability that the waiting time is between 5 and 7 minutes. In Figure 8.17,
the shaded area represents the desired probability.
Follow these steps to find probabilities for a uniform distribution.
1. On the homescreen, multiply the base times the height by typing 2 0.1.
Press ENTER to obtain the area (probability) under the uniform distribution
of 0.2, as shown in Figure 8.18.

Figure 8.17

Figure 8.18

88

8.3

8.3

Finding Probabilities for z-Scores

Normal Random Variables

The most commonly encountered type of continuous random variable is the normal random variable which has a specific form of a bell-shaped probability density curve called a normal curve. A normal random variable is also said to have
a normal distribution. Anynormal random variable is completely characterized
by specifying values for its mean, , and standard deviation .
Nature provides numerous examples of measurements that follow a normal curve.
The fact that so many different kinds of measurements follow a normal curve is
not surprising. On many attributes, the majority of people are somewhat close to
average, and as you move further from the average, either above or below, there
are fewer people with such values.
Features of Normal Curves and Normal Random Variables
As with any continuous random variable, the probability that a normal random
variable falls into a specified interval is equivalent to an area under its density
curve. Also, P (X = k) = 0, meaning that the probability is 0 that a normal random variable X exactly equals any specified value.
Some features shared by all normal curves and normal random variables (X) are:
1. The normal curve is symmetric and bell-shaped (but not all symmetric bellshaped density curves are normal curves).
2. P (X  ) = P (X  ) = 0.5, meaning that there are equal probabilities for
a measurement being less than the mean and greater than the mean.
3. P (X  d)P (X  d) for any positive number d. This means that the
probability that X is more than d units below the mean equals the probability
that X is more than d units above the mean.
4. The Empirical Rule holds:
a. P (    X  + ) r 0.68
b. P (  2  X  + 2) r 0.95
c. P (  3  X  + 3) r 0.997
Standardized Scores
We learned in Chapter 2 that a standardized score, also called a z-score, is the
distance between a specified value and the mean, measured in number of standard
deviations. We repeat the definition here using notation for random variables.
The formula for converting any value x to a z-score is
x
Value - Mean
z = Standard
deviation = 

Finding Probabilities for z-Scores

A normal random variable with mean = 0 and standard deviation  = 1 is said
89

Chapter 8

Random Variables

to be a standard normal random variable and to have a standard normal distribution. When we convert values for any normal random variable to z-scores, it is
equivalent to converting the random variable of interest to a standard normal random variable. We use the letter Z to represent a standard normal random variable.
Find a probability under the normal curve.
Follow these steps to find the probability under the normal curve that z is:
1. greater than 1.31.
2. les than 1.31
3. On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1.31, a , , and the upperbound 1 2nd EE 99, as shown
in Figures 8.19 and 8.20. The key EE is located above the , on the sixth
row, second column. Press ENTER to execute the command. The results are
shown in Figure 8.20.

Figure 8.19
The P (z > 1.31) = 0.0951.

Figure 8.20

4. Find the probability under the normal curve that z is less than 1.31.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1 2nd EE 99, a , , and the upperbound 1.31 as shown
in Figures 8.21 and 8.22. Be sure to use the grey negation key found on the
bottom row, column four. The key EE is located above the , on the sixth
row, second column. Press ENTER to execute the command. The results are
shown in Figure 8.22.

Figure 8.21
The P (z < 1.31) = 0.9049.

Figure 8.22

90

8.3

How to Solve General Normal Curve Problems

How to Solve General Normal Curve Problems

The TI-83 SE and TI-84 SE calculators can be used to find probabilities for any
general normal random variable. An important fact about nonnal random variables
is that any probability problem about a normal random variable can be converted
to a problem about a standard normal variable.
Example 8.24 - Probability That Height Is Less Than 62 Inches Assume that the
heights of college women follow a normal curve with = 6 5 inches and  =
2 .7inches, we can find probabilities associated with any possible range of
heights.For example, what is the probability that a randomly selected college
woman is 62inches or shorter? Equivalently, what proportion of college women
are 62 inchesor shorter?
Follow these steps to find the probability that a randomly selected college woman
is 62 inches or shorter.
1. Method 1: Transform the observation to a z  score.

