Project

You are required to complete a course project. The goal is to practice communicating statistical ideas clearly and critically—like you are reporting to colleagues whether a line of work is worth pursuing.

Types

Choose one of the following project types:

  • Research article review
    Pick a scholarly research article related to topics in statistical inference (e.g., estimation, testing, likelihood/Bayes, asymptotics, bootstrap, identifiability, robustness, missing data, multiple testing, etc.) _that does not duplicate a topic discussed in class. Your article must be different from your classmates’ choices.

  • Simulation or data analysis study
    Design and carry out a focused study that uses computation to investigate a statistical question. Examples:

    • Analyze a data set of interest, but go beyond any existing analysis,
    • Compare procedures (power, bias, coverage, MSE) under realistic departures from assumptions,
    • Illustrate a theorem/asymptotic approximation with finite-sample behavior,
    • Replicate and critique an empirical claim from a paper,
    • Build a small method “pipeline” (e.g., model selection + inference) and evaluate it.

Submission

  • Submit your project topic using the topic submission form by Friday, March 6, 2026. Revised topics are submitted using the same form.
  • After the topic deadline, check the posted topic list to confirm approval and double check the schedule. This list updates regularly but not immediately.
  • Prepare a 15-minute oral presentation (+5 minutes for questions) and a handout for the audience.

Presentations and observations

  • Project presentations are scheduled for April 20–22. You must assign yourself to a presentation slot using the Excel spreadsheet by Friday, March 27, 2026.
  • Each student must observe and assess TWO other presentations (using the Project Assessment Form).
  • Your project grade is likely to suffer if:
    • you are not on time for your presentation or observations,
    • you are not prepared,
    • you do not observe at least two presentations, and/or
    • your observer feedback is not significant.

Important to keep in mind

Your presentation should be from the perspective that you are reporting to your colleagues why this article/method/result is or is not worth pursuing further.

  • Your audience is smart, but not specialized in your exact topic. Provide background, definitions, and motivation.
  • Be critical: clearly state strengths and weaknesses, and what assumptions matter.
  • You have access to tools (including AI) that were not commonly available a few years ago—so I will expect more polish and deeper checking.
  • I may not be an expert in what you present, but I expect you to be, and I will ask probing questions.
  • Your grade will depend on both:
    • how clearly you present, and
    • how well you answer questions.

Guidelines

  • Article review.
    • Choose a topic not already covered in lecture in essentially the same way.
    • Prefer papers with enough technical content to support real discussion (not just an application report).
    • Your job is to explain the idea, demonstrate it on a simple example, and assess it.
  • Simulation/data study.
    • Keep the scope appropriate: one sharp question is better than many shallow ones.
    • Predefine your evaluation metrics (e.g., coverage, power, MSE, false discovery rate).
    • Use clear plots/tables and report uncertainty when relevant.

General presentation advice:

  • Keep the big picture in view; don’t get bogged down in algebra.
  • Choose one simple example that illustrates the main point.
  • Look at your audience while speaking; do not read your notes.
  • Motivate your talk: background, prior work, and why anyone should care.
  • Discuss pros/cons and failure modes.
  • Don’t use too many slides; about one per minute is usually enough.
  • Don’t overload slides; prioritize figures/tables over dense text.
  • Include key formulas/figures directly in your slides rather than making the audience hunt in the paper.
  • Use your own words. Do not plagiarize.
  • Practice beforehand. Time yourself.

You will be assessed by your observers and by me using this form:
Project Assessment Form