Fork me on GitHub

multilevelPSA

Visualizing Missing Data

Missingness is a common issue when analyzing large datasets. Moreover, logistic regression which is a common method to estimate propensity scores requires there to be no missing data. The multilevelPSA package includes a function missing.plot to provide a visualization to help understand the nature and extent of missingness within and across the levels of interest. The following example requires the pisa package in addition to the multilevelPSA package since the North America subset of PISA included in the multilevelPSA package does not have any missing data (missing data has been imputed using the mice package).

require(multilevelPSA)
require(pisa)
data(pisa.student)
data(pisa.psa.cols)
student = pisa.student[, c("CNT", pisa.psa.cols)]
student$CNT = as.character(student$CNT)
missing.plot(student, student$CNT)

plot of chunk missingplot


= Github page; = RSS XML Feed; = External website; = Portable Document File (PDF)
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License. Creative Commons License
Formulas powered by MathJax