Skip to contents

mlmr Public Beta Demo Guide

Source: DEMO.md

Thank you for trying mlmr.

mlmr is an open-source Shiny app and R toolkit for fitting, understanding, and reporting mixed-effects and multilevel models in R. The current version is an early public beta intended for demonstration, teaching, and structured feedback.

Install

install.packages("remotes")
remotes::install_github("MarcusHarrisUConn/mlmr")

Launch

mlmr::run_mlmr()

Five-Minute Demo

  1. Launch the app.
  2. On Data, keep the built-in example data selected.
  3. Skim Data > Structure and Data > Roles to see the grouping factors and variable roles.
  4. Open Model and review the preset outcome, predictors, centering choices, interactions, and random effects.
  5. Click Fit Example Model from Data > Overview, or open Estimate and click Fit Example Model.
  6. Open Results > Tables and inspect the APA-style tables.
  7. Open Results > Equations and inspect the level-by-level equations, combined equation, and Tau matrix.
  8. Open Results > Diagnostics and review model diagnostics.
  9. Open Report & Code and inspect the reproducible R code and LaTeX export.

Visual Walkthrough

The screenshots below show the built-in example workflow. They are intended for orientation and for beta testers who want to know what a successful fit should look like.

Data Overview

Data overview
Data overview

Model Builder

Model builder
Model builder

Results Dashboard

Results dashboard
Results dashboard

APA Tables

APA tables
APA tables

Equations

Equations
Equations

Report and Code

Report and code
Report and code

What to Evaluate

Please pay attention to:

  • whether the model-building workflow is understandable;
  • whether centering choices are clear;
  • whether random intercepts, random slopes, and Tau matrices are explained well;
  • whether equations look correct and manuscript-ready;
  • whether APA tables are useful;
  • whether the exported R code reproduces what the app did;
  • whether uploaded CSV, TSV/TXT, Excel, SPSS, SAS, or Stata files are easy to use;
  • whether the Model Readiness checklist catches problems before fitting;
  • whether diagnostics and warnings are interpretable.

Feedback

Please submit feedback through GitHub Issues:

https://github.com/MarcusHarrisUConn/mlmr/issues

Use the Demo feedback template when possible.

Helpful feedback includes:

  • the model you tried to fit;
  • whether you used example or uploaded data;
  • screenshots of confusing output;
  • generated R code if reproducibility was the issue;
  • warning or error messages;
  • what you expected to happen instead.

Current Limitations

The app is still in beta. Some advanced structures are scaffolded but not fully polished. Users should independently verify model specification, convergence, diagnostics, and interpretation before using results in production research.