Parental Reports for Twins

1. 1. After registration of newborn twins, which is before the age of 12 months for 89% of the twins, the first survey, sent to mothers, focuses on conception (spontaneous or otherwise), information from parents themselves (demographics, height, weight), pregnancy (e.g., duration, smoking and drinking behaviors of both parents during pregnancy), birth, and health of the children shortly after birth. For the majority of the twins, both parents were born in the Netherlands (90.7%). For the remaining families at least one parent was born in another Western county (3.9%), or in Turkey and Morocco (1.2%), and the other 4.2% were born in other non-Western countries. In 2005, all mothers registered with the NTR were mailed a two-page survey (referred as ‘NTR maternal survey’; Hoekstra et al., Reference Hoekstra, Willemsen, van Beijsterveldt, Montgomery and Boomsma2008). In this short survey we collected information on familial twinning and the twin pregnancy. Nearly 70% returned the survey (18,292 YNTR mothers, including 1,537 YNTR mothers who had never participated before in survey studies). For birth cohorts 2005 and later, we combined the original first survey and the ‘NTR maternal survey’ into a renewed survey-1. An online version of this survey-1 is in preparation.

2. 2. The survey at age 2 years (sent to mothers only) focuses on motor development, growth data as obtained from the Dutch National Health Services, breastfeeding, and handedness of twins and relatives. From birth cohort 2005 onwards, new items on well-being and temperament were added.

3. 3. At age 3, a survey is sent to both parents. The survey includes the CBCL 1.5-5 (Achenbach & Rescorla, Reference Achenbach and Rescorla2000), items on zygosity, health, child care, demographic characteristics, including religious background and parental education and occupation. From 2008 onwards, additional items on well-being, leisure activities, and sport activities of parents are included. To test if we could increase the response rate of the fathers, for three birth cohorts the father and the mother surveys were put into one booklet. This had no effect on the response rate of the fathers. If the mother returned the surveys, 75% of the fathers also completed a survey.

4. 4. The survey at age 5 is sent to mothers and includes a two-page section for fathers of twins about behavior problems. The content of this survey remained unchanged throughout the years and includes items on child care, growth, zygosity, toilet training, asthma and eczema, fine motor coordination, language development, and problem behaviors as assessed by short version of the Devereux Child Behavior checklist (DCB; Spivack and Spotts, Reference Spivack and Spotts1966; van Beijsterveldt et al., Reference van Beijsterveldt, Verhulst, Molenaar and Boomsma2004).

5. 5. At ages, 7, 9/10, and 12 years, information on the twins is collected from both parents and targets the development of emotional and behavior problems, as assessed by the Child Behavior Checklist 6-18 (CBCL; Achenbach & Rescorla, Reference Achenbach and Rescorla2001). From 2000 onwards, the Conners’ Parent Rating Scale — Revised (CPRS-R) was also included (Conners et al., Reference Conners, Sitarenios, Parker and Epstein1998; Conners, Reference Conners2001). In addition, these surveys included items on zygosity, medical conditions, vaccinations, tonsillectomy (from 2005), demographic characteristics, growth, sport and leisure activity of both children and parents, and additional items on well-being (from 2008). Due to the transition to a new administration database (Boomsma et al., Reference Boomsma, Willemsen, Vink, Bartels, Groot, Hottenga and van der Kleij2008a) and a shortage of staff, no surveys for ages 7 and 10 years were collected between 2008 and the last quarter of 2009.

Data Collection for Triplets

The NTR has registered 677 complete YNTR and 43 complete ANTR sets of triplets. Table 1 gives a summary of the data collection in triplets. The content of the triplet surveys is to a large extent similar to the content of the twin surveys. The survey at age 1 includes items on pregnancy, gestational age, smoking and alcohol consumption during pregnancy, and mode of delivery. Survey 2 includes information on breastfeeding, motor development, and growth. In 2008, a combination of survey 1 and 2 was sent to mothers of all triplets. The survey was sent to 535 mothers and was returned by 264 mothers. Since 2009, the combined survey 1 and 2 served as the first survey for newly registered triplets, and the number of returned surveys is now 273. The Child Behavior Checklist (Achenbach & Rescorla, Reference Achenbach and Rescorla2000, Reference Achenbach and Rescorla2001) was sent to parents of triplets aged 3 through 12 years, together with a short general questionnaire on parental and triplet characteristics (e.g., parental employment and religion, triplet school achievements and health).

