WEBVTT

00:00:00.095 --> 00:00:02.460
The following content is
provided under a Creative

00:00:02.460 --> 00:00:03.870
Commons license.

00:00:03.870 --> 00:00:06.910
Your support will help MIT
OpenCourseWare continue to

00:00:06.910 --> 00:00:10.560
offer high-quality educational
resources for free.

00:00:10.560 --> 00:00:13.460
To make a donation or view
additional materials from

00:00:13.460 --> 00:00:17.390
hundreds of MIT courses, visit
MIT OpenCourseWare at

00:00:17.390 --> 00:00:18.640
ocw.mit.edu.

00:00:27.090 --> 00:00:28.340
PROFESSOR: Good afternoon.

00:00:30.990 --> 00:00:32.229
Good afternoon.

00:00:32.229 --> 00:00:34.730
I want to remind you that
a week from today

00:00:34.730 --> 00:00:36.130
is the second exam.

00:00:36.130 --> 00:00:38.630
The format is identical to
what you had, 50 multiple

00:00:38.630 --> 00:00:42.940
choice questions; 30 from the
book, Chapters 5 to 7; 20 from

00:00:42.940 --> 00:00:44.595
Sacks and the lectures;
short answers.

00:00:48.990 --> 00:00:53.980
So talking about tests,
why do we have tests?

00:00:53.980 --> 00:00:55.330
You live in a world of tests.

00:00:58.140 --> 00:01:03.760
Everything from SATs to, on your
horizons, maybe GREs or

00:01:03.760 --> 00:01:06.310
MCATs or LSATs.

00:01:06.310 --> 00:01:10.850
So why do people have tests,
besides making it miserable to

00:01:10.850 --> 00:01:13.460
be under 30?

00:01:13.460 --> 00:01:16.320
It's the revenge of
the older people.

00:01:16.320 --> 00:01:17.570
Why do people have tests?

00:01:25.140 --> 00:01:25.410
Yeah?

00:01:25.410 --> 00:01:25.590
AUDIENCE: [INAUDIBLE].

00:01:25.590 --> 00:01:28.590
PROFESSOR: One of them
is just verify

00:01:28.590 --> 00:01:29.250
that you know something.

00:01:29.250 --> 00:01:34.400
And people will call it sort
of have you learned stuff?

00:01:34.400 --> 00:01:37.120
And the test is meant
to be an estimate of

00:01:37.120 --> 00:01:39.240
what have you learned?

00:01:39.240 --> 00:01:42.770
But a second form of testing,
what people sometimes will

00:01:42.770 --> 00:01:45.640
call high-stakes tests in a
slightly different way, are

00:01:45.640 --> 00:01:49.500
ones about selection and
prediction, which sometimes

00:01:49.500 --> 00:01:52.040
involve achievement or
aptitude, things

00:01:52.040 --> 00:01:53.970
like SATs or GREs.

00:01:53.970 --> 00:01:58.060
So achievements are meant to be
estimates of what you know,

00:01:58.060 --> 00:02:00.640
advanced placement tests, how
much math you know or physics.

00:02:00.640 --> 00:02:02.510
Content in a given area.

00:02:02.510 --> 00:02:05.950
And sort of a broader sense
of aptitude, SATs or GREs.

00:02:05.950 --> 00:02:07.630
Now, one thing that's different
about today's

00:02:07.630 --> 00:02:09.729
lecture on intelligence, it'll
be a little similar when we do

00:02:09.729 --> 00:02:13.830
personalities, up until now, to
a first approximation, most

00:02:13.830 --> 00:02:16.430
of the time we've talked about
what we view as general

00:02:16.430 --> 00:02:19.900
principles of the human mind
and brain across everybody.

00:02:19.900 --> 00:02:22.200
How vision works across
everybody, memory across

00:02:22.200 --> 00:02:25.010
everybody, thinking
across everybody.

00:02:25.010 --> 00:02:26.630
And now, we're switching
to a domain.

00:02:26.630 --> 00:02:29.390
And it doesn't have to be this
way, but this is the way the

00:02:29.390 --> 00:02:32.370
world has done it, where
we emphasized

00:02:32.370 --> 00:02:34.640
differences between people.

00:02:34.640 --> 00:02:37.620
So when people take tests, it's
not to see that everybody

00:02:37.620 --> 00:02:38.670
took the test.

00:02:38.670 --> 00:02:41.570
What they're going to do is, you
get an A, or a B, or a C,

00:02:41.570 --> 00:02:45.440
or you get a high SAT or a
medium SAT, and schools use

00:02:45.440 --> 00:02:50.260
the differences to
think about you.

00:02:50.260 --> 00:02:54.340
And so one question is, from an
empirical sense, are these

00:02:54.340 --> 00:02:55.910
tests pretty good at
predicting things?

00:02:55.910 --> 00:02:58.990
And the answer is
they're decent.

00:02:58.990 --> 00:03:00.940
They're not terrific.

00:03:00.940 --> 00:03:07.100
So the tests survive, like SATs
or GREs, because there's

00:03:07.100 --> 00:03:09.850
correlations between these kinds
of tests before you get

00:03:09.850 --> 00:03:13.090
to graduate school for example,
and how faculty rates

00:03:13.090 --> 00:03:16.260
students or how productive
people are in terms of

00:03:16.260 --> 00:03:19.170
research or as a graduate
student in this case, or the

00:03:19.170 --> 00:03:21.220
GPAs, or things like that.

00:03:21.220 --> 00:03:25.020
And the correlations are
about 0.3 to 0.4.

00:03:25.020 --> 00:03:27.390
To give you an intuitive
sense, the correlations

00:03:27.390 --> 00:03:29.990
between sex and height, whether
you're a man or a

00:03:29.990 --> 00:03:31.620
woman and height,
is about 0.4.

00:03:31.620 --> 00:03:34.100
So of course lots of women are
taller than lots of men.

00:03:34.100 --> 00:03:36.000
On average, men are
taller than women.

00:03:36.000 --> 00:03:38.190
It's a 0.4 correlation.

00:03:38.190 --> 00:03:41.240
So this is in that range.

00:03:41.240 --> 00:03:44.900
So tests are pretty
good predictors.

00:03:44.900 --> 00:03:47.020
Everybody says that for things
like predicting college

00:03:47.020 --> 00:03:50.190
performance or graduate school
performance, a combination of

00:03:50.190 --> 00:03:52.990
tests and grades are better.

00:03:52.990 --> 00:03:56.860
The difference is grades take
a lot of years to produce.

00:03:56.860 --> 00:04:00.680
Tests are one traumatic
outing.

00:04:00.680 --> 00:04:06.370
So tests are just a
part of our life

00:04:06.370 --> 00:04:09.460
in selection processes.

00:04:09.460 --> 00:04:11.260
So now what is intelligence?

00:04:11.260 --> 00:04:13.780
And it's a loaded word.

00:04:13.780 --> 00:04:15.240
I mean memory is kind
of a neutral word.

00:04:15.240 --> 00:04:16.680
Attention is a neutral word.

00:04:16.680 --> 00:04:18.120
Intelligence is a loaded word.

00:04:18.120 --> 00:04:22.490
I mean nobody wants to be
low on intelligence.

00:04:22.490 --> 00:04:25.670
And so psychologists have
come out of this way.

00:04:25.670 --> 00:04:27.620
But it's kind of funny, because
there's not going to

00:04:27.620 --> 00:04:29.610
be a really deep definition.

00:04:29.610 --> 00:04:31.890
Maybe it's a lot of different
things and how they average up

00:04:31.890 --> 00:04:33.070
or something like that.

00:04:33.070 --> 00:04:34.850
But people have said well,
it's something about the

00:04:34.850 --> 00:04:37.440
ability to solve problems, to
understand and learn complex

00:04:37.440 --> 00:04:41.390
material, to adapt to the
environment, mental quickness.

00:04:41.390 --> 00:04:43.790
Those are all senses you might
have of something that you

00:04:43.790 --> 00:04:46.360
think of as intelligence
in a person.

00:04:46.360 --> 00:04:50.280
But whether they're one thing,
many things, it's very hard to

00:04:50.280 --> 00:04:51.040
pull that out.

00:04:51.040 --> 00:04:53.680
And we don't know too much
about that really.

00:04:53.680 --> 00:04:55.970
And also I want remind you of
something we talked about, but

00:04:55.970 --> 00:04:57.860
it's especially important
in this field.

00:04:57.860 --> 00:05:00.810
We talked about an experiment
versus correlations.

00:05:00.810 --> 00:05:03.190
An experiment, you have to
have a dependent measure,

00:05:03.190 --> 00:05:03.930
something you measure.

00:05:03.930 --> 00:05:06.170
But critically and uniquely,
you have an independent

00:05:06.170 --> 00:05:08.720
variable, something
that you vary.

00:05:08.720 --> 00:05:11.010
That's what makes a study an
experiment, as opposed to a

00:05:11.010 --> 00:05:12.320
correlation.

00:05:12.320 --> 00:05:15.800
So studies of intelligence of
every kind are almost always

00:05:15.800 --> 00:05:19.140
correlational and almost
never experimental.

00:05:19.140 --> 00:05:21.360
That means they're always
subject to lots of

00:05:21.360 --> 00:05:23.830
interpretation and many
factors in the world.

00:05:23.830 --> 00:05:27.290
And that's very important
to keep in mind.

00:05:27.290 --> 00:05:30.730
And again, they tend to focus on
variation in people, rather

00:05:30.730 --> 00:05:34.275
than what is intelligence
itself.

00:05:34.275 --> 00:05:37.060
And again, here's another
correlation

00:05:37.060 --> 00:05:40.050
of height and weight.

00:05:40.050 --> 00:05:43.200
That's a stronger correlation.

00:05:43.200 --> 00:05:45.040
So where did intelligence
start with?

00:05:45.040 --> 00:05:46.600
It's a fascinating story.

00:05:46.600 --> 00:05:48.930
And by the way, let me tell
you the SAT story, on the

00:05:48.930 --> 00:05:50.680
sidebar, for a moment.

00:05:50.680 --> 00:05:55.060
So SATs, nowadays do people,
some people get ready a lot

00:05:55.060 --> 00:05:57.300
for SATs by courses you take
and things like that?

00:05:57.300 --> 00:05:59.570
I read about that.

00:05:59.570 --> 00:06:02.450
When I took it, people weren't
that organized.

00:06:02.450 --> 00:06:04.165
But getting into college wasn't
that competitive.

00:06:07.010 --> 00:06:11.250
And the original goals of SATs
was what, do you know?

00:06:11.250 --> 00:06:12.610
It's really interesting.

00:06:12.610 --> 00:06:15.480
Because it ends up being the
thing, they say well, because

00:06:15.480 --> 00:06:18.260
people who are maybe under more
supported circumstances,

00:06:18.260 --> 00:06:21.150
more financially advantageous
circumstances, get more help

00:06:21.150 --> 00:06:22.600
to get ready for the SAT.

00:06:22.600 --> 00:06:25.770
Those things might amplify SAT
differences in a way that's

00:06:25.770 --> 00:06:27.490
not really about the person
to start with.

