Most people agree that children of all backgrounds should be able achieve their full potential. This is exactly why all educational systems sort students into different classes or schools based on their abilities (ranging from age 10 in Germany to 16 in the UK and the US).1,2. These ability tracks can be used between and within schools to make it easier to adapt the learning environment to meet the needs of students with diverse abilities, rather than differentiating the curriculum within a single classroom.3. This implies that students are always sorted according to their potential ability. However, there are concerns that sorting may also be based on family background, especially if the tracking decisions are made at an early age. The empirical evidence supports the fact that the influence of family history on educational attainment is stronger when tracking is done at a younger stage, and the average performance is not higher (reviews of the literature). 2,4,5). These studies indicate that not all people can benefit from the benefits of tracking at a young stage.
Studies on inequality of educational opportunities often test whether delaying track to a later stage makes educational attainment less dependent on family background. Instead, they use parental socioeconomic stat (SES) measures for capturing family background and measures like IQ to capture potential abilities. This is problematic because of several reasons. One problem is that it is nearly impossible to measure the impact of family background on educational attainment in all possible ways. Researchers have used sibling similarity in educational achievement as an omnibus measure to address this problem.6. Sibling similarities do not reflect educational outcomes. They are also due to genetic sharing. Sibling similarities mix together the influences of family history and potential ability. Not all potential ability measures can be included. It is similar to family background that it is difficult, or even impossible, to measure all cognitive (e.g. verbal, information processing speed, and memory) abilities that are important in school.7,8. Family influence can be seen as early as prenatally. Therefore, every test that measures capture ability will reflect a part of the influence of family history.9.
This study uses a genetics toolbox in order to address many of these issues. To disentangle genetic influences and environmental influences shared between children from the same family as well as individual environmental influences, we analyze data from mono-and dizygotic twins. This is also known as the classical twin design. It allows us to estimate heritability and moderate environmental impact using a measured variable.10,11. It has been suggested that heritability, which is the variance explained genetically, is a good overall measurement for the opportunities people have to realize their potential. Shared environmental influence is a good overall measurement for the impact of family backgrounds12. If tracking occurs at a later stage, it increases the importance and decreases the importance family background. One would expect genetic influences to increase and shared environmental influences lower in educational systems where tracking occurs later.
This hypothesis has some caveats. First, genetic influences may not only reflect realizations of “positive” potential, but also “negative” potential that can hinder education (think, for example, of disorders), even though positive traits seem to be the most common.13,14. The shared environment also captures all environmental influences that make siblings closer. It should be considered a broad indicator of the various ways in which family history matters, including parental resources, behavior, and neighborhood characteristics. This is a strength as it avoids having to measure all of them. The downsides to this system are that it creates a black box and may capture information that one does not want. The shared environment would, for example, reflect differences in educational attainment between age cohorts, even if year of birth is not taken into account (ref. 15).
These issues are important to remember. However, others have suggested that genetic influence should be greater and shared environment influence lower in educational programs that promote equality of opportunity (e.g. refs. 16,17,18,19,20). This hypothesis is not supported empirically, and has not been tested for age tracking. The first question is: How does the influence genes and the shared environment affect educational attainment? This depends on the age atwhich children are tracked. Educational attainment is defined as the degree of secondary education. This is a strong determinant for final educational attainment.
It is important to understand why tracking young children can lead to inequalities of opportunity. The literature makes a distinction between primary and secondary effects of family history. Family background is important for education level first because it impacts how children perform (in terms test scores, etcetera).Primary effects). These performance differences aside, children from different backgrounds tend to have different levels of education, even though they perform equally. This is due to the fact that different educational decisions can cause them to diverge.Secondary effects). We analyze how age tracking affects these primary and second effects. This is done by separating genetic and shared environmental effects relating to performance from those net related to performance.
The first reason early tracking could amplify secondary effects is because it may increase the chance for children from disadvantageous backgrounds to attain lower levels education than would be expected given their potential. In this paper, we argue that children with disadvantaged backgrounds may be more likely to attain lower levels of education than would be expected based on their potential. It can be fatalBecause disadvantage is the driving force, a younger age for tracking decreases genetic influence and increases shared environmental influence, especially among high-performing students. Second, children with advantages could be able to achieve higher levels than their potential merits by being tracked early. We believe such CompensationBy advantage, it would be the opposite prediction. A lower tracking age reduces genetic influence and increases shared environment influence, especially among low-performing students. We are unsure if either of these mechanisms is responsible for secondary forms of inequality of opportunities. The second question we ask is how does the moderation between genetic and shared environment influence on educational attainment via tracking age relate to the performance level of children.
These questions can be answered by analyzing data N=8847 twins from The Netherlands Twin Register. The Netherlands is a great example of the effects of tracking. The tracking process begins at a young age, around 12 years old. It is broken down into five tracks. The three lowest tracks prepare students for senior secondary vocational training. A middle stream prepares students for tertiary vocational college. Finally, an academically-oriented track allows them to access university. Secondary schools have different tracking requirements. Some schools place students in one track right away when they start secondary school, based on the recommendation of their primary school teacher. Some schools don’t decide on the track level immediately, but they first place students in a class that has at least two levels. Secondary school teachers will observe students for 13 years in such a heterogeneous classroom to determine the best track level. This variation allows us study the effects of delaying definite track to a later age on inequality of opportunity.
The advantage to studying tracking within a country is that other aspects such as the education system and country are kept constant. However, it is possible for children to be tracked later or immediately. This is not always the case. This means that there is a possibility of a false association of interest (e.g., between delayed tracking and genetic or environmental influences on educational attainment) or suppression. The Dutch children and their parents choose the secondary school that offers their track level. There are no restrictions on choice: there are no catchment areas, no tuition fees, and often many schools within close proximity. Research shows that school quality, distance to school and student composition are all important criteria. Some studies also show that the weight of the criteria can be affected by parental SES and child performance.21,22,23. We don’t know of any study that explicitly assessed the importance of a homogeneous and heterogeneous class as a criterium. Herweijer, Vogels are different.24Asking parents how important they consider the number of school tracks can help you get to a better understanding. These results show that this criterion is not important and does not depend on parental SES, denomination, political views, or whether the child attends a heterogeneous school. This is somewhat encouraging, but homogeneous or heterogeneous classes could still be associated with other school characteristics that may lead to selectivity. Van Elk et al. 25Report for those with a MAVO Track recommendation (old label VMBO-t). Children from higher-SES areas and more urbanized areas are more likely to attend a heterogeneous school, but none of these effects are significant in multivariate analyses. Borghans et al. 26For those with a HAVO or VWO recommendation, higher-SES and better-performing children are less likely than others to attend a heterogeneous classroom. Although the evidence is not strong, we still examine whether two possible confounders, family SES and child’s school performance, can influence the decision to delay or not track.
The rest of the introduction will give a brief theoretical background about how genes and the environment influence child development. Next, we will discuss how the tracking age of an education system affects these environmental and genetic influences.
Modern industrialized societies have a heritability of cognitive abilities as measured by psychometric IQ tests. This means that 55% of variation in cognitive ability between children is due to genetic differences.27. High heritability doesn’t mean that cognitive or other forms are coded in genes and emerge with maturation.environment effects. Child Dev. 54, 424 (1983).” href=”http://www.nature.com/articles/s41539-022-00122-1#ref-CR28″ id=”ref-link-section-d2335215e577″>28. Genetic differences between people are only expressed through interactions with their immediate environment29. For example, a child might be gifted at reading but not if there is no book around. This will limit their genetic potential. Higher levels of heritability can be expected if children have better opportunities to develop their potential.
Transactional models suggest that children are selected for different environments partly based upon their genetic predispositions. These environments have an impact on their development.environment effects. Child Dev. 54, 424 (1983).” href=”http://www.nature.com/articles/s41539-022-00122-1#ref-CR28″ id=”ref-link-section-d2335215e588″>28,30,31In our case, this would indicate that children with a natural aptitude for reading are more inclined to pick up books than children without the aptitude (active genetic environment correlation). Parents are also more likely to purchase books for children who are eager to learn than for those with little interest (evocative genetic environment correlation). This means that children who have a predisposition to reading are more likely to be able to read than children who are less inclined. This means that genetic differences can be expressed in terms developmental differences if the environment is large enough to allow children to control their environment. Herd et al. 32It has been shown, for instance, that women became more like mens when their access to education was less restricted.
The organization of the education system is an important social context that can impact whether children reach their potential. It could be possible to optimize the learning environment for children according to their genetic predispositions by tracking them into different ability categories. In practice, however, it is possible that a child ends up in a particular track not only because of their genetic predispositions but also because of their socioeconomic background.33. The exposition below suggests that children realize their genetic potential earlier than they do now, and are less likely to attain higher education.
It is important to first clarify three points about the relationship between genes and family history before you do this. First, Jackson34Boudon stresses that primary effects also include transmissions of genetic material from parents and children.35Originally, emphasis was placed on sociocultural factors. These sociocultural factors may have a causal effect on children’s performance but not through the transmission genes. While parental SES can be causally influenced by parental genes, but not the other direction, and children’s performance via childrens genes, this would create a false association between parental SES with childrens performance. Primary (or secondary) The effects ofIt makes sense to exclude genetic transmission mechanisms from socioeconomic backgrounds. Second, and related to the previous, parental genes can influence both children’s and parents’ SES. Because of this, genes and family background can be correlated (called passive gene-environment correlation: For example, parents who are proficient in reading not only pass these genes onto their children but are also more likely have books in their household.36. It is more difficult to separate genetic and shared environment influences when there is geneshared correlation. However, we discuss in the methods section why it is not very problematic for us (see “Assumptions about the fitted twin models”) Third, based in part on the theoretical ideas we apply, it has been suggested that children have more options in high-than low-SES settings to choose and evoke their environment based their genetic proclivities.31. The genetic influences on cognition have been shown to be more prominent in children from high-SES families in the US than in the Netherlands or other Western European countries.37. Heritability is the same across parental SES for our measure of educational achievement.38. In our study, the potentially complicated matter of interaction between genes & parental SES is unlikely to play any role.
The Netherlands has a tracking system at the end primary school. It is based partly upon a national standardized score (CITO), and partly on teacher track recommendations.39. There is a strong correlation between CITO-scores, socioeconomic background, and CITO-scores40. Moreover, De Zeeuw et al. 38Research shows that children with a low genetic propensity to educational attainment still score high on CITO-test if their family is high-SES. This supports the idea of performance differences by family background not only reflecting differences in talent but also how conducive the home environment is to children reaching their potential.34. High-SES parents are more likely to have the financial resources to invest in education, e.g. Private tuition and learning materials can be provided by private tutors. This gives high-SES parents more cultural capital, which is beneficial for their children’s academic success.41. The more their parents influence their environment, the younger they become. Children grow up and have more control over their peer groups, academic courses, and other learning experiences (Tucker Drob et al. 2013, p. 351).31. This should allow for these choices to more easily be guided by genetic predispositions. This is consistent with the findings that cognitive variance due to shared environments decreases from approximately 60% in infancy to almost none in adolescence. Genetic variance, however, increases from 25% up to 70%31,42If children’s choices are not genetically controlled, more autonomy would lead to shared environmental effects that are non-shared and less genetic. Similar trends could be observed with educational performance. There is a lot of overlap between cognitive ability in the Netherlands and (genetic) shared environment.43. This would indicate that, the earlier children are sorted into different educational levels based on performance differences, the less educational attainment is dependent on genes, and the more on the shared environments (i.e. Primary effectsBecome stronger. Primary effects are not dependent on the influence that family background has on performance. They also depend on the effect of performance on attainment. We don’t know why the former would change with tracking aging.
Even taking into account educational performance differences, children of high socioeconomic backgrounds tend to go further than those from low socioeconomic backgrounds (Secondary effects)44,45. This is also true for the Netherlands: children who have the same CITO-score as their parents receive an average higher track recommendation and are more likely to attend a higher school the higher educated they are.46,47. Family background is more important than the family background for educational decisions made at a young age than at an older age.48,49. Children are more likely to be influenced by other factors than their true potential when they are younger.50. This could be detrimental to those from disadvantaged backgrounds and beneficial for those with advantaged backgrounds.
One example of hindering is when teachers would (unconsciously), underestimate children who don’t talk, dress, or behave according the cultural codes that most teachers know. These codes are often the codes of the higher class.41. They would also overestimate children of the higher classes who have this cultural capital (a form compensation). Dutch teachers indicate that 98% of students behave well and 43% consider their home situation. This is especially true for students whose CITO score is at the border between two tracks.51. It is hard to determine if these considerations are due merely to teacher bias, or if teachers are correct in seeing them to be important predictors of students’ success in a particular track.52. Teachers should expect to allow the home environment to weigh less in their track recommendations as children age more independently.
Different socioeconomic backgrounds can have different parenting styles when it comes to guiding their children in making the right choice. Relative risk aversion theory suggests that all parents care about avoiding downsizing for their children.53. For children from high-SES families, a move to one of the lower tracks would result in downward mobility. However, this does not apply to children from low-SES families. Therefore, high-SES parents are more likely to interfere in their children’s educational lives than low-SES ones. It has been shown that teachers who are high-SES parents often recommend a higher track to their children than teachers with lower incomes.51,52,54. Children younger than 10 years old are more dependent on their parents. This would make it more common for low-SES children (hampering) to enter a track lower than their true potential, and more likely that high-SES children (compensation) will enter a track higher than their true potential. Higher tracks are more difficult and therefore are perceived to be more risky. This is especially true for parents with low socioeconomic status who may not be familiar with the higher tracks. This uncertainty is magnified if tracking takes place at a younger age, as the timeframe ahead can be longer. This increases the risk of making poor decisions for those with lower socioeconomic backgrounds.55. As students age, parents and students gain more information about students’ abilities and success chances, and students are able to make their own decisions. This makes it easier for talented students from low-SES backgrounds to choose a high track, even if it is not familiar to their parents.56,57.
These arguments for secondary effect predict, similarly to the primary effects that tracking happens at a younger age. The entered secondary track level is more dependent on the shared environment, which curbs the expression of genetic differences. We expect the same applies to secondary educational attainment, as the entered track level is very important for secondary educational diplomas that children earn.
H1. H1.
age.
We have assumed, implicitly, that early tracking increases both hampering as well as compensation. However, it is possible that one of these processes could be the main driver of secondary results.
If children have higher educational attainment, the arguments against hampering are stronger. Teachers might find a lack in cultural capital particularly problematic for low-SES students who are enrolled at the highest education levels. The track that is most suitable for low-performing children will be familiar to their parents. However, the higher their performance, the more likely it becomes that low-SES parents won’t know the track that is appropriate. This suggests that the shared environment (families) is more important than the actual potential (genes), especially among high performers. If early tracking is most harmful, it is mainly among the high-performing children. Early track increases the importance of the shared environmental and lowers the importance of genes.
H2a. H2a. The moderating effects of age of monitoring on genetic and shared environment influences on educational attainment (net performance) are stronger if children perform better.
The argument for compensation, so that children with a favorable background are tracked higher than their actual level of performance, is more applicable if the performance levels of the children are lower58. A child with a low performance score is more likely to experience downward mobility. High-SES parents must take more steps to prevent this. However, high-SES children who perform well will not require compensatory measures. They will be on a high track anyway. In the case of compensation, it is important to consider the family background (or shared environment) as it can make a difference, especially for low performers. If early tracking is primarily used to increase compensatory actions, it would be expected that early tracking will increase the importance of the shared environmental and decrease the importance of genes, especially among children with low performance.
H2b. H2b.
It is not clear whether one would expect compensation (H2a), or mostly hampering (H2b), to be increased by the Netherlands’ immediate tracking compared to the Netherlands’ delayed tracking. It is possible that both are affected equally. In this case, the impact of tracking age in the Netherlands on environmental and genetic influences would not depend on children’s performance. Johnson et al. 59The US has lower educational attainment than Sweden, and the shared environmental effects are greater in the US for low intelligence. However, it is not as strong for high intelligence. These results could be interpreted as support for the compensation hypothesis (H2b), which suggests that the Swedish school system offers more equal opportunities to students than the US school system. This is in contrast to our argument that delayed tracking gives more opportunities than immediate tracking.
