c01 region-cluster variant (year-clustered GEE)

Usage

morie_otis_analyze_c01_ruhela_aggregate_region_cluster(data, out_dir = NULL)

morie_otis_analyze_c01_mrm_aggregate_region_cluster(data, out_dir = NULL)

Arguments

data: c01 data.frame.
out_dir: Optional.

Value

morie_otis_analysis_result.

Examples

# \donttest{
otis_c01 <- morie_synth_otis("c01", n = 120L, seed = 1L)
morie_otis_analyze_c01_ruhela_aggregate_region_cluster(otis_c01)
#> OTIS c01 -- Aggregate RF, year-clustered: Female -> Individuals in RC
#> =====================================================================
#>   Source file                            c01 (year-clustered variant)
#>   Dataset id                             c01-RC
#>   Treatment column                       T_female
#>   Outcome (count)                        NumberIndividuals_RestrictiveConfinement
#>   Year FE                                EndFiscalYear
#>   Covariate FEs                          none
#>   Cluster group (GEE)                    EndFiscalYear
#>   Cells                                  120
#>   Total outcome count                    4579
#>   Overdispersion (Poisson AIC - NB AIC)  1481.8
#> 
#> Aggregate Ruhela formulation -- Poisson + NB IRR + GEE clustered for treatment effect on count outcome:
#>   Family                                     Coefficient        IRR     95% CI          p-value   AIC     Notes
#>   ---------------------------------------------------------------------------------------------------------------------------
#>   Poisson                                    factor(T_female)1  0.898   [0.846, 0.953]  3.71e-04  2600.7  n=120, total_y=4579
#>   NB                                         factor(T_female)1  0.9117  [0.685, 1.214]  5.27e-01  1118.9  n=120, total_y=4579
#>   GEE-Poisson (cluster:EndFiscalYear, Exch)  factor(T_female)1  0.898   [0.737, 1.095]  2.87e-01  --      groups=7, n_obs=120
#> 
#> Year-clustered GEE variant of c01. With only 3 fiscal years (= 3 clusters) the GEE inference is unreliable; shipped for transparency, not as the primary inference.
#> 
#> IRR > 1 ==> treatment increases the count rate; IRR < 1 ==> treatment decreases it. Concordance between Poisson and NB indicates equidispersion; large gap (Poisson AIC > NB AIC) indicates overdispersion -- trust NB. 
# }