TABLE 1 Overview of the Returned Surveys for Triplets (Born Between 1986 and 2011)

*Number indicates triplet families (except for teacher and DHBQ self-reports).

Teacher Reports for Twins, Triplets, and Siblings of Twins and Triplets

To assess cognitive development and behavior and emotional problems of children in classroom settings, the NTR collects information from teachers for twins at ages 7, 9/10, and 12 from 1998 onwards. In addition, teacher data were collected in a few sub-groups of 5-year-olds (Polderman et al., Reference Polderman, Posthuma, De Sonneville, Verhulst and Boomsma2006b). First, parents of twins are asked if they consent to approaching the teacher of their children. In this step, we collect the name and addresses of teachers and schools. Initially, collection of the parental consent was by mail, but for 2 years the consents have been gathered via online forms. If parents agree, one or more teachers, depending on whether children go to the same class/school, are approached to fill in a survey. The teacher surveys include the Teacher Report Form from the Achenbach System of Empirically Based Assessment (ASEBA) instruments (Achenbach & Rescorla, Reference Achenbach and Rescorla2001), items on adaptive functioning, school achievement (de Zeeuw et al., Reference de Zeeuw, van Beijsterveldt, de Geus and Boomsma2012), and the Conners’ Teacher Rating Scales Revised, short version CTSR-R (Conners, Reference Conners2001). Recently, items were added on classroom size and on bullying and being bullied. From 2007 onward, teachers of any additional children in the family have been approached too (regardless of the age of the siblings, as long as they are in elementary school). After 14 years of data collection from teachers, the average response rate is around 65% but has decreased in recent years (Polderman et al., Reference Polderman, Huizink, Verhulst, van Beijsterveldt, Boomsma and Bartels2011). In 2012, we started with an Internet version of the teacher survey. When the survey is not completed, we send a paper-and-pencil version and a link to the Internet survey with a reminder letter. Also, to compare the willingness to respond to a paper-and-pencil survey and an Internet survey, we randomly split the group of teachers of a cohort of 9-year-old twins into two parts. Half of the teachers received a paper-and-pencil questionnaire and the other half were requested to fill in the same questionnaire online. The response rate seemed not to be overly affected by the survey approach. Initial results (before a reminder was sent) showed a response rate of 39.4% for the paper-and-pencil version and 47% for the online version. Starting in 2009, the teacher survey has also been collected for triplets (Lamb et al., Reference Lamb, Middeldorp, van Beijsterveldt, Vink, Haak and Boomsma2011). After obtaining parental consent to approach the children's teachers (62% responded positively), we obtained a teacher questionnaire for 256 triplets.

Self-Report Surveys From Twins, Triplets, and Their Siblings

In 2005 we started with the data collection of self-report surveys in YNTR twins, triplets aged 14, 16, and 18 years, and in their siblings. All multiples are asked to complete the Dutch Health and Behavior Questionnaire (DHBQ), which focuses on emotional and behavior problems assessed by the Youth Self-Report (Achenbach, Reference Achenbach1991), subjective well-being (Bartels & Boomsma, Reference Bartels and Boomsma2009a), lifestyle (Geels et al., Reference Geels, Bartels, van Beijsterveldt, Willemsen, van der Aa, Boomsma and Vink2011), exercise behavior (van der Aa, et al., Reference van der Aa, Boomsma, Rebollo-Mesa, Hudziak and Bartels2010a), sedentary behavior (van der Aa et al., Reference van der Aa, Bartels, te Velde, Boomsma, de Geus and Brug2012), and family functioning (van der Aa et al., Reference van der Aa, de Geus, van Beijsterveldt, Boomsma and Bartels2010b). Before inviting the twins and their siblings, parents are contacted to ask for permission to send their children a self-report survey. We also ask the parents to register non-twin siblings of the twins. Initially the parental consents were sent out at the same time as the surveys. From 2007 onwards, the parental consent form has been obtained separately. In addition, starting in 2007 the survey is sent only to 14- and 16-year-olds, and partly different elements of the DHBQ are used at ages 14 and 16. For instance, the 60-item short form of the Neuroticism Extraversion Openness Five Factor Inventory (NEO-FFI) was included for 16-year-olds (Costa & McCrae, Reference Costa and McCrae1992). At the same time, the DHBQ became available online. The invitation letter contained a hyperlink, which could be used to log in to the online survey. All participants who did not complete the online survey after 2 months were reminded by e-mail. The overall response rate is 47%, and is higher for the online than for the paper-and-pencil version. In 2009, twins and triplets from birth cohorts 1986 to 1991 and their siblings were invited to participate in the online ANTR survey-8 (see Willemsen et al., Reference Willemsen, Vink, Abdellaoui, den Braber, van Beek, Draisma and Boomsma2012). From 2010 onwards, 18-year-old twins and their siblings were invited to fill in a new version of the ANTR survey-8 and these cohorts are also invited for ANTR survey 9. Self-report data have additionally been collected in sub-samples of twins and siblings aged 11 and 12 who took part in studies of cognition, attention problems, or brain imaging (Bartels et al., Reference Bartels, Rietveld, van Baal and Boomsma2002a; Polderman et al., Reference Polderman, Gosso, Posthuma, van Beijsterveldt, Heutink, Verhulst and Boomsma2006a; van Soelen et al., Reference van Soelen, Brouwer, Peper, van Leeuwen, Koenis, van Beijsterveldt and Boomsma2012). A total of 1,714 twins and 127 of their siblings who enrolled in these studies have returned a self-report survey. The collection of data for the adolescent triplets is the same as in twins. In total, 295 triplets and 32 siblings filled in the DHBQ survey.

