0% found this document useful (0 votes)
6 views7 pages

Multiple Regression Analysis Guide

The document outlines a series of exercises related to multiple regression analysis, including graphical inspections, correlation assessments, regression equations, hypothesis testing, and evaluation of variance inflation factors (VIF). It provides detailed instructions for analyzing relationships between variables, interpreting regression coefficients, and utilizing ANOVA tables to determine model significance. Additionally, it emphasizes the importance of understanding predictive modeling and the implications of confidence intervals and prediction intervals.

Uploaded by

ashleylilyson
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
6 views7 pages

Multiple Regression Analysis Guide

The document outlines a series of exercises related to multiple regression analysis, including graphical inspections, correlation assessments, regression equations, hypothesis testing, and evaluation of variance inflation factors (VIF). It provides detailed instructions for analyzing relationships between variables, interpreting regression coefficients, and utilizing ANOVA tables to determine model significance. Additionally, it emphasizes the importance of understanding predictive modeling and the implications of confidence intervals and prediction intervals.

Uploaded by

ashleylilyson
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Examine the Multiple Regression Model Below and answer each of the questions.

This
exercise will prepare you for the questions on the final exam.

1) Graphical inspection and analysis of y, x relationship using scatterplot.

2) Utilize correlation matrix output below to determine if significant linear


relationships exist between y and x.
Method
Correlation Pearson
type
Number of 15
rows used

ρ: pairwise Pearson correlation

Correlations
HeatingOilGallons Temp(F)
Temp(F) -0.870
InsAttic -0.465 0.009

Pairwise Pearson Correlations


Sampl P-
e1 Sample 2 N Correlation 95% CI for ρ Valu
e
Temp(F) HeatingOilGallons 15 -0.870 (-0.956, -0.645) 0.000
InsAttic HeatingOilGallons 15 -0.465 (-0.789, 0.062) 0.081
InsAttic Temp(F) 15 0.009 (-0.506, 0.519) 0.975

3) Use Correlation output to answer the following questions

 What is the strength and nature of the correlation for temp and heating oil? __________

 What is the strength and nature of the correlation for Insulation and Heating Oil?
___________________

 What is the correlation for Insulation and Temp? ________

 Using the correlation matrix, assess whether multicollinearity is present. Explain


your process.______________________________

1|P age
Use ordinary least squares regression to develop predictive model. Identify the
regression coefficients and regression equation.
Regression Equation
HeatingOilGallon = 562.2 - 5.437 Temp(F) - 20.01
s InsAttic

Answer the following questions relating to regression equation above

 -What is the value of the constant bo? Interpret the constant

 -What is the value of the btemp? Interpret the value of the btemp

 -What is the value of bInsAttic? Interpret the value of bInsAttic

4) Using the coefficients table shown, test hypothesis test of regression


coefficients to determine if x variables are significant predictors of y. (Be able to
use a six-step process. Use both for T and pvalue techniques.)
Coefficients
T- P-
Term Coef SE Coef 95% CI Value Value
VIF
Constant 562.2 21.1 (516.2, 608.1) 26.65 0.000
Temp(F) -5.437 0.336 (-6.169, -4.704) -16.17 0.000 1.00
InsAttic -20.01 2.34 (-25.12, -14.91) -8.54 0.000 1.00

5) Using coefficients table above be able to compute confidence intervals (bi +/-
t*sb) for each coefficient (not constant).

6) Use the confidence intervals to test hypothesis and to determine average rate
of change in y for unitary change in x variable (net regression testing).

7) Evaluation of Variance Inflation Factors (VIF) values.


 Using coefficients table above, identify the VIF values for Temp and InsAttic

 What range can VIF values take?

 What value does VIF indicate severe multicollinearity?

 What is the corrective action for coefficients demonstrating VIF values in ≥ 10?

2|P age
8) Using ANOVA Table below, does the model explain a significant amount of
error in predicted y? (This is a six-step process).

9) Using ANOVA Table below, what amount of error is explained by the


independent variables in model? What is the official name of Rsqd?

10) What is the value of the standard error of the estimate S, Se or Syx?

11) Define the standard error of the


estimate

12) The standard error of the estimate is measured in what units of measure?

Model Summary
S R-sq R- PRESS R- AICc BIC
sq(adj) sq(pred)
26.0138 96.56% 95.99% 13158.8 94.43% 148.98 147.81

Analysis of Variance
Source DF Seq Contributio Adj SS Adj F-Value P-
SS n MS Value
Regression 2 228015 96.56% 228015 114007 168.47 0.000
Temp(F) 1 178624 75.64% 176938 176938 261.47 0.000
InsAttic 1 49390 20.92% 49390 49390 72.99 0.000
Error 12 8121 3.44% 8121 677
Total 14 236135 100.00%

13) Graphically test of assumptions for regression analysis. Use the


following graphs to test the respective assumptions for regression.

