posc: Probability of Outcome Superiority Curves
Computes, plots and compares probability of outcome superiority
curves (POSCs). For two people who differ by a given amount on a
continuous predictor, a POSC gives the probability that the person with
the higher predictor score also has the higher outcome, a probabilistic
and decision-oriented alternative to reporting a correlation coefficient.
Curves can be computed from a correlation under bivariate normality, from
a correlation matrix for an optimally weighted composite of predictors,
from a fitted linear model, or estimated directly from raw data with a
flexible spline model and bootstrap confidence bands.
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=posc
to link to this page.