|Home > Understanding Heritability
by Alan Feldman
High School Teacher of AP Psychology, AP History and Mathematics
Glen Rock High School
Glen Rock, New Jersey
Clearing Up the Confusion
Heritability is a term that appears in introductory college and AP-level psychology texts in the intelligence chapter. It also sometimes appears in a section discussing the relative influences of nature and nurture. Most of the time the discussion is rather brief, and frankly, some textbooks incorrectly describe the concept. It is crucial, however, that AP teachers correctly understand this concept and relate it accurately to their students. Do not be discouraged: you can fully understand the concept of heritability!
Heritability is a very elusive and difficult concept to understand. It is a correlational term that has only a positive value between 0 and 1 but can be stated as a percent ranging from 0 to 100 percent. For example, you may have heard the heritability of height is .9, or the heritability of intelligence is .7. But what does this mean?
To test yourself, decide whether the following statements are true or false. By the end of this article, you will find out if you were correct.
Explaining the Concept
- If a person has a disorder that has a heritability of 1, then the person will suffer from the disorder.
- The heritability of having fingers on each hand is 1 or close to 1.
- Heritability and inherited are nearly the opposite in meaning.
- In the mid-1950s in America, the heritability for wearing earrings was very high.
- The heritability of identical twins is 1.
- As the environment gets more similar for individuals of very different heritabilities, heritability increases.
- The heritability of a group of individuals with relatively similar heredities in very different environments is relatively low.
I will first explain heritability and then return to a discussion of these examples to clarify the concept.
Heritability statistics resulted from the work of Francis Galton, who was Charles Darwin's cousin. Heritability is a group statistic that makes no sense when applied to one person. Heritability is the extent to which differences in the appearance of a trait across several people can be accounted for by differences in their genes. Heritability does not reflect the extent to which traits are passed down from parent to offspring.
Heritability estimates are usually generated by twin studies. For example, if identical twins usually have similar IQs, but fraternal twins sometimes do not, the differences between the fraternal twins must have resulted from what is different between them that is not different for the identical twins. It is quite reasonable to assume that the difference is genetic. So, if IQ is very heritable, it means that individuals with the same genes have similar IQs and individuals with different genes have relatively different IQs.
Problems result when consideration is given to what heritability doesn't mean. When discussing an analysis of the heights of fraternal and identical twins, the behavioral geneticist Robert Plomin found the results indicated heritability for height to be 90 percent. This is often misinterpreted, for it does not mean at all that people's heights are determined 90 percent by genetics and 10 percent by the environment.
Heritability estimates are not measures of the importance of genes in the production of a trait, nor how modifiable a trait is to environmental influence. Heritability statistics do not in any way reflect the relative importance of genes in explaining traits. Heritability estimates only reflect what causes the variation in traits; they say nothing about what causes the traits themselves.
Variation and Causation
If this seems quite confusing -- good! You are on your way to understanding heritability. There is an important meaningful distinction between explaining the variation among traits and explaining the causation of the traits. Think of snowflakes and assume that both high humidity and low temperatures are essential for snowflakes to form. So, if on a particular day humidity is very high at the North Pole and very low at the South Pole, snow will only fall at the North Pole. This can fully account for the variation in snowfall between the North Pole and the South Pole on that day. However, this accounting for variation should not be taken to imply that low temperature is unimportant in causing snow; obviously temperature is a very important factor. When a factor does not vary across situations, it cannot account for the variations in outcomes across those situations, but this does not mean that this constant factor is unimportant in causing the outcome.
So, accounting for variation tells us little about causation. This is also possibly true when explaining our traits. It is very possible that genetic factors account for 90 percent of the differences seen in people's heights, but this does not mean that genetic factors are necessarily more important than environmental factors in causing people's heights.
Twenty-five years ago, the scientist Richard Lewontin devised an insightful demonstration to show this. Imagine planting ordinary, genetically diverse seeds into two radically different environments and then allowing them to grow to their full heights. One environment is very deprived, with just barely enough light, nutrients, and water for survival. The other environment is enriched with ideal amounts of light, water, and nutrients. All of the variation in height within each tray must be due to the genetic diversity of the seeds, since the seeds developed in identical environments, and therefore the variation observed in the heights of the plants within a tray cannot be attributed to differing environmental factors. So regardless of the environments in which the plants grew, heritability is 1, or 100 percent, within each tray. Yet obviously environmental factors played a large role in each individual tray, so even if heritability is 100 percent, the environment can have very powerful effects on the appearance of a trait.
To restate: since the environment for the seeds within each tray is identical, the genetic diversity of the seeds must account for the observed height differences, so the heritability within each tray is very high, probably close to 1. Yet the heritability between the trays is very low since the environments are so disparate and the genetic material is similar. The large observed differences in height between the trays are due to environmental differences.
Understanding The Bell Curve
Another example, this one from Stephen Jay Gould, is illuminating. Think about the relative heights of men in a poor village in an underdeveloped country 100 years ago. The average height for these malnourished men might be 5 feet 2 inches. The heritability in observed heights within this particular society can be quite high; men of tall fathers are on the average considerably taller than men of short fathers. However, this does not mean that a program of improved sanitation and nutrition could not significantly raise the average height of this group in a few generations.
This reflects one of the many significant errors of Herrnstein and Murray's controversial 1994 book, The Bell Curve. The error is to assume that genetic variation that can account for variation within a group is also the reason for the variation between groups. The converse is also true. Even if the heritability of a trait is 0, it can still be tremendously affected by genetic factors.
Imagine taking a bunch of genetically identical seeds (cloned if necessary) and scattering them in a variety of soil types. All the variation in the plants' heights would be accounted for by the environment (since the seeds are genetically identical), so the heritability of height here is 0. Yet, no one could reasonably argue that the genetic information in the seeds has no effect on the plants' heights. (Compare dandelion seeds and those of the California redwoods.)
So a trait can be importantly affected by genetic factors even if it is not heritable at all. If both fully heritable and fully nonheritable traits can be significantly affected by environmental and genetic factors, then how do they differ? In terms of the extent to which the environment and genes might affect traits, there is no difference. Heritability estimates tell us nothing about how genetic and environmental factors affect traits. So remember that when someone might mention or report that IQ is highly heritable; it does not mean or imply that environmental resources should not be employed to raise IQ scores. To say that high heritability estimates of IQ scores are less open to environmental interventions is false. It is false because environmental interventions can greatly affect even maximally heritable traits.
The Usefulness of Heritability Estimates
You might thus ask what the point or use is of heritability estimates. The answer is that they can estimate the sources of differences among people, but only for a particular population, at a particular time, and in particular circumstances. Heritability estimates cannot be generalized, because as the situation changes, the estimates can change drastically.
To illustrate this, think of the snowfall example again. When comparing the variation in snowfall between the North and South Poles, the relative humidity is essential. Yet when comparing snowfall in a variety of locations in a humid mountainous country near the equator, humidity is now virtually meaningless in explaining variation. In this case, the temperature differences account for the variation in snow.
Reviewing the Concept
Now let's return to the true/false statements from the start of this article:
1. If a person has a disorder that has a heritability of 1, then the person will suffer from the disorder.
This statement is most definitely false. Consider, for example, the disease phenylketonuria (PKU), which has a heritability of 1. This can result in mental retardation, yet the retardation can be prevented if phenylalanine is removed from the person's diet at birth, and if the person watches their diet very carefully.
2. The heritability of having fingers on each hand is 1 or close to 1.
This statement is also false. The heritability of having five fingers on each hand is very low, close to 0. This is because the source of the variation here is usually environmentally caused early on due to teratogens, and for adults due to accidents.
3. Heritability and inherited are nearly the opposite in meaning.
This statement is true. Although heritability might seem like it reflects how inherited a trait is, it doesn't do this at all. Actually, it is paradoxical. The more inherited a trait is, the less heritable it is. So if the liberal establishment ever gets schools to be relatively equal in terms of facilities, materials, and instructional staff, then the heritability of student achievement would increase. Although it might seem that inherited events are highly heritable, they are not.
4. In America in the mid-1950s, the heritability for wearing earrings was very high.
This statement is also true. Things that seem very noninheritable may have high heritability estimates. Fifty years ago, when virtually only women wore earrings, the heritability of wearing earrings was very high. This is because certain genes were usually found in earring wearers that were not found in non-earring wearers. Since heritability estimates are correlational, they tell us nothing about what causes a behavior. The gene whose presence is correlated with wearing earrings need not play any direct role in causing that behavior.
5. The heritability of identical twins is 1.
This statement is false. As a matter of fact, identical twins have a heritability of 0, since any variation in their behavior cannot be accounted for by genetic differences. (I know that some identical twins are not fully genetically identical -- but that topic is for another article.)
6. As the environment gets more similar for individuals of very different heritabilities, heritability increases.
This statement is true. This is because as the environment gets more similar, they become less of a source of variation for the individuals.
7. The heritability of a group of individuals with relatively similar heredities in very different environments is relatively low.
This statement is true as well. This is because as the heredities become more similar, they are less of a source of variation for the individuals in the group.
To sum up: it is always important to remember that the concept of heritability and the issue of modifiability are completely unrelated.
Gould, Stephen Jay. "Curveball." New Yorker 28 (November 1994): 139-149.
Griffin, G. A. Elmer, Diane Halpern, Frank Miele, Vincent M. Sarich, Michael Shermer, and Carol Tavris. Various articles in Skeptic 3 (1995) with theme "Race and I.Q.: What's Behind the Bell Curve?"
Moore, David. The Dependent Gene: The Fallacy of Nature vs. Nurture. New York: W. H. Freeman and Co., 2001.
Plomin, Robert. Nature and Nurture: An Introduction to Human Behavioral Genetics. Stamford, Connecticut: Wadsworth, 1996.
Ridley, Matt. The Agile Gene: How Nature Turns on Nature. (Originally published in hardcover as Nature via Nurture: Genes, Experience, and What Makes Us Human.) New York: Perennial, 2004.
Alan Feldman teaches AP Psychology at Perth Amboy High School in Perth Amboy, New Jersey. He is a veteran of numerous psychology summer programs, has been a Reader of the AP Psychology Examination since its inception in 1992, is an AP Psychology consultant for the College Board, and is currently on the AP Psychology Development Committee. He is also an adjunct instructor at Middlesex County College in Edison, New Jersey, where he teaches introductory, child development, social, and abnormal psychology courses. He served on the TOPSS Executive Board for four years. In 1994 Alan received the high school teaching award from Division Two of the American Psychological Association.