Subtitles section Play video
- There's this classic thought experiment
about two gardens.
So imagine one garden and it has
the perfect amount of sunlight,
and then you plant different flowers in that garden.
And you can look and say,
"Well, the differences between them
in these plants' heights is about the seeds that I planted.
They all have the perfect environmental conditions.
The differences between them are genetic."
Now imagine another garden, and it is neglected.
If you looked at that and you said,
"All those plants are stunted and dead,"
and then you said,
"Ah, the difference between these two gardens
is that these plants in the neglected garden
are genetically shorter than these plants
in the rich, well-resourced garden."
We can't then say that all the differences
between tall and short plants
across very different environments
are in fact due to genetics.
That's a really obvious thing when it comes to plants,
but it's a really difficult thing for people
to keep in mind when we're talking about individual
versus group differences.
This relationship between
heritability and response to the environment
is a complicated one.
'Heritability' is the proportion of variation
in a population that's due
to genetic differences between people.
It's a confusing term because people misinterpret heritable
to mean 'genetically determined,
unmodifiable by the environment, or innate.'
And it doesn't mean any of those things.
But it's hard for us scientists to describe why.
The thing about factoring in environment in genetic studies
is that we already face
an enormous, statistical power problem.
Do you have enough people?
Do you have enough observations?
Do you have enough data to be able to reliably say
that what you're picking up on is real
and not just random chance?
Genetic studies that ignore the environment
and are just looking for the "main effects"
of genes are already requiring sample sizes
in excess of one million people.
So when you start to factor in
gene-by-environment interaction,
the power problem gets even bigger.
Statistics is hard, biology is hard.
A lot of times people haven't had exposure to this
since high school, if at all.
In the larger conversation,
we're trying to make sense of information about race.
We're trying to make sense of information about genes,
about education, about inequality.
I think it's really understandable
that people have a hard time thinking
about how all these things go together.
'Portability' is this idea that what we've discovered
in one sample or one group of people,
how much can we port that over to a new group of people,
and trust that things are gonna work in the same way?
So we see this all the time with our social science effects.
A minimum wage change in one state
doesn't have the same effect as a minimum wage change
in another state.
What I think people often miss
is that genetic causes can operate in the same way.
So, just because a gene influences the probability
of an outcome in this group of people,
to the extent that that's probabilistic
rather than deterministic,
to the extent that that genetic effect
is moderated by the environment,
we're not gonna be able to expect it to be portable
when we study a new group of people.
When I say that something's heritable,
I mean that within a group of people
that have experienced
roughly-equivalent environmental conditions,
genetic differences between them are related
to the phenotypic differences between them.
The genes of the seeds are related to the height
of the flowers.
But that doesn't allow us to make the leap
that differences between groups
that have experienced
really different environmental legacies are due to genetics.