. Author manuscript; available in PMC: 2014 Nov 12.

Published in final edited form as: Eur J Cardiovasc Nurs. 2012 Apr 4;11(4):423–431. doi: 10.1016/j.ejcnurse.2011.03.008

Table 3.

Evaluation of Dietary Energy Density as a Mediating Variable.

Method of Analysis
Test for direct and indirect effects using simple linear regression models (Baron and Kenny)
	Coefficient	Standard error	T-statistic	P-value
b (WC^a, BDI-II^a)	.12	.05	2.33	.02
b(DED, BDI-II^a)	.02	.02	1.46	.15
b(WC^a, DED. BDI-II^a)	.35	.34	1.03	.31
b(WC^a, BDI-II^a. DED)	.11	.05	2.15	.03
Test for indirect effect using the normal distribution (Sobel)
	Value	Standard error	Z-statistic	P-value
Sobel	.01	.01	.73	.46
Test for indirect effect using the nonparametric bootstrap method
	Mean	Standard error	LL 95% CI	UL 95% CI
Bootstrap	.01	.01	−.01	.04

Note:

Square root transformed variable; BDI-II — Beck Depression Inventory II; DED — Dietary Energy Density; WC — Waist circumference; LL — Lower limit; UL — Upper limit; CI — Confidence interval.