00:06:27.490 --> 00:06:30.480
The original goal of the SATs
was to have a national test

00:06:30.480 --> 00:06:33.700
that was fair for everybody, so
that schools like Harvard,

00:06:33.700 --> 00:06:35.870
for example, that tended to
admit the same kind of

00:06:35.870 --> 00:06:37.180
students from the same
small set of

00:06:37.180 --> 00:06:38.440
schools in the Northeast.

00:06:38.440 --> 00:06:40.860
So now, we have a test for
everybody across the country.

00:06:40.860 --> 00:06:43.960
And we can admit people on their
merit, rather than on

00:06:43.960 --> 00:06:45.610
small, social connectivity
networks.

00:06:45.610 --> 00:06:46.560
Does that make sense?

00:06:46.560 --> 00:06:48.990
And yet, SATs are now viewed
as a suspicious

00:06:48.990 --> 00:06:49.710
tool in that regard.

00:06:49.710 --> 00:06:51.610
It was originally meant
to be democratizing.

00:06:51.610 --> 00:06:54.650
It's like everybody in the
country takes the same test.

00:06:54.650 --> 00:06:58.346
It's not whether you have a
connection through a network,

00:06:58.346 --> 00:06:59.770
to our university.

00:06:59.770 --> 00:07:01.460
IQs have the same story.

00:07:01.460 --> 00:07:04.720
For good reasons, and we'll talk
about that, IQ tests and

00:07:04.720 --> 00:07:08.870
interpretations of IQ don't
have a good history.

00:07:08.870 --> 00:07:11.760
But the original goal was
pretty reasonable.

00:07:11.760 --> 00:07:16.440
So in France in 1904, they began
universal elementary

00:07:16.440 --> 00:07:19.510
education, that every
child goes to class.

00:07:19.510 --> 00:07:21.920
And Alfred Binet, a physician,
wanted an objective way to

00:07:21.920 --> 00:07:24.135
identify children who
needed extra help.

00:07:24.135 --> 00:07:25.590
A perfectly good idea.

00:07:25.590 --> 00:07:26.790
A child comes to school.

00:07:26.790 --> 00:07:28.300
Some children are ready to go.

00:07:28.300 --> 00:07:31.520
Some children need a lot of
extra help because they have a

00:07:31.520 --> 00:07:32.600
difficulty.

00:07:32.600 --> 00:07:36.200
So he developed a set of tests
where he probed many different

00:07:36.200 --> 00:07:39.640
abilities, like copying a
drawing, repeating digits,

00:07:39.640 --> 00:07:41.680
recognizing coins, explaining
why a statement

00:07:41.680 --> 00:07:42.700
did not make sense.

00:07:42.700 --> 00:07:45.230
They wanted to probe different
little areas of abilities, so

00:07:45.230 --> 00:07:48.110
if a child is terrible at
something, or two things, the

00:07:48.110 --> 00:07:51.540
teacher knows this child needs
some extra help to get going.

00:07:51.540 --> 00:07:54.480
So that's a good use of
testing, I think by

00:07:54.480 --> 00:07:55.880
everybody's criterion.

00:07:55.880 --> 00:08:00.990
But then testing sort of
took off, IQ testing.

00:08:00.990 --> 00:08:03.280
IQ stands for intelligence
quotient.

00:08:03.280 --> 00:08:05.360
It's a test to give
typical children.

00:08:05.360 --> 00:08:08.470
They give it often for the most
commonly used IQ tests,

00:08:08.470 --> 00:08:09.580
at many different ages.

00:08:09.580 --> 00:08:11.930
They'll test a large number
of individuals.

00:08:11.930 --> 00:08:14.550
And they'll compare what
nowadays what they think of as

00:08:14.550 --> 00:08:17.430
your mental age to your
chronological age.

00:08:17.430 --> 00:08:19.950
So your chronological age is
how long you've lived.

00:08:19.950 --> 00:08:23.390
And your mental age is how
well you score on a test.

00:08:23.390 --> 00:08:26.780
The most famous IQ test, the
most widely used, takes about

00:08:26.780 --> 00:08:27.950
two hours of testing.

00:08:27.950 --> 00:08:30.040
It comes from David Wechsler,
the Wechsler Adult

00:08:30.040 --> 00:08:30.870
Intelligence Scale.

00:08:30.870 --> 00:08:32.780
One for children.

00:08:32.780 --> 00:08:34.630
And within that test, what's
within that test?

00:08:34.630 --> 00:08:37.909
Because when people talk about
tests I think, whatever kind

00:08:37.909 --> 00:08:41.270
of test it is, it can only
about as good as the test

00:08:41.270 --> 00:08:44.740
design itself, to the extent
that it's any good.

00:08:44.740 --> 00:08:46.930
A ridiculous test is a
ridiculous test, even if

00:08:46.930 --> 00:08:49.440
society gives it to you.

00:08:49.440 --> 00:08:51.470
But tests can be better
than that.

00:08:51.470 --> 00:08:56.160
So what is inside the most
widely used IQ tests,

00:08:56.160 --> 00:08:57.860
especially in clinical
environments?

00:08:57.860 --> 00:08:59.380
So let's talk about
one of them.

00:08:59.380 --> 00:09:00.790
It has a bunch of
little tests.

00:09:00.790 --> 00:09:01.810
One is vocabulary.

00:09:01.810 --> 00:09:04.900
They ask you to define
words, similarities.

00:09:04.900 --> 00:09:06.560
They'll ask you a series of
questions, like how are an

00:09:06.560 --> 00:09:08.710
airplane and a car alike?

00:09:08.710 --> 00:09:11.250
Very simple arithmetic
operations.

00:09:11.250 --> 00:09:13.980
Digit spans, how many digits
can you remember.

00:09:13.980 --> 00:09:16.110
Information, they'll give you
a series of questions, like

00:09:16.110 --> 00:09:17.990
who was Martin Luther
King, Jr.?

00:09:17.990 --> 00:09:20.700
Comprehension, they'll ask you
things about the world, like

00:09:20.700 --> 00:09:22.880
why are there taxes and
things like that.

00:09:22.880 --> 00:09:24.030
And you give an answer.

00:09:24.030 --> 00:09:25.830
Somebody scribbles down
when you say.

00:09:25.830 --> 00:09:27.090
And they score you.

00:09:27.090 --> 00:09:28.050
You get two points
or one point.

00:09:28.050 --> 00:09:29.080
We'll talk about that.

00:09:29.080 --> 00:09:30.700
And then they add up these
scores in the subtests.

00:09:33.540 --> 00:09:35.710
So in the nonverbal verbal
tests or the performance

00:09:35.710 --> 00:09:38.610
tests, they'll show you a
picture with a thing missing.

00:09:38.610 --> 00:09:39.430
Have you did that a kid?

00:09:39.430 --> 00:09:40.650
I love those things.

00:09:40.650 --> 00:09:43.270
What's missing in the picture
and stuff like that.

00:09:43.270 --> 00:09:45.460
Blocks, they'll have you
rearrange them into various

00:09:45.460 --> 00:09:46.790
geometric design.

00:09:46.790 --> 00:09:49.820
Picture arrangement, you'll
get cartoon panels and you

00:09:49.820 --> 00:09:52.080
have to put them in the correct
order to tell a story.

00:09:52.080 --> 00:09:53.960
And they measure how accurately
and quickly you do

00:09:53.960 --> 00:09:55.040
these things.

00:09:55.040 --> 00:09:57.130
There's nothing amazing about
any of these things.

00:09:57.130 --> 00:09:59.100
They're probing a variety
of little

00:09:59.100 --> 00:10:02.060
abilities that you have.

00:10:02.060 --> 00:10:04.720
And again, they will take the
mental age and chronological

00:10:04.720 --> 00:10:06.580
age, divide it, multiply
it by 100.

00:10:06.580 --> 00:10:09.750
And they'll standardize it, so
that 100 is the population

00:10:09.750 --> 00:10:11.040
mean in the United States.

00:10:11.040 --> 00:10:13.150
Other countries do these
things, as well.

00:10:13.150 --> 00:10:15.050
Is that OK?

00:10:15.050 --> 00:10:17.490
So I mean a test can only be
as possibly good as the

00:10:17.490 --> 00:10:18.440
content of it.

00:10:18.440 --> 00:10:20.490
And what it measures or doesn't
measure depends on the

00:10:20.490 --> 00:10:24.320
specific conflict of it.

00:10:24.320 --> 00:10:26.450
They'll scale the scores out,
so you get a normal

00:10:26.450 --> 00:10:29.220
distribution, with a
peak in the middle.

00:10:29.220 --> 00:10:30.800
The raw score is the
number you get.

00:10:30.800 --> 00:10:33.030
The standardized one is
adjusted for your age.

00:10:33.030 --> 00:10:34.820
And it's a pretty important
concept, because we'll come

00:10:34.820 --> 00:10:36.010
back to this in a moment.

00:10:36.010 --> 00:10:37.810
The raw score is your
actual score.

00:10:37.810 --> 00:10:39.970
But then they adjust
it for your age.

00:10:39.970 --> 00:10:43.080
Because you don't expect a
five-year-old to do as well as

00:10:43.080 --> 00:10:49.540
a 15-year-old, to do as well
as peak 18 to 23-year-olds.

00:10:49.540 --> 00:10:51.870
And then bad things happen
as you get older.

00:10:51.870 --> 00:10:54.460
I'll tell you a bit of
them about that too.

00:10:54.460 --> 00:10:56.950
2/3 of the people are at plus or
minus 1 standard deviation.

00:10:56.950 --> 00:10:58.600
They set it for that.

00:10:58.600 --> 00:11:01.260
More than two standard
deviations up or down, are

00:11:01.260 --> 00:11:04.600
about 5% on either end of the
distribution of scores.

00:11:04.600 --> 00:11:06.020
And they norm it by age.

00:11:06.020 --> 00:11:09.690
Again, you don't expect a
six-year-old to do something.

00:11:09.690 --> 00:11:10.860
But you get some funny things.

00:11:10.860 --> 00:11:12.460
Not that you have to worry
about this in detail.

00:11:12.460 --> 00:11:14.380
But it gives you a sense.

00:11:14.380 --> 00:11:18.820
So imagine you're a
five-year-old and you take the

00:11:18.820 --> 00:11:21.140
test, on the last day you're
a five-year-old.

00:11:21.140 --> 00:11:23.045
Or imagine you have your
birthday today and

00:11:23.045 --> 00:11:24.400
you're six years old.

00:11:24.400 --> 00:11:28.120
By moving from five to six, from
Wednesday to Thursday,

00:11:28.120 --> 00:11:31.410
you've moved up a year
for the scoring.

00:11:31.410 --> 00:11:33.410
Because they go by
year, not by day.

00:11:33.410 --> 00:11:36.940
So if you would have scored 120
on the IQ test yesterday,

00:11:36.940 --> 00:11:38.890
when they say what should
you have scored for a

00:11:38.890 --> 00:11:41.630
five-year-old, the very same
test, the next day,

00:11:41.630 --> 00:11:43.590
you would score 100.

