yeroon.net/lme4 is a webinterface for the popular CRAN package lme4 by Douglas Bates and Martin Maechler. It should make random effects / multilevel modelling available for applied researchers without the need for any specialized software. lme4 supports both hierarchical and crossed data and many distribution families, making it also a suitable tool for e.g. IRT and repeated measures data.
How to use
Learn how to use the application in 2 minutes by watching one of the demo video's:
- Demo 1: video, datafile ('fm1' and 'fm2' models from the lme4 package documentation)
- Demo 2: video, datafile (Example from Hox (2003), see also multilevel page)
The lme4 web application supports data in SPSS (*.sav), CSV and Tab delimited format. Once uploaded, variables appear in the treepanel and can be added to a model by drag-n-drop, or right-mousclicking. Multiple models can be built and kept in the workspace for easy comparison. When a model was successfully fit, correlation/covariance matrices and random effects plots can be opened with a single click. After the models have been defined, options appear to perform analysis of variance, or request a model report that is generated on the server using LaTeX.
The loading of the required packages and fitting of the models can be computationaly quite intensive. Furthermore the user is free to upload any data, and fit any type of model, which can also cause server problems. At the moment, these problems are avoided by limiting the size of the datafile, and maximum time of an R session. However, a more structural solution will be needed for this to work on a bigger scale.
Try it now!
This application is publicly available at the UCLA demo server. If you used, or tried to use the application, and you experienced any problems or have idea's for improvement, please drop me an email. The tool is still in development and feedback is really appreciated!