Bayesian Analysis Workshop
  • Home
  • Notebooks
    • 01 Bayesian Reasoning
    • 02 Priors
    • 03 Regression
    • 04 Binary and Ordinal Models
    • 05 Hierarchical Models
    • 06 Model Checking
    • 07 Reporting
  • Frequentist Workflow
  • Setup
  • Facilitator Guide
  • Reporting Templates
  • Data
    • Download Datasets
    • Education Intervention
    • Frequentist Workflow
    • Teacher Adoption
    • Student Engagement

Bayesian Analysis for Applied Researchers

Bayesian Analysis Workshop
Home Notebooks Frequentist Workflow Data Setup Facilitator Guide Reporting Templates
Workshop Resources
Setup Guide Frequentist Workflow Data Downloads Facilitator Guide Reporting Templates Source on GitHub
Spring 2026 cohort · open
Applied Bayesian Analysis Workshop

Bayesian Analysis
for Applied Researchers

A practical, scenario-based workflow for moving from p-values to posterior reasoning — priors, regression, multilevel models, model checking, and reporting.

Built for researchers who already use frequentist methods. Each module pairs a research scenario with the modeling decisions an applied analyst actually faces: which likelihood, which prior, which check, and which sentence to write when the posterior comes back.

Start Notebook 1 → View Reporting Templates →
Seven notebooks
Applied research scenarios
Reporting-ready interpretations
Workshop Route 6 stages
1
Frequentist question
Start with your research question.
2
Prior
Make assumptions explicit.
3
Likelihood
Choose the right model for your data.
4
Posterior
Update beliefs with the data.
5
Model check
Evaluate and refine your model.
6
Reporting claim
Write the right sentence.
Facilitated by
Portrait of Moeketsi Mosia

Moeketsi Mosia

Bayesian methods·Applied education research
✦

Guiding you through a practical workflow from modeling decisions to reporting claims.

The Seven-Notebook Pathway

A progression from p-values to posterior reporting

Frequentist habits
Defensible Bayesian claims
What changes in your reasoning

Three shifts that change what you can actually claim.

SHIFT 01
p < .05 → P(θ | data)

From p-values to posterior probabilities

A p-value tells you the data is surprising under a null you didn't believe. A posterior tells you what to believe now — and how strongly.

SHIFT 02
default flat → elicited prior

From default assumptions to explicit priors

Every analysis carries assumptions. Bayesian analysis asks you to write them down — and gives you the tools to stress-test them before they touch the data.

SHIFT 03
β̂ ± SE → 95% CrI

From coefficients to defensible research claims

A point estimate compresses everything into two numbers. A posterior keeps the shape — so reporting keeps the nuance reviewers and stakeholders are looking for.

© 2026 · Moeketsi Mosia · MIT licensed Built in the open · github / bayesian-analysis-workshop