- Similar results are sometimes reported based on comparisons of degree of contact between twins in a given pair and their subsequent trait similarity. If the degree of within-pair similarity for a given trait is not very different between twin pairs with high versus low contact, it is concluded that the EEA is valid for that trait.
- Even if the bias induced by violating the EEA is significant, it is random and so does not systematically lead to either inflated or deflated heritability estimates (Barnes et al. 2014, \cite{Littvay_2012}).
- The differences in environmental similarity between MZ and DZ twins are an effect of the MZ twins' greater genetic similarity, not a cause of their greater phenotypic similarity. In other words, greater genetic similarity between MZ twins -> greater similarity in environments between MZ twins. For instance, Evans & Martin (2000) argue, "...it appears that the more similar parental treatment of MZ vs. DZ twins occurs in response to the greater similarity of actions initiated by MZ pairs" [Emphasis in original]. Joseph et al. (2015) refer to this as "Argument A" and criticize it as circular.
- The EEA is valid with respect to causes of the trait being studied, so it doesn't matter if it's invalid with respect to anything else. This is often framed as an emphasis of the need for trait-relevant violations of the EEA to be found in order to invalidate twin studies of the specific trait in question. Bouchard & McGue (2002, p. 9), for instance, argue that "...the issue is not whether MZ twins experience more environmental similarity than DZ twins, but rather whether they are more likely to share trait-relevant features of their environments" [Emphasis mine].\cite{Bouchard_2002} This argument is often made as part of argument 1, as twin researchers typically acknowledge, to their credit, that the EEA is wrong insofar as MZ twins are generally treated more similarly than are their DZ counterparts. So the way this argument (#6) works is generally that researchers conduct studies aimed specifically at testing the validity of the EEA with regard to a specific trait, and they typically find no evidence that the EEA is invalid for trait-relevant environmental features, and voila! The EEA is valid (except for the many things for which it might not be, but those don't count here, of course!).
- Because of argument #6 (just above this one), the EEA should be evaluated on a case-by-case basis, rather than assumed to be so uniformly false as to completely invalidate all twin studies of all traits. Bouchard & McGue (2002, p. 9) exemplify this argument when they write, "Good scientific practice...requires that the [equal environments] assumption be repeatedly tested for each trait under investigation and particular findings that depend on the assumption be replicated in designs that do not make the assumption. Large studies with multiple kinships can specifically test such assumptions"\cite{Bouchard_2002}. Similarly, Kohler et al. (2011) argue that "...the validity of this assumption [i.e. the EEA] needs to be evaluated in the specific context".\cite{Kohler2011}
- Studies of misclassified twins vindicate the EEA by demonstrating that any EEA violations do not affect variation in whatever trait is being studied. It is not uncommon for these studies to be cited by scholars attempting to defend the twin method against its critics, as discussed below.
- Studies of twins reared apart vindicate the CTD and the EEA it relies on, because twins reared apart studies don't rely on the EEA, but they still get similar results. Thus this is supposed to prove that any EEA violations don't significantly bias heritability estimates generated by the CTD. Here is an example: "...studies of twins raised together have been validated by studies of twins reared apart, suggesting that the shared environment does not exert substantially enhanced influence on MZ twins."\cite{Cesarini_2008}
- Other, non-twin studies produce very similar heritability estimates as classical twin studies do, despite not making the same assumptions (e.g. the EEA). This is said to provide an independent source of evidence to support the accuracy of the heritability estimates from twin studies. For instance, Barnes et al. (2014) inform us that "...findings garnered from studies using the classical twin design, the adoption design, and the twins reared apart design largely converge, revealing that genetic influences explain approximately 50 percent of the variance in antisocial phenotypes".\cite{BARNES_2014} Molecular genetic research on unrelated subjects, in contrast to old-school genetic research on related subjects like twins, has been
Notably, argument #6 can be, and often is, used by BGists to shift the burden of proof off of themselves (who are claiming that trait X is Y% heritable or whatever) and onto their critics (who are arguing that the EEA, and thus the twin method itself, is invalid). As Bouchard put it, "Causal status must be demonstrated, not assumed" \cite{Bouchard_1993}.
Responses to responses to critics
How do critics of twin studies respond to these responses?
Point 1
Critics may respond to point 1 by saying that the studies purporting to find support for the EEA are seriously methodologically flawed (e.g. Beckwith & Morris 2008, Charney 2017, Pam et al. 1996). They may also point out that two-thirds of MZ twins are "monochorionic", meaning that they share a placenta, so their environments are more equal than those of all DZ twins even before birth--something it is not possible to control for in twin studies (Charney 2017).
With regard to this last point, recent evidence presented by behavior geneticists indicates that the biasing effects of chorionicity on heritability estimates in twin studies are small and do not affect most phenotypes anyway (van Beijesterveldt et al. 2016). Other BG researchers argue that "... monochorionic MZ twins [are] more likely to differ in birth weight (result of the ‘transfusion syndrome’ and the following differences in availability of placental blood). To the extent that data are available, chorionicity and its consequences do not seem to threaten the logic of the twin method"\cite{Viding_2008}.