= P (Z  1.11)
P (x  62) = P Z  6265
2.7
Find the probability under the normal curve that z is less than 1.11.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1 2nd EE 99, a , , the upperbound 1.11, and ) as
shown in Figures 8.21 and 8.22. Be sure to use the grey negation key found on
the bottom row, column four. The key EE is located above the , on the sixth
row, second column. Press ENTER to execute the command. The results are
shown in Figure 8.22.

Figure 8.23
Figure 8.24

6265
The P (x  62) = P Z  2.7
= P (Z  1.11) = 0.1335. In other
words, about 13% of college women are 62 inches or shorter.
2. Method 2: Enter the lowerbound, upperbound, , and  in terms of the x variable.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1 2nd EE 99, a , , the upperbound 62, the value for ,
65, the value for , 2.7, and ) as shown in Figures 8.25 and 8.26. Be sure

91

Chapter 8

Random Variables

to use the grey negation key found on the bottom row, column four. The key
EE is located above the , on the sixth row, second column. Press ENTER
to execute the command. The results are shown in Figure 8.26.

Figure 8.25
Figure 8.26
The P (x  62) = 0.1333. In other words, about 13% of college women are
62 inches or shorter. Observe that Method 2 is good only for the z = x

formula and is not valid for any other z formula.
Example 8.2 - Proportion of Women Who Are Taller Than 68 Inches If we
assume that college womens heights follow a normal curve with = 6 5
inchesand  = 2 .7 inches, we can find probabilities associated with any possible
range ofheights. Suppose we want to find the proportion of college women who
are tallerthan 68 inches.
Follow these steps to find the proportion of college women who are taller than 68
inches.
1. Method 1: Transform the observation to a z  score.

= P (Z > 1.11)
P (x > 68) = P Z > 6865
2.7
Find the probability under the normal curve that z is more than 1.11.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1.11 a , , the upperbound of 1 2nd EE 99, and ) as shown
in Figures 8.27 and 8.28. The key EE is located above the , on the sixth
row, second column. Press ENTER to execute the command. The results are
shown in Figure 8.28.

Figure 8.27

The P (x > 68) = P Z >

6865
2.7

Figure 8.28
= P (Z > 1.11) = 0.1335. In other

words, about 13% of college women are 68 inches or taller.

92

8.3

How to Solve General Normal Curve Problems

2. Method 2: Enter the lowerbound, upperbound, , and  in terms of the x variable.

On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 68, a , , the upperbound 1 2nd EE 99, the value for , 65,
the value for , 2.7, and ) as shown in Figures 8.29 and 8.30. The key EE
is located above the , on the sixth row, second column. Press ENTER to
execute the command. The results are shown in Figure 8.30.

Figure 8.29
Figure 8.30
The P (x > 68) = 0.1333. In other words, about 13% of college women are 68
inches or taller. Observe that Method 2 is good only for the z = x
 formula
and is not valid for any other z formula.
Example 8.24 - Continued Proportion of Women Between 62 and 68 Inches
TallIf we assume that college womens heights follow a normal curve with = 6 5
inches and  = 2 .7 inches, we can find probabilities associated with any possible
range of heights. Suppose we want to find the proportion of college women who
between 62 and 68 inches tall.
Follow these steps to find the proportion of college women who are taller than 68
inches.
1. Method 1: Transform the observations
to z  scores.

6265
P (x  62) = P Z

2.7
= P (Z  1.11)

= P (Z  1.11)
P (x  68) = P Z  6865
2.7
Find the probability under the normal curve that z is between z = 1.11 and
z = 1.11.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 1.11 a , , the upperbound of 1.11, and ) as shown in Figures 8.31 and 8.32. The key EE is located above the , on the sixth row, second column. Press ENTER to execute the command. The results are shown

93

Chapter 8

Random Variables

in Figure 8.32.

Figure 8.31
Figure 8.32

6865
The P (62  x  68) = P 6265
= P (1.11  Z  1.11) =

Z

2.7
2.7
0.7330. In other words, about 73% of college women are between 62 and 68
inches tall.
2. Method 2: Enter the lowerbound, upperbound, , and  in terms of the x variable.
On the homescreen, press 2nd DISTR , selecting 2: normalcdf( , type the
lowerbound of 62, a , , the upperbound 68, the value for , 65, the value
for , 2.7, and ) as shown in Figures 8.33 and 8.34. The key EE is located
above the , on the sixth row, second column. Press ENTER to execute the
command. The results are shown in Figure 8.34.

Figure 8.33
Figure 8.34
The P (62  x  68) = 0.7335. In other words, about 73% of college women
are between 62 and 68 inches tall. Observe that Method 2 is good only for the
z = x
 formula and is not valid for any other z formula.

8.4

Finding Percentiles
In some problems, we want to know what value of a variable has a given percentile
ranking. For example, we may want to know what pulse rate is the 25th percentile
of pulse rates for men. Notice that the word percentile refers to the value of a variable. The percentile rank corresponds to the cumulative probability (area to the
left under the density curve) for that value.
Suppose that the 25th percentile of pulse rates for adult males is 64 beats per
minute. This means that 25% of men have a pulse rate below 64. The percentile is
64 beats per minute (a value of the variable) and the percentile rank is 25% or .25
(a cumulative probability).
Example 8.26 - The 75th Percentile of Systolic Blood Pressures Suppose that
the blood pressures of men aged 18 to 29 years old can be described with a normal curve having mean = 1 2 0 and standard deviation  = 1 0 . What is the
75th
94

8.4

Finding Percentiles

percentile? In other words, what is the blood pressure value x such that P (Blood
pressure  x) = 0.75?
Follow these steps to find the 75th percentile of systolic blood pressures..
1. Method 1: Find the value of z  for which P (Z  z  ) = p.
In order to find the value of z  for which P (Z  z  ) = p, we use the invNorm
function requiring the area to the left of z  . Therefore for the 75th percentile,
the area to the left of the 75th percentile is 0.75, as shown in Figure 8.35.. Take
the following steps:
On the homescreen, press 2nd DISTR , located on the fourth row, column
four, above VARS . Select 3: invNorm( .Type 0.75 ) ENTER , as shown in
Figure 8.36. The results are shown in Figure 8.37, indicating the appropriate
z  for the 75th percentile is 0.67, rounded to 2 decimal places.

Figure 8.35
Figure 8.36

Figure 8.37

On the homescreen, compute x = z   + . Type 0.67 10 + 120, pressing

ENTER to execute the command. The results are shown in Figure 8.38.

Figure 8.38
The 75th percentile is 126.7 or about 127. P (Blood pressure  126.7) =
0.75. In other words, about 75% of the blood pressures of men aged 18 to 29
years old are below 127.
2. Method 2: Enter the area, , and  in terms of the x variable.
On the homescreen, press 2nd DISTR , selecting 3: invNorm( .Type the area,
0.75, the value for , 120, the value for , 10, and ) as shown in Figures 8.38
and 8.39. Press ENTER to execute the command. The results are shown in

95

Chapter 8

Random Variables

Figure 8.39.

Figure 8.39
The 75th percentile is 126.7 or about 127. P (Blood pressure  126.7) =
0.75. In other words, about 75% of the blood pressures of men aged 18 to 29
years old are below 127.

96

Chapter 9
Understanding Sampling
Distributions: Statistics as
Random Variables
This chapter introduces the reasoning that allows researchers to make conclusions
about entire populations using relatively small samples of individuals. The secret
to understanding how things work is to understand what kind of dissimilarity we
should expect to see among different samples from the same population.
This chapter serves as an introduction to the reasoning that allows researchers to
make conclusions about entire populations on the basis of a relatively small sample
of individuals. The basic idea is that we must work backwards, from a sample to a
population. We start with a question about a population like: How many teenagers
are infected with HIV? At what average age do left-handed people die? What
is the average income of all students at a large university? We collect a sample
from the population about which we have the question and measure the variable of
interest. We can then answer the question of interest for the sample. Finally, based
on statistical theory, we will be able to determine how close our sample answer is
to what we really want to know, the true answer for the population.
After reading this chapter you should be able to:
1. Simulate the sampling distribution for a sample proportion.
2. Simulate the sampling distribution for a sample mean.
3. Determine areas and probabilities for a Students t-distribution.
Keystrokes Introduced
1. 2nd LIST I OPS>5: seq(expression, variable, begin, end[,increment] )
returns a list.
2. MATH I I -PRB>7: randBin(numtrials,prob,[,numsimulations] ) generates and displays a random real number from a specified Binomial distribution.
3. MATH I I -PRB>6: randNorm(,,numtrials ) generates and displays a
random real number from a Normal distribution specified by and  for a
specified number of trials.
4. 2nd DISTR 5: tcdf(lowerbound,uppebound,df ) computes the Students tdistribution probability between lowerbound and upperbound for.the specified
df (degrees of freedom).

97

Chapter 9

9.1

Understanding Sampling Distributions: Statistics as Random Variables

Sampling Distribution for One Sample Proportion

In this Section we cover sampling distributions for one sample proportion. However, the module includes substantial discussion and explanation that should help
you understand sampling distributions in general.
Suppose we conduct a binomial experiment with n trials and get successes on x of
the trials. Or, suppose we measure a categorical variable for a representative sample of 11individuals, and x of them have responses in a certain category. In each
case, we can compute the statistic pb= the sample proportion = nx , the proportion
of trials resulting in success, or the proportion in the sample with responses in the
specified category. If we repeated the binomial experiment or collected a new sample, we would probably get a different value for the sample proportion.
A result given in Section 8.7 of the text is that with sufficiently large n, a binomial random variable is also approximately a normal random variable. A binomial
random variableX counts the number of times an event happens in n trials, but
the approximate normality also applies to the proportion, pb = nx . Dividing each
possible value of X by the sample size n does not change the shape of the n distribution of possible values. In other words, the sampling distributionfor a sample
proportion is approximately a normal distribution.
Example 9.4 - Possible Sample Proportions Favoring a Candidate
This sample size in this example has been changed from 2400 in the text to a sample size of 24 in order to make this practical for a TI calculator.
Suppose that of all voters in the United States, 40% are in favor of Candidate X
for president. Pollsters take a sample of 24 voters. What proportion of the sample
would be expected to favor Candidate X? The rule tells that that the proportion of
the sample who favor Candidate X is a random variable that has a normal distribution. The mean and standard deviation for the distribution are:
Mean = p q
= 0.40 (40%)
q
s.d.(b
p) =

p(1p)
n

0.4(10.4)
24

= 0.1

Follow these steps to simulate the sampling distribution for this sample proportion.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,

98

9.1

Sampling Distribution for One Sample Proportion

L5, L6. Press ENTER to execute the command.

2. Place a sequence of random real numbers from a binomial distribution into list
L1.
On the homescreen press 2nd LIST I to select the LIST OPS menu.
a. Select 5: seq( .
Press MATH I I I to the PRB menu. Select 7: randBin(. Type
24 , 0.4 ) x,T,, n , 1 , 100 , 1 ) STO 2nd L1 . The x,T,, n key
is located on the 3rd row, end column. Press ENTER to execute the command, as shown in Figure 9.1.

Figure 9.1

Figure 9.2

3. Obtain the numerical summaries of the number of voters in favor of Candidate

X from the sample of 24 voters.
Press STAT I to obtain the STAT CALC menu.
a. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the
number of voters in favor of Candidate X. The output from the TI calculator is displayed in Figure 9.2.
Observe that in this random sample the mean of the number of voters in favor of Candidate X is 9.62 and the standard deviation is 2.39. The random
sample that you produce may have a different mean and standard deviation.
4. Set up the plot for the histogram of the right handspan measurements for the
females.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the third icon in the first row, the histogram. Press ENTER . Use
the down arrow key to select L1 as the list, 2nd L1 . Use the down
arrow key to enter 1 as the Freq:. The settings for Plot 1 are shown in
Figure 9.3.

