equivalence_test: Conduct an equivalence test of the distributional parallel...

In soichiroy/orddid: Difference-in-Differences Design for Ordinal Outcome

Description

equivalence_test() implements an equivalance test to assess the distributional parallel trends assumption using the data from the pre-treatment periods.

Usage

equivalence_test(object, alpha = 0.05, threshold = NULL)

Arguments

object	A fitted object from ord_did.
alpha	The level of a test. This value should take between 0 and 1. Default is 0.05.
threshold	An equivalance threshold. If left as NULL, the data-driven threshold, estimated in calc_threshold, is used for the test.

Value

equivalence_test() returns a list of class ‘orddid.test’, which contains the following items:

tv	A vector of point-wise deviation between q1(v) and q0(v).
tv_var	A vector of variances for each t(v).
tmax	A maximum deviation of q1(v) and q0(v).
v_range	A range of v used to evaluate q1 and q0.
Uv	Point wise 1 - alpha level upper confidence interval.
Lv	Point wise 1 - alpha level lower confidence interval.
Umax	Maximum upper bound.
Lmin	Minimum lower bound.
Upvalue	Point-wise pvalues associated with the upper bounds.
Lpvalue	Point-wise pvalues associated with the lower bounds.
pvalue	P-value of the test.
zscore	Z-score of the test.
reject	Decision of the equivalance test. If TRUE, the test rejects the null of non-equivalance.

soichiroy/orddid documentation built on Oct. 3, 2020, 5:10 a.m.