Longitudinal Survey Data

The structure of the age-specific data collection is presented in Table 2. Because not all cohorts have yet reached a specific age, the numbers for these cohorts in the final column are still growing for children aged 2–18. Up until 2012, 31,879 mothers of 63,758 twins returned the first survey. Of these twins, there are still 56,724 actively registered (89%). Participants can leave the register at any time and if possible, we ask them for their reason to do so. The following reasons were given: unwilling to participate any longer (77.8%), disabled/ill (0.8%), deceased (2.9%), temporarily inactive (0.4%), and moved to a new address without notifying the NTR (18.1%). We will try to trace this last group and, if found, they usually become active participants again.

TABLE 2 Overview of Longitudinal Data Collection in Parents of Twins, By Self-Reports (After Age 12), and of Teachers of Young Twins and Their Siblings

>12* indicates that at least one self-report is returned (age >12 years); ^number indicates individuals (twins and siblings for self- and teacher report); age 18+ indicates returned survey could be YNTR-DHBQ or ANTR-Q8.

The procedure for the collection of surveys is as follows: Parents who did not return surveys within 2–4 months receive a written reminder. If resources allow, persistent non-responders are contacted by phone. From 2010 onwards, we sent a reminder by e-mail if an e-mail address is known. In Figure 1 and in Table 2 it can be seen that with increasing ages the sample sizes decrease. The primary reason is of course that not all of the birth cohorts have yet reached a specific age (see also Figure 1) and due to attrition of participating families (Bartels et al., Reference Bartels, van Beijsterveldt, Derks, Stroet, Polderman, Hudziak and Boomsma2007). In Table 3, an overview is given of the response rate for each age. Here, rates are averaged over all years. In recent years the participation rate has decreased. It is also important to note that a substantial proportion of the families can miss participation for a particular survey, but that these families often participate again in later surveys. In Table 3 an overview is given of the number families with such ‘partial’ participation. Sometimes families do not want to participate at a specific age, and they agree to be approached at a later age, but the majority of these families have moved to a new address that was retraced only after a survey had been closed. Low response rates may reflect bias and threaten the generalizability of the results. A number of studies have explicitly addressed potential bias. Using the ‘NTR maternal survey’ data collected by Hoekstra et al. (Reference Hoekstra, Willemsen, van Beijsterveldt, Lambalk, Montgomery and Boomsma2010), we compared characteristics of twin mothers between responders and non-responders. Non-response was associated with offspring zygosity (more monozygotic (MZ) mothers participated), maternal smoking during pregnancy, being younger at twin birth, and having a lower educational level. No differences were found for the use of fertility treatments, familial twinning, or religion. Importantly, the pattern of non-response did not differ between MZ and dizygotic (DZ) mothers. For internalizing and externalizing problems, Bartels et al. (Reference Bartels, van Beijsterveldt, Derks, Stroet, Polderman, Hudziak and Boomsma2007) compared the level of problem behaviors in 3-year-old twins between three groups: a group that participated at ages 3, 7, 10, and 12, a group that participated only at age 3, and a group with temporary non-participation. No significant mean differences in externalizing and internalizing behavior problems were found between the three groups. Around age 14, parental consent is obtained before approaching the 14- and 16-year-old twins and siblings. Around 40% of the parents return these forms. Comparisons of parental and twins’ characteristics between responders and non-responders of the self-report revealed that responders were more likely to have returned multiple earlier surveys, and that non-response was associated with a lower socio-economic status, lower educational levels, higher prevalence of mothers and fathers who smoked during pregnancy, and more problem behavior in the children. However, the differences were small, significance reflecting mainly the large sample sizes. Non-response was associated with more externalizing behaviors at age 7 and 12 (with effect sizes of 0.17 and 0.24), but not at age 3 (Bartels et al., Reference Bartels, van der Aa, van Beijsterveldt, Middeldorp and Boomsma2011). For internalizing problems, no differences were found between the two groups. It is important to note that the children in 13% of the families that did not respond to the parental consent completed an ANTR survey 8 after having reached age 18. Thus, non-response in adolescence does not signal permanent dropout.

