In order to answer the questions, you will have to analyse the datasetattached via usingGRETL or other statistic software.

The data are a small selection of international indicators collated by the World Bank. A detailed description of each variable is available in the data file by clicking on Data > Read info(orData > Dataset info, depending on thegretlversion you have).

1. Graph, as a scatterplot, the relationship between life expectancy (lifexp) and Gross Domestic Product (GDP) per person (gdp_pc), an indicator of income per person. Can you tell anything about the relationship between life expectancy and income per person from this graph?
2. Create a new variable: the log of GDP per person (l_gdp_pc). Graph the relationship between life expectancy and the log of GDP per person, and compare this graph with the graph you generated in your answer to Question 1. For which of the two scatterplots is a linear model more appropriate? Explain your answer.
3. Now calculate the correlations between life expectancy and the two measures of GDP per person (in levels and logs). What do they tell you? Are they more informative than scatterplots between the variables?
4. Estimate a simple (univariate) regression model in which life expectancy (lifexp) is a function of log of GDP per person (l_gdp_pc). Then:

1.  
   1. Write down the model that you have estimated using your results for the coefficients.
   2. Interpret and comment on the statistical significance of the slope coefficient of log GDP per person.
   3. Comment on the statistical quality of the overall model.
2. Estimate a multiple (multivariate) regression model in which life expectancy (lifexp) is a function of log of GDP per person (l_gdp_pc), prevalence of HIV (hiv), progress to secondary education (educ_sec), and health expenditure as a percentage of GDP (health_gdp). Then:

. Write down the model that you have estimated and its estimated coefficients.

1.  
   1. What do the magnitude and significance of the estimated coefficients tell you?
   2. Comment on the statistical quality of the model.
   3. Comment on the differences between the univariate and multivariate models.

Taskasksto estimate the relationship between life expectancy and income per person.

Questions 1, 2 and 3 ask to explore the data using charts and descriptive statistics. Make sure you read the questions carefully to decide which variable to put in the Y-axis (the dependent variable) and in the X-axis (the independent variable).

Questions 4 and 5 ask you to estimate two regression equations that model the relation between life expectancy and income per person, and to interpret, and in question5, to compare the results.