G = 2 sum n_i log(n_i / e_i) (Hedderich, Sachs & Reynarowych 2023, eq 7.48),
chi-square on k - 1 - m df for k counts against expected counts or
probabilities with m estimated parameters, or on (r - 1)(c - 1) df for an r x
c table against independence (DescTools::GTest). Zero cells contribute
zero. The book's Hardy-Weinberg example prints 1.92; its counts give 1.19.
Usage
lrgtst(observed, expected = NULL, n_estimated = 0)
Arguments
- observed
Counts (vector) or an r x c matrix.
- expected
Optional expected counts or probabilities.
- n_estimated
Parameters estimated from the data (vector case).
Value
Named list: statistic, df, p_value, expected.
References
Agresti, A. (2013). Categorical Data Analysis, Sec. 3.2.
Examples
lrgtst(c(18, 55, 27), c(0.207, 0.496, 0.297), n_estimated = 1)$statistic
#> [1] 1.189479