TABLE 3 Overview of Response Rate and Response Pattern of Parental Surveys for Twins at Ages 1 to 18 for Birth Cohorts 1986–2011

*The first number refer to the absolute response rate; the second is conditional on least one survey returned. ^For survey 2 (Q1), there is only one previous survey. **Indicates that at least one parental survey is returned.

Summary of YNTR Results

The longitudinal genetically informative data collection and the in-depth phenotyping in young Dutch twins resulted in a large number of scientific papers on a range of topics. We showed that genetic factors not only contribute to variation in problem behaviors in childhood, but also to the continuity of behavior and emotional problems during childhood (e.g., Boomsma et al., Reference Boomsma, van Beijsterveldt, Bartels and Hudziak2008b; Rietveld et al., Reference Rietveld, Hudziak, Bartels, van Beijsterveldt and Boomsma2004). In general, we found that shared environment accounted for less of the phenotypic variation and stability in behavior problems than genetic factors, especially when recognizing that part of this shared environment is actually represented by stability in rater bias (Bartels et al., Reference Bartels, van Beijsterveldt, Derks, Stroet, Polderman, Hudziak and Boomsma2007). By collecting longitudinal data on problem behavior as rated by teachers, stability will be further investigated in future studies. Using standard measures across YNTR and ANTR, such as the measurements of the ASEBA, it is feasible to assess problem behavior across the life span. Recently, we investigated attention problems in a quasi-longitudinal design spanning age 3 through 60 years, and found that the heritability of attention problems is higher than previously suggested, with early genetic and environmental factors having long-lasting effects (Kan et al., in press). A similar approach was used to analyze the genetic architecture of anxious-depressed symptom scores across the life span (Nivard et al., submitted). The question of a potential change in genetic architecture of traits across childhood and adolescence was also addressed for lifestyle traits. As an example, we point to our studies on leisure time exercise behavior (Huppertz et al., Reference Huppertz, Bartels, van Beijsterveldt, Boomsma, Hudziak and de Geus2012; van der Aa et al., Reference van der Aa, de Geus, van Beijsterveldt, Boomsma and Bartels2010a) that showed a very strong shift from shared environmental determination to genetic determination across the age span of 7–19. At age 7, both MZ and DZ correlations were around 0.90, suggesting a strong contribution of the environment of the twins to childhood exercise behavior. During adolescence, the MZ correlations decrease a bit to about 0.80, but DZ correlations drop especially in boys. At age 17–19, MZ correlations are still as high as 0.70, but DZ correlation by this time has decreased to 0.34 for females and 0.48 for males (Huppertz et al., Reference Huppertz, Bartels, van Beijsterveldt, Boomsma, Hudziak and de Geus2012; van der Aa et al., Reference van der Aa, de Geus, van Beijsterveldt, Boomsma and Bartels2010a), suggesting that late adolescent exercise behavior reflects heritable factors, which we hypothesize to include exercise ability and the acute psychological response to exercise (de Geus & de Moor, Reference de Geus, de Moor, Bouchard and Hoffman2011).

The wealth of item-level data contained in the NTR offers possibilities for multivariate item-level analyses. Franić et al. (2012) applied multivariate genetic item analyses to the items of the Internalizing grouping of the Child Behavior Checklist 4-18 (CBCL 4-18), and showed that the indeterminacy of the phenotypic factor structure of internalizing problems in children as measured by the CBCL (in the sense of a three- and a four-factor model both providing a satisfactory account of the data) is a result of several different models pertaining to genetic and environmental influences, giving rise to the observed covariance structure. While additive genetic influences conform to a two-factor model, the individual-specific environmental influences form a four-dimensional structure. Interestingly, shared environmental influences form a unidimensional structure, therefore affecting Internalizing items across the board. This not only offers important insights for the test constructor, but may also aid diagnosis (e.g., by elucidating the etiology of comorbidity of anxiety and depression; genetically, these two syndromes appear to form a unidimensional structure, while unique environmental influences appear to drive their differentiation). In addition, the results of this type of analysis may assist studies aimed at identifying genetic variation relevant to phenotypic individual differences, by providing insights needed to optimally define the phenotype.

