Our Knowledge of Averages is Below Average!

In 1954 Darrell Huff published How to Lie with Statistics. It is still in print (W. W. Norton). No lie!

One reason this book has legs is that Huff made everyday statistical concepts come alive and understandable through clever examples and illustrations. Over the years I have enjoyed using this (obviously dated) cartoon from his book to illustrate the concept of “average”:


A naïve couple selecting the ideal location for their first home might pick Oklahoma City for the average temperature of 60.2 degrees Fahrenheit (about 16 Celsius). Expecting daily, chilly weather that grew neither cold nor hot, they would be in for a shock when they shivered in -17° F (-27° C) during February and staggered in 113° F (45° C) later in August!

The lesson: Be not fooled by the “Flaw” of Averages. Over the past few weeks, a slew of queries have come my way prompted by a recent meta-analysis led by Tinca J. C. Polderman of Vrije Universiteit (or VU—Free University) of Amsterdam. Professor Polderman’s team covered 2,748 individual research reports that included nature/nurture estimates on 17,804 traits. Note: In my world, a “trait” refers to a behavioral personality disposition, such as extraversion or ambition, but in Dr. Polderman’s world a “trait” is used far more inclusively—not only dispositions, but skin quality, height, intelligence, and disease-proneness. For convenience of reporting, these thousands of traits were clustered into 28 groups. The groups ranged from social values to metabolism rates and included behavioral traits, mental abilities, skin condition, and aging patterns. In other words, they covered the whole person. Their findings: Across all traits—or aspects of the person, 49% of our personality is inherited, while 51% is acquired through the influence of environment. Or, nurture wins by a hair over nurture.

BUT!!! Beware the Flaw of Averages. Some of the 28 groups show extremely high heritability (neurological, cardiovascular, personality disorders, structure of the eyeball, etc.) while others show much lower heritability (activities, alcohol-related disorders, height, weight, etc.). While it is true that every one of the 28 groups had a significant genetic component, the degree of heritability ranged from very low to very high. So to say that personality is half nature and half nurture is to ignore the importance of extreme variation in heritability of elements within this meta-analysis. To say that everything about us is half learned implies that everything about us is reasonably easy to unlearn, or change. That is not the case, however, as some parts of our person are far more resistant to change because of their hard-wired nature. Change advocates risk creating frustration and guilt by encouraging folks to undergo changes in parts of their personality without taking into account the specific heritabilities of those parts.

At the Center for Applied Cognitive Studies, where we are in the business of personality assessment, we focus primarily on the Five-Factor Model of personality, or the Big Five. Across the Big Five supertraits and their constituent subtraits, on “average” about 60% of one’s level of the traits is inherited. However, that 60% “average” masks a higher heritability for Extraversion than for Accommodation/Agreeableness. And within each of the five supertraits (Need for Stability, Extraversion, Originality/Openness, Accommodation/Agreeableness, and Consolidation/Conscientiousness), the subtraits that comprise them show a wide range of heritability, such that E2: Sociability/Gregariousness shows greater heritability than E3: Activity Level. For example, growing up in a physical culture environment (early morning yoga, a 10-mile run, ending with a dip in a mountain lake) could turn someone who otherwise might have turned out to be a couch potato into a physically active worker who doesn’t sit still for long.

And, to further complicate the notion of “average,” even a specific subtrait such as “sociability” comprises dozens if not hundreds of specific correlated behaviors, such as preference for loud/quiet environments, memory for names, ease of talking, preference for reading/writing, comfort with solitude, demonstrativeness, ease of making conversation, level of tact, level of humility, sense of humor, and so forth. And, some of these behaviors (e.g., preference for loud/quiet) have higher heritabilities than others (e.g., level of tact).

So why do we even bother to estimate the level of heritability of the various elements of personality? For me, the importance of knowing that every aspect of personality has some genetic basis is that it helps me and others to avoid two major mistakes of the 20th century. First is the mistake of Adolf Hitler, who acted on the assumption that personality is 100% inherited and therefore not changeable. And second,  the mistake of behaviorists such as B. F. Skinner, who reacted to Hitler by assuming that they could train anyone to be anything, given the proper resources.

mendel chart

Today, the more reasonable outlook is interactionist—our inheritance provides a starting point, and we build on it for a lifetime. When we attempt to change a behavior and experience strong resistance, either we are fighting a losing battle with our genes, or our methods are wrong. I suppose that we will never give up looking for better methods to change people, but I urge my colleagues to focus those efforts on how to change disabling qualities such as addiction, rather than on how to change our basic, inborn temperaments that simply make us interestingly different from one another. Don’t try to increase your sociability score—rather, build a satisfying life around your comfort with relative solitude!

John Milton once said something to the effect that we are free not to sin. That is also true with our personalities—we are free to accept who we are, to not try to change. But when who we are runs afoul of the law or our associates, we are free to figure out how we need to change or remain the same.

I cannot stop without mentioning a profound ethical dilemma for the 21st century. As scientists discover specific genes that are associated with specific personality elements, other scientists will attempt to devise methods to manipulate personality elements by modifying or otherwise affecting those genes. We need guidelines for reasonable use of this genetic knowledge. Is surgery to make one more sociable to be permitted?

  1. No comments yet.
  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: