11-1. 11-2 Chapter Eleven Simple Linear Regression Analysis McGraw-Hill/Irwin Copyright © 2004 by The McGraw-Hill Companies, Inc. All rights reserved.

11-11-11

11-11-22

Chapter Eleven

Simple Linear Regression Analysis

McGraw-Hill/Irwin Copyright © 2004 by The McGraw-Hill Companies, Inc. All rights reserved.

11-11-33

Simple Linear Regression

11.1 The Simple Linear Regression Model

11.2 The Least Squares Point Estimates

11.3 Model Assumptions, Mean Squared Error, Std. Error

11.4 Testing Significance of Slope and y-Intercept

11.5 Confidence Intervals and Prediction Intervals

11.6 The Coefficient of Determination and Correlation

11.7 An F Test for the Simple Linear Regression Model

*11.8 Checking Regression Assumptions by Residuals

*11.9 Some Shortcut Formulas

11-11-44

11.1 The Simple Linear Regression Model

εxββ=εμy= y|x 10

y|x = + 1x + is the mean value of the dependent variable y when the value of the independent variable is x.

is the y-intercept, the mean of y when x is 0.

1 is the slope, the change in the mean of y per unit change in x.

is an error term that describes the effect on y of all factors other than x.

AverageHourly Weekly FuelTemperature Consumption

Week x (deg F) y (MMcf)1 28.0 12.42 28.0 11.73 32.5 12.44 39.0 10.85 45.9 9.46 57.8 9.57 58.1 8.08 62.5 7.5

11-11-55

The Simple Linear Regression Model Illustrated

11-11-66

11.2 The Least Squares Point Estimates

n

xxxxSS

n

yxyxyyxxSS

SS

SSb

iiixx

iiiiiixy

xx

xy

2

22

1

)(

)()(

xbby 10 ˆ

n

xx

n

yyxbyb ii 10

Estimation/Prediction Equation:

Least squares point estimate of the slope 1

Least squares point estimate of the y-intercept 0

11-11-77

Example: The Least Squares Point Estimates

0.1279

355.1404

6475.179

355.14048

)8.351(76.16874

6475.1798

)7.81)(8.351(11.3413

1

22

2

xx

xy

iixx

iiiixy

SS

SSb

n

xxSS

n

yxyxSS

15.84

)98.43)(1279.0(2125.10

98.438

8.351

2125.108

7.81

10 xbyb

n

xx

n

yy

i

i

Slope b1 y-Intercept b0

y x x2 xy12.4 28.0 784.00 347.2011.7 28.0 784.00 327.6012.4 32.5 1056.25 403.0010.8 39.0 1521.00 421.209.4 45.9 2106.81 431.469.5 57.8 3340.84 549.108.0 58.1 3375.61 464.807.5 62.5 3906.25 468.75

81.7 351.8 16874.76 3413.11

Gas of MMcf10.720.1279(40)-15.84 xbby 10 ˆPrediction (x = 40)

11-11-88

11.3 The Regression Model Assumptions

Assumptions about the model error terms, ’s

Mean Zero The mean of the error terms is equal to 0.

Constant Variance The variance of the error terms is, the same for all values of x.

Normality The error terms follow a normal distribution for all values of x.

Independence The values of the error terms are statistically independent of each other.

εxββ=εμy= y|x 10Model

11-11-99

Regression Model Assumptions Illustrated

11-11-1010

Mean Square Error and Standard Error

Mean Square Error, point estimate of residual variance 2

2

n-

SSEMSEs

2n-

SSEMSEs Standard Error, point estimate of

residual standard deviation

y x pred y - pred (y - pred)2

12.4 28.0 12.2588 0.1412 0.01993711.7 28.0 12.2588 -0.5588 0.31225712.4 32.5 11.6833 0.7168 0.51373110.8 39.0 10.8519 -0.0519 0.0026949.4 45.9 9.9694 -0.5694 0.3242059.5 57.8 8.4474 1.0526 1.1080098.0 58.1 8.4090 -0.4090 0.1672897.5 62.5 7.8463 -0.3462 0.119889

SSE 2.568011

Example 11.6 The Fuel Consumption Case

0.428

6

568.22

2

n-

SSEMSEs

0.6542 428.02ss

22 )ˆ( iii yyeSSE Sum of Squared Errors

11-11-1111

11.4 Significance Test and Estimation for Slope

xx

bb SS

ss

s

bt=

1

1

where1

Test Statistic

If the regression assumptions hold, we can reject H0: 1 = 0 at the level of significance (probability of Type I error equal to ) if and only if the appropriate rejection point condition holds or, equivalently, if the corresponding p-value is less than .

