OPSR (development version)

OPSR 1.0.1

opsr_simulate() now takes seed argument to enable reproducibility.
plot.opsr() and opsr_te() define type arg list to behave similarly to predict.opsr().

print.summary.opsr() now shows GOF indicators.
Added timeuse_data from the TimeUse+ study.
extract.opsr(): R2 for each regime in texreg functions (and don’t repeat GOFs if beside = TRUE).
Nicer summary printing and setting fixed values in null model to NA.
Added .loglik arg to opsr() to return log-likelihood vector.
Detects and warns on singularity issues.
Resolved memory leak.
Corrected outcome-specific R2 computations.
Added type = "Xb" and type = "correction" in predict.opsr().
Works for only 2 treatment regimes (i.e., the Tobit-5 model).
Smoothing constant (used in continuity correction) can be passed to predict.opsr() via delta arg.
Added opsr_te() for treatment effect computations.
Added plot() method for treatment effect visualization.
Improved predict()’s argument argument checking and error messages as well as the man pages.
Passing the null model as sole argument to anova() will raise an error.
Added vignette corresponding to the JSS submission.