PValue: Calculate a simple bounded p-value for the treatment effect...

In hollina/scul: Synthetic Control Using Lasso (SCUL)

Calculate a simple bounded p-value for the treatment effect reporting the larger end of the bound

Description

Determines the rank of the mean absolute value of the standardized treatment effect of the target product relative to the placebo distribution. There are many alternative ways to use the placebo distribution to calculate a mean value (e.g., mean p-value or another test statistic). This is a simple example of one such way.

Usage

PValue(
  x.PlaceboPool.full = SCUL.inference$y.placebo.StandardizedDifference.Full,
  x.PlaceboPool.CohensD = SCUL.inference$y.placebo.CohensD,
  TreatmentBeginsAt = SCUL.input$TreatmentBeginsAt,
  OutputFilePath = SCUL.input$OutputFilePath,
  CohensD = SCUL.input$CohensDThreshold,
  y.actual = SCUL.output$y.actual,
  y.scul = SCUL.output$y.scul,
  StartTime = SCUL.input$TreatmentBeginsAt,
  EndTime = nrow(SCUL.output$y.scul)
)

Arguments

x.PlaceboPool.full	A (T by L), where L<=J) data frame containing all products that are included in the placebo distribution Default is SCUL.inference$y.placebo.StandardizedDifference.Full
x.PlaceboPool.CohensD	A (1 by L) data frame containing all pre-period Cohen's D fit statistic for each placebo unit. Default is SCUL.inference$y.placebo.CohensD,
TreatmentBeginsAt	An integer indicating which row begins treatment. Default is SCUL.output$TreatmentBeginsAt.
OutputFilePath	Output file path. Default is SCUL.input$OutputFilePath.
CohensD	A real number greater than 0, indicating the Cohen's D threshold at which fit is determined to be "poor". The difference is in standard deviation units. Default is SCUL.input$CohensDThreshold.
y.actual	The actual (target) data. Default is SCUL.output$y.actual.
y.scul	Synthetic data created by SCUL procedure. Default is SCUL.output$y.scul.
StartTime	The begining time period for which the average pseduo treatment effect is calculated. T
EndTime	The end time period for which the average pseduo treatment effect is calculated. T

Value

list The bounds for the rank based p-value based upon rank of mean absolute value of post-treatment standardized effect against null distribution.

hollina/scul documentation built on Oct. 15, 2023, 5:43 p.m.