The large data collections have made it possible to investigate variation in individual differences in behaviors that are less common in the population, or where information about prevalence is lacking. For example, for gender identity disorder (when there is a disassociation between the biological sex and feelings of gender identity), information about the prevalence in children is scarce and derived from adult clinical data (Zucker & Bradley, Reference Zucker and Bradley1995). On the basis of two items in the CBCL, we found that the prevalence of cross-gender behavior was 3.2% and 5.2% for 7-year-old boys and girls, respectively, and decreased to 2.4% and 3.3% for 10-year-old boys and girls, respectively. Surprisingly, the prevalence rate of cross-gender behavior of girls with a male co-twin was lower than of girls with a female co-twin. Genetics played an important role in the development of cross-gender behaviors (van Beijsterveldt et al., Reference van Beijsterveldt, Hudziak and Boomsma2006). Another advantage of the large data collections is that we can select discordant MZ twins, even for traits that are highly heritable. Comparisons of discordant MZ twins offer an alternative to the traditional case–control study, since cases and controls are perfectly matched for age, sex, and genetic background. For example, Lehn et al. (Reference Lehn, Derks, Hudziak, Heutink, van Beijsterveldt and Boomsma2007) used MZ twins discordant for attention problems to investigate specific environmental influences on attention problems and found low birth weight and delayed motor development as risk factors for attention problems. Using the same twins, Ehli et al. (Reference Ehli, Abdellaoui, Hu, Hottenga, Kattenberg, van Beijsterveldt and Davies2012) recently found that copy number variations (CNVs) were associated with attention problems and found evidence for the presence of two de novo CNVs.

The increasing number of genotyping data in combination with longitudinal survey data can be used for testing of gene-environment interactions. In a large twin study on asthma liability and tests for gene-environment interactions, we found that being a boy, born in the 1990s, premature birth, longer incubator time, and child care outside the home all increased the risk for asthma (van Beijsterveldt & Boomsma, Reference van Beijsterveldt and Boomsma2008). Although our findings revealed almost no evidence for gene-environment interaction, the risk factors that were identified could be useful in studies with measured genotypes. An important risk factor could be birth cohort, and since we recruit children from 1986, the YNTR asthma data could be very informative.

Collecting biological material for DNA may be more difficult in children than in adults. Recently, we found that DNA from buccal samples seems to be a reliable alternative for blood sampling. Utilizing Affymetrix SNP 6.0 microarrays from 372 DNA samples of children, we found high concordance in genotype and CNVs across blood and buccal samples (Scheet et al., Reference Scheet, Ehli, Xiao, van Beijsterveldt, Abdellaoui, Althoff and Boomsma2012). Genotype concordance was very high (R ² > 99%) between co-twins from 43 MZ pairs, suggesting that DNA from buccal samples perform as well as DNA from blood samples on the current generation of microarray technologies.