00:11:43.590 --> 00:11:45.910
Now, that's the extreme case.

00:11:45.910 --> 00:11:47.290
But you understand what
happens, right?

00:11:47.290 --> 00:11:49.130
Because here's how
well you did.

00:11:49.130 --> 00:11:50.440
That might have been
pretty good.

00:11:50.440 --> 00:11:53.950
120 is above the average
for a five-year-old.

00:11:53.950 --> 00:11:57.190
But the next day, if you're
a six-year-old, now for a

00:11:57.190 --> 00:11:59.140
six-year-old, it's
just average.

00:11:59.140 --> 00:12:02.850
So age has a big effect
on these things.

00:12:02.850 --> 00:12:06.810
Here's something sad for aging,
but good for you.

00:12:06.810 --> 00:12:08.850
Be smart now.

00:12:08.850 --> 00:12:11.920
A 40-year-old who
scores average

00:12:11.920 --> 00:12:16.280
would be a real undergrad.

00:12:16.280 --> 00:12:18.560
If I score like you--

00:12:18.560 --> 00:12:19.560
no, let me rephrase this.

00:12:19.560 --> 00:12:22.830
If you were compared to a
40-year-old, you would have

00:12:22.830 --> 00:12:25.320
100 IQ, you would
have a 230 IQ.

00:12:25.320 --> 00:12:28.310
Because age knocks down
the score as well.

00:12:28.310 --> 00:12:30.230
So things go up and
then things go

00:12:30.230 --> 00:12:33.210
down, in this problem.

00:12:33.210 --> 00:12:36.670
And usually they think of a
score below 70 as clinically

00:12:36.670 --> 00:12:38.850
concerning.

00:12:38.850 --> 00:12:41.400
And the top 2%, you get to join
Mensa and do whatever

00:12:41.400 --> 00:12:42.130
Mensa people do.

00:12:42.130 --> 00:12:44.310
Which I don't know exactly,
but it sounds interesting.

00:12:46.950 --> 00:12:49.710
An important thing about tests
of any kind, are two huge

00:12:49.710 --> 00:12:51.220
concepts, reliability
and validity.

00:12:51.220 --> 00:12:53.690
So reliability refers to the
idea that if you were tested

00:12:53.690 --> 00:12:56.900
more than once, you would
get the same score.

00:12:56.900 --> 00:12:58.650
The better IQ tests
work a fair bit to

00:12:58.650 --> 00:13:00.360
try to achieve that.

00:13:00.360 --> 00:13:03.190
Validity is the deeper
question.

00:13:03.190 --> 00:13:05.360
Does the test validly measure
what it's supposed to measure?

00:13:05.360 --> 00:13:08.180
So if we had you flip a coin and
said that's your IQ, you'd

00:13:08.180 --> 00:13:09.830
say, that's ridiculous.

00:13:09.830 --> 00:13:11.210
That's not a valid test.

00:13:11.210 --> 00:13:13.090
We could measure your height.

00:13:13.090 --> 00:13:14.830
That would be extremely
reliable measure.

00:13:14.830 --> 00:13:16.390
You'll have the same height
today as tomorrow, if we did a

00:13:16.390 --> 00:13:17.340
decent job.

00:13:17.340 --> 00:13:20.350
That wouldn't be a valid measure
of anything about any

00:13:20.350 --> 00:13:22.580
mental ability, never mind
something like intelligence.

00:13:22.580 --> 00:13:25.050
So those are two important
concepts in

00:13:25.050 --> 00:13:27.620
testing, but very different.

00:13:27.620 --> 00:13:30.300
Now, why do people think IQ
tests are interesting, besides

00:13:30.300 --> 00:13:33.330
just this weird curiosity
about intelligence?

00:13:33.330 --> 00:13:36.450
Because IQ tests correlate with
GPA in high school and

00:13:36.450 --> 00:13:39.910
college, job success, salary,
stable marriages, staying out

00:13:39.910 --> 00:13:42.340
of jail, and how
long you live.

00:13:42.340 --> 00:13:43.380
Again, it's a correlation.

00:13:43.380 --> 00:13:45.645
It could be many different
things.

00:13:45.645 --> 00:13:48.250
The thing I want to talk you out
of is the idea that if you

00:13:48.250 --> 00:13:52.985
have a 200 IQ today, you've got
something stored up that

00:13:52.985 --> 00:13:54.590
will fend off disease.

00:13:54.590 --> 00:13:56.850
It's not that, necessarily.

00:13:56.850 --> 00:13:59.290
It's that there's lots of things
in your life that got

00:13:59.290 --> 00:14:00.600
you to that score.

00:14:00.600 --> 00:14:02.710
And all those things are likely
to persist in your

00:14:02.710 --> 00:14:07.250
life: good health, support of
education, interest maybe in

00:14:07.250 --> 00:14:09.230
taking on challenging tasks.

00:14:09.230 --> 00:14:10.110
You had those before.

00:14:10.110 --> 00:14:11.400
You had them while you
took the test.

00:14:11.400 --> 00:14:14.440
You continue to have
them after.

00:14:14.440 --> 00:14:16.660
Another way to think about it
though, which is both a plus

00:14:16.660 --> 00:14:19.390
and minus, depending on how
you look at it, IQ scores

00:14:19.390 --> 00:14:21.670
account for approximately
25% of the

00:14:21.670 --> 00:14:23.260
variation on these things.

00:14:23.260 --> 00:14:24.220
So that's not huge.

00:14:24.220 --> 00:14:27.020
That means there's many, many,
many other factors, besides

00:14:27.020 --> 00:14:28.890
whatever IQ is capturing.

00:14:28.890 --> 00:14:31.250
But it's about as big a chunk
as you can get for a couple

00:14:31.250 --> 00:14:32.300
hours of testing.

00:14:32.300 --> 00:14:35.290
And we know that other things
like personality, variation,

00:14:35.290 --> 00:14:38.210
education, culture, all those
things blend in on how a

00:14:38.210 --> 00:14:40.775
person will do on all
these outcomes.

00:14:43.630 --> 00:14:44.720
So I'm going to go
back and forth.

00:14:44.720 --> 00:14:45.850
Because on the one hand, we
say there's a lot of stuff

00:14:45.850 --> 00:14:46.210
besides IQ.

00:14:46.210 --> 00:14:48.010
But IQ has some interestingness,
otherwise

00:14:48.010 --> 00:14:49.040
nobody would talk about it.

00:14:49.040 --> 00:14:50.260
And it would just disappear
from the world.

00:14:50.260 --> 00:14:55.790
So on average, on average, what
does 15 IQ points mean?

00:14:55.790 --> 00:14:58.640
So on average, a person with
100 IQ in the US will go to

00:14:58.640 --> 00:15:00.660
high school and a year or two
in a community college.

00:15:00.660 --> 00:15:04.560
On average, these are averages
over huge variety of people.

00:15:04.560 --> 00:15:07.120
You get 15 more IQ points, and
on average, you're likely to

00:15:07.120 --> 00:15:09.630
finish college and have
a white collar job.

00:15:09.630 --> 00:15:11.890
Have 15 fewer points, and you're
not likely to complete

00:15:11.890 --> 00:15:12.700
high school.

00:15:12.700 --> 00:15:14.570
And you'd have to have a job
that doesn't require

00:15:14.570 --> 00:15:16.270
educational background.

00:15:16.270 --> 00:15:20.870
So 15 point on averages
correlate with things.

00:15:20.870 --> 00:15:24.450
Here's another one, because
everybody in the end is a

00:15:24.450 --> 00:15:25.930
little bit curious.

00:15:25.930 --> 00:15:28.680
In your own life, how much of
what you're good at or not

00:15:28.680 --> 00:15:32.210
good at, of any of us, comes
from our genes, from our

00:15:32.210 --> 00:15:32.970
environment?

00:15:32.970 --> 00:15:34.790
Where does that come from?

00:15:34.790 --> 00:15:37.930
What are the things that
support us or limit us?

00:15:37.930 --> 00:15:42.970
So here's one study that took
women from the same family.

00:15:42.970 --> 00:15:45.800
So they're all sisters
from the same family.

00:15:45.800 --> 00:15:48.690
They're trying to control a
little bit of the genetics,

00:15:48.690 --> 00:15:50.220
because they share genes
within the family.

00:15:50.220 --> 00:15:51.280
They're not twins.

00:15:51.280 --> 00:15:53.360
And definitely shared
environment.

00:15:53.360 --> 00:15:55.760
They're in the same household.

00:15:55.760 --> 00:15:59.320
And within those families, they
look at the IQ scores.

00:15:59.320 --> 00:16:02.580
So here's the average group,
above average, well above

00:16:02.580 --> 00:16:05.760
average, a little below average,
a lot below average.

00:16:05.760 --> 00:16:09.330
And they say that in these
women, all coming from the

00:16:09.330 --> 00:16:13.120
same household, here's their
average income and here's the

00:16:13.120 --> 00:16:17.500
average rate of illegitimate
births, which is one measure

00:16:17.500 --> 00:16:21.540
that people think is kind of
a social, on average not a

00:16:21.540 --> 00:16:23.500
desired outcome, to have a
child without a family

00:16:23.500 --> 00:16:23.990
supporting.

00:16:23.990 --> 00:16:26.180
It's not going to help a person
succeed in the world to

00:16:26.180 --> 00:16:28.570
be a single mother.

00:16:28.570 --> 00:16:32.370
And so you can see, the salary
pretty much moves with the IQ

00:16:32.370 --> 00:16:33.320
on average.

00:16:33.320 --> 00:16:36.890
And the rate of illegitimate
births pretty much

00:16:36.890 --> 00:16:38.620
moves with that too.

00:16:38.620 --> 00:16:41.730
So again, it's an average.

00:16:41.730 --> 00:16:44.510
But there's some relationship
between these things.

00:16:44.510 --> 00:16:49.980
Now, if you pick up the
newspaper today and read about

00:16:49.980 --> 00:16:55.110
nuclear disasters in Japan and
problems in the Middle East

00:16:55.110 --> 00:16:57.750
and other things like that, you
would say, it's hard to

00:16:57.750 --> 00:16:59.880
say the world is getting
smarter.

00:16:59.880 --> 00:17:02.010
But you may or may not be
surprised to hear of the

00:17:02.010 --> 00:17:02.940
so-called Flynn effect.

00:17:02.940 --> 00:17:05.450
This was kind of a surprise when
it was discovered, which

00:17:05.450 --> 00:17:06.430
is basically this.

00:17:06.430 --> 00:17:09.369
IQ scores, where they're
measured, go up around the

00:17:09.369 --> 00:17:10.400
world all the time.

00:17:10.400 --> 00:17:14.089
By IQ test scores, it's as if
the world is getting smarter

00:17:14.089 --> 00:17:17.859
all the time, which is kind
of an optimistic thing.

00:17:17.859 --> 00:17:20.240
Whether it's getting wiser
is a good question.

00:17:20.240 --> 00:17:23.030
But by IQ score, smarter
all the time.

00:17:23.030 --> 00:17:26.630
So that's hidden for many years,
because every few years

00:17:26.630 --> 00:17:29.320
these tests get renormed
for the groups.

