How difficult is it to change the typical distributions of a trait, such as aggression, substance abuse, suicide attempts, as the distributions differ between males and females? Nature versus Nurture debates build off this question in two ways. One is the matter of fixity versus flexibility in the development of traits in individuals over their life course. The other is the relative degrees of hereditary versus environmental influences on the variation of the trait between versus within groups. (“Groups” here refers to males or females, but the question might be extended to socially defined racial or socio-economic groups.) What have these two Nature-Nurture issues got to do with each other?
One answer, on which the rest of the post elaborates:
If a certain three assumptions are made, then partitioning of the variation of the trait within each group can be interpreted as indicating the relative degrees of hereditary versus environmental influences on the variation of the trait, even if the specific genes making up the hereditary influence are yet to be identified. Once they are identified, the hypothesis that those genes are involved in variation between the averages of the groups can be examined. If the hypothesis is confirmed, then we have gained knowledge about the differences between the averages of the groups. The ways that this knowledge about difference is pertinent to the other Nature-Nurture issue—fixity versus flexibility in the development of traits—depends on whether the study of difference is seen as a way to probe development or as something that takes development as given.
1. First a clear view is required of what it means to partition variation in a trait. As a previous post describes, partitioning of variation originated in agriculture between the two World Wars. Let us “imagine many genetically defined and reproducible varieties of plants raised in many locations. (Genetically defined is easy to visualize if individuals of the variety are clones of each other, as they are when a plant is grown from a cutting of another plant. It can also be thought of as genealogically defined, e.g., each individual is a grandchild of the same two pairs of grandparents.) In each location multiple individuals—”replicates”—are raised. The individuals in all the locations for all the varieties are all measured for, say, height, and there is variation among those measures. The average or mean height for each variety can be calculated and there is variation among those variety means. [T]he ratio of variation among the variety means [to] the total variation across the whole data set” can be calculated. The same can be done for location means and for the means of the variety-location combinations. That’s what it means to partition variation. [diagram] “With humans, we can replicate a variety in the form of identical or fraternal twins, so, with only two individuals in each genetically defined variety, there will be many, many [variety-location combinations] for which there is no measured value. That means that the [partitioning of variation] must be more approximate.” [diagram]
2. Now, heritability is the technical name for the ratio of variation among variety means to the overall variation for the trait. “Notice that heritability is derived from measurements of an observed trait; there has been no mention of genetic or environmental factors involved in producing the trait.” This name is a source of confusion given the colloquial but completely logically unrelated idea that a trait is heritable if a parent passes on to an offspring genes that determine the trait. Indeed, it is commonplace even for researchers experienced in partitioning variation, to state something like the following (from wikipedia): “Heritability is the proportion of observed differences on a trait among individuals of a population that are due to genetic differences.” If “genetic differences” is read as differences in the genes related to the trait, then this is either simply mistaken (Taylor 2014) or an indication that a certain assumption has been made, namely, that “a gradient of a measurable genetic factor (or composite of factors)… run[s] through the differences among variety means” (Taylor 2014, 28), albeit with the identity of those genetic factors unknown. [diagram]
3. A prior assumption of “underlying homogeneity” is needed for that assumption to be plausible. To explain what this means, consider one way to estimate heritability for a human trait, namely, comparing the similarity of identical twins, who share all their genes, with the similarity of fraternal twins, who share a smaller %; in both cases, the twins are raised together. Even if the similarity between twins or a set of close relatives is associated with the similarity of yet-to-be-identified genetic factors, the factors may not be the same from one set of relatives to the next, or from one environment to the next. In other words, the underlying factors may be heterogeneous. [diagram] We need to assume that that is not the case for there to be “a gradient of a measurable genetic factor (or composite of factors)… run[ning] through the differences among [the] means of the twins.”
4. Thus far partitioning of variation is something applied to data within a given group, such as twins in a defined human population, not between groups. The third assumption is that, if there is a gradient on a measurable genetic factor within group A, and another one within group B, then it is plausible that differences between the groups—strictly, between the averages for the two groups—are related to the genetic factors of the within-group gradients. [diagram (where gf’s & ef’s = genetic and environmental factors)] To see that it is not logically necessary for within-group causes to extend to explanation of differences between the averages of groups, consider the following analogy (Taylor 2014, 40):
Lindman’s (1992) textbook illustrates a cautionary note about [partitioning of variation] using high school students’ test scores in algebra viewed in relation to their teacher and school as an example. The students within a school are randomly assigned to a teacher in that school. Lindman notes that a significant difference between the means for the locations (schools) “is likely to be interpreted as due to differences in physical facilities, administration, and other factors that are independent of the teaching abilities of the teachers themselves… [However, d]ifferences between teachers in different schools are part of the [average location or school difference], and the observed differences between schools could be due entirely to the fact that some schools have better teachers [or] some schools have smarter children attending them” (Lindman 1992, 194).
Lindman could have added that the observed differences between schools could be due to any combination of factors, such as students respond worse to teachers whose attention is distracted because their school’s administrators insist more on detailed documentation of student performance. [Partitioning of variation] cannot help researchers hypothesize about the difference in the average scores from one school to the next when the teachers are replicated (in their students’ test scores) only within schools.
5. If the three assumptions are made, then partitioning of the variation of the trait within each group can be interpreted as indicating the relative degrees of hereditary versus environmental influences on the variation of the trait, even if the specific genes making up the hereditary influence are yet to be identified. Once they are identified, the hypothesis that those genes are involved in variation between the averages of the groups can be examined. If the hypothesis is confirmed, then we have gained knowledge about the differences between the averages of the groups.
6. The original question then becomes how is this knowledge about difference pertinent to the other Nature-Nurture issue, namely, fixity versus flexibility in the development of traits in individuals over their life course? As noted in Taylor (2014, 2-3), the study of difference may or may not shed light on development:
Biologists often study abnormalities in order to gain insight about typical processes of development. In Swyer syndrome, for example, a child with XY chromosomes has female, not male, external genitalia. The influence of estrogen and progesterone on development is illuminated by the finding that, if these hormones are administered at puberty, breasts can develop and regular menstruation can occur (Michala et al. 2008). Variation under a more or less normal range of conditions can also help biologists understand typical development. The age at which a baby comes to walk, for example, varies according to whether it sleeps on its back or stomach, has been swaddled or carried around in a sling, and so on, and this variation has been related to the timing and degree to which the baby uses its upper body muscles (Fausto-Sterling 2014).
Difference can, however, be studied without providing much insight about development. For example, the eyes of fruit flies, normally red, are sometimes white. Biologists identified the location on the chromosomes that corresponds to the white-eye mutation early in the 1900s and later investigated the pigment-formation metabolic pathway and the enzymes involved as fruit fly eyes develop the normal or mutant color. Such knowledge says little about how eyes develop. Even on the narrower question of how eyes get to have color, a lot has to be known already about the development of the eye as a whole to make sense of how and when during the fly’s development the enzyme produces color in the eye.
Which line of research on difference would follow from confirmation of the hypothesis—difference as a way to probe development or difference that takes development as given?
Future posts will take up this last question as well as the intuition that the two Nature-Nurture issues have something to do with each other in the context of calling the three assumptions into question.
Fausto-Sterling, A. (2014). “Letting go of normal.” Boston Review (March/April), http://www.bostonreview.net/wonders/fausto-sterling-motor-development (viewed 21 June 2014).
Lindman, H. R. (1992). Analysis of Variance in Experimental Design. New York: Springer-Verlag.
Michala, L., D. Goswami, et al. (2008). “Swyer syndrome: Presentation and outcomes.” BJOG: An International Journal of Obstetrics & Gynaecology 115(6): 737-741.
Taylor, Peter J. (2014) Nature-Nurture? No: Moving the Sciences of Variation and Heredity Beyond the Gaps.