Welcome to the homepage of simPM!

simPM: SIMulation-based power analysis for Planned Missing designs.

simPM is an R package that automates the Monte Carlo simulation-based search for optimal planned missing (PM) and ‘post hoc’ planned missing (PHPM) designs. This R package is developed and maintained by Yi Feng & Dr. Gregory R. Hancock from the University of Maryland.

The story behind simPM

“Oh %&$#!, they cut my funding!”

We have heard that too often, unfortunately, from researchers who rely on external funding to support their longitudinal studies. There are a variety of reasons that may lead to the adjusted (shrunken) budget announced by the funding agency. Further, what often adds to the challenge is that the researchers are asked to provide a revised study plan to convince that their project can be continued, showing satisfactory inferential validity and statistical power, albeit with the reduced budget.

“What can I do except for firing my consultant and (poor) RAs?”

Planned missing data methods present a very promising solution for such challenging situations, providing many practical and methodological advantages. However, careful planning is critical such that the planned missing design can preserve enough information and statistical power.

simPM was created to help researchers survive the unexpected funding cut in the course of a longitudinal study. It can be used to find an optimal ‘post hoc’ planned missing design that allows the researchers to complete the study at a reduced cost, while maintaining satisfactory level of statistical power for testing the focal parameters.

What does simPM do?

By automizing the simulation-based power analysis for planned missing designs in longitudinal context, simPM can free the researchers from manually configuring the possible PM designs, determining their eligibility, setting up the simulations, and summarizing the results over replications, which can be tedious and time-consuming work especially when there is a large number of plausible PHPM designs to be evaluated.

How to install simPM?

The source code of this R package is made public on the author’s Github page. To install simPM on your local machine, please run the following code

install.packages("devtools")
library(devtools)
install_github("YiFengEDMS/simPM")

Citations

Feng, Y. & Hancock, G. R. (in press). Oh no! They cut my funding! Using ‘post hoc’ planned missing data designs to salvage longitudinal research . Child Development.

Feng, Y., & Hancock, G. R. (2019, April). Oh %&$#!, they cut my funding: Using planned missing data methods to salvage longitudinal research. Paper presented at the annual meeting of the American Educational Research Association (AERA), Division D: Measurement & Research Methodology, Toronto, ON, Canada.

  • Note: The full text of this working paper is available here. The final form post peer review has been accepted for publication and is currently in press in Child Development.

Feng, Y. & Hancock, G. R. (2019). simPM: SIMulation-based power analysis for Planned Missing designs. R package version 0.0.0.9000. https://yifengedms.github.io/simPM/

Questions or Suggestions?

Send an email to yifeng94@umd.edu. We are happy to hear about your thoughts!