. Author manuscript; available in PMC: 2010 Feb 22.

Published in final edited form as: Struct Equ Modeling. 2008;15(1):23–51. doi: 10.1080/10705510701758166

TABLE 1.

Average Type I Error and Power, Mediated Effects

		Sample Size
Effect Type	Test	50	100	200
Null two-path, two zero paths	Standard z	.00020^a	.00030^a	.00000^a
	M	.00400^a	.00350^a	.00250^a
	Percentile bootstrap	.00440^a	.00250^a	.00220^a
	Bias-corrected bootstrap	.01360^a	.01130^a	.01160^a
Null two-path, one zero path	Standard z	.02050^a	.02540	.03140
	M	.05020	.04910	.04940
	Percentile bootstrap	.04200	.04620	.05070
	Bias-corrected bootstrap	.06900	.06760	.07130
Null two-path, overall	Standard z	.01410^a	.01750^a	.02150^a
	M	.03570	.03480	.03470
	Percentile bootstrap	.03020	.03250	.03540
	Bias-corrected bootstrap	.05160	.04990	.05260
Null three-path, three zero paths	Standard z	.00000^a	.00000^a	.00000^a
	Empirical M	.00000^a	.00000^a	.00000^a
	Percentile bootstrap	.00150^a	.00000^a	.00000^a
	Bias-corrected bootstrap	.00600^a	.00150^a	.00400^a
Null three-path, two zero paths	Standard z	.00000^a	.00000^a	.00007^a
	Empirical M	.00364^a	.00350^a	.00386^a
	Percentile bootstrap	.00307^a	.00143^a	.00214^a
	Bias-corrected bootstrap	.01736^a	.01650^a	.01343^a
Null three-path, one zero path	Standard z	.00860^a	.02013^a	.02847^a
	Empirical M	.05413	.05393	.05627
	Percentile bootstrap	.03467	.04433	.04587
	Bias-corrected bootstrap	.08093^a	.07820^a	.07860^a
Null three-path, overall	Standard z	.00416^a	.00974^a	.01381^a
	Empirical M	.02784	.02768	.02897
	Percentile bootstrap	.01826^a	.02210^a	.02316^a
	Bias-corrected bootstrap	.04739	.04539	.04435
Nonzero two-path	Standard z	.48360	.66690	.73090
	M	.58050	.71140	.78430
	Percentile bootstrap	.53510	.69740	.76920
	Bias-corrected bootstrap	.57820	.72080	.80460
Nonzero three-path	Standard z	.19480	.38380	.50240
	Emp-M	.35520	.47540	.58720
	Percentile bootstrap	.29080	.45120	.55500
	Bias-corrected bootstrap	.37400	.50020	.61960

Proportion outside Bradley (1978) robustness interval.