99

Chapter 9

Understanding Sampling Distributions: Statistics as Random Variables

5. Enter the function to superimpose the normal curve on the histogram.

Press Y= , row 1, column 1, to enter the function, as shown in Figure 9.4. Press
p
(30/2.39 (2))e((1/2)(x  9.62)2 /2.392 ). Observe that in this random
sample the mean of the number of voters in favor of Candidate X is 9.62 with
a standard deviation is 2.39, are entered into the function to determine the yvalues of the graph. The 30 is a scaling factor designed to make the plot of
the histogram and the normal curve coincide. Other scaling factors can be explored. You may choose to replace the mean of 9.62 and the standard deviation
of 2.39 with the mean and standard deviation from your random sample. The
left and right parentheses are located on row 6. Press 2nd ,  is located on
the 5th row, right column above the ^ key. Press 2nd e, e is located on the
8th row, left column above the LN key. Be sure to use the grey negation key
when you enter (1/2). The function is shown in Figure 9.4.
6. Set the Window viewing variables in order to view the graph.
Press WINDOW , row 1, column 2. Set Xmin to 11, Xmax to 27; Xscl to 1;
Ymin to -5, being sure to use the grey negation key. Set Ymax to 31; Yscl to
1; Xres to 1. These settings are illustrated in Figure 9.5
7. View the graph.
Press GRAPH , to view the graph, as shown in Figure 9.6.

Figure 9.3

Figure 9.4

Figure 9.5
Figure 9.6
8. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
9. Clear the function.
Press Y= , and Press Y= and press CLEAR to remove all functions For

100

9.2

Sampling Distribution for One Sample Mean

each line that is not blank, place the cursor on the function and press CLEAR
Press 2nd QUIT .
Figure 9.6 indirectly shows how a sampling distribution provides information
about the accuracy of a sample statistic. In that example, we learned that with
a sample size of n =24 voters it is nearly certain that the proportion of voters
favoring Candidate X in the sample will be within 3(0.1) = 0.3 of the true
population proportion.

9.2

Sampling Distribution for One Sample Mean

In this Section we cover sampling distributions for one sample mean. However, the
module for one mean includes substantial discussion and explanation that should
help you understand sampling distributions in general.
Suppose a population consists of thousands or millions of individuals, and we are
interested in estimating the mean of a quantitative variable. It we sample 25 peopleandcompute the mean of the variable for that sample,how close will that sample
mean be to the population mean we are trying to estimate? Each time we take a
sample we will get a different sample mean. Can we say anything about what we
expect those means to be?
For example, suppose we are interested in estimating the average weight loss for
everyone who attends a national weight-loss clinic for ten weeks. Suppose, unknown to us, the distribution of weight losses for everyone in this population is approximately normal with a mean of 8 pounds and a standard deviation of 5 pounds.
Conditions for the Sampling Distribution of the Mean to Be Approximately
Normal
As with sample proportions, statisticians understand what to expect for the possible distribution of sample means in repeated sampling from the same population.
Technically called the sampling distribution of the sample mean, we call this rule
the Normal Curve Approximation Rule for Sample Means, or simply the Rule for
Sample Means to convey what it says. Unlike the equivalent rule for proportions,
it is not always necessary to have a large sample for this rule to work.If the populationof measurementsis bell-shaped,then the result holds for all sample sizes.The
Rule for Sample Means applies in both of the following types of situations:
Situation 1 The population of the measurements of interest is bell-shaped
and-a-random sample of any size is measured.
Situation 2 The population of measurements of interest is not bell-shaped,
but a large random sample is measured.

101

Chapter 9

Understanding Sampling Distributions: Statistics as Random Variables

definition The Normal Curve Approximation Rule for Sample Means can
be defined as follows:
Let = mean for the population of interest.
Let  =standard deviation for the population of interest.
Let x = mean for the sample = sample mean.
If numerous random samples of the same size n are taken, the distribution
of possible values of X is approximately normal, with
Mean=

Standarddeviation= s.d.(x) = 
n
This approximate normal distribution is called the sampling distribution of
x or the sampling distribution of the mean.
Technical Note: The n observations in each sample must all be independent,
which they will be if random samples are used.
Example 9.7 - Hypothetical Mean Weight Loss
For our hypothetical weight-loss example, the population mean and standard deviation were = 8 pounds and  = 5 pounds, respectively, and we were taking random
samples of size 25. The mean and standard deviation for the distribution are:
Mean = = 8 pounds
s.d.(b
x) = sn = s525 = 1.0
Follow these steps to simulate the sampling distribution for this mean.
1. Preparations:
a. Turn off allY= functions.
Press Y= and press CLEAR to remove all functions For each line that
is not blank, place the cursor on the function and press CLEAR Press
2nd QUIT .
b. Clear all lists in the Stat editor.
Press STAT , selecting 4: ClrList. Enter each list name: L1, L2, L3, L4,
L5, L6. Press ENTER to execute the command.
2. Place a sequence of random real numbers from a normal distribution with mean
= = 8 pounds and s.d.(b
x) = 1.0 into list L1.
On the homescreen press MATH I I I to the PRB menu.

102

9.2 Sampling Distribution for One Sample Mean

a. Select 6: randNorm( . Enter the mean = = 8, s.d.(b
x) = 1.0, and the
number of samples, n = 500: 8 , 1 , 500 ) STO 2nd L1 . Press
ENTER to execute the command, as shown in Figure 9.7.

Figure 9.7

Figure 9.8

3. Obtain the numerical summaries of the sampling distribution of these sample

means.
Press STAT I to obtain the STAT CALC menu.
a. Select 1: 1-Var Stats and press ENTER . Press 2nd L1 to select the
means stored in list L1. The output from the TI calculator is displayed in
Figure 9.8.
Observe that in this random sample the mean of the means is 8.06 and the
standard deviation is 0.969, reasonably close to the theoretical values. The
random sample that you produce may have a different mean and standard
deviation.
4. Set up the plot for the histogram of the sampling distribution of the sample
means.
Press 2nd STAT PLOT accessing the StatPlot menu.
(i) Press ENTER , selecting Plot 1. Place the cursor on ON and press
ENTER . Use the down arrow key and the right arrow key to select
the third icon in the first row, the histogram. Press ENTER . Use
the down arrow key to select L1 as the list, 2nd L1 . Use the down
arrow key to enter 1 as the Freq:. The settings for Plot 1 are shown in
Figure 9.9.
5. Enter the function to superimpose the normal curve on the histogram.
Press Y= , row 1, column 1, to enter the function, as shown in Figure 9.10.
p
Press (80/0.97 (2))e((1/2)(x  8.06)2 /0.972 ). Observe that in this example, the mean of the sample means is 8.06 with a standard deviation is 0.97,
and are entered into the function to determine the y-values of the graph. The
80 is a scaling factor designed to make the plot of the histogram and the normal curve coincide. Other scaling factors can be explored. You may choose to
replace the mean of 8.06 and the standard deviation of 0.97 with the mean and
standard deviation from your random sample. The left and right parentheses

103

Chapter 9

Understanding Sampling Distributions: Statistics as Random Variables

are located on row 6. Press 2nd ,  is located on the 5th row, right column
above the ^ key. Press 2nd e, e is located on the 8th row, left column above
the LN key. Be sure to use the grey negation key when you enter (1/2).
The function is shown in Figure 9.10.
6. Set the Window viewing variables in order to view the graph.
Press WINDOW , row 1, column 2. Set Xmin to 4.5, Xmax to 12.5; Xscl to
1; Ymin to -2, being sure to use the grey negation key. Set Ymax to 300; Yscl
to 1; Xres to 1. These settings are illustrated in Figure 9.11
7. View the graph.
Press GRAPH , to view the graph, as shown in Figure 9.12.

Figure 9.9

Figure 9.10

Figure 9.11
Figure 9.12
8. Turn off all plots and return the graph window to standard viewing.
Press 2nd STAT PLOT , selecting PlotsOff and press ENTER . Press ZOOM
and select 6: ZStandard to restore the default graph window settings.
9. Clear the function.
Press Y= , and Press Y= and press CLEAR to remove all functions For
each line that is not blank, place the cursor on the function and press CLEAR
Press 2nd QUIT .
Figure 9.12 indirectly shows how a sampling distribution provides information
about the accuracy of a sample statistic. In this example, we learned that with
a sample size of n =25 the weight losses are approximately normal. From the
Empirical Rule, we know the following facts about possible sample means in
this situtation, based on intervals extending 1, 2. and 3 standard deviations
from the mean of 8:
a. There is a 68%chance that the sample mean will be between 7and 9.
b. There is a 95%chance that the sample mean will be between 6 and 10.

104

9.3

Areas and Probabilities for Students t-Distribution

c. It is almost certain that the sample mean will be between 5 and 11.

