prob.ordered.verdicts: Absorption probabilities for ordered-category jury models

In sate: Scientific Analysis of Trial Errors (SATE)

Absorption probabilities for ordered-category jury models

Description

Compute the probability that an ordered-category Markov chain on jury vote counts will eventually absorb at each unanimous verdict, starting from any transient (non-unanimous) composition. Internally, this constructs the transition matrix with 'transition.matrix.ordered()' **using its defaults**, i.e., equal cut lines (no lambda weighting).

Usage

prob.ordered.verdicts(jury_n, verdict_options, digits = NULL, collab = TRUE)

Arguments

jury_n	Integer. Number of jurors.
verdict_options	Character vector of ordered verdict labels (e.g., 'c("NG","Lesser","G")'). Order matters: left = most lenient.
digits	Integer. Number of digits to round in the returned matrix. Default '3'.
collab	Logical. If 'TRUE' (default), attach human-friendly row/column labels: rows are verdict names; columns are starting states (transients first, then unanimities).

Details

Let P be the transition matrix returned by 'transition.matrix.ordered(jury_n, verdict_options)', with meta attributes providing: - 'T': number of transient states, - 'K': number of absorbing states (equal to 'length(verdict_options)'), - 'states': list of length 'T + K' of count vectors (per state), - 'n': the jury size, - 'verdict_options': the verdict labels.

Value

A numeric K by T+K matrix of absorption probabilities. Rows index absorbing verdicts in 'verdict_options'. Columns index starting states: first all transient compositions, then each unanimity composition (one per verdict). If 'collab = TRUE', row/column names are added.

See Also

[transition.matrix.ordered]

Examples

library(sate)

# Three-verdict ordered model with a 12-person jury:
prob.ordered.verdicts(12, c("NG", "M2", "M1"))

# Probability of ultimately unanimous "Lesser" starting from A=6, B=4, C=2:
prob.ordered.verdicts(12, c("A","B","C"), digits = 3)

sate documentation built on March 6, 2026, 1:07 a.m.