pt_designer: Declare a process tracing design from a causal model

Description Usage Arguments Examples

Description

Warning – do not use redesign with design produced from this designer

Usage

1
2
pt_designer(true_model, estimation_model = true_model, ops, query,
  strategy = TRUE, N = 1, sims = 2000)

Arguments

true_model

A model: not available to applied researcher

estimation_model

A model: as supposed by applied researcher

ops

Operations that form part of the query

query

The query

strategy

Data gathering strateg – a vector of TRUES or FALSES indicating which nodes are observed

N

Population for which inferences might be made. Estimand averages over these.

sims

Precision for posterior calculation

Examples

1
2
3
4
5
model <- pt_model()
ops = list(c(0, NA, NA, NA, NA), c(1, NA, NA, NA, NA))
query = function(x) (x[[1]][5] ==1) & (x[[2]][5] == 0)
design <- pt_designer(model, ops = ops, query = query, sims = 500)
draw_estimates(design)

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