studysim 0.5.0

studysim 0.5.0, simulated medical studies for teaching in jamovi (menu Simulate). Fixes and improvements from a critical review of 0.4.

Answer key

• True model with direct and total effects; variables that act only through others are shown as such.
• Expected estimates adjusted as the causal diagram says, next to crude and fully adjusted ones, with estimand labels.
• Trials: risk per arm, risk difference, NNT, risk ratio, and marginal vs conditional OR/HR.
• Expected values averaged over several large studies, with their Monte Carlo interval.
• Numeric effect modifiers centred (by default at the population mean).

Simulated study

• Biases on purpose: recall bias, controls selected by exposure, a collider, non-adherence (ITT, as treated, per protocol), informative loss to follow-up, immortal time; in diagnostic studies, spectrum and partial verification, and accuracy at a cut-off.
• Other outcomes as rows of the Variables list, drawn in the causal diagram.
• Wide layout (one row per person) for crossover and longitudinal trials.
• Export codebook, and Export class: one CSV and codebook per student key, zipped, with a register and one answer-key report.
• Teacher code that locks a library's key; effect sizes on the diagram off by default.
• CSV files for spreadsheets (UTF-8 BOM; semicolons and decimal commas by language, or as chosen).
• Events-per-parameter warning when a model has fewer than 10.

Compare designs: replicate studies per design: mean estimate, bias, standard error, power, CI coverage and the sampling distribution.

Fixes

• Design options left empty now take the scenario's own values (the panel's defaults used to override them).
• Continuous outcomes no longer pile up at their bounds.
• Error messages translated in all four analyses.

Install: download the .jmo for your CPU (x64 Intel/AMD, arm64 Apple silicon) and use Modules (⊕) → Sideload in jamovi 28.4 or newer. English, Spanish and Catalan.

studysim-scenario-skill.zip: a Claude skill that turns a derived scenario into a teaching scenario.

Tutorial: https://victor-moreno.github.io/jamovi-studysim/ (new page: analysing each design in jamovi)

studysim 0.4.0

studysim 0.4.0, simulated medical studies for teaching in jamovi (menu Simulate). Everything new since 0.1.0 (0.2 and 0.3 were not released separately).

Designs (16): cross-sectional, ecological, case-control, matched case-control, cohort, nested case-control, case-cohort, prognosis with competing risks; diagnostic accuracy (markers with a target AUC, marker panels) and method comparison (Bland-Altman); randomised (with interim looks), cluster randomised, 2×2 factorial, 2×2 crossover, longitudinal with dropout, before-after with regression to the mean.

Simulated study

• Several outcomes for the same people (linked outcomes, illness-death).
• Realism: missing data (MCAR, MAR, MNAR), measurement error, detection limits and assay batches, data-entry errors, extra null variables.
• Grouping of numeric variables (quantiles, cuts), log2/log10 and rescaling.
• Answer key with the true model and what crude and adjusted analyses of the design should find, now with its Monte Carlo uncertainty; a teacher's answer-key report (HTML).

Compare designs: one scenario under several designs drawn from the same population, with estimates and 95% CI against the truth, a forest plot and all the studies in one dataset.

Derive scenario: missing covariates by complete cases, mean, single or multiple imputation; codes such as 99 read as missing; structural values written as rules.

Library: 65 scenarios across 21 clinical domains, most replicating a real study (open data, teaching datasets or published tables, including papers chosen for critical reading). Browse them in the scenario catalogue.

Fixes

• Answer key with log-scale outcomes that have zeros.
• MAR missingness in method comparison.
• Case-control rows now in random order.
• Student key ignores capitals and spaces.
• Follow-up shown in each outcome's own time units.

Install: download the .jmo for your CPU (x64 Intel/AMD, arm64 Apple silicon) and use Modules (⊕) → Sideload in jamovi 28.4 or newer. English, Spanish and Catalan.

studysim-scenario-skill.zip: a Claude skill that turns a derived scenario into a teaching scenario.

Tutorial: https://victor-moreno.github.io/jamovi-studysim/

studysim 0.1.0

First release of studysim, simulated medical studies for teaching in jamovi (menu Simulate).

• Simulated study: scenario → outcome → design (case-control, matched case-control, cohort, cross-sectional, randomised trial, 2×2 crossover, prognosis); measured variables and grouping (quantiles or cuts); missing data; seed and student keys; data dictionary, descriptives, causal diagram and answer key; dataset opened in a new window or exported as CSV.
• Library: list, check, explore and export scenarios; custom libraries in YAML, Excel or JSON, with their own translations.
• Derive scenario: estimate a scenario from a real dataset and check it by simulation.
• Six built-in scenarios; English, Spanish and Catalan.

Install: download the .jmo for your CPU (x64 Intel/AMD, arm64 Apple silicon) and use Modules (⊕) → Sideload in jamovi 28.4 or newer.

studysim-scenario-skill.zip: a Claude skill that turns a derived scenario into a teaching scenario (see ai/ and the tutorial).

Tutorial: https://victor-moreno.github.io/jamovi-studysim/