Pearson chi-square + Cramer's V on every meaningful 2-way slice of the c-series datasets. Honour to Prof. Doob's chi-square tradition in Canadian corrections research.
Usage
morie_otis_analyze_c_chi2(
datasets,
contingency_value = "NumberIndividuals_RestrictiveConfinement",
out_dir = NULL
)Examples
# \donttest{
otis_c03 <- morie_synth_otis("c03", n = 120L, seed = 1L)
otis_c04 <- morie_synth_otis("c04", n = 120L, seed = 1L)
morie_otis_analyze_c_chi2(list(c03 = otis_c03, c04 = otis_c04))
#> OTIS c-series -- chi^2 + Cramer's V family on demographic contingency tables
#> ============================================================================
#> Contingency value column NumberIndividuals_RestrictiveConfinement
#> Slices tested 7
#>
#> Contingency chi^2 and Cramer's V on c-series slices (value = NumberIndividuals_RestrictiveConfinement):
#> Dataset Slice n chi^2 dof p Cramer's V min cell
#> --------------------------------------------------------------------------------------------------------
#> c03 Race x Gender 4395 334.099 9 1.50e-66 0.2757 78
#> c04 Race x Region_MostRecentPlacement 4395 1637.186 36 2.95e-321 0.3052 0
#> c05 Religion x Region_MostRecentPlacement 0 n/a 0 n/a n/a 0
#> c06 Age_Category x Region_MostRecentPlacement 0 n/a 0 n/a n/a 0
#> c07 Alert_Type x Gender 0 n/a 0 n/a n/a 0
#> c08 Religion x Gender 0 n/a 0 n/a n/a 0
#> c09 Age_Category x Gender 0 n/a 0 n/a n/a 0
#>
#> Cramer's V is the canonical effect-size measure for chi^2 independence on 2-way contingency tables: V~0 -> independence, V->1 -> strong association. p<.05 with V<0.1 ==> statistically detectable but practically tiny.
# }