10  Further information

Packages of interest

The following non exhaustive list provides a few examples of commands and packages that tackle common types of analysis which might be relevant to users of large scale social surveys:

  • Further regression analysis: the glm() command can be used for fitting a large number of regression models including Poisson and multinomial logistic regression. The packages lme4 and nlme include functions to fit respectively linear and non linear multilevel models, also known as mixed models.

  • For users interested in latent variable modelling the factanal() command from the stats package enables to conducts factor analysis. Other resources are available in the poLCA (Latent Class Analysis), ltm (Latent Trait modelling), sem (Structural Equation Modelling) packages. The Lavaan package also provides a wide range of functions for structural equation modelling including models with categorical outcomes.

  • For those conducting longitudinal and time series analysis the statsand the ‘tseries’ packages contains useful functions. The packages survival and eha deal with event history and survival analysis, whereas plm is designed for panel data and fixed and random effects models.

  • Using R for creating maps is now common among social scientists and geographers with packages such as rmaps, sp, rgdal, rgeos and ggmaps

Additional online resources

There are hundreds of web sites dedicated to R, in addition to CRAN and the main R help list, R-Help with its searchable archives. A few common ones are listed below: