Multiple Regression
-
Upload
fufu-zein-fuad -
Category
Documents
-
view
220 -
download
1
description
Transcript of Multiple Regression
![Page 1: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/1.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-1
Basic Business Statistics(8th Edition)
Chapter 14Introduction to
Multiple Regression
![Page 2: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/2.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-2
Chapter TopicsChapter Topics
The multiple regression modelThe multiple regression model Residual analysisResidual analysis Influence analysisInfluence analysis Testing for the significance of the Testing for the significance of the
regression modelregression model Inferences on the population Inferences on the population
regression coefficientsregression coefficients Testing portions of the multiple Testing portions of the multiple
regression modelregression model
![Page 3: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/3.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-3
0 1 1 2 2i i i k ki iY b b X b X b X e
Population Y-intercept
Population slopes Random Error
The Multiple Regression ModelThe Multiple Regression Model
A relationship between A relationship between one dependent one dependent and two or and two or more more independentindependent variables is a variables is a linear functionlinear function
Dependent (Response) variable for sample
Independent (Explanatory) variables for sample model
1 2i i i k ki iY X X X
Residual
![Page 4: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/4.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-4
Population Multiple Regression Model
X2
Y
X1YX = 0 + 1X 1i + 2X 2i
0
Y i = 0 + 1X 1i + 2X 2i + i
ResponsePlane
(X 1i,X 2i)
(O bserved Y )
i
X2
Y
X1YX = 0 + 1X 1i + 2X 2i
0
Y i = 0 + 1X 1i + 2X 2i + i
ResponsePlane
(X 1i,X 2i)
(O bserved Y )
i
Bivariate model
![Page 5: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/5.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-5
Sample Multiple Regression Model
X2
Y
X1
b0
Y i = b0 + b1X 1 i + b2X 2 i + e i
ResponsePlane
(X 1i, X 2i)
(O bserved Y)
^
e i
Y i = b0 + b1X 1 i + b2X 2 i
X2
Y
X1
b0
Y i = b0 + b1X 1 i + b2X 2 i + e i
ResponsePlane
(X 1i, X 2i)
(O bserved Y)
^
e i
Y i = b0 + b1X 1 i + b2X 2 i
Bivariate model
Sample Regression PlaneSample Regression Plane
![Page 6: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/6.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-6
Simple and Multiple Simple and Multiple Regression ComparedRegression Compared
Coefficients in a simple regression pick up the impact of that variable plus the impacts of other variables that are correlated with it and with the dependent variable.
Coefficients in a multiple regression net out the impacts of other variables in the equation.
![Page 7: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/7.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-7
Simple and Multiple Simple and Multiple Regression Compared: ExampleRegression Compared: Example
Two simple regressionsTwo simple regressions:
Multiple regression:
0 1
0 1
Oil Temp
Oil Insulation
0 1 2Oil Temp Insulation
![Page 8: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/8.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-8
Multiple Linear Multiple Linear Regression EquationRegression Equation
Too complicated
by hand! Ouch!
![Page 9: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/9.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-9
Interpretation of Interpretation of Estimated CoefficientsEstimated Coefficients
Slope (Slope (bbii)) Estimated that the average value of Estimated that the average value of YY changes changes
by by bbii for each one unit increase in for each one unit increase in XXii holding all holding all other variables constant (ceterus paribus)other variables constant (ceterus paribus)
Example: if Example: if bb11 = -2, then fuel oil usage ( = -2, then fuel oil usage (YY) is ) is expected to decrease by an estimated two expected to decrease by an estimated two gallons for each one degree increase in gallons for each one degree increase in temperature (temperature (XX11) given the inches of insulation ) given the inches of insulation ((XX22))
Y-intercept (Y-intercept (bb00)) The estimated average value of The estimated average value of YY when all when all XXii = 0 = 0
![Page 10: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/10.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-10
Multiple Regression ModelMultiple Regression Model: Example
Oil (Gal) Temp Insulation275.30 40 3363.80 27 3164.30 40 1040.80 73 694.30 64 6
230.90 34 6366.70 9 6300.60 8 10237.80 23 10121.40 63 331.40 65 10
203.50 41 6441.10 21 3323.00 38 352.50 58 10
(0F)
Develop a model for estimating Develop a model for estimating heating oil used heating oil used for a single family for a single family home in the month of January based home in the month of January based on on average temperature average temperature and and amount amount of insulation in inches.of insulation in inches.
