Skip to content

Simulation and power analysis of panel/hierarchical data that allows for independently generating effects by cross-section (between-subject) and case (within-subject), including auto-correlation in either dimension.

saudiwin/panelsim

Repository files navigation

README

This package enables the simulation of panel data with independent effects in the cross-section and cases (over time). The code is based on that used to simulate panel data in "Identification and Interpretation of Within-unit and Cross-sectional Variation of Panel Data Models" by Jonathan Kropko and Robert Kubinec. The original code used to reproduce the paper's results can be found in the simulation_code.R file (and associated linked files) in the top-level directory of this repo.

Independently generating these two dimensions of variance in a panel data set permits much more robust consideration of how one-way and two-way fixed effects and random effects models handle TSCS/panel data. The package also enables the simulation of autocorrelation in either the over-time dimension and the correlation (spatial) dimension. This package is still under development.

About

Simulation and power analysis of panel/hierarchical data that allows for independently generating effects by cross-section (between-subject) and case (within-subject), including auto-correlation in either dimension.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages