. 2021 Feb 1;51(3):237–249. doi: 10.1007/s10519-020-10035-7

Table 2.

The expected covariance matrices in simulations 1–3

	MZ 1	MZ 2
MZ 1	σ²_A + σ²_C* + σ²_F + 2σ_AF + σ²_E	σ²_A + σ²_C* + σ²_F + 2σ_AF
MZ 2	σ²_A + σ²_C* + σ²_F + 2σ_AF	σ²_A + σ²_C* + σ²_F + 2σ_AF + σ²_E
	DZ 1	DZ 2
DZ 1	σ²_A + σ²_C* + σ²_F + 2σ_AF + σ²_E	½σ²_A + σ²_C* + σ²_F + 2σ_AF
DZ 2	½σ²_A + σ²_C* + σ²_F + 2σ_AF	σ²_A + σ²_C* + σ²_F + 2σ_AF + σ²_E
	σ_F² = 2m²(σ²_A + σ²_C* + σ²_F + 2σ_AF+ σ²_E) = 2m²σ²_Ph
	σ_AF = (mσ_A)/(1−σ_F m)

Parameter m is the regression coefficient in regression of parental phenotype on F in twins (shared environmental factor attributable to cultural transmission)

σ²_F is shared environmental variance due to cultural transmission (see Keller et al. 2009, for the derivation)

σ²_C* is shared environmental variance, not due to cultural transmission

σ²_E is unshared environmental variance

σ²_A additive genetic variance

σ_AF covariance of A and F (see Keller et al. 2009, for the derivation)

Note in fitting the model we estimate σ²_C (i.e., σ²_F + σ²_C*)