00:17:29.320 --> 00:17:30.670
Does that makes sense?

00:17:30.670 --> 00:17:33.210
So they're given again and
100 is the average.

00:17:33.210 --> 00:17:35.310
And they're given again and
100 is the average.

00:17:35.310 --> 00:17:38.040
But if they ignore this
averaging and just say, across

00:17:38.040 --> 00:17:39.970
these tests that only change a
little bit, some of them only

00:17:39.970 --> 00:17:42.760
change a little bit, if we just
take the scores, look at

00:17:42.760 --> 00:17:45.606
1930, 1940.

00:17:45.606 --> 00:17:48.770
And it just keeps going up.

00:17:48.770 --> 00:17:51.340
So Flynn is the person who
discovered that the world is

00:17:51.340 --> 00:17:54.320
getting smarter by
raw IQ tests.

00:17:56.970 --> 00:17:59.640
And one of tests they use, I'll
show you an example of

00:17:59.640 --> 00:18:01.420
this, it's another widely
used test called Raven's

00:18:01.420 --> 00:18:03.510
Progressive Matrices, IQs
go up about three

00:18:03.510 --> 00:18:05.770
points every 10 years.

00:18:05.770 --> 00:18:10.160
So that if you were born in
1930, a person would have an

00:18:10.160 --> 00:18:15.490
IQ of 100, their child 108,
and their grandchild, 120,

00:18:15.490 --> 00:18:16.810
with a standard deviation
higher.

00:18:16.810 --> 00:18:20.590
So it's like smarter,
smarter, smarter.

00:18:20.590 --> 00:18:24.860
I don't recommend to go home and
tell your parents that any

00:18:24.860 --> 00:18:26.610
debate you have with them is
pointless, because you're

00:18:26.610 --> 00:18:29.390
smarter than they are.

00:18:29.390 --> 00:18:32.180
But if push comes to shove,
you can look up the Flynn

00:18:32.180 --> 00:18:34.760
effect and have a discussion
about that.

00:18:34.760 --> 00:18:39.240
So somebody who has an IQ today,
just flip it around.

00:18:39.240 --> 00:18:41.360
If you have a 100, your
grandparent would

00:18:41.360 --> 00:18:44.176
have a IQ of 82.

00:18:44.176 --> 00:18:46.400
By this reasoning, people
in 1900 would have

00:18:46.400 --> 00:18:47.650
a mean IQ of 70.

00:18:50.330 --> 00:18:52.580
And nobody knows why it's going
up all over the world.

00:18:52.580 --> 00:18:54.010
There's all kinds of
ideas about maybe

00:18:54.010 --> 00:18:55.650
nutrition is going up.

00:18:55.650 --> 00:18:57.950
Maybe reasoning abilities
are being pushed.

00:18:57.950 --> 00:18:59.350
Maybe educational things.

00:18:59.350 --> 00:19:01.730
I mean it's impossible
scientifically to know why,

00:19:01.730 --> 00:19:05.450
because it's happening in such
a global, giant scale.

00:19:05.450 --> 00:19:08.410
But some people have said, the
tests don't go up uniformly

00:19:08.410 --> 00:19:09.120
across the tests.

00:19:09.120 --> 00:19:12.320
Which gives us a hint about the
way the world is changing

00:19:12.320 --> 00:19:15.840
and how intelligence tests in
some way become a barometer of

00:19:15.840 --> 00:19:17.200
what mental abilities are

00:19:17.200 --> 00:19:20.710
increasingly prized by societies.

00:19:20.710 --> 00:19:23.670
So there's relatively small
gains across years in

00:19:23.670 --> 00:19:26.950
vocabulary or general knowledge
or arithmetic.

00:19:26.950 --> 00:19:29.530
Where there are large gains
are what you might call

00:19:29.530 --> 00:19:32.850
abstract thinking tasks,
like similarities.

00:19:32.850 --> 00:19:34.730
So let's focus on that
for a moment.

00:19:34.730 --> 00:19:36.670
So a similarities question
might be like this.

00:19:36.670 --> 00:19:39.810
In what way are dogs
and rabbits alike?

00:19:39.810 --> 00:19:43.870
If you answer both mammals,
you get two points.

00:19:43.870 --> 00:19:46.090
That's the best you can do.

00:19:46.090 --> 00:19:48.870
And that's what we would call an
abstract or taxonomic one.

00:19:48.870 --> 00:19:51.140
You're abstracting across
the two things.

00:19:51.140 --> 00:19:53.190
Or you could give an answer
like, you use

00:19:53.190 --> 00:19:55.660
dogs to hunt rabbits.

00:19:55.660 --> 00:19:57.150
That's not a wrong answer.

00:19:57.150 --> 00:20:00.090
That's considered a
one-point answer.

00:20:00.090 --> 00:20:01.920
You kind of seem to know what a
dog is, you kind of seem to

00:20:01.920 --> 00:20:02.410
know a rabbit.

00:20:02.410 --> 00:20:05.120
But you didn't give the best
answer by this test criterion.

00:20:05.120 --> 00:20:07.250
Because it's a functional
description of these things,

00:20:07.250 --> 00:20:09.980
not a abstract, conceptual
one.

00:20:09.980 --> 00:20:13.210
So let's think about different
societies, besides the Western

00:20:13.210 --> 00:20:16.500
industrialized one that
we represent mostly.

00:20:16.500 --> 00:20:22.250
In Liberia, they ask people to
sort baskets of food, tools,

00:20:22.250 --> 00:20:24.710
containers, and clothing.

00:20:24.710 --> 00:20:28.330
So most of us, Western educated
in a very broad

00:20:28.330 --> 00:20:31.350
sense, would put the things of
food together, the things of

00:20:31.350 --> 00:20:32.000
tools together.

00:20:32.000 --> 00:20:34.100
Those are the abstract
categories.

00:20:34.100 --> 00:20:36.370
The people from this tribe
put them in terms

00:20:36.370 --> 00:20:37.300
of functional pairings.

00:20:37.300 --> 00:20:39.870
They put the potato next to the
knife they use to peel the

00:20:39.870 --> 00:20:42.250
potato, right?

00:20:42.250 --> 00:20:44.250
And here's the fun thing
in this study.

00:20:44.250 --> 00:20:47.040
They said, what would
a fool do with this?

00:20:47.040 --> 00:20:49.720
And they said oh, well, a fool
would put all the food things

00:20:49.720 --> 00:20:51.610
together and all the
utensils together.

00:20:51.610 --> 00:20:54.060
But that's pointless to do that,
because what are you

00:20:54.060 --> 00:20:55.310
going to do with all the food
things over here, when the

00:20:55.310 --> 00:20:56.500
utensils are over here.

00:20:56.500 --> 00:20:57.330
You understand?

00:20:57.330 --> 00:20:59.500
It's what our society
emphasizes as a

00:20:59.500 --> 00:21:01.830
useful way of thinking.

00:21:01.830 --> 00:21:05.540
So it could be arbitrary.

00:21:05.540 --> 00:21:08.460
But it's not arbitrarily
rewarded and prized in the

00:21:08.460 --> 00:21:14.000
world of education, that
we live in here.

00:21:14.000 --> 00:21:17.970
So intelligence tests attach
value at a specific time and

00:21:17.970 --> 00:21:18.700
specific test.

00:21:18.700 --> 00:21:21.180
Scores are changing over time.

00:21:21.180 --> 00:21:24.230
Different cultures are valuing
different things.

00:21:24.230 --> 00:21:25.480
It's constantly renormed.

00:21:28.200 --> 00:21:30.340
And because of that, here's
another example of how the

00:21:30.340 --> 00:21:31.110
renorming is.

00:21:31.110 --> 00:21:34.360
It used to be for some number of
years, that 70 was used as

00:21:34.360 --> 00:21:37.080
a cut-off for estimates of,
broadly speaking, mental

00:21:37.080 --> 00:21:38.120
retardation.

00:21:38.120 --> 00:21:41.040
And why that matters is that a
person for example would get

00:21:41.040 --> 00:21:45.140
certain kinds of medical
services provided to them.

00:21:45.140 --> 00:21:48.530
When a new version of the WISC,
of the Wechsler test,

00:21:48.530 --> 00:21:51.980
came out in 1991, so a renormed
version, so that

00:21:51.980 --> 00:21:56.360
means the scores are going up,
if every state in the country

00:21:56.360 --> 00:21:58.970
had simultaneously used that
new test when it came out,

00:21:58.970 --> 00:22:01.110
which they didn't quite, but
if they had, the number of

00:22:01.110 --> 00:22:04.440
people rated as retarded,
diagnosed as retarded, would

00:22:04.440 --> 00:22:06.630
have doubled in a single day.

00:22:06.630 --> 00:22:09.560
Because the norms aren't
pushing the numbers up.

00:22:09.560 --> 00:22:10.220
OK?

00:22:10.220 --> 00:22:13.620
So the number of people scoring
very poorly would have

00:22:13.620 --> 00:22:16.710
doubled in a single day because
of the renorming.

00:22:19.320 --> 00:22:21.730
It's very important to realize
that all these things, what's

00:22:21.730 --> 00:22:25.270
really in the test, what's
valued by a sort of society or

00:22:25.270 --> 00:22:28.030
culture, and the norming things
are really influencing

00:22:28.030 --> 00:22:29.855
a lot of how you get
to these numbers.

00:22:33.110 --> 00:22:37.540
So ideas about intelligence,
roughly in this field, have

00:22:37.540 --> 00:22:39.040
come out of what people call

00:22:39.040 --> 00:22:40.900
psychometric analysis of subtests.

00:22:40.900 --> 00:22:42.850
They give you the scatter
of tests.

00:22:42.850 --> 00:22:45.780
And the thing that struck many
people is how for a given

00:22:45.780 --> 00:22:48.570
individual, there's correlations
among the tests.

00:22:48.570 --> 00:22:52.280
A person who does well on one
test, tends to do well on

00:22:52.280 --> 00:22:55.270
another test, and tends to
do well on a third test.

00:22:55.270 --> 00:23:00.590
And a researcher named Spearman
gave this correlation

00:23:00.590 --> 00:23:03.730
among tests, the letter "g,"
the name g, to stand for

00:23:03.730 --> 00:23:07.660
general intelligence, and "s"
for specific intelligence.

00:23:07.660 --> 00:23:10.445
And since then, many people have
said well really is there

00:23:10.445 --> 00:23:12.680
one broad sense of intelligence,
are there

00:23:12.680 --> 00:23:15.060
multiple separate kinds
of intelligence?

00:23:15.060 --> 00:23:16.370
And people debate that
in the field.

00:23:16.370 --> 00:23:19.240
And probably somewhat the answer
is it depends what

00:23:19.240 --> 00:23:21.630
you're measuring for and the
purpose of the measurement.

00:23:21.630 --> 00:23:24.130
But a good one to look at,
because it's very clear, is

00:23:24.130 --> 00:23:27.600
the difference between fluid and
crystallized intelligence.

00:23:27.600 --> 00:23:29.810
So fluid intelligence is
what we think of as

00:23:29.810 --> 00:23:31.290
novel problem solving.

