Pearson and adjusted residuals of a contingency table

(n_ij - e_ij) / sqrt(e_ij) and the adjusted residual (n_ij - e_ij) / sqrt(e_ij (1 - n_i. / n) (1 - n_.j / n)) under independence (Hedderich, Sachs & Reynarowych 2023, eqs 7.357, 7.359; chisq.test residuals and stdres).

Usage

ctresid(table)

Arguments

table

r x c matrix of counts.

Value

Named list: expected, pearson, adjusted (matrices), statistic, df, p_value.

References

Haberman, S. J. (1973). Biometrics 29, 205-220.

Examples

ctresid(matrix(c(14, 22, 32, 18, 16, 8, 8, 2, 0), 3, byrow = TRUE))$adjusted
#>           [,1]       [,2]      [,3]
#> [1,] -3.386826 -0.2605251  3.647351
#> [2,]  1.623996  0.8119979 -2.435994
#> [3,]  3.269696 -0.9341987 -2.335497