Table 5 Variance components for Educational Attainment in our Nuclear Twin and Family Design models.

From: Conventional twin studies overestimate the environmental differences between families relevant to educational attainment

Ref #

Model name

A

N

S

T

P

rGE

E

PA-1

ASTPE

51% (32–56%)

 

10% (0–19%)

16% (8–27%)

0% (0–2%)

0% (0–9%)

23% (21–26%)

PA-1.1

ASTPE (no rGE)

51% (46–56%)

 

10% (0–18%)

16% (8–26%)

0% (0–0%)

 

23% (21–26%)

PA-1.2

ASTE

51% (46-56%)

 

10% (0–18%)

16% (8–26%)

  

23% (21–26%)

PA-2

ANTPE

56% (8–56%)

0% (0-0%)

 

24% (16–32%)

0% (0–13%)

−2% (NA–12%)

23% (21–26%)

PA-2.1

ANTPE (no rGE)

52% (47–56%)

1% (0–10%)

 

24% (16–30%)

0% (0-0%)

 

23% (20–26%)

PA-2.2

ATPE

52% (37–60%)

  

25% (20–32%)

0% (0–0%)

0% (−1–8%)

23% (21–26%)

PA-2.3

ATPE (no rGE)

52% (47–56%)

  

25% (20–30%)

0% (0-0%)

 

23% (21–26%)

PA-2.4

ANTE

52% (47–56%)

0% (0–10%)

 

24% (16–30%)

  

23% (21–26%)

PA-2.5

ATE

52% (47–56%)

  

25% (20–30%)

  

23% (21–26%)

PA-3

ANSTE

51% (46–56%)

0% (0–8%)

10% (0–18%)

16% (7–26%)

  

23% (21–26%)

PA-3.1

ASTE

51% (46–56%)

 

10% (0–18%)

16% (8–26%)

  

23% (21–26%)

SH-1

ASTPE

36% (25–49%)

 

11% (1–19%)

16% (7–27%)

4% (1–7%)

10% (7–12%)

23% (21–26%)

SH-2

ANTPE

39% (5–52%)

0% (0–0%)

 

25% (17–32%)

3% (2–17%)

10% (7–12%)

23% (20–26%)

SH-2.1

ATPE

39% (28–52%)

  

25% (17–32%)

3% (1–7%)

10% (7–12%)

23% (20–26%)

SH-3

ANSTE

69% (63–76%)

0% (0–2%)

6% (0–15%)

4% (0–14%)

  

21% (19–24%)

SH-3.1

ASTE

69% (61–76%)

 

6% (0–15%)

4% (0–14%)

  

21% (19–24%)

SH-3.2

ANSE

70% (63–76%)

0% (0–3%)

9% (2–16%)

   

22% (19–24%)

SH-3.3

ASE

70% (63–76%)

 

9% (2–16%)

   

22% (19–24%)

  1. PA phenotypic assortment model, SH social homogamy model, A additive genetic influence, N non-additive genetic influence (dominance), S environmental influences shared by all siblings, T environment shared by twins only, P effects of parental education, rGE passive gene-environment correlation, E nonshared environment.
  2. Results for both Phenotypic Assortment (PA) models and Social Homogamy (SH) models are provided. Baseline models in which N, S, or P have been fixed to zero for model-identification purposes appear in the top row of each section. Our overall best fitting model (PA-ASTE) is in boldface and appears twice under PA-1.2 and PA-3.1. Our best fitting SH model (SH-ASTPE) is italicised. 95% confidence intervals are in parentheses. Alternative results where non-additive genetic influences were characterised as epistasis are provided in Supplementary Table 3, but differences were negligible.
Back to article page