00:23:31.290 --> 00:23:32.440
You have a situation
you were never in

00:23:32.440 --> 00:23:34.640
before, what do you do?

00:23:34.640 --> 00:23:36.930
Crystallized intelligence
are facts you know,

00:23:36.930 --> 00:23:40.250
information you know.

00:23:40.250 --> 00:23:43.120
So here's the conceptual
idea that maybe general

00:23:43.120 --> 00:23:45.470
intelligence is a thing.

00:23:45.470 --> 00:23:48.250
And then it shows up on
different tests, with the

00:23:48.250 --> 00:23:49.680
different specific
intelligences.

00:23:49.680 --> 00:23:51.090
That's like maybe verbal
or arithmetic.

00:23:54.230 --> 00:23:58.710
But let me show you the fluid
and crystallized one.

00:23:58.710 --> 00:24:01.060
So the fluid one we said were
first mental processes that

00:24:01.060 --> 00:24:02.840
apply for a novel situation.

00:24:02.840 --> 00:24:06.110
And that declines a lot with
age and adulthood.

00:24:06.110 --> 00:24:10.090
So these lines going down
here, I'm about here and

00:24:10.090 --> 00:24:11.150
you're about here.

00:24:11.150 --> 00:24:14.930
Who should be grading who in
this course, you might ask?

00:24:14.930 --> 00:24:17.980
The only reason that I get to
grade you is because these

00:24:17.980 --> 00:24:19.190
lines keep going up here.

00:24:19.190 --> 00:24:21.970
This is how many
facts you know.

00:24:21.970 --> 00:24:25.150
So the plus of living longer
is that you get

00:24:25.150 --> 00:24:27.510
to know more stuff.

00:24:27.510 --> 00:24:30.170
The risk of living longer is you
don't use it as flexibly

00:24:30.170 --> 00:24:32.410
under novel situations.

00:24:32.410 --> 00:24:34.100
So a pretty strong distinction,
at least with

00:24:34.100 --> 00:24:37.040
aging, between so-called fluid
intelligence and crystallized

00:24:37.040 --> 00:24:38.140
intelligence.

00:24:38.140 --> 00:24:40.990
Here's another study of exactly
the same thing.

00:24:40.990 --> 00:24:44.960
It's a striking ski slope of
decline in fluid intelligence.

00:24:44.960 --> 00:24:50.840
Look at this, ernt, from 20
to 80, it just goes down

00:24:50.840 --> 00:24:52.650
relentlessly.

00:24:52.650 --> 00:24:54.680
But when it comes to
crystallized intelligence,

00:24:54.680 --> 00:24:57.810
things like vocabulary, that
kind of hangs in there, until

00:24:57.810 --> 00:25:00.530
you have some big
health problem.

00:25:00.530 --> 00:25:01.740
So at least those two
kinds seem to

00:25:01.740 --> 00:25:05.210
have a different basis.

00:25:05.210 --> 00:25:08.110
And people have thought about
different ways to think about

00:25:08.110 --> 00:25:08.770
intelligence.

00:25:08.770 --> 00:25:11.030
And I'll just tell you that in
general you get positive

00:25:11.030 --> 00:25:12.340
correlations among the tests.

00:25:12.340 --> 00:25:14.650
And so often people think
about this general

00:25:14.650 --> 00:25:15.280
intelligence.

00:25:15.280 --> 00:25:16.990
The correlation among
the tests is

00:25:16.990 --> 00:25:18.240
the best single predictor.

00:25:20.670 --> 00:25:23.560
Now, there's lots of responses
to ideas about intelligence.

00:25:23.560 --> 00:25:26.980
One of them is the study of
emotional intelligence.

00:25:26.980 --> 00:25:28.410
Peter Salovey is the researcher

00:25:28.410 --> 00:25:30.240
most involved in that.

00:25:30.240 --> 00:25:31.850
And we know this in
everyday life.

00:25:31.850 --> 00:25:36.730
We know smart people, what we
might call book smart people,

00:25:36.730 --> 00:25:40.070
who seem unwise in how they
relate to people, how they

00:25:40.070 --> 00:25:43.360
perceive emotions, how they
facilitate thought with

00:25:43.360 --> 00:25:45.930
emotion, how they understand or
manage their emotions, how

00:25:45.930 --> 00:25:47.470
they relate to other people.

00:25:47.470 --> 00:25:52.530
And so it's common sense to
think of that as something

00:25:52.530 --> 00:25:53.950
that varies across people.

00:25:53.950 --> 00:25:56.210
It's very hard to
measure that.

00:25:56.210 --> 00:25:59.400
I mean, part of the tyranny of
tests is what can be measured,

00:25:59.400 --> 00:26:00.120
gets measured.

00:26:00.120 --> 00:26:01.900
What's hard to measure,
is hard to measure.

00:26:01.900 --> 00:26:05.640
It's very hard to measure, in
a laboratory, in an hour, a

00:26:05.640 --> 00:26:06.900
person's emotional
intelligence.

00:26:06.900 --> 00:26:07.770
People have tried.

00:26:07.770 --> 00:26:10.020
It's just hard to do.

00:26:10.020 --> 00:26:12.310
Here's another area that's
gotten a lot of attention,

00:26:12.310 --> 00:26:14.560
from Howard Gardner
at Harvard.

00:26:14.560 --> 00:26:17.320
Kind of fighting back against
this idea of a single form of

00:26:17.320 --> 00:26:19.880
intelligence as captured by
g, the idea of multiple

00:26:19.880 --> 00:26:21.050
intelligences.

00:26:21.050 --> 00:26:23.760
And looking at things, the
associations in patients and

00:26:23.760 --> 00:26:24.840
development.

00:26:24.840 --> 00:26:27.380
That we can talk about
linguistic intelligence,

00:26:27.380 --> 00:26:30.540
spatial intelligence, music
intelligence, bodily

00:26:30.540 --> 00:26:31.140
intelligence.

00:26:31.140 --> 00:26:33.180
Great athletes have great
bodily intelligence.

00:26:33.180 --> 00:26:35.430
Intrapersonal or interpersonal,
it's kind of

00:26:35.430 --> 00:26:36.710
like emotional.

00:26:36.710 --> 00:26:38.020
Existential.

00:26:38.020 --> 00:26:39.250
So he said, there's all these
different kinds of

00:26:39.250 --> 00:26:40.720
intelligence.

00:26:40.720 --> 00:26:43.030
And each person is a
profile of these.

00:26:43.030 --> 00:26:45.340
You can be really strong in one,
and medium in another,

00:26:45.340 --> 00:26:47.210
and not so strong in another.

00:26:47.210 --> 00:26:49.700
The problem with this whole
approach is, yes, it's true

00:26:49.700 --> 00:26:51.520
that people have a
huge variety of

00:26:51.520 --> 00:26:52.640
facets of their humanity.

00:26:52.640 --> 00:26:55.480
This whole course, I feel like
every lecture we do, we just

00:26:55.480 --> 00:26:57.070
turn a person a little
bit to study another

00:26:57.070 --> 00:26:59.440
facet of their humanity.

00:26:59.440 --> 00:27:03.600
But it's hard as heck to
measure interpersonal

00:27:03.600 --> 00:27:05.880
intelligence or naturalistic
intelligence.

00:27:05.880 --> 00:27:07.670
Now, maybe somebody will figure
out how to do that.

00:27:07.670 --> 00:27:10.300
It's just hard to measure those
things, even if it feels

00:27:10.300 --> 00:27:13.430
like something conceptually
that's there.

00:27:13.430 --> 00:27:17.130
So this idea, which has had a
lot of resonance to people not

00:27:17.130 --> 00:27:20.580
liking very simple-minded ideas
of intelligence, very

00:27:20.580 --> 00:27:23.610
unitary ones, it's still not
gained much traction in

00:27:23.610 --> 00:27:24.860
practical use.

00:27:27.060 --> 00:27:29.165
So let me talk a bit about some
of the things that people

00:27:29.165 --> 00:27:31.240
have said, well, if we talk
about this intelligence, what

00:27:31.240 --> 00:27:33.590
might be some mental mechanisms
or neural

00:27:33.590 --> 00:27:36.400
mechanisms that support
or underlie it.

00:27:36.400 --> 00:27:39.110
So some things that go with
it are mental speed.

00:27:39.110 --> 00:27:42.410
And you know this, kind of
quickness of thought.

00:27:42.410 --> 00:27:44.910
OK, not even big thoughts, just
quickness feels like it

00:27:44.910 --> 00:27:46.860
helps a person solve problems.

00:27:46.860 --> 00:27:49.000
So for example there's a
moderately high correlation

00:27:49.000 --> 00:27:52.250
between IQ, complicated tests,
of all those tests, and how

00:27:52.250 --> 00:27:55.390
quickly you push a button
when the light comes on.

00:27:55.390 --> 00:27:59.360
Now, we never say, wow, Sally
or Fred is really smart.

00:27:59.360 --> 00:28:00.950
They can push a button
when the light

00:28:00.950 --> 00:28:03.090
comes on really fast.

00:28:03.090 --> 00:28:04.890
Let's go to them and get the
answers for the problem set.

00:28:07.490 --> 00:28:10.540
But that would go with the idea
that simply being fast is

00:28:10.540 --> 00:28:13.560
kind of a useful thing for
acting on the world.

00:28:13.560 --> 00:28:16.090
Or how quickly people decide
whether two lengths, two

00:28:16.090 --> 00:28:18.040
lines, are the same length
or a different length.

00:28:18.040 --> 00:28:20.810
Very simple thing, just speed.

00:28:20.810 --> 00:28:23.110
And another idea is working
memory capacity.

00:28:23.110 --> 00:28:25.790
How much information you can
hold in your head at any one

00:28:25.790 --> 00:28:28.000
time that's relevant
to the problem or

00:28:28.000 --> 00:28:29.510
task in front of you.

00:28:29.510 --> 00:28:31.220
So people have tried to explore
these in certain ways.

00:28:31.220 --> 00:28:32.950
And I'm going to talk to you
about a couple of brain

00:28:32.950 --> 00:28:34.970
studies that have tried
to look at this.

00:28:34.970 --> 00:28:38.730
So again, we talked last time
about frontal cortex and

00:28:38.730 --> 00:28:39.430
different parts of it.

00:28:39.430 --> 00:28:41.340
And especially the
dorsal-lateral prefrontal

00:28:41.340 --> 00:28:44.680
cortex is involved
in thinking.

00:28:44.680 --> 00:28:48.670
And so people have taken tasks
and said, what happens when

00:28:48.670 --> 00:28:49.720
you think about them?

00:28:49.720 --> 00:28:53.190
So here's one, an easy item
on the so-called Cattell

00:28:53.190 --> 00:28:54.710
Culture-Fair Test.

00:28:54.710 --> 00:28:57.920
So which of these things
goes into here.

00:28:57.920 --> 00:28:59.170
There's harder ones.

00:29:01.910 --> 00:29:06.350
Or here's a slightly adapted
version of the Raven's

00:29:06.350 --> 00:29:07.160
Progressive Matrices.

00:29:07.160 --> 00:29:08.290
I'll let you think about
that for a moment.

00:29:08.290 --> 00:29:09.980
I don't remember which it is.

00:29:09.980 --> 00:29:13.170
So you're told here's some
rows and columns.

00:29:13.170 --> 00:29:17.380
One of these eight answers
logically fits in here.

00:29:17.380 --> 00:29:21.100
And it's your job to figure
out which one.

00:29:21.100 --> 00:29:22.630
Let's see.

00:29:22.630 --> 00:29:24.890
How do we like--

00:29:24.890 --> 00:29:25.170
three.

00:29:25.170 --> 00:29:27.935
I like three too.

00:29:27.935 --> 00:29:29.330
Does that three seem OK?

00:29:32.110 --> 00:29:35.560
This is a medium-hard item
on this kind of a test.

00:29:35.560 --> 00:29:39.290
And then they do these kind
of map of results.

00:29:39.290 --> 00:29:41.060
This is what's fun about
the Raven's test, the

00:29:41.060 --> 00:29:41.630
item you just saw.

00:29:41.630 --> 00:29:43.950
Which is they sort of looked
at lots of other tests, so

00:29:43.950 --> 00:29:47.710
specifically spatial tests,
specifically verbal tests,

00:29:47.710 --> 00:29:49.720
specifically math tests.

00:29:49.720 --> 00:29:53.530
And they said if we have to give
you one test only, one

00:29:53.530 --> 00:29:57.400
test only, what test would best
predict your performance

00:29:57.400 --> 00:29:59.410
on all the other tests?

00:29:59.410 --> 00:30:03.130
And the answer is Raven's
or IQ tests like that.

00:30:03.130 --> 00:30:05.500
So of course if I give you
a math test, that better

00:30:05.500 --> 00:30:06.810
predicts your math
performance.

00:30:06.810 --> 00:30:08.760
If I give you a verbal test,
that better predicts your

00:30:08.760 --> 00:30:09.680
verbal performance.

00:30:09.680 --> 00:30:11.890
If I gave you a spatial test,
that better predicts your

00:30:11.890 --> 00:30:13.520
special performance.

00:30:13.520 --> 00:30:15.550
But if I want to give you one
test that gives me the best

00:30:15.550 --> 00:30:18.820
prediction of your performance
on other tests, it's things

00:30:18.820 --> 00:30:21.530
like these IQ tests
that do that.

00:30:21.530 --> 00:30:24.380
And that's just empirical.

00:30:24.380 --> 00:30:26.240
That's not arguing about
values or what

00:30:26.240 --> 00:30:27.460
intelligence is.

00:30:27.460 --> 00:30:29.760
A lot of people who work in this
field, they don't even

00:30:29.760 --> 00:30:31.080
care about these arguments.

00:30:31.080 --> 00:30:34.370
They just say, I just want
a test that predicts.

00:30:34.370 --> 00:30:37.020
It's your job, if you wanted
to figure out why it is.

00:30:37.020 --> 00:30:39.330
It's the genes, environment,
education, whatever.

00:30:39.330 --> 00:30:42.100
I just want a test
that predicts.

00:30:42.100 --> 00:30:44.290
So from that perspective,
these tests

00:30:44.290 --> 00:30:47.540
have a strong history.

00:30:47.540 --> 00:30:48.830
And what turns on in the
brain when you solve

00:30:48.830 --> 00:30:49.700
this kind of a task.

00:30:49.700 --> 00:30:52.245
Well, the frontal
cortex a lot.

00:30:52.245 --> 00:30:52.830
The parietal cortex.

00:30:52.830 --> 00:30:54.200
But especially, frontal
cortex.

00:30:54.200 --> 00:30:57.280
So it goes with our idea that
thinking is very much tied to

00:30:57.280 --> 00:30:58.990
prefrontal cortical function.

00:30:58.990 --> 00:31:02.440
Here's another example of a
study that said, let's compare

00:31:02.440 --> 00:31:04.680
what happens in the brain when
you have test items that are

00:31:04.680 --> 00:31:09.190
hard, which is the outlier one
here, that require a lot of g,

00:31:09.190 --> 00:31:12.060
or items that are easy, that
require very little g.

00:31:12.060 --> 00:31:14.990
Will up here in the brain, what
happens when you turn on

00:31:14.990 --> 00:31:16.610
your g, so to speak?

00:31:16.610 --> 00:31:19.730
And again, the prominent areas
tend to be, for both spatial

00:31:19.730 --> 00:31:21.365
and verbal, prefrontal cortex.

00:31:24.610 --> 00:31:26.350
One more.

00:31:26.350 --> 00:31:28.500
the n-back task, we
use that a lot.

00:31:28.500 --> 00:31:30.520
Where your job is,
letters come on.

00:31:30.520 --> 00:31:33.180
And in this one, one-back, you
say, what do I see a letter

00:31:33.180 --> 00:31:35.380
that's the same letter
as the one before?

00:31:35.380 --> 00:31:36.810
So, for example--

00:31:36.810 --> 00:31:37.910
two-back, sorry.

00:31:37.910 --> 00:31:40.970
So you see, R, R, M, X, this
is two-back, because an M

00:31:40.970 --> 00:31:42.210
occurred two items ago.

00:31:42.210 --> 00:31:43.370
Three-backs are really hard.

00:31:43.370 --> 00:31:44.970
Because the letters are going,
boom, boom, boom, boom.

00:31:44.970 --> 00:31:47.490
And what's hard about it is,
the letter in front of you,

00:31:47.490 --> 00:31:49.780
you have to say does it match
the one I saw three letters

00:31:49.780 --> 00:31:51.910
previously, plus remember that,
because it's going to be

00:31:51.910 --> 00:31:54.270
the target for the one that
comes three after.

00:31:54.270 --> 00:31:55.070
It's hard.

00:31:55.070 --> 00:31:55.910
I can tell you it's hard.

00:31:55.910 --> 00:31:56.720
You can be smart.

00:31:56.720 --> 00:31:58.500
It's pretty hard.

00:31:58.500 --> 00:32:02.660
And for this task in general,
the harder it gets, the more

00:32:02.660 --> 00:32:06.580
you turn on the dorsal-lateral
prefrontal cortex.

00:32:06.580 --> 00:32:08.965
And so they did this test with
students in Washington St.

00:32:08.965 --> 00:32:12.670
Louis, Louis, related it to
performance on the Raven's.

00:32:12.670 --> 00:32:16.790
And sure enough in the brain,
it's again prefrontal cortex.

00:32:16.790 --> 00:32:20.430
So a bunch of studies have found
all these tests involve

00:32:20.430 --> 00:32:21.290
a fair bit of the brain.

00:32:21.290 --> 00:32:24.623
But they seem to really draw
upon the capacities that you

00:32:24.623 --> 00:32:27.450
and I have in prefrontal
cortex.

00:32:27.450 --> 00:32:30.930
And one more note, in brain
imaging when we talk about

00:32:30.930 --> 00:32:33.100
things, sometimes it's better
to have more activation and

00:32:33.100 --> 00:32:34.570
sometimes less activation.

00:32:34.570 --> 00:32:36.830
And you have to know the
behavior to know that.

00:32:36.830 --> 00:32:39.660
So whenever you see somebody
saying oh, this person, this

00:32:39.660 --> 00:32:41.960
group, or whatever, had more
brain activity or less brain

00:32:41.960 --> 00:32:44.920
activity, you don't know until
you know the rest of the

00:32:44.920 --> 00:32:47.440
behavior, whether more
is good or bad.

00:32:47.440 --> 00:32:49.680
Sometimes when people get really
good at something, they

00:32:49.680 --> 00:32:52.880
use less of their brain, because
they can do it so

00:32:52.880 --> 00:32:54.240
efficiently.

00:32:54.240 --> 00:32:57.560
So here's an example of people
with lower and higher IQ doing

00:32:57.560 --> 00:32:58.370
an n-back task.

00:32:58.370 --> 00:33:01.440
And you can see that the people
with lower IQ have much

00:33:01.440 --> 00:33:04.460
more activation than the
people with higher IQ.

00:33:04.460 --> 00:33:07.150
It's as if these people have
to use all their brain

00:33:07.150 --> 00:33:08.950
resources to accomplish this.

00:33:08.950 --> 00:33:10.910
And these people only
have to use some.

00:33:10.910 --> 00:33:12.160
It's as if it were that.

00:33:14.600 --> 00:33:18.890
So here comes one of the
classic topics of

00:33:18.890 --> 00:33:20.140
the fate of a human.

00:33:20.140 --> 00:33:22.380
How much is your genes at birth
and how much is the

00:33:22.380 --> 00:33:24.230
environment that you're in.

00:33:24.230 --> 00:33:26.610
And if we were really to
understand that in a deep,

00:33:26.610 --> 00:33:29.120
deep scientific way, we would
know the genes are related to

00:33:29.120 --> 00:33:32.630
intelligence; the experiential
factors in the world, family,

00:33:32.630 --> 00:33:37.910
education, emotional support,
and so on; how they interact.

00:33:37.910 --> 00:33:39.860
All those things, we know
practically nothing about.

00:33:39.860 --> 00:33:41.230
I'll show you one
tiny example.

00:33:41.230 --> 00:33:42.750
We know practically nothing
about them in

00:33:42.750 --> 00:33:44.270
any depth at all.

00:33:44.270 --> 00:33:48.010
So we mostly end up using twin
studies to make the best

00:33:48.010 --> 00:33:49.710
estimates we can.

00:33:49.710 --> 00:33:51.410
Which doesn't tell you
what the factors are.

00:33:51.410 --> 00:33:54.710
It just tells you how much is
a twin who is identical more

00:33:54.710 --> 00:33:56.740
or less like a twin
who is fraternal.

00:33:56.740 --> 00:33:59.340
OK, we'll come back to that.

00:33:59.340 --> 00:34:01.750
But I do want to tell you sort
of an exciting direction

00:34:01.750 --> 00:34:04.950
that's slightly getting at the
genetic story of this.

00:34:04.950 --> 00:34:06.570
So you may know, many
of you more

00:34:06.570 --> 00:34:08.810
than I do, about genetics.

00:34:08.810 --> 00:34:11.350
But single nucleotide
polymorphisms are DNA

00:34:11.350 --> 00:34:12.800
variations.

00:34:12.800 --> 00:34:15.860
These occur relatively
commonly.

00:34:15.860 --> 00:34:17.699
They're estimated to make
up about 90% of

00:34:17.699 --> 00:34:19.260
human genetic variation.

00:34:19.260 --> 00:34:21.250
And they're constant from
generation to generation.

00:34:21.250 --> 00:34:24.250
So people have studied a few
of these variations among

00:34:24.250 --> 00:34:26.860
people and related them
to performance.

00:34:26.860 --> 00:34:29.290
So let me remind you of
something from last lecture,

00:34:29.290 --> 00:34:32.210
the Wisconsin Card
Sorting Task.

00:34:32.210 --> 00:34:36.250
Where you sorted these cards
mysteriously and we said that

00:34:36.250 --> 00:34:39.239
if you have a frontal lesion,
you do poorly on this task.

