Course Catalogue

View Specification Document | Reading List

This core course will cover the following topics:

1. Linear Regression & predicting continuous data

2. Analysis of Count data (Poisson, negative binomial, zero-inflated models)

3. Cox (and parametric) Survival analysis including competing risks

4. Logistic regression & predicting categorical data

5. Introduction to Time Series regression and forecasting

6. Hierarchical models and clustered data

7. Introduction to causal inference methods (I) (Propensity Score Matching and IV)

8. Introduction to causal inference methods (I) (Difference in difference, synthetic controls, regression discontinuity)

9. Handling missing data

10. Introduction to Bayesian analysis

12 x 1 hour lectures and 10 hours of tutorials

None

None

None

One written assignment of 2000 words max. (70%)

Two computing exercise submissions in the form of short answer exercises (2 x 15%). Will include syntax used.

To provide students with a solid grounding in modern and cutting edge statistical methods

To equip students with the appropriate skills necessary to analyse routinely collected health and administrative data using the statistical software packages R and Stata

To provide students with the skills necessary to choose an appropriate method of analysis for a given statistical problem.

To give students experience of carrying out advanced statistical analyses in R and/or Stata

By the end of this course students will be able to:

1. Critically assess the output of univariable and multivariable linear regression models and demonstrate how to develop and refine the models.

2. Understand the distinctions between zero inflated, Poisson and negative binomial regression and critically appraise their role in model development of count data

3. Understand the distinction between semi-parametric Cox and parametric survival models and critically appraise the output of both

4. Critically assess the design conditions as to when to select the most appropriate survival analysis model

5. Interpret and critically appraise the output from a univariate and multivariable logistic regression model

6. Understand and critically appraise the design conditions as to when to use time series methods for data analysis

7. Critically assess the design conditions as to when to use hierarchical models in any given statistical context and critically appraise their output

8. Demonstrate a critical awareness of the techniques used for causal inference and critically appraise the output from subsequent models.

9. Understand the consequences of missing data and critically appraise the methods used to solve missing data problems

10. Critically appraise the design conditions as to when to use Bayesian methods for data analysis and critically appraise the output from Bayesian models.

Students must submit at least 75% by weight of the components (including examinations) of the course's summative assessment.