1
The sample size consisted of 1984 twins (Range=12-21 years, M=16.94, SD= .58). In total there were
693 males and 972 females), including 774 MZ (monozygote) and 1210 DZ (dizygote) twins. The mean average
score for internalizing problems for MZ firstborns was 9.8 (SD=7.64) ,for second borns 9.6 (SD=7.65) . The
average mean for externalizing problems for MZ twins was 7.1(SD= 5.24) and for second borns 6.7(SD=5.18).
The mean average score for internalizing problems for DZ firstborns was 9.7 (SD=7.55) for second borns 10.1
(SD=7.90).For externalizing problems the mean average score for firstborns in DZ twins was 7.5(SD=5.35) ,
and for second borns 7.6(SD=5.20). There can be concluded that first and second born within MZ and DZ twins
score approximately the same on internalizing and externalizing problems. Based on the twin correlations of the
Saturated Model in table2 the expectation for the estimates of the proportion of variance explained by A is
37% , by D is 18% and by E is 45%. There was expected more influences of the unique environmental effects
(E) than of the additive genetic effects (A) and the least influences of the dominant genetic effects (D). Based on
the correlations (table2) for MZ (r=.55) and DZ (r=.24) twins there was decided to use an ADE model
(rMZ>2rDZ). There can be concluded that the ADE provides a good fit to the data as it does not fit significantly
worse than the Saturated model, p=.634 (table1). The estimated covariances of the ADE model, suggests that the
covariance between internalizing and externalizing problems is for 53% due to additive genetic effects (A), for
35% due to dominant genetic effects (D) and for 43% due to unique environmental effects (E). Since the ADE
model has less parameters than the saturated model, the AE sub model was compared with the ADE model. The
AE model provides a good fit to the data as it does not fit significantly worse than the ADE model, p=.600.
Since the AE model fits good and has less parameters (table1) than the ADE model, further analysis were done
with the AE model. The model without the additive genetic correlation (modelAErA0) does fit significantly
worse than the AE model, p<.001. Additive genetic effects are important in explaining the covariance between
internalizing and externalizing problems. The AErA1 model in which the additive genetic correlation is 1 does
fit significantly worse than the AE model, p<.001. Additive genetic effects don’t completely overlap between
internalizing and externalizing problems. The AErE0 model, (unique environmental correlation) is significantly
larger than 0 as the model fits significantly worse than the AE model, p<.001. Unique environmental effects are
important in explaining the covariance between internalizing and externalizing problems and the unique
environmental effects on internalizing problems do substantive overlap with the ones in externalizing problems.
Based on the results in table1, there can be concluded that the AE model compared to the other (sub) models,
provides the best fit to the data. The contribution of A to the variance of internalizing problems was 54% , 95%
CI(45%-61%), the contribution of E to the variance of internalizing problems was 46%, 95%CI(40%-55%). The
contribution of A to the variance of externalizing problems is 55% , 95% CI(47%-62%) ,the contribution of E
to the variance of externalizing problems is 45%, 95%CI (38%-53%). The contribution of A to the covariance
between internalizing and externalizing problems is 57%, 95% CI (40%-72%) and the contribution of E to the
covariance between internalizing and externalizing problems is 43%, 95% CI (28%-60%). The correlations
between the A and E variance components are, rA=.37; there is little overlap between additive genetic factors
that influence the differences in internalizing and externalizing problems and rE= .33; there is little overlap
between the unique environmental factors that influence the difference in externalizing and internalizing
problems. Based on all the results there can be concluded that the phenotypic differences of internalizing and
externalizing problems is due to an overlap in additive genetics and the phenotypic differences of internalizing
and externalizing problems are also due to unique environmental effects.
The sample size consisted of 1984 twins (Range=12-21 years, M=16.94, SD= .58). In total there were
693 males and 972 females), including 774 MZ (monozygote) and 1210 DZ (dizygote) twins. The mean average
score for internalizing problems for MZ firstborns was 9.8 (SD=7.64) ,for second borns 9.6 (SD=7.65) . The
average mean for externalizing problems for MZ twins was 7.1(SD= 5.24) and for second borns 6.7(SD=5.18).
The mean average score for internalizing problems for DZ firstborns was 9.7 (SD=7.55) for second borns 10.1
(SD=7.90).For externalizing problems the mean average score for firstborns in DZ twins was 7.5(SD=5.35) ,
and for second borns 7.6(SD=5.20). There can be concluded that first and second born within MZ and DZ twins
score approximately the same on internalizing and externalizing problems. Based on the twin correlations of the
Saturated Model in table2 the expectation for the estimates of the proportion of variance explained by A is
37% , by D is 18% and by E is 45%. There was expected more influences of the unique environmental effects
(E) than of the additive genetic effects (A) and the least influences of the dominant genetic effects (D). Based on
the correlations (table2) for MZ (r=.55) and DZ (r=.24) twins there was decided to use an ADE model
(rMZ>2rDZ). There can be concluded that the ADE provides a good fit to the data as it does not fit significantly
worse than the Saturated model, p=.634 (table1). The estimated covariances of the ADE model, suggests that the
covariance between internalizing and externalizing problems is for 53% due to additive genetic effects (A), for
35% due to dominant genetic effects (D) and for 43% due to unique environmental effects (E). Since the ADE
model has less parameters than the saturated model, the AE sub model was compared with the ADE model. The
AE model provides a good fit to the data as it does not fit significantly worse than the ADE model, p=.600.
Since the AE model fits good and has less parameters (table1) than the ADE model, further analysis were done
with the AE model. The model without the additive genetic correlation (modelAErA0) does fit significantly
worse than the AE model, p<.001. Additive genetic effects are important in explaining the covariance between
internalizing and externalizing problems. The AErA1 model in which the additive genetic correlation is 1 does
fit significantly worse than the AE model, p<.001. Additive genetic effects don’t completely overlap between
internalizing and externalizing problems. The AErE0 model, (unique environmental correlation) is significantly
larger than 0 as the model fits significantly worse than the AE model, p<.001. Unique environmental effects are
important in explaining the covariance between internalizing and externalizing problems and the unique
environmental effects on internalizing problems do substantive overlap with the ones in externalizing problems.
Based on the results in table1, there can be concluded that the AE model compared to the other (sub) models,
provides the best fit to the data. The contribution of A to the variance of internalizing problems was 54% , 95%
CI(45%-61%), the contribution of E to the variance of internalizing problems was 46%, 95%CI(40%-55%). The
contribution of A to the variance of externalizing problems is 55% , 95% CI(47%-62%) ,the contribution of E
to the variance of externalizing problems is 45%, 95%CI (38%-53%). The contribution of A to the covariance
between internalizing and externalizing problems is 57%, 95% CI (40%-72%) and the contribution of E to the
covariance between internalizing and externalizing problems is 43%, 95% CI (28%-60%). The correlations
between the A and E variance components are, rA=.37; there is little overlap between additive genetic factors
that influence the differences in internalizing and externalizing problems and rE= .33; there is little overlap
between the unique environmental factors that influence the difference in externalizing and internalizing
problems. Based on all the results there can be concluded that the phenotypic differences of internalizing and
externalizing problems is due to an overlap in additive genetics and the phenotypic differences of internalizing
and externalizing problems are also due to unique environmental effects.
