Table 3.

Association between the number of CAG repeats in the AR gene and the mean of the left and right hand 2D:4D ratios

	ALSPAC	QIMR	Meta-analysis
All individuals
	N = 5328	N = 498	N = 5826
Mean	0.014 (0.016), P = 0.19	0.040 (0.052), P = 0.22	0.016 (0.015), P = 0.14
High	0.012 (0.015), P = 0.21	0.014 (0.046), P = 0.38	0.012 (0.014), P = 0.20
Low	0.013 (0.016), P = 0.21	0.041 (0.057), P = 0.24	0.015 (0.015), P = 0.16
Male
	N = 2615	N = 231	N = 2846
Mean	−0.002 (0.020), P = 0.54	−0.099 (0.083), P = 0.88	−0.007 (0.019), P = 0.65
Female
	N = 2713	N = 287	N = 3000
Mean	0.046 (0.028), P = 0.05	0.125 (0.072), P = 0.04	0.056 (0.026), P = 0.02
High	0.030 (0.025), P = 0.12	0.068 (0.058), P = 0.12	0.036 (0.023), P = 0.06
Low	0.040 (0.026), P = 0.06	0.112 (0.082), P = 0.09	0.047 (0.025), P = 0.03

Displayed are beta (SE) and P-values in each of the cohorts and the combined estimates from the fixed effects, inverse-variance weighted meta-analysis. ‘Mean’ refers to analyses involving the average CAG repeat length, ‘High’ refers to analyses involving the highest length repeat and ‘Low’ refers to analyses involving the lower length repeat. One-tailed P-values testing for a positive association between CAG repeat length and 2D:4D ratio are presented.