![Page 11: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/11.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-11
1 2ˆ 562.151 5.437 20.012i i iY X X
Sample Multiple Regression Equation: Sample Multiple Regression Equation: ExampleExample
CoefficientsIntercept 562.1510092X Variable 1 -5.436580588X Variable 2 -20.01232067
Excel Output
For each degree increase in temperature, the estimated average amount of heating oil used is decreased by 5.437 gallons, holding insulation constant.
For each increase in one inch of insulation, the estimated average use of heating oil is decreased by 20.012 gallons, holding temperature constant.
0 1 1 2 2i i i k kiY b b X b X b X
![Page 12: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/12.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-12
Venn Diagrams and Venn Diagrams and Explanatory Power of RegressionExplanatory Power of Regression
Oil
Temp
Variations in oil explained by temp or variations in temp used in explaining variation in oil
Variations in oil explained by the error term
Variations in temp not used in explaining variation in Oil SSE
SSR
![Page 13: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/13.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-13
Venn Diagrams and Venn Diagrams and Explanatory Power of RegressionExplanatory Power of Regression
Oil
Temp
2
r
SSR
SSR SSE
(continued)
![Page 14: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/14.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-14
Venn Diagrams and Venn Diagrams and Explanatory Power of RegressionExplanatory Power of Regression
Oil
TempInsulation
Overlapping Overlapping variation in both Temp and Insulation are used in explaining the variationvariation in Oil but NOTNOT in the estimationestimation of nor
12
Variation NOTNOT explained by Temp nor Insulation SSE
![Page 15: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/15.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-15
Coefficient of Coefficient of MultipleMultiple Determination Determination
Proportion of total variation in Y explained by all X variables taken together
Never decreasesNever decreases when a new when a new XX variable variable is added to modelis added to model Disadvantage when comparing models
212
Explained Variation
Total VariationY k
SSRr
SST
![Page 16: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/16.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-16
Venn Diagrams and Venn Diagrams and Explanatory Power of RegressionExplanatory Power of Regression
Oil
TempInsulation
212
Yr
SSR
SSR SSE
![Page 17: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/17.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-17
AdjustedAdjusted Coefficient of Multiple Coefficient of Multiple DeterminationDetermination
Proportion of variation in Y explained by all X variables adjusted for the number of X variables used
Penalize excessive use of independent variables
Smaller than Useful in comparing among models
2 212
11 1
1adj Y k
nr r
n k
212Y kr
![Page 18: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/18.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-18
Coefficient of Multiple DeterminationCoefficient of Multiple Determination
Regression StatisticsMultiple R 0.982654757R Square 0.965610371Adjusted R Square 0.959878766Standard Error 26.01378323Observations 15
Excel Output
SST
SSRr ,Y 2
12
Adjusted r2
reflects the number of explanatory variables and sample size
is smaller than r2
![Page 19: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/19.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-19
Interpretation of Coefficient of Interpretation of Coefficient of Multiple DeterminationMultiple Determination
96.56% of the total variation in heating oil 96.56% of the total variation in heating oil can be explained by difference in can be explained by difference in temperature and amount of insulationtemperature and amount of insulation
95.99% of the total fluctuation in heating oil can 95.99% of the total fluctuation in heating oil can be explained by difference in temperature and be explained by difference in temperature and amount of insulation after adjusting for the amount of insulation after adjusting for the number of explanatory variables and sample sizenumber of explanatory variables and sample size
2,12 .9656Y
SSRr
SST
2adj .9599r
![Page 20: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/20.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-20
Using The Model to Make PredictionsUsing The Model to Make Predictions
Predict the amount of heating oil used for Predict the amount of heating oil used for a home if the average temperature is 30a home if the average temperature is 300 0
and the insulation is six inches.and the insulation is six inches.
