Skip to contents

Coerce a wide data frame to a schedule of potential outcomes

as_schedule.Rd

Attaches the "schedule" class and metadata attributes to a data frame that is already structured as a schedule of potential outcomes — i.e., one column per treatment level containing the potential outcomes for each unit. To convert long-format experimental data instead, see experiment_to_schedule.

Usage

as_schedule(data, treatment, potential_outcomes)

Arguments

data

A data frame structured as a schedule of potential outcomes, with one column per treatment level.

treatment

The unquoted name of the column in data that contains the treatment assignments, which must be a factor.

potential_outcomes

A tidyselect expression selecting the columns that contain the potential outcomes, one per treatment level (e.g., c(Ycontrol, Ytreatment) or starts_with("Y")).

Value

A data frame with class "schedule" and metadata attributes: treatment, treatment_levels, treatment_counts, and potential_outcome_cols.

See also

experiment_to_schedule

Examples

sched <- data.frame(
  id = 1:6,
  treatment = factor(c("control", "control", "treatment",
                       "treatment", "control", "treatment")),
  Ycontrol   = c(10, 12, 15, 18, 11, 17),
  Ytreatment = c(11, 13, 16, 19, 10, 17)
)
as_schedule(sched, treatment, c(Ycontrol, Ytreatment))
#>   id treatment Ycontrol Ytreatment
#> 1  1   control       10         11
#> 2  2   control       12         13
#> 3  3 treatment       15         16
#> 4  4 treatment       18         19
#> 5  5   control       11         10
#> 6  6 treatment       17         17