NEWS

sate 2026-03-05

Minor Changes:

Improved reliability of compare.jury.stats by preventing unintended rounding effects in calculations.
Added optional simulation mode (simulate) to key stats and graph functions, with sensible defaults and user override support.
Strengthened input checks and error messages (missing arguments, invalid ranges/types, and required dataset fields).
Improved output clarity (including P(G) labeling) while keeping backward compatibility.
Expanded test coverage for core probability/statistical behavior and edge cases.
Updated function documentation for clearer usage and argument expectations.

Major Changes:

Updated get_pG_by_k function. Implemented analytic method for obtaining guilty/death verdict probabilities given jury size. The analytic solution is based on matrix algebra, not simulation or use of while loop. Improved function used by analysis functions for faster and more precise results.
Added transition.matrix function. It returns a transition probability matrix for jury deliberating choice between NG and G verdicts.
Added transition.matrix.ordered function. It generates a transition probability matrix for jury deliberating choice among three or more ordered verdict options.
Added prob.ordered.verdicts function. It obtain the final (absorbing) state probabilities for jury deliberating among three or more ordered verdict options. This function calls the transition.matrix.ordered function (which may also be called as stand alone function).
Added prob_ord_from_pool function. Given verdict preferences in population from which jurors are selected, it returns deliberated verdict probabilities for trial jury deliberating among three or more ordered verdict options. This function calls the transition.matrix.ordered and prob.ordered.verdicts functions (which may also be called as stand alone functions).

Minor Changes:

Updated weights_for_population function to handle survey data without missing respondent data.

Minor Changes:

Added compare.juror.stats function for comparing and evaluating difference between juror verdict preferences in actual and hypothetical trial conditions.
Added three functions to assist in calculating sampling weights to represent key demographic characteristics of a state or USA. Apply the encode.cloud.respondents.variables function to respondent information reported by Cloud Research. Use target.population.demographics to obtain demographic statistics for target population. Use weights_for_population to calculate sampling weights that balance respondent demographics relative to target population.
Added basic.plot.grid function to create basic plot element that can then be customized.
Improved reporting of estimated values in tables.
Improved documentation of functions.

Minor Changes:

Update the P(G|k) values supplied by the helper function get.model.values. Updated values based on matrix algebra rather than computer simulation. Differences extremely minor, but additional jury sizes added.
Introduced get_pG_by_k function. This function allows user to calculate P(G|k) values for juries with 3 or more members.
Improved formatting of results table returned by as.jury.stats, compare.jury.stats,
sim.as.jury.stats, and sim.compare.jury.stats functions.
Added margin of error (MOE) to results tables reported by as.jury.stats, compare.jury.stats, sim.as.jury.stats, and sim.compare.jury.stats functions.
Fixed minor bug with randomly draw proportions falling outside [0, 1] interval.
Improved documentation of functions.

Minor Changes:

Added observed.deliberation dataset, a compilation of observed deliberation outcomes that is useful to estimate relationship between initial jury poll and trial verdicts.

Major Changes:

Major revision of all existing functions.
Graphing functions introduced, allowing users to plot the estimated probability of guilty verdict in single trial condition, or plot the effect of a trial error as difference between actual and hypothetical trial conditions.
Deliberation process modeled using a deliberate function which reduces standard errors of estimates. Users have option to estimate values using empirical model and standard errors generated through computational methods.
Account for selection of small trial jury from large jury pool.
Introduced optional arguments to account for varying jury size and peremptory strikes.

Major Changes: