Asymptotic linear-by-linear association test of a table of two ordered classifications (the Mantel test of trend).
Equivalent to coin::lbl_test() in R. As coin's asymptotic test, except for tables with a two-level factor.
Z = sqrt(N - 1) r, r the Pearson correlation of the row and column scores over the N observations: the linear
statistic sum x_ij u_i v_j standardised by its permutation (conditional) mean and variance. The scores are 1..k
by default for every factor (the level index; empty rows and columns keep their place). coin leaves a two-level
factor unordered and codes it by an indicator of its first level, which turns the sign of Z, and refuses 2 x 2
tables; here two-level factors are scored like the others and a 2 x 2 table is tested (Z^2 is then the 'N - 1'
chi-squared statistic). p-values from the normal distribution: "greater" is positive association.
The test result, with statisticName"Z"; see LblTestResult.
Throws
RangeError for negative or non-finite counts, a total of at most 1, scores of the wrong length or not
finite, and tables whose row or column scores do not vary over the observations (Z undefined).
Asymptotic linear-by-linear association test of a table of two ordered classifications (the Mantel test of trend). Equivalent to
coin::lbl_test()in R. As coin's asymptotic test, except for tables with a two-level factor.Z = sqrt(N - 1) r, r the Pearson correlation of the row and column scores over the N observations: the linear statisticsum x_ij u_i v_jstandardised by its permutation (conditional) mean and variance. The scores are 1..k by default for every factor (the level index; empty rows and columns keep their place). coin leaves a two-level factor unordered and codes it by an indicator of its first level, which turns the sign of Z, and refuses 2 x 2 tables; here two-level factors are scored like the others and a 2 x 2 table is tested (Z^2 is then the 'N - 1' chi-squared statistic). p-values from the normal distribution:"greater"is positive association.