Point 2
In the case of Barnes et al. (2014) and Felson (2014), this point is supported by arguing that heritability estimates are not significantly reduced by controlling for environmental similarity. For instance, Barnes et al. (2014) write, "...numerous studies examining the potential moderating effects of environmental similarity on h2 and c2 estimates have found that violations of the EEA result in statistically nonsignificant parameter deviations". They also cite both previous studies on the EEA and its effects on heritability estimates, as well as their own simulations, to conclude that "...when the EEA is violated, estimates are inflated by between 1 and 5 percentage points." [Emphasis in original.]
Point 3
Often made along with point 2, this is an important and common argument made by BGists to dismiss arguments by critics about the EEA. Often this will involve citing studies such as that of Kendler et al. (1994), which found that "... mothers' and fathers' approach to raising twins had no significant influence on twin resemblance for the four examined psychiatric disorders."\cite{Kendler1994} One example of such a citation is in a recent paper by Gillespie et al. (2017), which states that "A number of studies have reported that violations [of the EEA] do not significantly predict twin pair resemblance for a number of complex behaviours and phenotypes." There's a lot of citations at the end of this sentence in the original paper, and needless to say, Kendler et al. (1994) is one of these citations.\cite{Gillespie_2017}
How has it been addressed/responded to in the past? First I should outline the methodology used in these sorts of studies. Basically, they attempt to examine the relationship between environmental similarity and behavioral trait similarity in twins. If the EEA is not only false but actually biases twin study results, the argument goes, then twins whose environments are more similar should be more similar with regard to behavior as well (i.e. a positive, strong correlation between the two similarities). Conversely, it would support the EEA (or at least suggest that its violation doesn't really matter) if there was no strong relationship (or even a negative one) between these measures.
So on to the responses of critics:
Many studies supposedly showing that greater environmental similarity isn't related to greater behavioral similarity may in fact show the opposite. As Beckwith & Morris noted in 2008,
"...a closer look at the studies in support of the EEA finds that many actually contain significant data arguing that “trait-relevant” EEA was indeed violated. For example, one study found that DZ twins who were more likely to be confused in appearance were more likely to have similar scores for activity and impulsivity traits. MZ and DZ twins as a whole who were more similar in appearance were more likely to share all four measures of personality traits assessed."\cite{Beckwith_2008}
Also, some of these studies were based on yet another non-EEA assumption: that physical resemblance -> environmental similarity, and that, if the EEA is "meaningfully" false (or the "trait-relevant" EEA is false, to put it differently), then environmental similarity -> behavioral similarity. Thus, some of these studies assume that if more physically similar twins aren't also more behaviorally similar, then the trait-relevant EEA must be true. Why is this a questionable line of reasoning? As Pam et al. noted in 1996, "...bonds between MZ twins are not only the product of a physical resemblance which induces others to treat them as a unit, but may also arise due to intimate contact and common experience - even if the twins can be told apart on sight!"\cite{Pam_1996}
Point 6: This common argument claims that the only way the EEA being violated is a problem is when the violation is with regard to something etiologically relevant to the trait being studied. Joseph et al. (2015) dub it "Argument B", arguing that limiting assessments of the EEA to "trait-relevant" factors (as argument #6 does) obscures other important environmental factors for which MZ twins' environments are more similar (Joseph et al. 2015). The fundamental reason this argument doesn't hold up is that there's no really reliable way for researchers to know which variables are "trait-relevant" i.e. actually cause the trait being studied in a given context. "Even those researchers who do attempt to measure environmental influences, as the authors of one study point out, choose out of “a virtually endless array of possible environmental characteristics.”"\cite{Cronk2002}, quoted in \cite{Beckwith_2008} (p. 786)
Point 8: Misclassified twin studies such as a recent one by Conley et al.\cite{Conley2013} are often cited to rebut the argument that the EEA is false and, therefore, the results of classical twin studies are invalid. Here are some examples:
- "...personality and cognitive differences between MZ and DZ twins persist even among DZ twins whose zygosity has been miscategorized by their parents, indicating that being mistakenly treated as an identical twin by one’s parents is not sufficient to generate observed differences in similarity"\cite{Cesarini_2008}
- "in the relatively rare cases where parents miscategorize their twins as MZ instead of DZ (or the converse), differences in correlations of cognitive ability and personality persist"\cite{CESARINI_2010}
- (This is my personal favorite because it's the most brazen) "Perhaps the most empirically rigorous method of assessing the validity of the EEA is through the use of misclassified twin samples."\cite{BARNES_2014}
How exactly do these studies work? Fosse et al. (2015) explain: "The assumption in these studies is that if it is the parents’ treatment of the twins and not the twins’ genes that is etiologically important, then fraternal twins that are misclassified as identical should correlate higher for mental disorders than twins correctly classified as fraternal, whereas identical twins misclassified as fraternal should correlate lower than correctly classified identical twins."\cite{Fosse_2015} In other words, consider the table below, where all four letter values are correlations. If the EEA is invalid, then c should be higher than d, whereas a should similarly be higher than b.
Actually MZ | Actually DZ |
Perceived MZ | a | c |
Perceived DZ | b | d |