Estimators

DefaultDID <: DiffinDiffsEstimator

Default difference-in-differences estimator selected based on the context.

Reg <: DiffinDiffsEstimator

RegressionBasedDID <: DiffinDiffsEstimator

Estimation procedure for regression-based difference-in-differences.

A StatsSpec for this procedure accepts the following arguments:

Key	Type restriction	Default value	Description
`data`			A `Tables.jl`-compatible data table
`tr`	`DynamicTreatment{SharpDesign}`		Treatment specification
`pr`	`TrendOrUnspecifiedPR{Unconditional,Exact}`		Parallel trend assumption
`yterm`	`AbstractTerm`		A term for outcome variable
`treatname`	`Symbol`		Column name for the variable representing treatment time
`subset`	`Union{BitVector,Nothing}`	`nothing`	Rows from `data` to be used for estimation
`weightname`	`Union{Symbol,Nothing}`	`nothing`	Column name of the sample weight variable
`vce`	`Vcov.CovarianceEstimator`	`Vcov.CovarianceEstimator`	Variance-covariance estimator
`treatintterms`	`TermSet`	`TermSet()`	Terms interacted with the treatment indicators
`xterms`	`TermSet`	`TermSet()`	Terms for covariates and fixed effects
`contrasts`	`Union{Dict{Symbol,Any},Nothing}`	`nothing`	Contrast coding to be processed by `StatsModels.jl`
`drop_singletons`	`Bool`	`true`	Drop singleton observations for fixed effects
`nfethreads`	`Int`	`Threads.nthreads()`	Number of threads to be used for solving fixed effects
`fetol`	`Float64`	`1e-8`	Tolerance level for the fixed effect solver
`femaxiter`	`Int`	`10000`	Maximum number of iterations allowed for the fixed effect solver
`cohortinteracted`	`Bool`	`true`	Interact treatment indicators by treatment time
`solvelsweights`	`Bool`	`false`	Solve the cell-level least-square weights with default cell partition
`lswtnames`	Iterable of `Symbol`s	`tuple()`	Column names from `treatcells` defining the cell partition used for solving least-square weights