R2MLwiN-package: Running MLwiN from within R

Description

R2MLwiN is an R command interface to the MLwiN multilevel modelling software package, allowing users to fit multilevel models using MLwiN (and also WinBUGS / OpenBUGS) from within the R environment.

Arguments

New features in version 0.8-3

Support for model comparison tables via texreg-package and memisc-package have been added to R2MLwiN version 0.8-3. For an example of using texreg-package see e.g. demo(MCMCGuide04).

Important differences between version 0.8-0 and earlier versions

A number of wide-ranging changes, including a new model-fitting syntax more in keeping with that conventionally used in R, were introduced in R2MLwiN version 0.8-0.

The demos, which replicate both the User's Guide to MLwiN (Rasbash et al, 2012) and MCMC Estimation in MLwiN (Browne, 2012) manuals, provide practical demonstrations of many of these changes. See demo(package = "R2MLwiN") for a list of demo titles; to run one type e.g. demo(UserGuide03) or view a demo's script via file.show(system.file("demo", "UserGuide03", package = "R2MLwiN")).

The Formula is now specified via a formula object (with some differences in specification: see runMLwiN). So, for example, previously a 2-level model random intercept model would be specified by e.g. normexam ~ (0|cons + standlrt) + (2|cons) + (1|cons), levID = c('school', 'student'), with normexam the response variable, cons a constant of ones forming the intercept, which is allowed to vary at level 1 (student) and level 2 (school), and standlrt included as a predictor in the fixed part of the model. Whilst back-compatibility is preserved (i.e. this specification will currently still work) the same model can now be more parsimoniously specified via normexam ~ 1 + standlrt + (1 | school) + (1 | student). As well examples in the demos, see runMLwiN and Formula.translate for further info.
As a means of specifying cross-classified, multiple membership or CAR models, xclass is now deprecated. Instead, cross-classified models are specified via xc = TRUE, multiple membership models are specified via mm, and CAR models are specified via car, in the list of estoptions. mm and car can be a list of variable names, a list of vectors, or a matrix. See runMLwiN for further details.
Multiple membership/CAR information can now be specified using matrices. df2matrix and matrix2df functions have also been added to convert such information between data.frame and matrix formats.
As a means of specifying common (i.e. the same for each category) or separate (i.e. one for each category) coefficients in ordered multinomial and multivariate response models, c (for common) and s (for separate) have been replaced by the employment of square brackets after the relevant variable to indicate a common coefficient is to be fitted (a separate coefficient will be fitted otherwise). Within these square brackets needs to be placed a numeric identifier indicating the responses for which a common coefficient is to be added (see runMLwiN for further details). E.g. what would have been previously specified, within the Formula object, as ... (0s|cons + ravens) + (0c|fluent{1, 0}) ... would now be specified by ... 1 + ravens + fluent[1] ....
When added as a predictor, a variable encoded as a factor is automatically handled as categorical, replacing the previous use of square brackets after the variable name.
A number of generic s4 methods have been added to improve compatibility with statistical functions which use them (e.g. see stats4-package). So, for example, the addition of a logLik means a likelihood ratio test can now be conducted on two mlwinfitIGLS-class objects using the lrtest function, e.g. lrtest(mymodel1, mymodel2). See help(package = "R2MLwiN") for the index listing these various methods.

Maintainer

Zhengzheng Zhang zhengzheng236@gmail.com

Author

Zhang, Z., Charlton, C.M.J., Parker, R.M.A., Leckie, G., and Browne, W.J. (2016) Centre for Multilevel Modelling, University of Bristol.

References

R2MLwiN

Zhang, Z., Parker, R.M.A., Charlton, C.M.J., Leckie, G. and Browne, W.J. (2016) R2MLwiN: A Package to Run MLwiN from within R. Journal of Statistical Software, 72(10), 1-43. doi:10.18637/jss.v072.i10

MLwiN software and manuals

Browne, W.J. (2012) MCMC Estimation in MLwiN, v2.26. Centre for Multilevel Modelling, University of Bristol.

Rasbash, J., Charlton, C., Browne, W.J., Healy, M. and Cameron, B. (2009) MLwiN Version 2.1. Centre for Multilevel Modelling, University of Bristol.

Rasbash, J., Charlton, C. and Pillinger, R. (2012) Manual Supplement to MLwiN v2.26. Centre for Multilevel Modelling, University of Bristol.

Rasbash, J., Steele, F., Browne, W.J. and Goldstein, H. (2012) A User's Guide to MLwiN Version 2.26. Centre for Multilevel Modelling, University of Bristol.

OpenBUGS

Thomas, A., O'Hara, B., Ligges, U. and Sturtz, S. (2006) Making BUGS Open. R News, 6, 12:17.

WinBUGS

Spiegelhalter, D.J., Thomas, A. and Best, N.G. (1999) WinBUGS Version 1.2 User Manual. MRC Biostatistics Unit.

Examples

Run this code

if (FALSE) {
library(R2MLwiN)
# NOTE: if MLwiN not saved in location R2MLwiN defaults to, specify path via:
# options(MLwiN_path = 'path/to/MLwiN vX.XX/')
# If using R2MLwiN via WINE, the path may look like this:
# options(MLwiN_path = '/home/USERNAME/.wine/drive_c/Program Files (x86)/MLwiN vX.XX/')

data(tutorial, package = "R2MLwiN")

(mymodel <- runMLwiN(normexam ~ 1 + standlrt + (1 + standlrt | school) + (1 | student),
                     estoptions = list(EstM = 1), data = tutorial))

## The R2MLwiN package includes scripts to replicate all the analyses in
## Rasbash et al (2012) A User's Guide to MLwiN Version 2.26 and
## Browne, W.J. (2012) MCMC estimation in MLwiN Version 2.26.
## The MLwiN manuals are available online, see:
## http://www.bristol.ac.uk/cmm/software/mlwin/download/manuals.html

## For a list of demo titles
demo(package = 'R2MLwiN')

## Take MCMCGuide03 as an example
## To view file
file.show(system.file('demo', 'MCMCGuide03.R', package='R2MLwiN'))

## To run the demo
demo(MCMCGuide03)
}

Run the code above in your browser using DataLab