1. Which data set will you be working with?
CO2 Emissions by Country in The World Bank.
2. What is your dependent variable and what is your main or key independent variable?
Dependent variable： CO2 emissions by country
Independent variable：population
3. What is your research question?
How does population affect the amount of CO2 emissions per country between the years 1960 and 2014?
4. Why is it an important question (i.e., why should we care)?
Analyzing the relationship population between the amount of CO2 emissions by country helps in determining the manner its effects are distributed. The aim is to devise means at which this emission can be addressed.
5. Do you think there will be a need to collect additional data? If so, what variable do you have in mind? You do not have to have a full answer to this question yet, but do your best.
I think there is a need to collect additional data of GDP in every country. Part of the research was to understand the pattern of distribution of CO2. However, the main aim was to explore ways in which its adverse effects could be countered. Therefore, knowing the GDP of every country would also aid in addressing any negative effects that come of it.
6. Based on your research question, what is your hypothesis about the relationship between the main independent variable and the dependent variable?
Hypothesis: The rate of CO2 emissions by country may not be directly proportional to the respective country’s size of the population.
7. Give your reasons why you have this hypothesis; that is, please try to give your thoughts on the underlying economic theory provides a reason for your hypothesis.
The amount of CO2 produced by a country is dependent on factors such as weather changes and other aspects such as plantations. However, a huge population size may not imply any variance to the level of CO2. A country with a small size of population may have a higher level of this element than those with more population.
1. What does the simple regression suggest about the relationship between your x and y variables? Is the slope coefficient statistically significant? What is the size? How do you interpret it?
Source | SS df MS Number of obs = 12,248
-------------+---------------------------------- F(1, 12246) = 37067.11
Model | 103764.011 1 103764.011 Prob > F = 0.0000
Residual | 34280.9068 12,246 2.79935544 R-squared = 0.7517
-------------+---------------------------------- Adj R-squared = 0.7516
Total | 138044.917 12,247 11.2717333 Root MSE = 1.6731
------------------------------------------------------------------------------
lnco2 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lnPop | .9880223 .0051318 192.53 0.000 .9779631 .9980815
_cons | -6.35125 .0833884 -76.16 0.000 -6.514704 -6.187796
To measure the impact of population size on carbon dioxide emissions, we established model (1):
（1）
The dependent variable is the logarithm of carbon dioxide emissions (lnco2), and the independent variable is the logarithm of the population, which is the error term.
From the specific regression results, we can get:
(1) The adjusted R2 value is 0.7516, which indicates that the independent variable lnpopulation can explain the information of dependent variable lnco 275.16%.
(2) The value of F statistic is 37067.11, and its corresponding P value is less than 0.05 (in this case, the significance level is controlled at 5%).
(3) From the specific results, the independent variable lnpopulation has a significant positive effect on the dependent variable lnco2. At this time, its t statistic value is 192.53, its corresponding P value is less than 0.05, which shows that the regression coefficient value of the variable lnpopulation is not equal to 0, that is, it is meaningful in statistical sense, and its corresponding regression coefficient number. The value is 0.988, that is, when the variable lnpopulation changes 1 unit, the variable lnco2 changes 0.988 units.
2. Describe the additional control variables that you included and why you included them in your regression. Now explain the results: What does the table suggest about how they affect your dependent variable (or not) and how do you interpret the coefficients and the t-statistics?
Source | SS df MS Number of obs = 7,365
-------------+---------------------------------- F(3, 7361) = 30919.34
Model | 62263.3061 3 20754.4354 Prob > F = 0.0000
Residual | 4941.02965 7,361 .671244349 R-squared = 0.9265
-------------+---------------------------------- Adj R-squared = 0.9264
Total | 67204.3358 7,364 9.12606407 Root MSE = .8193
------------------------------------------------------------------------------
lnco2 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
lnPop | .1871355 .0066558 28.12 0.000 .1740882 .2001827
lnManu | .418834 .016292 25.71 0.000 .386897 .450771
lnGDP | .3929029 .0179113 21.94 0.000 .3577917 .4280142
_cons | -11.79659 .107627 -109.61 0.000 -12.00757 -11.58561
------------------------------------------------------------------------------
In the model (1), we only considered the influence of lnpopulation on the dependent variable lnco2. There are many factors affecting carbon emissions, such as the level of economic development, the level of manufacturing expansion and so on. Therefore, considering these factors may have a significant impact on carbon emissions, this paper studies the effect of lnpopulation on carbon emissions. Model (1) based on the establishment of the model (2):
（2）
From the regression results, we can see that:
(1) The adjusted R2 value is 0.9264, which shows that the independent variables lnpopulation, in Manu and lnGDP can explain the 92.64% of information of the dependent variable lnco2, which is higher than the adj.R2 value obtained by the model (1). The results further show that other factors are affecting the carbon emissions besides the population.
(2) The value of the F statistic is 30919.34, and its corresponding P value is less than 0.05 (in this case, the significance level is controlled at 5%). This result shows that the whole model (2) is remarkable.
(3) From the regression results, we know that the variable lnpopulation has a significant positive effect on the dependent variable lnco2. At this time, its t statistic value is 28.12, its corresponding P-value < 0.05, indicating that the value of the regression coefficient corresponding to the variable lnpopulation is not equal to 0, that is, it is significant in the statistical sense, the same below. The corresponding regression coefficient is 0.187 that is when the variable lnpopulation changes by one unit and the variable lnco2 changes by 0.187 units, which indicates that with the increase of population, carbon emissions are also rising. The variable lnManu has a significant positive effect on the dependent variable lnco2, and its corresponding regression coefficient is 0.419, that is, when the variable lnManu changes by one unit, the variable lnco2 changes by 0.419 units, which indicates that with the development of manufacturing industry, carbon emissions are also increasing. The variable lnGDP has a significant positive effect on the dependent variable lnco2. Its corresponding regression coefficient is 0.393, that is, when the variable lnGDP changes by one unit, the variable lnco2 changes by 0.393 units, which shows that with the development of the economy, carbon emissions are also rising. From the estimated results, the growth of carbon emissions relates to economic development, manufacturing development and population growth.
3. What happens to the size and significance of your key independent variable across the two regressions? Does it change in an appreciable manner when you include additional controls? If so, why do you think this is? If not, why might this be?
The results are tiny different when more variables are added in the model. This may happen due to the growth rate. Because whenever the population increases the number of industries increase due to which the waste products also improve and the carbon dioxide emission will be increased. In the model 1 and model 2, we can see that the coefficient value of the variable ln population has changed from 0.988 to 0.187, which indicates that the value of the coefficient has changed to some extent, but the direction of its effect on the dependent variable carbon emissions has not changed, both of them are positive effects. On the other hand, it also says that the value of the coefficient has changed from 0.988 to 0.187. Compared with the results of the model (1), the effect of the independent variable lnpopulation on the dependent variable lnco two is weakening in the model (2). When additional control variables are added to the model, although the value of the regression coefficient will change, the direction of its role should not change. Therefore, I think that when we added additional control variables to the model, they will break in an obvious way.
01
2
lnco=c+clnpopulation
+e
0124
2ln
lnco=c+clnpopulationcManucGDP
+++e
mt19941i24
mypaper.docx
0
