Quantitative UX Essential Pack: From Basics to Advanced

A five-session, hands-on workshop for UX researchers, designers, and product managers who want to make confident, evidence-based decisions. No statistics background is needed.

Instructor: Mohsen Rafiei, Ph.D., Quant UXR Lead and Assistant Professor of Cognitive Psychology

Offered by: Perceptual User Experience Lab (PUXLab)

Format: Five live online sessions of 2.5 hours each, hands-on throughout

Tools: Jamovi, JASP, R, and AI assistants.

Contact: Admin@puxlab.com

Session 1

Descriptive and Parametric Statistics: Seeing the Story Behind the Numbers

This session covers how to turn raw survey responses into summaries that can be defended in front of stakeholders. Topics include the mean, median, and standard deviation, together with the conditions under which each measure misleads; types of variables; frequency counts and cross-tabulations; what a p-value actually tells the researcher and what it does not; 95% confidence intervals; and effect sizes, which answer whether an observed difference is large enough to matter. Participants learn to compare designs using t-tests and ANOVA.

Session 2

Non-Parametric Tests: Making Sense of Messy, Real-World Data

This session examines what a five-user test can and cannot tell us. It explains how small samples exaggerate effects, reverse their direction, and produce unstable p-values, and it presents the study designs that protect against these problems. Participants review the data shapes UX researchers most often encounter, such as task times, success rates, and error counts, and learn the robust tests built for them: Welch's t-test, the Mann-Whitney U test, the Wilcoxon signed-rank test, the Kruskal-Wallis test, chi-square, and Fisher's exact test.

Session 3

Bayesian Statistics: Updating Evidence Like a Designer

This session introduces a different way of thinking about evidence: state what you expect, then let the data update it. Topics include priors; credible intervals, which state directly what stakeholders want to know; and Bayes factors, which can provide evidence in favor of no difference, something p-values cannot do. The session also covers Bayesian A/B testing, in which monitoring results as they arrive is statistically legitimate.

Session 4

Regression Analysis: What Drives User Outcomes and the Foundation Behind Predictive AI

This session moves from asking whether a difference exists to asking what drives an outcome, and by how much. It covers simple, multiple, and logistic regression; mixed-effects models for repeated measures; and a decision guide that maps each type of outcome to the appropriate member of the regression family. Participants learn to read any regression output table, to avoid the five mistakes that most often affect beginners, and to translate coefficients into business language.

Session 5

Modeling and AI: Machine Learning and LLMs for Smarter UX Decisions

This session clarifies what artificial intelligence, machine learning, and generative AI actually are, and why large language models are not magic. It covers AI methods for open-ended feedback, including topic modeling, semantic clustering, and aspect-based sentiment analysis; predicting and segmenting users on the basis of behavior; and the situations in which AI should not be used. Participants receive a practical trust checklist for verifying any AI-generated analysis before it reaches a stakeholder.

What Participants Will Take Away

•   The ability to choose the right analysis for any UX question and to explain it in plain language

•   A complete toolkit that works even with five users and a Friday deadline

•   Slides, in-depth attendee guides, curated practice datasets, and reporting templates for every method covered

•   A certificate of completion

Prerequisites: none. If you can read a bar chart, you are ready.

For upcoming cohort dates and registration, please contact Admin@puxlab.com.