The regression of deforestation on population density (Sample size, N=70)

Assume you regress deforestation (Y) on population density (X), and obtain below regression output

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.659150127
R Square 0.43447889
Adjusted R Square 0.426162404
Standard Error 0.703122104
Observations 70

ANOVA
df SS MS F Significance F
Regression 1 25.82797 25.82797 52.24308 5.50331E-10
Residual 68 33.61789 0.494381
Total 69 59.44586

Coefficients Standard Error t Stat P-value Lower 95% Upper 95%

Intercept 0.599964899 0.112318 5.341646 1.15E-06 0.375837116 0.824092682
Pop dens 0.000842327 0.000117 7.227937 5.5E-10 0.00060978 0.001074874

Excel output

. reg forestloss popdens

Source SS df MS Number of obs = 70

F(1, 68) = 52.24
Model 25.8279706 1 25.8279706 Prob > F = 0.0000
Residual 33.6178894 68 .494380726 R-squared = 0.4345
Adj R-squared = 0.4262
Total 59.44586 69 .861534203 Root MSE = .70312

forestloss Coefficient Std. err. t P>|t| [95% conf. interval]

popdens .0008423 .0001165 7.23 0.000 .0006098 .0010749

_cons .5999649 .1123184 5.34 0.000 .3758371 .8240927

Excel output

Hypothesis testing

You can test whether the slope coefficient is statistically significant at 5% level of significance using t-
statistic or the significance of R-squared using F-statistic

For individual coefficient, you can use confidence interval, t-ratio or P-value

Steps

Hypothesis: H0: 2  0
 H1: 2  0

Decision rule: Reject Ho if P-value < 0.05 or t-value > t-crit (1.995469), obtained by
=[Link].2T(0.05,68) in excel , the probability is 0.05 and degree of freedom is N-K=68 where N=70
is the number of observation and K=2 is the number of estimated coefficients(Intercept and slope)

ˆ2  0
Decision: Since p<0.05 and t= = (0.000842327/0.000117)=7.227937>1.995469 (t-
Se2

critical value, you can use either statistical table or excel) , we reject H0 at a 5 % level of
significance. Note that using the confidence interval approach, we also reject the null since
95% confidence interval [0.00060978; 0.001074874] does not contain zero.

If the confidence interval does not contain the value claimed by the null hypothesis, then our
sample result is different enough from the claimed value, and we therefore reject H0. However,
if the confidence interval contains the value claimed by the null hypothesis, then our sample
result is close enough to the claimed value, and we therefore do not reject H0.

Conclusion: There is enough statistical evidence to suggest that 2 is significantly different to

zero.

Below is the statistical table

At 5% level of significance (two-tails , why two tails , the hypothesis is H0: 2  0 and H1: 2  0) with
68 degree of freedom, you can see that we don’t have exactly value for 68, the value closer to 68 is 60,
hence we can use critical value of 2.

Testing the significance of R-squared.

R-squared measures the strength of the relationship between your model and the dependent
variable. However, it is not a formal test for the relationship. The F-test of overall significance
is the hypothesis test for this relationship. If R2 = 0 then X does not have any explanatory power
for Y. The test of the hypothesis R2 = 0 can therefore be interpreted as a test of whether the
regression explains anything at all. In the case simple regression, this test is equivalent to a test
of β (slope coefficient) = 0. (see details page 83).

The following are the steps:

Hypothesis: H0: R2=0, or 2 =0 (explanatory variable X has explanatory power on Y)

 H1: X does not have explanatory power on Y.

2
Decision rule: Reject Ho if P-value < 0.05 or F-statistic > F critical, F-statistic=  N 2 K  R
1  R   K  1
R 2 /  K  1
which is equivalent to , K-is the explanatory variable plus the intercept term.
1  R 2  /  N  K 
Hence, the critical value is obtained in excel: =[Link] (probability, deg_freedom1,
deg_freedom2) , degree of freedom 1 is N-K and degree of freedom K-1 ( simple regression
we have two coefficients(intercept and one slope coefficient) hence K-1=1.

Decision: Since p<0.05 and F statistic >F critical, you can use either statistical table or excel)
,we reject H0 at a 5 % level of significance.