## PB HLTH 240C — Final Project

## Timeline

- Thursday, November 13th — Abstract/Proposal: email a single pdf per group to Davide and Oleg
- Thursday, December 11th, 12:30pm — Oral presentation
- Thursday, December 11th, 12:30pm — Report: email a single pdf per group to Davide and Oleg

## Guidelines for abstract

- Provide title and list of authors. Do not exceed 1 page.

## Guidelines for report

### Length of the report

- Not to exceed 15 pages of text.

### Structure of report

- Title and list of authors
- Abstract
- Introduction
- Methods
- Data: (real or simulation) study design, data structure
- Results
- Discussion: conclusions and open questions
- References
- Tables and Figures: numbered, captions, legends, …

### Subject-matter

- Introduction and precise statement of question.
- Avoid uncessary jargon.

### Translation of subject-matter question into statistical question

- State clearly and precisely the data structure, model (i.e., set of data generating distributions), and parameters of interest.

### Literature review

### Methods

- For each method under consideration, describe precisely the estimation/testing approach for the parameter of interest.
- What are underlying assumptions? Are these plausible?

### Simulation study

- It can be tricky to find suitable models for simulation.
- State clearly and precisely the data generating model.
- Argue that your chosen model is appropriate and realistic.
- What are good measures of performance: risk, ROC?

### “Real” datasets

- Benchmark for comparison: What is “truth”?
- What are good measures of performance: risk, ROC?

### Precise use of statistical terminology and notation

- E.g. Parameter vs. estimator, statistical significance, bias, variance.

Advertisements