. 2015 Dec;62:129–137. doi: 10.1016/j.psyneuen.2015.08.005

Table 4.

Mediation Analysis.

			β		%	Mediation model
X	Y	M	c	c’	Attenuation	F (df)	Lower CI	Upper CI
START^a	g_fR	gMD	−.34^d	−.29^d	14.71	5.53 (2,67)	−.2573	−.0071
START^a	g_fR	WMH^b	−.28^c	−.28^c	0.00	3.20 (2,74)	−.0748	.1030
START^a	SpeedR	gMD	−.42^e	−.36^e	14.29	9.70 (2,67)	−.3147	−.0254
START^a	SpeedR	WMH^b	−.35^d	−.32^d	8.57	6.53 (2,74)	−.2665	−.0036
START^a	MemoryR	gMD	−.42^e	−.36^d	14.29	8.40 (2,67)	−.3178	−.0047
START^a	MemoryR	WMH^b	−.35^d	−.31^d	11.43	6.67 (2,74)	−.2648	−.0005
END^a	g_fR	WMH^b	−.36^d	−.35^d	2.78	5.60 (2,77)	−.0890	.0562
END^a	SpeedR	WMH^b	−.35^d	−.31^d	11.43	7.08 (2,77)	−.2203	−.0074
END^a	MemoryR	WMH^b	−.36^d	−.33^d	8.33	7.32 (2,76)	−.2286	−.0013

X: independent variable, Y: outcome variable, M: mediator, c: path from X to Y (listwise, may differ slightly from pairwise associations reported in Table 2), c’: path from X to Y accounting for M. Bold type face indicates significant mediation effect (confidence intervals do not include 0; Preacher and Hayes, 2008). START: cortisol levels at the start of cognitive testing, END: cortisol levels at the end of cognitive testing, g_f: general cognitive ability, speed: speed of processing, memory: memory ability, R: variable is a residual, corrected for age 11 cognitive ability. All tests of mediation are one-tailed and bias-corrected.

Log transformed.

Square root transformed.

p < .05.

p < .01.

p < .001.