top of page

R Package ‘MicroMacroMultilevel

Lu, J.G., Page-Gould, E., & Xu, N.R. (2017)

     

To date, most multilevel methodologies can only unbiasedly model macro-micro situations, wherein group-level explanatory variables (e.g., city temperature) are used to predict an individual-level outcome variable (e.g., citizen personality). In contrast, this R package enables researchers to unbiasedly model micro-macro situations, wherein individual-level explanatory variables (e.g., team member personality) are used to predict a group-level outcome variable (e.g., team performance). For a detailed example, see Akinola, Page-Gould, Mehta, & Lu (PNAS, 2016).

In most micro-macro multilevel modeling, it is statistically biased to directly regress the group-level outcome variable on the unadjusted group means of individual-level explanatory variables (Croon & van Veldhoven, 2007). 

This R package is useful because it enables unbiased micro-macro multilevel modeling by producing the best linear unbiased predictors (BLUP) of the group means (i.e., the adjusted group means). 

Reference manualhttps://cran.r-project.org/web/packages/MicroMacroMultilevel/MicroMacroMultilevel.pdf

bottom of page