Two-phase sampling design

Description

Declares a two-phase (double) sampling design: phase-1 variables are measured on the whole cohort and used to stratify a phase-2 subsample on which the expensive covariate is then measured. The analysis weights phase-2 records by the inverse of their phase-2 selection probability (survey / calibration estimators).

Usage

two_phase(phase1, phase2)

Arguments

phase1 A non-empty character vector naming the phase-1 stratification column(s) measured on the full cohort.
phase2 A single character string naming the 0/1 (or logical) column flagging phase-2 selection.

Value

A matchatr_design object of type “two_phase” carrying the phase-1 strata and phase-2 selection columns.

See Also

matcha()

Other design constructors: case_cohort(), counter_matched(), matched_cc(), nested_cc(), unmatched_cc()

Examples

library("matchatr")

two_phase(phase1 = "stratum", phase2 = "in_phase2")
<matchatr_design>
 Type:       Two-phase
 Phase 1:    stratum
 Phase 2:    in_phase2
 Weights:    design (two-phase)