Estimating the natural indirect effect and the mediation proportion via the product method

Table 1 A comparison of the current work with several previous literature evaluating the empirical performance of the product method in mediation analysis under four data types: Case #1, continuous outcome and continuous mediator; Case #2, continuous outcome and binary mediator; Case #3, binary outcome and continuous mediator; and Case #4, binary outcome and binary mediator

	Natural Indirect Effect							Mediation Proportion
Literatures	Case #1	Case #2	Case #3			Case #4		Case #1	Case #2	Case #3			Case #4
			Approx.	Exact	P.A.	Approx.	Exact			Approx.	Exact	P.A.	Approx.	Exact
Current work	B.V.I	B.V.I	B.V.I	B.V.I		B.V.I	B.V.I	B.V.I	B.V.I	B.V.I	B.V.I		B.V.I	B.V.I
Barfield et al. (2017)	T	T	T			T
Biesanz, Falk, and Savalei (2010)	I.T
Fritz and MacKinnon (2007)	T
Gaynor et al. (2019)					B.I		B.I
MacKinnon, Warsi, and Dwyer (1995)	B.V							B.V
MacKinnon et al. (2002)	T
MacKinnon, Lockwood, and Williams (2004)	I
Rijnhart et al. (2019)			B.V			B.V				B.V			B.V
Samoilenko, Blais, and Lefebvre (2018)						B.V.I	B.V.I

¹Note: The B, V, I, and T denote the bias, variance, confidence interval, and hypothesis testing, respectively. If one of those appears in one cell, it indicates that this operating characteristic has been covered in this literature. In Cases #3 and #4, Approx., Exact, and P.A. denote the approximate expression, exact expression, and the probit approximation expression, respectively (See Table 2 for their specific formulas)

ISSN: 1471-2288