00:34:39.239 --> 00:34:43.460
You get stuck with one kind of
problem solving or habit,

00:34:43.460 --> 00:34:47.650
answer even when it's wrong,
in a new circumstance.

00:34:47.650 --> 00:34:51.270
So what people have done is look
at one particular gene

00:34:51.270 --> 00:34:56.830
that's related to this enzyme
that modulates dopamine

00:34:56.830 --> 00:34:59.290
function in the prefrontal
cortex.

00:34:59.290 --> 00:35:02.205
So it's a risk factor
for schizophrenia.

00:35:02.205 --> 00:35:05.710
It metabolizes dopamine
in prefrontal cortex.

00:35:05.710 --> 00:35:11.350
We vary with which version
of this gene we have.

00:35:11.350 --> 00:35:13.570
And here's the performance
on the card sorting task.

00:35:13.570 --> 00:35:14.730
And let me pick these.

00:35:14.730 --> 00:35:16.060
These are healthy people.

00:35:16.060 --> 00:35:19.550
If you have this version of the
gene or this pairing or

00:35:19.550 --> 00:35:23.240
this pairing, your card sorting
task is going up.

00:35:23.240 --> 00:35:25.620
So this is kind of amazing
that we're close to that,

00:35:25.620 --> 00:35:26.860
where a single gene
is varying.

00:35:26.860 --> 00:35:28.110
Now, the score doesn't
go up a lot.

00:35:28.110 --> 00:35:29.960
It accounts a small percent
of the variance.

00:35:29.960 --> 00:35:33.890
But it's one gene having an
identifiable correlation with

00:35:33.890 --> 00:35:37.580
performance on a
demanding task.

00:35:37.580 --> 00:35:39.160
And you can do this n-back
kind of task

00:35:39.160 --> 00:35:40.450
we just talked about.

00:35:40.450 --> 00:35:42.470
And then you can look in the
brain as people are performing

00:35:42.470 --> 00:35:46.360
this, divided by which version
of this gene they have.

00:35:46.360 --> 00:35:52.060
And what you get is more
activation in the people with

00:35:52.060 --> 00:35:56.750
this complement, than
this complement.

00:35:56.750 --> 00:35:58.690
Sort of consistent with the idea
they're working harder.

00:35:58.690 --> 00:36:01.200
There's the same group of people
who are performing less

00:36:01.200 --> 00:36:02.710
well in the card sorting task.

00:36:02.710 --> 00:36:05.210
So there are some genes that
have been identified that are

00:36:05.210 --> 00:36:07.930
associated with what we may call
broadly, intelligence.

00:36:07.930 --> 00:36:11.030
But that one gene only accounts
for a tiny bit of the

00:36:11.030 --> 00:36:13.430
overall score on these
kinds of things.

00:36:13.430 --> 00:36:15.840
And we don't know much more
about it than that.

00:36:15.840 --> 00:36:18.170
So how do people even
begin to think

00:36:18.170 --> 00:36:19.910
about nature and nurture?

00:36:19.910 --> 00:36:21.520
And I like to think about
this in two ways.

00:36:21.520 --> 00:36:23.830
One way is narrowly IQ tests,
because there's been so much

00:36:23.830 --> 00:36:24.840
work in that area.

00:36:24.840 --> 00:36:27.340
But it's the deep question
of how do we get

00:36:27.340 --> 00:36:29.260
to be who we are?

00:36:29.260 --> 00:36:31.900
So again, they mostly look at
twin studies, so-called

00:36:31.900 --> 00:36:35.100
behavioral genetics, and
estimated heritability, how

00:36:35.100 --> 00:36:38.302
much of variation is
due to genetics.

00:36:38.302 --> 00:36:39.810
And I'll come back to
this height thing.

00:36:39.810 --> 00:36:42.300
But to give you a sense, height
is thought to be 90%

00:36:42.300 --> 00:36:44.480
inheritable in the US.

00:36:44.480 --> 00:36:46.970
Tall parents on average will
average will have taller kids.

00:36:46.970 --> 00:36:49.030
Less tall parents,
less tall kids.

00:36:49.030 --> 00:36:51.860
And in twin studies, they
contrast monozygotic twins,

00:36:51.860 --> 00:36:54.210
twins that are identical or
have the same genes, with

00:36:54.210 --> 00:36:58.960
dizygotic ones, who share some
genes, but not nearly as many.

00:36:58.960 --> 00:37:01.250
Of course, they share the
same environment a lot.

00:37:01.250 --> 00:37:04.000
Have you guys seen the two twins
talking to each other

00:37:04.000 --> 00:37:06.220
video on YouTube?

00:37:06.220 --> 00:37:07.530
If you haven't seen it,
it's kind of fun.

00:37:07.530 --> 00:37:09.270
The two twins are babbling.

00:37:09.270 --> 00:37:11.930
But they're so demonstrative,
it's as if they were talking

00:37:11.930 --> 00:37:12.810
to each other in babble.

00:37:12.810 --> 00:37:13.913
It's kind of fun, if
you haven't see it.

00:37:13.913 --> 00:37:15.163
It's very cute.

00:37:18.390 --> 00:37:20.850
Here is it again.

00:37:20.850 --> 00:37:25.390
The fraternal twins and here's
the identical twins, before

00:37:25.390 --> 00:37:27.160
they're born.

00:37:27.160 --> 00:37:30.220
This has been studied
a fair bit.

00:37:30.220 --> 00:37:34.690
Here's the correlations between
IQ scores and people

00:37:34.690 --> 00:37:35.510
of varying relations.

00:37:35.510 --> 00:37:37.340
So let's work down that list.

00:37:37.340 --> 00:37:39.160
Unrelated persons,
reared together.

00:37:39.160 --> 00:37:41.350
So perhaps an adopted
person and his or

00:37:41.350 --> 00:37:42.750
her sibling or something.

00:37:42.750 --> 00:37:45.730
They're correlated at
about what, 0.2.

00:37:45.730 --> 00:37:48.030
Foster parent and child, it
goes up a little bit.

00:37:48.030 --> 00:37:49.830
Parent and child living
together, so that's a

00:37:49.830 --> 00:37:52.170
biological parent and child
living together, that goes up

00:37:52.170 --> 00:37:53.010
a fair bit.

00:37:53.010 --> 00:37:55.110
Brother and sister reared
apart, back down.

00:37:55.110 --> 00:37:56.400
Brother and sister
reared together.

00:37:56.400 --> 00:38:00.940
So you could look at this
as kind of shared genes.

00:38:00.940 --> 00:38:02.920
And this might be something
like environment.

00:38:02.920 --> 00:38:03.510
Does make sense?

00:38:03.510 --> 00:38:04.950
Intuitively.

00:38:04.950 --> 00:38:07.640
That's assuming there's no
interaction between genes and

00:38:07.640 --> 00:38:09.760
the environment, which
we know must happen.

00:38:09.760 --> 00:38:12.470
So when I do this I'm treating
them as if they were additive

00:38:12.470 --> 00:38:13.990
and independent.

00:38:13.990 --> 00:38:16.260
Everything we know about genes
and the environment are

00:38:16.260 --> 00:38:18.010
tremendously interactive.

00:38:18.010 --> 00:38:19.760
We just don't have insight
into that in any

00:38:19.760 --> 00:38:22.850
deep way, in humans.

00:38:22.850 --> 00:38:24.220
Here's the really
stunning one.

00:38:24.220 --> 00:38:28.430
Identical twins reared apart,
70.7 correlation.

00:38:28.430 --> 00:38:30.640
Identical twins reared apart.

00:38:30.640 --> 00:38:33.570
That suggests a huge
influence of genes.

00:38:36.170 --> 00:38:39.750
Identical twins reared
together, 90.9.

00:38:39.750 --> 00:38:41.830
So this is the finding
that stunned people.

00:38:41.830 --> 00:38:45.020
So look at identical
twins reared apart.

00:38:45.020 --> 00:38:48.670
They're much more like each
other than a brother and

00:38:48.670 --> 00:38:51.490
sister reared together, who
share all the same environment

00:38:51.490 --> 00:38:54.410
of education, parental support,
and other things.

00:38:54.410 --> 00:38:57.070
So I'm going to come
back to this.

00:38:57.070 --> 00:39:00.190
But as best people can estimate,
and these might be

00:39:00.190 --> 00:39:02.540
radically replaced as we deepen
our understanding of

00:39:02.540 --> 00:39:05.460
the actual genes and the actual
environmental factors,

00:39:05.460 --> 00:39:07.370
but people are estimating
it's about half

00:39:07.370 --> 00:39:10.150
and half for IQ tests.

00:39:10.150 --> 00:39:13.210
And we discussed the bit of
arbitrariness in that.

00:39:13.210 --> 00:39:14.600
So about half and half.

00:39:14.600 --> 00:39:16.370
Let me show you a couple of
environmental influences that

00:39:16.370 --> 00:39:17.960
have been identified.

00:39:17.960 --> 00:39:23.160
And the first two, one suspects
has to do on average

00:39:23.160 --> 00:39:25.470
with something about the
healthiness of the environment

00:39:25.470 --> 00:39:26.920
the child was born into.

00:39:26.920 --> 00:39:31.830
The third one is kind of fun
just to talk about because you

00:39:31.830 --> 00:39:34.950
may wonder if you have a sibling
or no sibling, what

00:39:34.950 --> 00:39:37.120
does it mean that you were first
born, or second born, or

00:39:37.120 --> 00:39:38.960
third born, at least
statistically.

00:39:38.960 --> 00:39:40.300
So let's do the first two.

00:39:40.300 --> 00:39:44.570
Here's breast feeding, 3,000
people were followed from

00:39:44.570 --> 00:39:46.720
birth to young adulthood, who
had breast feeding, for at

00:39:46.720 --> 00:39:48.670
least six months, correlated
with a five to

00:39:48.670 --> 00:39:50.590
seven point IQ gain.

00:39:50.590 --> 00:39:53.820
So we said, 15 points really
seems to mean something in

00:39:53.820 --> 00:39:54.740
educational outcome.

00:39:54.740 --> 00:39:56.610
So 5 to 7 is half of that.

00:39:56.610 --> 00:39:59.560
So that's something.

00:39:59.560 --> 00:40:00.580
So let's talk about
this for a minute.

00:40:00.580 --> 00:40:04.890
You could say, and I can tell
you as a parent who fairly

00:40:04.890 --> 00:40:09.130
recently had children, all the
mothers in the US that I know

00:40:09.130 --> 00:40:10.480
were going like, I've got
to breast feed and

00:40:10.480 --> 00:40:13.420
get my kid to MIT.

00:40:13.420 --> 00:40:15.250
Two more months of breast
feeding and this kid could go

00:40:15.250 --> 00:40:18.200
all the way to tenured
faculty.

00:40:18.200 --> 00:40:21.380
I can tell you there's a lot of
pressure on mothers, if you

00:40:21.380 --> 00:40:22.600
don't know this.

00:40:22.600 --> 00:40:23.900
Not my mother.