3|P age
Plot of residuals for fit value. Discuss
homoscedasticity/constant variance
assumption

Examine Residuals on plot and discuss


if assumption of homoscedasticity or
constant variance is satisfied.

Plot(s) of residuals for independent


variable. Discuss
homoscedasticity/constant variance
assumption

Examine Residuals on plot and discuss


if assumption of homoscedasticity or
constant variance is satisfied.

Plot(s) of residuals for independent


variable. Discuss
homoscedasticity/constant variance
assumption

Examine Residuals on plot and discuss


if assumption of homoscedasticity or
constant variance is satisfied.

4|P age
Regression Equation
Using Regression Model for Prediction:
HeatingOilGallons = 562.2 - 5.437 Temp(F) - 20.01 InsAttic

Settings

Setting
Variable
Temp(F) 35
InsAttic 6.5

Prediction
Fit SE 95% CI 95% PI
Fit
241.791 6.95807 (226.630, 256.951) (183.119, 300.462)

14) Use of model for prediction for x value within relevant range of x.

 -What is the predicted/fit value of y? Interpret the meaning of the fit? Include
settings for x variables in your interpretation.

 -What are the bounds of PI interval? Interpret the meaning of the PI interval? Include
settings for x variables in your interpretation.

 -What are the bounds of the of CI interval? Interpret the meaning of the CI interval?
Include settings for x variables in your interpretation.

Use ordinary least squares regression to develop predictive model. Identify


regression coefficients and regression equation.
Regression Analysis: Value versus Heating, Fireplace

5|P age
Method
Categorical predictor (1, 0)
coding

Regression Equation
Fireplace
0 Value = 50.09 + 16.19 Heating

1 Value = 53.94 + 16.19 Heating

Coefficients
Term Coef SE 95% CI T-Value P- VI
Coef Value F
Constant 50.09 4.35 (40.61, 59.57) 11.51 0.000
Heating 16.19 2.57 (10.58, 21.80) 6.29 0.000 1.00
Fireplace
1 3.85 1.24 (1.15, 6.56) 3.10 0.009 1.00

15) Why are two regression equations provided in the dummy variable model?
Explain each?

16) Using the coefficients table shown above, test hypothesis the regression
coefficients to determine if x variables are significant predictors of y. (Be able to
use a six-step process for T and pvalue techniques.)

 -What is the value of the constant bo? Interpret the constant?

 -What is the value of the bheating? Interpret the value of the bheating?

 -What is the value of bfireplace? Interpret the value of bfireplace ?

Using ANOVA Table below test the model for significance (six step process)
Model Summary
S R-sq R- PRESS R- AICc BI
sq(adj) sq(pred) C
2.26260 81.11% 77.96% 102.233 68.56% 75.72 74.55

Analysis of Variance

Source DF Seq Contributio Adj Adj F- P-


SS n SS MS Value Value
Regression 2 263.70 81.11% 263.70 131.852 25.76 0.000
Heating 1 214.37 65.93%7 202.36 202.357 39.53 0.000
Fireplace 1 49.33 15.17% 49.33 49.330 9.64 0.009
Error 12 61.43 18.89% 61.43 5.119
Total 14 325.14 100.00%

6|P age
17) Using ANOVA table above, does the model explain a significant
amount of error in predicted y?

18) Using ANOVA table above, what amount of error is explained by the
independent variables in model?

19) What is the official name of


Rsqd?
Use of Model for Prediction of Y home value:

Regression Equation
Value = 50.09 + 16.19 Heating + 0.000000 Fireplace_0
+ 3.85 Fireplace_1
Settings

Setting
Variable
Heating
Prediction 1
Fireplace 1
Fit SE Fit 95% CI 95% PI
70.1293 1.86976 (66.0554, 74.2032) (63.7341, 76.5245) X

X denotes an unusual point relative to predictor levels used to fit the model.

20) Use of model for prediction for x value within relevant range of x.

 -What is the predicted/fit value of y? Interpret the meaning of the fit? Include
settings for x variables in your interpretation.

 -What are the bounds of PI interval? Interpret the meaning of the PI


interval? Include settings for x variables in your interpretation.

 -What are the bounds of the of CI interval? Interpret the meaning of the CI
interval? Include settings for x variables in your interpretation.

7|P age

You might also like