biqq_E_Var: Function to identify possible case selection strategies and...

In macartan/biqq: Package for BIQQ analysis as described in Jacobs, Humphreys (2016)

Description

Assesses the set of strategies that could be implemented by looking for clues in k cases given existing X,Y data, recorded in a 4-vector, 'XY_base'. The output lists the possible strategies, assesses the probability of different data realizations that coudl result from each strategy, the posterior variance that arises in each case, and the overall expected posterior variance from each strategy under consideration.

Usage

biqq_E_Var(k = 1, XY_base = rep(1, 4), strategies = NULL,
  f = my_loss, name = "Results", biqq_model = biqq, ...)

Arguments

k	A scalar saying how many clues are to be sought in total
XY_base	is the data before considering K data
strategies	optional: if only a subset of strategies are examined strategies can be indicated of the form "0-0-1-1" meaning seek one clue for 10 cases and one for 11 cases, or more generally c("1-1-0-0", "0-0-1-1"). Use this option with caution to be sure that naming of strategies is correct and unique

Value

A list

Examples

biqq_E_Var(k = 1, XY_base = rep(1,4), chains = 2)
biqq_E_Var(k = 1, XY_base = rep(1,4), chains = 2, strategies = c("1-0-0-1"))
biqq_E_Var(k = 2, XY_base = rep(1,4), chains = 2, strategies = c("1-0-0-1"))

macartan/biqq documentation built on May 6, 2019, 6:03 p.m.