The predicted heating oil used is 278.97 gallons
1 2
ˆ 562.151 5.437 20.012
562.151 5.437 30 20.012 6
278.969
i i iY X X
![Page 21: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/21.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-21
Residual PlotsResidual Plots
Residuals vs. May need to transform May need to transform YY variable variable
Residuals vs. May need to transform variableMay need to transform variable
Residuals vs. May need to transform variableMay need to transform variable
Residuals vs. time May have autocorrelationMay have autocorrelation
Y
1X
2X1X
2X
![Page 22: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/22.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-22
Residual Plots: ExampleResidual Plots: Example
Insulation Residual Plot
0 2 4 6 8 10 12
No discernable patternNo discernable pattern
Temperature Residual Plot
-60
-40
-20
0
20
40
60
0 20 40 60 80
Re
sid
ua
ls
May be some non-May be some non-linear relationshiplinear relationship
![Page 23: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/23.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-23
Influence AnalysisInfluence Analysis
To determine observations that have influential To determine observations that have influential effect on the fitted modeleffect on the fitted model
Potentially influential points become candidates Potentially influential points become candidates for removal from the modelfor removal from the model
Criteria used areCriteria used are The hat matrix elements The hat matrix elements hhii
The Studentized deleted residuals The Studentized deleted residuals ttii**
Cook’s distance statistic Cook’s distance statistic DDii
All three criteria are complementaryAll three criteria are complementary Only whenOnly when all three criteria all three criteria provide provide
consistent results should an observation be consistent results should an observation be removedremoved
![Page 24: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/24.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-24
The Hat Matrix Element The Hat Matrix Element hhii
If , Xi is an Influential Point Xi may be considered a candidate for
removal from the model
2
2
1
1 i
i n
ii
X Xh
n X X
2 1 /ih k n
![Page 25: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/25.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-25
The Hat Matrix Element The Hat Matrix Element hhii : :Heating Oil ExampleHeating Oil Example
Oil (Gal) Temp Insulation h i
275.30 40 3 0.1567363.80 27 3 0.1852164.30 40 10 0.1757
40.80 73 6 0.246794.30 64 6 0.1618
230.90 34 6 0.0741366.70 9 6 0.2306300.60 8 10 0.3521237.80 23 10 0.2268121.40 63 3 0.2446
31.40 65 10 0.2759203.50 41 6 0.0676441.10 21 3 0.2174323.00 38 3 0.1574
52.50 58 10 0.2268
No No hhii > > 0.40.4 No observation appears to No observation appears to be a candidate for removal be a candidate for removal from the modelfrom the modelNOTE: 2(2+1)/ 15= 6/15=NOTE: 2(2+1)/ 15= 6/15=0,40,4
15 2
2 1 / 0.4
n k
k n
![Page 26: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/26.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-26
The Studentized Deleted Residuals The Studentized Deleted Residuals ttii**
: difference between the observed difference between the observed and and predicted based on a model that includes predicted based on a model that includes all observations except observation all observations except observation ii
: standard error of the estimate for a model that includes all observations except observation i
An observation is considered influential if An observation is considered influential if is the critical value of a two-tail test at is the critical value of a two-tail test at
10% level of significance10% level of significance
* 1
i
i
ii
et
S h
iY
iY
iS
ie
*2i n kt t
2n kt
![Page 27: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/27.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-27
The Studentized Deleted Residuals The Studentized Deleted Residuals ttii
** :Example :Example
Oil (Gal) Temp Insulation t i*
275.30 40 3 -0.3772363.80 27 3 0.3474164.30 40 10 0.824340.80 73 6 -0.187194.30 64 6 0.0066
230.90 34 6 -1.0571366.70 9 6 -1.1776300.60 8 10 -0.8464237.80 23 10 0.0341121.40 63 3 -1.853631.40 65 10 1.0304
203.50 41 6 -0.6075441.10 21 3 2.9674323.00 38 3 1.168152.50 58 10 0.2432
2 11
15 2
1.7957n k
n k
t t
t10* and t13
* are influential points for potential removal from the model
*10t
*13t
