Skip to contents

Computes standardized mean differences (SMD) between two groups for each covariate in X. Returns one row per covariate with group means, standard deviations, raw and absolute SMD, and conventional balance thresholds.

Usage

smd_balance_table(group, X, w = NULL)

Arguments

group

Integer (or coercible) binary group indicator. Must have exactly two distinct non-missing values.

X

Data frame of baseline covariate columns.

w

Optional numeric case-weight vector. When supplied, weighted group means (wmean_0, wmean_1) and weighted SMD (wsmd) are added. Weighted SMD uses the unweighted pooled SD as the standardizer (Rubin simplification).

Value

A data.frame of class c("smd_balance_table", "data.frame") with one row per covariate and columns: attribute, n_group_0, n_group_1, mean_0, sd_0, mean_1, sd_1, smd, abs_smd, balanced_020 (abs_smd < 0.20), balanced_010 (abs_smd < 0.10), wmean_0, wmean_1, wsmd, wabs_smd, wbalanced_020, wbalanced_010 (weighted variants; NA when w = NULL).

Details

SMD is a conventional companion diagnostic, not the oda balance objective. The primary oda balance assessment uses oda_balance_table. This function is intended for comparison with non-ODA balance reports.

No p-values are computed. SMD is a descriptive statistic. For a variable with zero within-group variance in both groups, smd is NA.

Examples

group <- c(rep(0L, 30), rep(1L, 30))
X     <- data.frame(age = c(rep(45, 30), rep(55, 30)),
                    score = rnorm(60, 50, 10))
smd_balance_table(group, X)
#>   attribute n_group_0 n_group_1   mean_0     sd_0   mean_1     sd_1       smd
#> 1       age        30        30 45.00000  0.00000 55.00000 0.000000        NA
#> 2     score        30        30 47.15789 10.40146 49.05927 7.815423 0.2066766
#>     abs_smd balanced_020 balanced_010 wmean_0 wmean_1 wsmd wabs_smd
#> 1        NA           NA           NA      NA      NA   NA       NA
#> 2 0.2066766        FALSE        FALSE      NA      NA   NA       NA
#>   wbalanced_020 wbalanced_010
#> 1            NA            NA
#> 2            NA            NA