0:

0:

0:

1

1

1

a

a

a

H

H

H

2/2/

2/

or

isthat,

tttt

tt

tt

tt

t, t/2 and p-values are based on n – 2 degrees of freedom.

Alternative Reject H0 if: p-Value

tofrightondistributit underarea Twice

tofleftondistributit underArea

tofrightondistributit underArea

100(1-)% Confidence Interval for 1

][12/1 bstb

11-11-1212

Significance Test and Estimation for y-Intercept

xxb

b SS

x

nss

s

bt=

20 1

where0

0

Test Statistic

If the regression assumptions hold, we can reject H0: 0 = 0 at the level of significance (probability of Type I error equal to ) if and only if the appropriate rejection point condition holds or, equivalently, if the corresponding p-value is less than .

0:

0:

0:

0

0

0

a

a

a

H

H

H

2/2/

2/

or

isthat,

tttt

tt

tt

tt

t, t/2 and p-values are based on n – 2 degrees of freedom.

Alternative Reject H0 if: p-Value

tofrightondistributit underarea Twice

tofleftondistributit underArea

tofrightondistributit underArea

100(1-)% Conf Interval for 0

][02/0 bstb

11-11-1313

Example: Inferences About Slope and y-Intercept

Example 11.7 The Fuel Consumption Case Excel Output

Regression StatisticsMultiple R 0.948413871R Square 0.899488871Adjusted R Square 0.882737016Standard Error 0.654208646Observations 8

ANOVAdf SS MS F Significance F

Regression 1 22.980816 22.980816 53.694882 0.000330052Residual 6 2.567934 0.427989Total 7 25.548750

Coefficients Standard Error t Stat P-valueIntercept 15.83785741 0.801773385 19.75353349 0.000001092Temp -0.127921715 0.01745733 -7.327679169 0.000330052

Coefficients Standard Error Lower 95% Upper 95%Intercept 15.83785741 0.801773385 13.87598718 17.79972765Temp -0.127921715 0.01745733 -0.170638294 -0.085205136

Tests

Intervals

11-11-1414

11.5 Confidence and Prediction Intervals

] valueDistancety[ /2s

t is based on n-2 degrees of freedom

] valueDistance+1ty[ /2s

Prediction (x = x0)

010ˆ xbby Distance Value

xxSS

xx

n

20 )(1

100(1 - )% confidence interval for the mean value of y, y|xo

If the regression assumptions hold,

100(1 - )% prediction interval for an individual value of y

11-11-1515

Example: Confidence and Prediction Intervals

Example 11.7 The Fuel Consumption Case

Minitab Output (predicted FuelCons when Temp, x = 40)

Predicted Values Fit StDev Fit 95.0% CI 95.0% PI 10.721 0.241 ( 10.130, 11.312) ( 9.014, 12.428)

11-11-1616

11.6 The Simple Coefficient of Determination

The simple coefficient of determination r2 is

variationTotal

n variatioExplainedr2

(SSE)SquaresofSumErrorˆ= variationdUnexplaine

(SSR) SquaresofSumRegressionˆ= variationExplained

(SSTO) SquaresofSumTotal = variationTotal

2

2

2

)y(y

)yy(

)y(y

ii

i

i

variation dUnexplaine variation Explained variation Total

r2 is the proportion of the total variation in y explained by the simple linear regression model

11-11-1717

The Simple Correlation Coefficient

The simple correlation coefficient measures the strength of the linear relationship between y and x and is denoted by r.

negative is if

and positive, is if

12

12

brr=

brr=

Where, b1 is the slope of the least squares line.

