Each deck guides studio time for its lesson: questions answered on students' devices, work in pairs and small groups, peer activities, and discussion as a class. A deck can also be presented without a room, worked through on your own with a recorded walkthrough, or printed as a facilitator guide. All 8 decks are ready.
Lessons
8 of 8 decks readyL · 01
A Structured Approach to Data AnalysisDrawing the research question as a causal diagram and choosing the adjustment set, a collider made by choosing the sample, the coding faults in a raw file and their corrections, verifying and cleaning the file in one script, and the first one-predictor associations.
L · 02
Data Cleaning and Descriptive AnalysesTypes of data error and the three missing-data mechanisms, screening and editing impossible values and codes, the IQR and z-score rules, complete cases against multiple imputation, and choosing the summaries and layout of a Table 1.
L · 03
Linear and Logistic RegressionChoosing a two-variable test, fitting a line by least squares and seeing the t-test as a regression, reading and testing the adjusted model of food insecurity and HbA1c, checking it with the four diagnostic panels and a variance inflation factor, fitting a logistic regression for poor control with its odds ratios and predicted probabilities, and choosing and comparing models, including an interaction with insulin use.
L · 04
Generalized Linear ModelsChoosing a model for the outcome, then logistic, ordinal, multinomial, Poisson and negative binomial regression and offsets, with the R code explained one argument at a time, run live, and run again in pairs on students' laptops.
L · 05
Modelling Dependent DataRecognising clustered and repeated data, measuring clustering with the intracluster correlation and the design effect, fitting and checking linear and logistic mixed models and generalized estimating equations, and analysing a randomized trial measured four times.
L · 06
Exploratory Data Analysis and VisualizationChoosing a display for the question, finding top-coding, heaping and missing-value codes in base R plots, building boxplots, smoothers and facets in ggplot2 one layer at a time, reading a correlation heatmap, and combining panels into a captioned figure.
L · 07
Measurement and PsychometricsScoring a scale with reversed items and checking the total, estimating test-retest reliability, agreement and internal consistency, weighing validity evidence and the size of an important change, and judging an exploratory and a confirmatory factor model, with item response curves.
L · 08
Mediation, Moderation and Path AnalysisNaming the role of a third variable on a causal diagram, testing whether an association differs between groups with an interaction, splitting an association into direct and indirect parts with a bootstrap, and comparing path models in lavaan, with the conditions and wording each result needs.
Students: join studio time. Open the join page and enter the four-letter code on the main screen, or scan the QR code.
Join studio time