You are required to undertake an empirical project involving the analysis: The Social protection Survey Assignment, UOS, UK

The Social Protection Survey (EPS) is the largest and oldest longitudinal survey that exists in Chile, with a sample of around 16,000 respondents distributed in all regions of the country. The questionnaire includes questions on labour and social security with detailed information in areas such as education, health, social security, job training, wealth and assets, family history, and household information.

 In this dataset there are 9 variables:

1. folio – a participant ID
2. Age (in years)
3. Female (0= male/1= female)
4. Edu_yrs (years of education)
5. Ins (insurance: 1 = private, 2 = public, 3 = none)
6. Hlstat (Health status 1= best, 5= worst)
7. R_incomeall (income in Chilean peso)
8. Survey (year of data survey)
9. N_depend (number of dependents)

For this assessment, you are required to undertake an empirical project involving the analysis and interpretation of a secondary dataset to test a health economic hypothesis. Learning outcomes to be assessed - you should be able to:

• Show an in-depth understanding of the application of statistical techniques to health data.

• Show an in-depth understanding of the use of quasi experimental designs.
• Critically examine different regression models.
• Apply statistical techniques using standard econometric software.
• Develop an economic hypothesis and understand the appropriate statistical methods required to test the hypothesis.

You should conduct your analysis in R/RStudio.

Your analysis should be limited to the latest wave of data for the year 2009 only and you should employ cross sectional methods.

You should make sure to cover the following steps in your analysis:

• Develop a health economic hypothesis of a relationship between two of the variables in the dataset, supported by some outside literature or empirical research (make sure to include appropriate references).
• Describe and justify how you will test this hypothesis - you should employ both bivariate and multivariate methods (Note: remember only data for year 2009).
• Consider the level of missing data and appropriate methods for handling it.
• Present relevant descriptive statistics (Note: remember only data for year 2009) · Perform appropriate modelling/statistical tests to refute or support the hypothesis.
• Critically examine the results of your analyses.
• Consider the strengths and limitations of your methods and how other study design or analytical techniques could help address your hypothesis

o What is this analysis not able to tell us? Think about causality.

o Is there potential bias?

o Are there other longitudinal methods that could be used?