![Page 28: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/28.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-28
Cook’s Distance Statistic Cook’s Distance Statistic DDii
is the Studentized residualis the Studentized residual
If , an observation is considered is considered influentialinfluential
is the critical value of the is the critical value of the FF distribution at a 50% level of significance distribution at a 50% level of significance
2
2 1i i
ii
SR hD
h
1i
i
YX i
eSR
S h
1, 1i k n kD F
1, 1k n kF
![Page 29: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/29.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-29
Cook’s Distance Statistic Cook’s Distance Statistic DDii : Heating Oil : Heating Oil ExampleExample
Oil (Gal) Temp Insulation D i
275.30 40 3 0.0094363.80 27 3 0.0098164.30 40 10 0.049640.80 73 6 0.004194.30 64 6 0.0001
230.90 34 6 0.0295366.70 9 6 0.1342300.60 8 10 0.1328237.80 23 10 0.0001121.40 63 3 0.308331.40 65 10 0.1342
203.50 41 6 0.0094441.10 21 3 0.4941323.00 38 3 0.082452.50 58 10 0.0062
No Di >> 0.835 No observation No observation appears to be candidate appears to be candidate for removal from the for removal from the modelmodelUsing the three criteria, there Using the three criteria, there is insufficient evidence for the is insufficient evidence for the removal of any observation removal of any observation from the modelfrom the model
1, 1 3,12
15 2
0.835k n k
n k
F F
![Page 30: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/30.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-30
Testing for Overall SignificanceTesting for Overall Significance
Show if there is a linear relationship between Show if there is a linear relationship between all of the all of the XX variables together and variables together and YY
Use F test statisticUse F test statistic Hypotheses:Hypotheses:
HH00: : ……kk = = 0 (no linear relationship)0 (no linear relationship) HH11: at least one : at least one ii ( at least one independent ( at least one independent
variable affects variable affects Y Y )) The null hypothesis is a very strong statementThe null hypothesis is a very strong statement Almost always reject the null hypothesisAlmost always reject the null hypothesis
![Page 31: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/31.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-31
Testing for Overall SignificanceTesting for Overall Significance
Test statistic:
where F has p numerator and (n-p-1) denominator degrees of freedom
(continued)
all /
all
SSR pMSRF
MSE MSE
![Page 32: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/32.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-32
Test for Overall SignificanceTest for Overall SignificanceExcel Output: ExampleExcel Output: Example
ANOVAdf SS MS F Significance F
Regression 2 228014.6 114007.3 168.4712 1.65411E-09Residual 12 8120.603 676.7169Total 14 236135.2
p = 2, the number of explanatory variables n - 1
p value
Test StatisticMSR
FMSE
![Page 33: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/33.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-33
Test for Test for OverallOverall Significance SignificanceExample SolutionExample Solution
F0 3.89
H0: 1 = 2 = … = p = 0
H1: At least one i 0 = .05df = 2 and 12
Critical Value(s):
Test statistic:
Decision:
Conclusion:
Reject at = 0.05
There is evidence that There is evidence that at least one independent at least one independent variable affects Yvariable affects Y
= 0.05
F 168.47(Excel Output)
![Page 34: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/34.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-34
Test for Significance: Test for Significance: IndividualIndividual Variables Variables
Show whether there is a linear Show whether there is a linear relationship between the variable relationship between the variable XXii and and YY
Use tt Test Statistic Hypotheses:Hypotheses:
H0: i 0 (No linear relationship) H1: i 0 (Linear relationship between Xi and
Y)
![Page 35: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/35.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-35
““T” Test StatisticT” Test StatisticExcel Output: ExampleExcel Output: Example
Coefficients Standard Error t StatIntercept 562.1510092 21.09310433 26.65093769X Variable 1 -5.436580588 0.336216167 -16.16989642X Variable 2 -20.01232067 2.342505227 -8.543127434
t Test Statistic for X1 (Temperature)
t Test Statistic for X2 (Insulation)
i
i
b
btS
