Create Treatment and Control Groups — create_treatment_control

Creates matched treatment and control groups for impact evaluation based on event identification results.

Usage

create_treatment_control_groups(
  event_data,
  control_conditions = NULL,
  matching_variables = NULL,
  exact_match_variables = NULL,
  control_ratio = 1,
  caliper = 0.1,
  exclude_future_treated = TRUE,
  replacement = FALSE
)

Arguments

event_data: Data.table from identify_treatment_events()
control_conditions: Optional conditions for control group eligibility. Similar format to treatment_conditions in identify_treatment_events()
matching_variables: Character vector of variables to use for matching. Default: NULL (no matching, uses random eligible controls). Note: If provided, this serves as a reminder but actual matching should be done using propensity_score_matching() after group creation
exact_match_variables: Character vector of variables requiring exact matches. Default: NULL
control_ratio: Numeric. Ratio of control to treatment units. Default: 1
caliper: Numeric. Maximum distance for propensity score matching. Default: 0.1
exclude_future_treated: Logical. Exclude future-treated units from control group? Default: TRUE
replacement: Logical. Sample control units with replacement? Default: FALSE

Value

A data.table with treatment and control group assignments including:

group_assignment: Factor: "treatment" or "control"
match_id: Matching pair identifier
propensity_score: Estimated propensity score (if matching used)
match_distance: Distance to matched unit
control_weight: Weight for control observations

Examples

if (FALSE) { # \dontrun{
# Create treatment and control groups
groups <- create_treatment_control_groups(
  event_data = identified_events,
  matching_variables = c("age", "sector", "region"),
  exact_match_variables = c("gender"),
  control_ratio = 2
)
} # }