Table 3.

Main effects of age, race, and SES, quadratic effect of age, and interaction effects between race/SES and age (listwise deletion N = 1636).

	Unstandardized b (Standard Error), p-value	95% Confidence Interval	Effect Size (r)
Main effects
Age	.02 (.001), p < .001	[.013, .018]	.26
Black (vs. White)	.39 (.077), p < .001	[.238, .542]	.12
Asian (vs. White)	−.29 (.070), p < .001	[−.420, −.152]	.10
SES	−.19 (.025), p < .001	[−.243, −.145]	.18
Female (vs. Male)	.11 (.044), p = .011	[.026, .200]	.06
Quadratic and Interaction effects
Age x Age	−.0005 (.0001), p < .001	[−.0007, −.0002]	.08
Black (vs. White) x Age	.02 (.006), p < .001	[.011, .033]	.09
Asian (vs. White) x Age	−.01 (.005), p = .006	[−.023, −.004]	.06
Black (vs. White) x Age x Age	.0004 (.0006), p = .530	[−.001, .002]	.01
Asian (vs. White) x Age x Age	−.001 (.0005), p = .011	[−.0023, −.0003]	.06
SES x Age	−.006 (.002), p < .001	[−.009, −.003]	.09
SES x Age x Age	−.0002 (.0001), p = .139	[−.0006, .0000]	.03
Single model
Age	.02 (.002), p < .001	[.015, .022]	.22
Black (vs. White)	.46 (.137), p < .001	[.194, .733]	.08
Asian (vs. White)	−.11 (.126), p = .372	[−.360, .135]	.02
SES	−.12 (.046), p = .010	[−.206, −.025]	.06
Female (vs. Male)	.10 (.045), p = .028	[.010, .186]	.05
Age x Age	−.0006 (.0002), p < .001	[−.001, −.0003]	.08
Black (vs. White) x Age	.01 (.006), p = .024	[.002, .027]	.05
Asian (vs. White) x Age	−.01 (.005), p = .072	[−.019, .001]	.04
Black (vs. White) x Age x Age	.000 (.0006), p = .896	[−.001, .001]	.00
Asian (vs. White) x Age x Age	−.001 (.0005), p = .034	[−.002, −.000]	.05
SES x Age	−.004 (.002), p = .018	[−.008, −.001]	.06
SES x Age x Age	−.0003 (.0002), p = .080	[−.001, .000]	.04

Note. Continuous predictors were mean-centered, and binary predictors were effect-coded. SES refers to socioeconomic status. Effect sizes are expressed in semi-partial correlation coefficients (i.e., the square root of delta R²) of the specific terms. Main effects of age, race, SES, and sex were always entered simultaneously as predictors of inflammation in the same block. For the quadratic and interaction effects section, each quadratic effect, linear interaction effect, or quadratic interaction effect was entered as a separate model (but both race codes entered when testing and race interaction effects). For the single model section, all main, quadratic, and interaction effects were entered simultaneously as predictors of inflammation.