![Page 36: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/36.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-36
t t Test : Example Solution Test : Example Solution
HH00: : 1 1 = 0 = 0
HH11: : 11 0 0
df = 12
Critical Value(s):
Test Statistic:
Decision:
Conclusion:
Reject Reject HH00 at at = 0.05 = 0.05
There is evidence of a There is evidence of a significant effect of significant effect of temperature on oil temperature on oil consumption.consumption.t0 2.1788-2.1788
.025
Reject H0 Reject H0
.025
Does temperature have a significant effect on monthly Does temperature have a significant effect on monthly consumption of heating oil? Test at consumption of heating oil? Test at = 0.05 = 0.05.
tt Test Statistic = -16.1699 Test Statistic = -16.1699
![Page 37: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/37.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-37
Venn Diagrams and Venn Diagrams and Estimation of Regression ModelEstimation of Regression Model
Oil
TempInsulation
Only this information Only this information is used in the is used in the estimation of estimation of
2
Only this Only this information is information is used in the used in the estimation ofestimation of
1This This information is information is NOT used in NOT used in the the estimation of estimation of nor nor1 2
![Page 38: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/38.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-38
Confidence Interval Estimate Confidence Interval Estimate for the Slopefor the Slope
Provide the 95% confidence interval for the population slope 1 (the effect of temperature on oil consumption).
11 1n p bb t S
Coefficients Lower 95% Upper 95%Intercept 562.151009 516.1930837 608.108935X Variable 1 -5.4365806 -6.169132673 -4.7040285X Variable 2 -20.012321 -25.11620102 -14.90844
-6.169 1 -4.704
The estimated average consumption of oil is reduced by between 4.7 gallons to 6.17 gallons per each increase of 10 F.
![Page 39: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/39.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-39
Contribution of a Single Independent Contribution of a Single Independent VariableVariable
Let Xk be the independent variable of be the independent variable of interestinterest
Measures the contribution of Measures the contribution of XXkk in in explaining the total variation in explaining the total variation in YY ( (SSTSST))
kX
| all others except
all all others except
k k
k
SSR X X
SSR SSR X
![Page 40: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/40.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-40
Contribution of a Single Independent Variable kX
1 2 3
1 2 3 2 3
| and
, and and
SSR X X X
SSR X X X SSR X X
Measures the contribution of in explaining SST
1X
From ANOVA section of regression for
From ANOVA section of regression for
0 1 1 2 2 3 3i i i iY b b X b X b X 0 2 2 3 3i i iY b b X b X
![Page 41: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/41.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-41
Coefficient of Partial Determination ofCoefficient of Partial Determination of
Measures the proportion of variation in Measures the proportion of variation in the dependent variable that is explained the dependent variable that is explained by by XXkk while controlling for (holding while controlling for (holding constant) the other independent constant) the other independent variables variables
2 all others
| all others
all | all others
Yk
k
k
r
SSR X
SST SSR SSR X
kX
![Page 42: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/42.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-42
Coefficient of Partial Determination forCoefficient of Partial Determination forkX
(continued)
1 221 2
1 2 1 2
|
, |Y
SSR X Xr
SST SSR X X SSR X X
Example: Two Independent Variable Model
![Page 43: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/43.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-43
Venn Diagrams and Coefficient of Partial Venn Diagrams and Coefficient of Partial Determination forDetermination for
kX
Oil
TempInsulation
1 2|SSR X X
21 2
1 2
1 2 1 2
|
, |
Yr
SSR X X
SST SSR X X SSR X X
=
![Page 44: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/44.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-44
Contribution of a Subset of Contribution of a Subset of Independent VariablesIndependent Variables
Let Let XXss be the subset of independent be the subset of independent variables of interestvariables of interest
