NEWS.md

In snsf-data/ERforResearch: Expected Rank for Research Evaluation

ERforResearch 4.0.0

• (!!!) switch from rjags to runjags for get_mcmc_samples().
• all function and documentations were updated due to the change to runjags.
• (!!!) important naming changes to align all package-communications: from application to proposal, voter to assessor and section to panel.
• colored points in plot_er_distribution().

ERforResearch 3.1.1.9000

• added an option in the ER plotting function to allow for not having the PM, and skipping it in the plot

ERforResearch 3.1.1

• added an option in the ER plotting function to allow for not having the PM, and skipping it in the plot
• documentation updates

ERforResearch 3.1.0

• documentation updates
• added parameter rhat_threshold in get_mcmc_samples() to allow the user to change the threshold used to decide whether the chains converged.
• added option maximal_testing in mcmc function
• add a default continuous model with a nine point scale
• added na.rm = TRUE whenever max(rhat) calculated

ERforResearch 3.0.0.9000

• nothing to add

ERforResearch 3.0.0

Include Reviewer comments:

• changed name of get_inits_overdispersed_two_chains() to get_inits_overdispersed_four_chains() and implemented the new default of four chains with different RNGs.
• get_mcmc_samples() function defaulted to four chains with four different samplers and fixed (random but dependent on seed) starting values.
• updated the test of get_mcmc_samples() function:
• changed Rhat threshold to 1.01 instead of 1.1
• update the updating loop (the multiplier of n.adapt, burnin and iterations)
• the get_max_rhat() function was deleted
• added option to compute effective sample size and MCMC error in get_mcmc_sample()
• the default variables in get_mcmc_sample() were changed to being able to sample from all the models discussed in the paper (Heyard et al (2021)), with the default model
• rankability computation etc was deleted an documentation of get_er_from_jags() updated

ERforResearch 2.0.0

• add max_iter option in get_er_form_jags()
• new option to not add the rank based on pm in the get_er_from_jags() function
• fix get_mcmc_samples()
• added maximum number of iterations to get_mcmc_samples()
• update get_sankey_plot_br()
• added get_sankey_plot_br() function
• get_mcmc_samples() now tests whether chains converge, if not, the adaptation, burnin and iteration phases are increases.

ERforResearch 1.0.0

• get_mcmc_samples() now tests whether chains converge, if not, the adaptation, burnin and iteration phases are increases.
• all functions were adapted accordingly
• possibility to add overdispersed initial values in get_mcmc_samples()
• functions for convergence diagnostic: get_inits_overdispersed_two_chains() and get_max_rhat()
• default model can now also have an ordinal outcome variable and a heterogenous residual structure

ERforResearch 0.1.0.9009

• allow model to account for heterogenous residuals

ERforResearch 0.1.0.9008

• allow model to account for heterogenous residuals
• extended get_er_from_jags() to be used with ordinal outcomes
• added ordinal extension in get_mcmc_samples() function

ERforResearch 0.1.0.9007

• added an option to choose between inner and outer CrI for funding decision
• update get_right_data_format()

snsf-data/ERforResearch documentation built on May 12, 2024, 9:25 p.m.