Representativeness and Participation of Twins

A recurring question for twin researchers concerns the representativeness of twins. We compared heritability estimates for growth at birth and in early childhood (Mook-Kanamori et al., Reference Mook-Kanamori, van Beijsterveldt, Steegers, Aulchenko, Raat, Hofman and Jaddoe2012) between Dutch twins and singleton offspring and found comparable heritability estimates. At birth the heritability estimates for body length were 26% for singletons and 27% for twins, and for weight heritability estimates were 26% in singletons and 29% in twins. At age 3, the heritability estimates for height were 63% and 72% for respectively singletons and twins, while the heritability for weight for twins was 71% and higher than for singletons (42%). Regarding physical and psychological development, comparisons of twins and singletons revealed no large differences. Compared to singletons, twins have a growth delay during early childhood (van Dommelen et al., Reference van Dommelen, de Gunst, van der Vaart, van Buuren and Boomsma2008), but catch up their growth during childhood (Estourgie-van Burk et al., Reference Estourgie-van Burk, Bartels, van Beijsterveldt, Delemarre-van de Waal and Boomsma2006). At age 18 twins attained the same height and weight as their peers in the general population (Estourgie-van Burk et al., Reference Estourgie-van Burk, Bartels, Boomsma and Delemarre-van de Waal2010). Compared to their siblings, twins were as tall as their siblings, but were leaner. Motor development of twins, as measured by achievement of motor milestones, showed no deviations compared to a normative database of singletons (Brouwer et al., Reference Brouwer, van Beijsterveldt, Bartels, Hudziak and Boomsma2006). In childhood, the developmental trajectories of externalizing problems are similar in twins and singletons (Robbers et al., Reference Robbers, Bartels, van Oort, van Beijsterveldt, van der Ende, Verhulst and Huizink2010). The course of development trajectories of internalizing problems are broadly the same, but after age 9, twins showed less internalizing problems than singletons. Small differences in the cognitive abilities have been reported between twins and singletons at age 6, but these differences were less than one IQ-point and disappeared at later grades (Webbink et al., Reference Webbink, Posthuma, Boomsma, de Geus and Visscher2008). In a within-family design, we found that during primary school twins had lower ratings on arithmetic, reading, and language than their non-twin siblings. However, the differences observed between twins and their non-twin siblings in educational achievement seemed to be dependent on the birth order within the family. Twins who were first in birth order within the family had the same, or even higher, ratings as their non-twin sibling, while twins with a sibling who was first in birth order within the family had lower ratings than their non-twin siblings (de Zeeuw et al., Reference de Zeeuw, van Beijsterveldt, de Geus and Boomsma2012).

To keep study samples representative, it is important to retain families in the register and in the studies. Not all families contact the register when they move to a new address and not all families are motivated to fill out all surveys. One way to motivate families is by informing the participants about ongoing research studies and results. This is done in various ways. We have a Web site (http://www.tweelingenregister.org/) and a yearly newsletter TWINFO that is mailed to all families and also published on the Web site (http://www.tweelingenregister.org/twinfo/). Recently, the NTR also started to use social media. This includes Facebook (https://www.facebook.com/NederlandsTweelingenRegister), Twitter (https://twitter.com/NTR_VU), and Hyves (http://ntr-vu.hyves.nl). With social media, information can be shared on a more informal level, updates on study results can be given on a regular basis, and contacts with the participants can be established quickly (Stroo et al., Reference Stroo, de Moor, Lighart, Brouwer, Willemsen, Boomsma and Vink2012). In addition, the NTR has started building a personal portal for participants. The purpose is to provide customized insight into their own results (e.g., personality profile), in addition to the general information.

Concluding Remarks

Critics of twin studies often argue that the utility of twin samples is limited to the determination of genetic and environmental influences on specific phenotypes. As already presented, the NTR provides more knowledge than important behavioral genetic findings. The NTR has already provided findings using molecular genetic, neuroimaging, endophenotypic, and family study methodologies to provide hints on the links between the genome, brain, and behavior. A little understood, but equally important gain from the NTR has been the non-genetic exploration of the validity, stability, sensitivity, and specificity of new nosologic studies. Through the use of the NTR, advances have been made in understanding the quantitative nature of obsessive-compulsive behavior (Hudziak et al., Reference Hudziak, van Beijsterveldt, Althoff, Stanger, Rettew, Nelson and Boomsma2004; van Grootheest et al., Reference van Grootheest, Cath, Beekman and Boomsma2007), pediatric bipolar disorder (Boomsma et al., Reference Boomsma, Rebollo, Derks, van Beijsterveldt, Althoff, Rettew and Hudziak2006b), oppositional defiant disorder (Derks et al., Reference Derks, Dolan, Hudziak, Neale and Boomsma2007), and aggression (Ligthart et al., Reference Ligthart, Bartels, Hoekstra, Hudziak and Boomsma2005). Although it is true that behavioral genetic analyses have yielded insights into the genetic architecture of each of these traits, additional approaches such as latent class, latent trait, and item level analyses have helped advance criteria-based understanding of each of these common disorders. In every case, other groups have picked up on these ‘diagnostic’ conceptualizations and advanced them in other, non-twin, databases. The NTR, specifically because of its longitudinal developmental infrastructure, provides an unparalleled opportunity to study the resilience of children and adolescence. As time goes forward, investigators will not only be able to investigate the genetic and environmental contributions to the development of adolescent and young adult medical and psychiatric illness, but also shed light on why some children recover from early psychopathology to lead lives well. The NTR and its capacity to continue to collect both phenotypic and biological data will provide an excellent resource to better understand the biological relations between the environment, the (epi)genome, and medical outcomes.