Syllabus (public companion)

Orientation and course shape — the LMS is authoritative for graded specifics

This is a public companion to the syllabus. It describes the course’s purpose, shape, and rhythm. It is not the contract of record: official dates, grading weights, policies, and submission details live in Blackboard (the LMS), which governs. Where this page and the LMS differ, the LMS wins.

Course description

This course introduces Bayesian statistics as a coherent framework for reasoning under uncertainty, updating uncertainty with data, and using models to predict and decide. It is applied and computational, but not a software-training course or a formula-memorization course. We move from discrete Bayes’ rule and diagnostic reasoning to priors, likelihoods, posteriors, posterior prediction, simple conjugate models, simulation-based summaries, Bayesian regression, model checking, and introductory hierarchical ideas. Markov chain Monte Carlo appears as a computational tool when useful; MCMC theory is not the center of the term.

Who it is for

A mathematically serious undergraduate audience that has completed an introductory/applied statistics course and an undergraduate probability course (or instructor permission). You should arrive having seen conditional probability, independence, random variables, expectation and variance, common distributions, sampling variability, confidence intervals, p-values, and simple linear regression. No prior course in mathematical statistics is assumed; no prior programming course is assumed.

Learning outcomes

By the end of the course a student who has done the work should be able to:

  1. Explain Bayesian reasoning as updating uncertainty in light of evidence.
  2. Use Bayes’ rule in discrete settings and connect it to Bayesian modeling.
  3. Identify, write, and interpret the prior, likelihood, posterior, and posterior predictive distributions in an applied model.
  4. Work with simple models (Beta-Binomial, Gamma-Poisson, Normal) at an undergraduate level.
  5. Use probability notation and simulation to summarize posterior uncertainty.
  6. Calculate and interpret posterior probabilities, posterior means/medians, credible intervals, and predictive intervals.
  7. Evaluate how prior choices affect posterior conclusions and explain prior sensitivity plainly.
  8. Use posterior predictive thinking to check whether a model is adequate.
  9. Fit and interpret basic Bayesian regression for numerical and binary outcomes.
  10. Describe introductory hierarchical/multilevel ideas such as partial pooling and shrinkage.
  11. Distinguish parameter, predictive, model, and decision uncertainty.
  12. Compare Bayesian and classical summaries responsibly.
  13. Create reproducible work with local R, VS Code, and Quarto.
  14. Communicate Bayesian results to a non-technical audience with appropriate caution.

Weekly rhythm

A lecture-format course meeting three times per week (Mon/Wed/Fri, 50 minutes). Most non-exam weeks follow a predictable shape: a concept day with a short exit ticket, a computation/application day with a short exit ticket, and a quiz/application day with a short Friday quiz followed by an applied case or project work. Doing the small weekly work is the most reliable path through the bigger assessments.

Assessment shape (indicative — not a contract)

Assessment is built from small, frequent evidence plus a few larger pieces. The categories and their approximate emphasis are below; the authoritative weights, dates, drop rules, and grading details are in the LMS, and are described here only to orient you to the shape of the work.

Category What it is Rough emphasis
Exit tickets brief in-class concept/interpretation checks (most Mon/Wed) small
Weekly quizzes short Friday quizzes on a new scenario small
Homework roughly biweekly; conceptual + modeling + reproducible Quarto/R work the largest single category
Midterm exam concept-and-computation, mid-term moderate
Project a final-third applied Bayesian analysis with a reproducible report moderate
Final exam cumulative; emphasizes synthesis and interpretation moderate

Low scores in the small categories are dropped per the LMS policy. Nothing on this page sets a weight, a due date, or a grade.

Software and reproducibility

Local R, VS Code, and Quarto; selected R packages for Bayesian modeling, visualization, and posterior simulation (final list posted in the LMS). Reproducibility is part of the statistical work: code, output, graphics, and written interpretation should support each other. See the setup page.

AI use (summary)

AI assistants may be used as study tools (concept explanations, practice problems, debugging, checking understanding) but may not produce submitted work, and are prohibited during quizzes and exams unless explicitly allowed. When you use an AI assistant on a graded assignment, include a brief AI Use Note: Tool · Purpose · Verification — the verification line is load-bearing. The authoritative AI policy is in the LMS.

Materials

  • Primary open text: Bayes Rules! (Johnson, Ott, Dogucu; CC BY-NC-SA 4.0), used freely online.
  • Supplemental open/library readings as announced.
  • A computer able to run the course software (personal, lab, or MAC fallback).
  • Blackboard Ultra for submissions, the gradebook, announcements, and dates.

Where things live

  • This site — public notes, labs, and resources (the instructional spine).
  • Blackboard — submissions, gradebook, official dates, graded prompts, keys, and rubrics (authoritative).
  • Local software — R, VS Code, Quarto, and instructor-provided files for reproducible work.

The dated, graded, official syllabus of record is in the LMS. This page is an orientation companion and may be updated for clarity.