Table 2.

Predicting AQ total. Fixed effect estimates β (and standard errors). P-values derived from Wald tests. M1-M3 denote models adjusting for increasing numbers of covariates. Pubertal stage and timing models split to avoid multicollinearity.

		Models with pubertal stage			Models with pubertal timing
		M1 Baseline model	M2 M1 + maternal age	M3 M2 + time of amniocentesis	M1 Baseline model	M2 M1 + maternal age	M3 M2 + time of amniocentesis
M1	pT	-0.97(1.76)	-0.33(1.95)	-1.87(4.31)	-1.04(1.74)	-0.43(1.92)	-1.98(4.20)
		p = 0.59	p = 0.87	p = 0.67	p = 0.56	p = 0.83	p = 0.64
	Sex	7.48(3.19)	5.22(3.46)	5.64(8.47)	6.11(3.01)	4.33(3.24)	6.38(7.84)
		p = 0.02	p = 0.14	p = 0.51	p = 0.05	p = 0.19	p = 0.42
	Rater	-3.61(0.83)	-3.53(0.91)	-3.09(1.25)	-3.67(0.83)	-3.54(0.92)	-3.13(1.26)
		p < 0.001	p < 0.001	p = .02	p < 0.001	p < 0.001	p = 0.02
	pT x Sex	1.72(3.67)	2.15(3.94)	9.50(8.48)	-0.60(3.49)	-1.00(3.71)	3.47(8.43)
		p = 0.64	p = 0.59	p = 0.27	p = 0.87	p = 0.79	p = 0.69
	Pubertal Stage	2.59(1.79)	1.56(2.00)	-1.67(2.62)	–	–	–
		p = 0.15	p = 0.44	p = 0.53
	pT x Pubertal Stage	3.94(2.42)	5.65(2.61)	10.19(3.31)	–	–	–
		p = 0.11	p = 0.04	p = 0.003
	Pubertal Timing	–	–	–	2.28(2.14)	2.06(2.21)	-1.05(2.63)
					p = 0.29	p = 0.36	p = 0.69
	pT x Pubertal Timing	–	–	–	5.98(2.58)	7.28(2.74)	11.15(3.27)
					p = 0.02	p = 0.008	p = 0.001
M2	Maternal Age at Birth	–	-0.01(0.16)	-0.12(0.20)	–	0.03(0.16)	-0.05(0.20)
			p = 0.96	p = 0.56		p = 0.83	p = 0.79
M3	Time of Amniocentesis (gestational weeks)	–	–	-2.07(0.61)	–	–	-2.02(0.59)
				p = 0.001			p = 0.001
	# Observations	170	139	94	170	139	94
	# Participants	95	73	49	95	73	49
	Log Likelihood	-551.02	-446.96	-297.37	-549.91	-445.71	-296.48
	AIC	1120.04	913.91	616.74	1117.83	911.42	614.95
	BIC	1148.26	943.26	644.72	1146.05	940.77	642.93

pT = Prenatal testosterone; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.

Bonferroni-adjusted p-threshold = .0167