Exact unconditional test for association in 2x2 tables

Described in Chapter 4 "The 2x2 Table"

Exact_unconditional_test_2x2(n, statistic = "Pearson", gamma = 1e-04)

Arguments

n: the observed counts (a 2x2 matrix)
statistic: 'Pearson' (Suissa-Shuster test default), 'LR' (likelihood ratio), ' unpooled' (unpooled Z), or 'Fisher' (Fisher-Boschloo test)
gamma: parameter for the Berger and Boos procedure (default=0.0001 gamma=0: no adj)

Value

An object of the contingencytables_result class, basically a subclass of base::list(). Use the utils::str() function to see the specific elements returned.

Note

Somewhat crude code with maximization over a simple partition of the nuisance parameter space into 'num_pi_values' equally spaced values (1000, hardcoded). This method could be improved with a better algorithm for the maximization however, it works well for most purposes. plot() the results to get an indication of the precision. A refinement of the maximization can be done with a manual restriction of the parameter space.

Table orientation

In most study designs, rows designate a grouping of the data, for instance, into treatment or exposure groups, while the columns indicate a dichotomous outcome, usually with the first column representing the outcome of interest. This designation of rows and columns may not be relevant in all study designs, please see the introduction to chapter 4 for details.

Examples

Exact_unconditional_test_2x2(tea)
#> The Suissa-Shuster exact unconditional test: P = 0.28916
Exact_unconditional_test_2x2(perondi_2004)
#> The Suissa-Shuster exact unconditional test: P = 0.02821
Exact_unconditional_test_2x2(lampasona_2013)
#> The Suissa-Shuster exact unconditional test: P = 0.05238
Exact_unconditional_test_2x2(ritland_2007)
#> The Suissa-Shuster exact unconditional test: P = 0.04992