Scribd
Upload
19 views14 pages

Lecture 8

Uploaded by

Alper Kara
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
19 views14 pages

Lecture 8

Uploaded by

Alper Kara
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

Lecture 8

Statistical aspects of regression


and
Multiple regression
From Lecture 7
• We have cross-sectional data of 70 tropical countries

Forest loss Population density Crop ch. Pasture ch.


Number of observations 70 70 70 70
Max 5.3 2,769 39.7 33.5
Min 0 0.89 -2.9 -20.1
Mean 1.1 643.5 6.6 3.0
Std dev 0.93 730.85 7.96 8.50

Regression Statistics
R-Square 0.43
Adjusted R Square 0.43
Coeff S.E t P-value Lower 95% Upper 95%
Intercept 0.59 0.112 5.34 0.000 0.37 0.82
Pop dens 0.0008 0.0001 7.23 0.000 0.0006 0.001
Confidence intervals
• Confidence intervals provide interval estimates
• We are 95% confident that  is greater than 0.0006 and less than 0.001. This
indicates that the marginal effect on deforestation is greater than 0.0006 and
less than 0.001.
• The more confident you wish to be about an interval the wider it becomes
so 99% confident intervals will always be wide than 95% confidence intervals.

Regression Statistics
R-Square 0.43
Adjusted R Square 0.43
Coeff S.E t P-value Lower 95% Upper 95%
Intercept 0.59 0.112 5.34 0.000 0.37 0.82
Pop dens 0.0008 0.0001 7.23 0.000 0.0006 0.001
Hypothesis testing

Regression Statistics
R-Square 0.43
Adjusted R Square 0.43
Coeff S.E t P-value Lower 95% Upper 95%
Intercept 0.59 0.112 5.34 0.000 0.37 0.82
Pop dens 0.0008 0.0001 7.23 0.000 0.0006 0.001
Hypothesis testing
Hypothesis testing

Regression Statistics
R-Square 0.43
Adjusted R Square 0.43
Coeff S.E t P-value Lower 95% Upper 95%
Intercept 0.59 0.112 5.34 0.000 0.37 0.82
Pop dens 0.0008 0.0001 7.23 0.000 0.0006 0.001
Multiple regression model
Explaining house prices

Variable Observations Mean Std. Dev. Min Max


Sale price 546 68,121.6 26,702.7 25,000 190,000
Lot size 546 5,150.3 2,168.2 1,650 16,200
Number of bedrooms 546 2.9 0.7 1 6
Number of baths 546 1.3 0.5 1 4
Number of stories 546 1.8 0.8 1 4
Explaining house prices
R-squared = 0.5355

Sale price Coefficient Std. Err. t t P>t [95% Confidence Interval]


Lot size 5.42 0.36 14.70 0.000 4.7 6.1
Number of bedrooms 2,824.6 1,214.8 2.33 0.020 438.2 5,210.9
Number of baths 17,105.17 1,734.4 9.86 0.000 13,698.1 20512.2
Number of stories 7,634.8 1,007.9 7.57 0.000 5,654.8 9614.9
Constant -4,009.5 3,603.1 -1.11 0.266 -11,087.3 3068.2
Omitted variables bias

• If we omit explanatory variables that should be present in the regression and if


these omitted variables are correlated with explanatory variables that are
included, then the coefficients on the included variables will be wrong.

Multicollinearity

• Problem that arises if some or all the explanatory variables are highly correlated
with one another
Omitted variables bias

• If we omit explanatory variables that should be present in the regression and if


these omitted variables are correlated with explanatory variables that are
included, then the coefficients on the included variables will be wrong.

Sale price Coefficient Std. Err. t P>t [95% Confidence Interval]

Lot size 5.9 .39 15.06 0.000 5.18 6.74


Number of bedrooms 5,976.8 1,271.8 4.70 0.000 3,478.5 8,475.1
Number of stories 9,636.3 1,071.4 8.99 0.000 7,531.6 11,740.9
Constant 2,246.7 3,848.9 0.58 0.560 -5,313.9 9,807.3
Multicollinearity

• Problem that arises if some or all the explanatory variables are highly correlated
with one another

Sale price Lot size Number of bedrooms Number of baths Number of stories
Sale price 1.0000
Lot size 0.5358 1.0000
Number of bedrooms 0.3664 0.1519 1.0000
Number of baths 0.5167 0.1938 0.3738 1.0000
Number of stories 0.4212 0.0837 0.4080 0.3241 1.0000

You might also like