Measures the contribution of the subset Measures the contribution of the subset xxss in in explaining explaining SSTSST
| all others except
all all others except
s s
s
SSR X X
SSR SSR X
![Page 45: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/45.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-45
Contribution of a Subset of Independent Contribution of a Subset of Independent Variables: ExampleVariables: Example
Let Let XXss be be XX11 and and XX33
1 3 2
1 2 3 2
and |
, and
SSR X X X
SSR X X X SSR X
From ANOVA section of regression for
From ANOVA section of regression for
0 1 1 2 2 3 3i i i iY b b X b X b X 0 2 2i iY b b X
![Page 46: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/46.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-46
Testing Portions of ModelTesting Portions of Model
Examines the contribution of a Examines the contribution of a subset subset XXss of explanatory variables of explanatory variables to the relationship with to the relationship with YY
Null hypothesisNull hypothesis: Variables in the subset do not
significantly improve the model when all other variables are included
Alternative hypothesisAlternative hypothesis: At least one variable is significant
![Page 47: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/47.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-47
Testing Portions of ModelTesting Portions of Model
Always one-tailed rejection regionone-tailed rejection region Requires comparison of two regressions
One regression includes everything Another regression includes everything
except the portion to be tested
(continued)
![Page 48: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/48.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-48
Partial Partial FF Test For Contribution of Test For Contribution of Subset of Subset of XX variables variables
HypothesesHypotheses: H0 : Variables Xs do not significantly improve
the model given all others variables included H1 : Variables Xs significantly improve the
model given all others included Test StatisticTest Statistic:
with df = m and (n-p-1) m = # of variables in the subset Xs
| all others /
allsSSR X m
FMSE
![Page 49: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/49.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-49
Partial Partial FF Test For Contribution of A Single Test For Contribution of A Single
HypothesesHypotheses: H0 : Variable Xj does not significantly improve
the model given all others included H1 : Variable Xj significantly improves the
model given all others included Test StatisticTest Statistic:
With df = 1 and (n-p-1) m = 1 here
jX
| all others
alljSSR X
FMSE
![Page 50: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/50.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-50
Testing Portions of Model: ExampleTesting Portions of Model: Example
Test at the Test at the = .05 level to = .05 level to determine whether the determine whether the variable of average variable of average temperature significantly temperature significantly improves the model given improves the model given that insulation is includedthat insulation is included.
![Page 51: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/51.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-51
Testing Portions of Model: ExampleTesting Portions of Model: Example
H0: X1 (temperature) does not improve model with X2 (insulation) included
H1: X1 does improve model
= .05, df = 1 and 12
Critical Value = 4.75
ANOVASS
Regression 51076.47Residual 185058.8Total 236135.2
ANOVASS MS
Regression 228014.6263 114007.313Residual 8120.603016 676.716918Total 236135.2293
(For X1 and X2) (For X2)
Conclusion: Reject H0; X1 does improve model
1 2
1 2
| 228,015 51,076261.47
, 676.717
SSR X XF
MSE X X
![Page 52: Multiple Regression](https://reader035.fdocuments.us/reader035/viewer/2022081512/5695d1051a28ab9b0294d570/html5/thumbnails/52.jpg)
© 2002 Prentice-Hall, Inc. Chap 14-52
When to Use the When to Use the F testF test
The The FF test for the inclusion of a single test for the inclusion of a single variable after all other variables are variable after all other variables are included in the model is IDENTICAL to included in the model is IDENTICAL to the the tt test of the slope for that variable test of the slope for that variable
The only reason to do an The only reason to do an F F test is to test test is to test several variables togetherseveral variables together