9.3

Areas and Probabilities for Students t-Distribution

Because Students t-distribution differs for each possible df value, we cant summarize the probability areas in one table like we could for the standard normal
distribution. We would need a separate table for each possible df value. Instead,
tables for the t-distribution are tailored to specific uses.
Many calculators and computer software programs provide probabilities (areas)
for specified (t-values and t-values for specified areas. For example, the TI-83 and
TI-84 calculate the Students t-distribution probability between a lowerbound and
an upperbound for a specified degrees of freedom, df. In other words, it provides
P(t> k).
Example 9.7 Standardized Mean Weights In Section 9.3, we considered four
hypothetical samples of n = 25 people who were trying to lose weight at a clinic.
Weplayed the role of the all-knowing sage and assumed we knew that = 8
and = 5. If the value for is correct, then the standardized statistic
= x8
t = x
ss
ss
n

25

has a t-distribution with df = 25  1 = 24. If we were to generate thousands of

random samples of size 25 and draw a histogram of the resulting standardized tstatistics, they would adhere to this t-distribution.
In practice, we do not draw thousands of samples and we do not know . Suppose
we speculated that = 8 pounds and drew one random sample, the first one given
in Table 9.1 of the text, for which x = 8.32 pounds and s = 4.74 pounds. Are the
sample results consistent with the speculation that = 8 pounds? In other words.
is a sample mean of 8.32 pounds reasonable to expect if = 8 pounds?
The standardized statistic is
= x8
= 8.328
= 0.34
t = x
4.74
ss
ss
s
n

25

25

Follow these steps to find the probability of observing a test statistic of t = 0.34,
or greater.
1. Calculate the probability of t = 0.34 or greater.
Press 2nd DISTR , located on row 4, column 4, above VARS ,to obtain the
distribution function menu.
a. Use the down arrow key, H , selecting 5: tcdf( , the Students-t cumulative distribution probability function, as shown in Figure 9.13. Press
ENTER. Enter the lowerbound, upperbound, and degrees of freedom, df.
Type 0.34 , 100 , 24 ) , as shown in Figure 9.14. Press ENTER to execute
the command. The output from the TI calculator is displayed in Figure

105

Chapter 9

Understanding Sampling Distributions: Statistics as Random Variables

9.15.

Figure 9.13

Figure 9.14

Figure 9.15

This statistic,t = 0.34, tells us that the sample mean of 8.32 is only about
0.3684 of a standard error above 8, which is certainly consistent with a
population mean weight loss of 8 pounds.
Variations in Finding Areas for a Students t-distribution
Variation 1: Follow these steps to find the probability of observing a test statistic
of t = 0.34, or less, given that the degrees of freedom is 24.
1. Calculate the probability of t = 0.34 or less.
Press 2nd DISTR , located on row 4, column 4, above VARS ,to obtain the
distribution function menu.
a. Use the down arrow key, H , selecting 5: tcdf( , the Students-t cumulative distribution probability function, as shown in Figure 9.16. Press
ENTER. Enter the lowerbound, upperbound, and degrees of freedom, df.
Type 0.34 , 100 , 24 ) , as shown in Figure 9.14. Press ENTER to execute
the command. The output from the TI calculator is displayed in Figure
9.15.

Figure 9.16

Figure 9.17

Figure 9.18

This TI calculator output tells us that the probability of finding a value of

t = 0.34 or less, given that the degrees of freedom is 24, is about 0.6316
or 63.16% of the time.
Variation 2: Follow these steps to find the probability of observing a test statistic between t = 1.17 and t = +2.27, given that the degrees of freedom is
9.
1. Calculate the probability of observing a value of t between t = 1.17 and
t = +2.27.

106

9.3

Areas and Probabilities for Students t-Distribution

Press 2nd DISTR , located on row 4, column 4, above VARS ,to obtain the
distribution function menu.
a. Use the down arrow key, H , selecting 5: tcdf( , the Students-t cumulative
distribution probability function, as shown in Figure 9.19. Press ENTER.
Enter the lowerbound, upperbound, and degrees of freedom, df . Type 1.17 , 2.27 , 9 ) , as shown in Figure 9.20. Press ENTER to execute the
command. The output from the TI calculator is displayed in Figure 9.21.

Figure 9.19

Figure 9.20

Figure 9.21

This TI calculator output tells us that the probability of finding a value of

t between t = 1.17 and t = +2.27, given that the degrees of freedom is
9 is about 0.8393 or 83.93% of the time.

107

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189