Regression StatisticsMultiple R 0.948413871R Square 0.899488871Adjusted R Square 0.882737016Standard Error 0.654208646Observations 8

ANOVAdf SS MS F Significance F

Regression 1 22.980816 22.980816 53.694882 0.000330052Residual 6 2.567934 0.427989Total 7 25.548750

Example 11.15Fuel Consumption Excel Output

948414.0899489.0

899489.0548750.25

980816.222

r

r

11-11-1818

Different Values of the Correlation Coefficient

11-11-1919

11.7 F Test for Simple Linear Regression Model

To test H0: = 0 versus Ha: 0 at the level of significance

Test Statistic:

Explained variation

(Unexplained variation)/(n-2)

F(model)

Reject H0 if F(model) > For p-value <

Fis based on 1 numerator and n-2 denominator degrees of freedom.

11-11-2020

Example: F Test for Simple Linear Regression

Test Statistic:

695.53)28/(567904.2

980816.22

2)-)/(n variationed(Unexplain

variationExplainedF(model)

Example 11.17 The Fuel Consumption Case Excel Output

ANOVAdf SS MS F Significance F

Regression 1 22.980816 22.980816 53.694882 0.000330052Residual 6 2.567934 0.427989Total 7 25.548750

F-test at = 0.05 level of significance

Reject H0 at level of significance, since

Fis based on 1 numerator and 6 denominator degrees of freedom.

05.000033.0value-p

and99.5695.53F(model) 05.F

11-11-2121

*11.8 Checking the Regression Assumptions by Residual Analysis

For an observed value of y, the residual is

where the predicted value of y is calculated as

y)predictedy(observedˆ yye

xbby 10ˆ

If the regression assumptions hold, the residuals should look like a random sample from a normal distribution with mean 0 and variance 2.

Residual Plots

Residuals versus independent variablesResiduals versus predicted y’sResiduals in time order (if the response is a time series)Histogram of residualsNormal plot of the residuals

11-11-2222

Checking the Constant Variance Assumption

Example 11.18: The QHIC CasePlot: Residual versus x and predicted responses

11-11-2323

Checking the Normality Assumption

Example 11.18: The QHIC CasePlots: Histogram and Normal Plot of Residuals

11-11-2424

Checking the Independence Assumption

Plots: Residuals versus Fits (to check for functional form, not shown) Residuals versus Time Order

11-11-2525

Combination Residual Plots

Example 11.18: The QHIC Case Minitab OutputPlots: Histogram and Normal Plot of Residuals, Residuals versus Order (I Chart), Residuals versus Fit.

-300 -200 -100 0 100 200 300

0123456789

Residual

Freq

uenc

y

Histogram of Residuals

0 10 20 30 40-500

0

500

Observation Number

Res

idua

l

I Chart of Residuals

2

2

X=0.000

3.0SL=396.3

-3.0SL=-396.3

0 20040060080010001200140016001800

-300

-200

-100

0

100

200

300

Fit

Res

idua

lResiduals vs. Fits

-2 -1 0 1 2

-300

-200

-100

0

100

200

300

Normal Plot of Residuals

Normal Score

Res

idua

l

Residual Model Diagnostics

11-11-2626

*11.9 Some Shortcut Formulas

xx

xyyy

xx

xy

yy

SS

SSSS=SSE

SS

SSSSR

SSSSTO

2

2

variationdUnexplaine

variationExplained

variationTotal

n

yyyySS

n

xxxxSS

n

yxyxyyxxSS

iiiyy

iiixx

iiiiiixy

2

22

2

22

)(

)(

)()(

where

11-11-2727

Simple Linear RegressionSummary: 11.1 The Simple Linear Regression Model 11.2 The Least Squares Point Estimates 11.3 Model Assumptions, Mean Squared Error, Std.

Error 11.4 Testing Significance of Slope and y-Intercept 11.5 Confidence Intervals and Prediction Intervals 11.6 The Coefficient of Determination and

Correlation 11.7 An F Test for the Simple Linear Regression

Model*11.8 Checking Regression Assumptions by

Residuals*11.9 Some Shortcut Formulas