00:40:23.900 --> 00:40:27.110
But now mothers hear, do this
for your kid, do this for your

00:40:27.110 --> 00:40:30.570
kid, do this for your kid,
because of results like these.

00:40:30.570 --> 00:40:33.810
So one possibility is there's
something in breast milk.

00:40:33.810 --> 00:40:35.000
That would be the
direct mechanism

00:40:35.000 --> 00:40:36.620
that's making kids smarter.

00:40:36.620 --> 00:40:38.580
Give me another possibility that
has nothing to do with

00:40:38.580 --> 00:40:39.220
the breast milk?

00:40:39.220 --> 00:40:39.580
Yeah?

00:40:39.580 --> 00:40:42.940
AUDIENCE: Maybe mothers who
breast feed, also tend to be

00:40:42.940 --> 00:40:44.380
mothers who are more nurturing
to their children.

00:40:44.380 --> 00:40:45.640
PROFESSOR: It could be
more nurturing to

00:40:45.640 --> 00:40:46.780
children in many ways.

00:40:46.780 --> 00:40:50.050
They could be mothers who can
stay home and spend some time

00:40:50.050 --> 00:40:52.480
with the children versus being
under very pressured economic

00:40:52.480 --> 00:40:54.910
circumstances that
prohibit that.

00:40:54.910 --> 00:40:56.420
It could be a lot of things.

00:40:56.420 --> 00:40:57.300
Does that make sense?

00:40:57.300 --> 00:40:57.950
We don't know.

00:40:57.950 --> 00:41:00.720
Nobody has shown yet that it is
the breast milk that goes

00:41:00.720 --> 00:41:03.870
into your brain and pushes up
your brain, in any sense.

00:41:03.870 --> 00:41:05.330
It's just a correlation.

00:41:05.330 --> 00:41:08.750
And of course, if you're a
parent, I'll do the breast

00:41:08.750 --> 00:41:11.180
feeding and let the scientists
figure it out in the future.

00:41:11.180 --> 00:41:12.240
Here's another one.

00:41:12.240 --> 00:41:14.200
There's the birth weight.

00:41:14.200 --> 00:41:15.440
So again, you could say
the same thing.

00:41:15.440 --> 00:41:19.160
I mean certainly having very low
birth weight is a big risk

00:41:19.160 --> 00:41:20.680
factor for lots of things.

00:41:20.680 --> 00:41:24.240
But birth weight again goes
with many factors in the

00:41:24.240 --> 00:41:26.230
environment that are supportive
for a healthy

00:41:26.230 --> 00:41:28.070
pregnancy versus not.

00:41:28.070 --> 00:41:29.510
So it doesn't mean that
it's the birth weight.

00:41:29.510 --> 00:41:31.640
It could be all the other
things that were in the

00:41:31.640 --> 00:41:35.280
environment during the
pregnancy, at birth, and

00:41:35.280 --> 00:41:38.580
continue, on average, in the
environment of that child,

00:41:38.580 --> 00:41:40.740
throughout their childhood
and young education.

00:41:40.740 --> 00:41:41.840
Does that make sense?

00:41:41.840 --> 00:41:44.280
If they're born in a very
unsupportive medical

00:41:44.280 --> 00:41:47.370
environment, they might have
a low birth weight.

00:41:47.370 --> 00:41:49.380
And unless their parents win
the lottery or something,

00:41:49.380 --> 00:41:51.840
they're likely to be in a
difficult environment for all

00:41:51.840 --> 00:41:55.500
their educational
years to come.

00:41:55.500 --> 00:41:58.320
So people have tried to make
clear, because it's so easy to

00:41:58.320 --> 00:42:00.430
get into sort of little mental
traps on genes versus

00:42:00.430 --> 00:42:02.330
environment, how genetic
diversity affects

00:42:02.330 --> 00:42:03.670
heritability.

00:42:03.670 --> 00:42:05.000
So there's a couple different
ways to do it.

00:42:05.000 --> 00:42:05.890
Let's do this one.

00:42:05.890 --> 00:42:09.160
So imagine you had cloned
tomato plants.

00:42:09.160 --> 00:42:12.520
That means they all have the
exact same genes in this

00:42:12.520 --> 00:42:13.790
hypothetical example.

00:42:13.790 --> 00:42:17.020
But somebody has poor soil,
somebody has good soil.

00:42:17.020 --> 00:42:19.130
Well, you'll get big
differences.

00:42:19.130 --> 00:42:21.470
And the heritability
will be zero.

00:42:21.470 --> 00:42:23.880
Environment can have
huge influences.

00:42:23.880 --> 00:42:28.160
But if you have genetically
diverse plants, then you're

00:42:28.160 --> 00:42:29.210
going to have an interaction.

00:42:29.210 --> 00:42:32.200
So some plants will do OK,
even in poor soil.

00:42:32.200 --> 00:42:35.110
Some plants will do most
spectacularly in good soil and

00:42:35.110 --> 00:42:36.090
less spectacularly alike.

00:42:36.090 --> 00:42:38.120
Environment by genes'
interactions.

00:42:38.120 --> 00:42:39.660
And you can see this in so
many different ways.

00:42:39.660 --> 00:42:40.430
So here's one example.

00:42:40.430 --> 00:42:43.730
So we said height is 90%
heritable within a society.

00:42:43.730 --> 00:42:47.440
But Japanese men in the US, or
people with Japanese extract,

00:42:47.440 --> 00:42:50.600
are three inches taller
than people in Japan.

00:42:50.600 --> 00:42:53.630
So they still have whatever
Japanese genes are.

00:42:53.630 --> 00:42:54.470
They have Japanese genes.

00:42:54.470 --> 00:42:57.240
But they're in the US getting
the US diet of McDonald's.

00:42:57.240 --> 00:43:00.000
And it's pushing them
up three inches.

00:43:03.450 --> 00:43:04.660
Height is heritable.

00:43:04.660 --> 00:43:06.980
But environment is a big
part of the story.

00:43:06.980 --> 00:43:09.370
So it's just both.

00:43:09.370 --> 00:43:12.560
Now, here's a fun study.

00:43:12.560 --> 00:43:16.930
And a reason that if you're
a younger sibling and you

00:43:16.930 --> 00:43:19.420
desperately want two IQ points,
you might knock off

00:43:19.420 --> 00:43:20.570
your older sibling.

00:43:20.570 --> 00:43:21.700
I don't recommend this.

00:43:21.700 --> 00:43:23.310
This is tongue-in-cheek.

00:43:23.310 --> 00:43:25.580
But it's kind of fun study,
just thinking about birth

00:43:25.580 --> 00:43:27.160
order more in a fun way.

00:43:27.160 --> 00:43:28.520
Well, look at the whole
thing, we're talking

00:43:28.520 --> 00:43:29.950
about from 99 to 104.

00:43:29.950 --> 00:43:32.520
So it's five points we're
talking there.

00:43:32.520 --> 00:43:35.410
But in some countries, they sort
of test everybody going

00:43:35.410 --> 00:43:37.180
into the military.

00:43:37.180 --> 00:43:39.730
And I think this was
done in Denmark.

00:43:39.730 --> 00:43:40.980
And what they said is.

00:43:43.900 --> 00:43:45.510
These are all children.

00:43:45.510 --> 00:43:46.910
These are the IQ scores.

00:43:46.910 --> 00:43:50.130
And these are the IQ scores
that you would have.

00:43:50.130 --> 00:43:52.040
But what happens if your
older sibling has

00:43:52.040 --> 00:43:54.246
passed away, sadly.

00:43:54.246 --> 00:43:56.250
You get those points.

00:43:56.250 --> 00:43:58.770
And here's a person who
has an older sibling.

00:43:58.770 --> 00:44:00.010
And the older sibling
passed away.

00:44:00.010 --> 00:44:01.460
And they moved up
those points.

00:44:01.460 --> 00:44:03.970
That is, if you're a younger
sibling, you have a few lower

00:44:03.970 --> 00:44:07.230
IQ points, on average.

00:44:07.230 --> 00:44:09.410
But kind of strikingly, if
your older sibling passes

00:44:09.410 --> 00:44:13.110
away, sadly; you seem to get
those IQ points as you move up

00:44:13.110 --> 00:44:14.580
the birth order.

00:44:14.580 --> 00:44:15.840
Yeah?

00:44:15.840 --> 00:44:17.620
AUDIENCE: Is this if they pass
away before birth, or --?

00:44:17.620 --> 00:44:18.280
PROFESSOR: Yes.

00:44:18.280 --> 00:44:19.700
Yeah, when they're young.

00:44:19.700 --> 00:44:20.880
It's too late.

00:44:20.880 --> 00:44:21.520
Sorry.

00:44:21.520 --> 00:44:25.550
For all you in here, too late.

00:44:25.550 --> 00:44:27.420
It's only a couple
points also.

00:44:27.420 --> 00:44:29.060
But this got into Science,
because people are always

00:44:29.060 --> 00:44:30.070
wondering about--

00:44:30.070 --> 00:44:32.490
there's been a huge curiosity
about birth order

00:44:32.490 --> 00:44:33.740
consequences.

00:44:41.670 --> 00:44:44.200
But let's talk a little bit
about birth order again.

00:44:44.200 --> 00:44:46.340
Because there's a second really
interesting point.

00:44:46.340 --> 00:44:48.720
So here's Raven's score.

00:44:48.720 --> 00:44:52.440
And here is if you're a lone
child, the second of two

00:44:52.440 --> 00:44:55.590
children, the third of three,
the fourth of four, the eighth

00:44:55.590 --> 00:44:56.860
of eight, to ninth of nine.

00:44:56.860 --> 00:44:58.300
You can see the scores
keep going down.

00:44:58.300 --> 00:45:00.920
It's not dramatic, but
larger families.

00:45:00.920 --> 00:45:04.350
And you can have a million ideas
about why that would be.

00:45:04.350 --> 00:45:06.770
About why it is the larger your
family and the further

00:45:06.770 --> 00:45:09.810
you're down on the list, you
lose a few IQ points for every

00:45:09.810 --> 00:45:11.560
sibling who's older.

00:45:11.560 --> 00:45:17.190
So here's one idea from Robert
Zajonc, who's a very creative

00:45:17.190 --> 00:45:17.820
psychologist.

00:45:17.820 --> 00:45:18.960
We're not sure that
it's right.

00:45:18.960 --> 00:45:21.390
But I think it's really
fun to think about.

00:45:21.390 --> 00:45:22.550
So here is this idea.

00:45:22.550 --> 00:45:25.370
He said, I'm going to make up
an idea, which is this.

00:45:25.370 --> 00:45:28.216
The first born, maybe he is
exposed only to adults.

00:45:28.216 --> 00:45:31.980
The adults are talking to them
about the views of the day and

00:45:31.980 --> 00:45:32.570
all kinds of things.

00:45:32.570 --> 00:45:35.330
The second born has two
adults, let's say

00:45:35.330 --> 00:45:36.670
in a typical family.

00:45:36.670 --> 00:45:39.650
But also has an older sibling.

00:45:39.650 --> 00:45:41.600
So the intellectual environment
will be the

00:45:41.600 --> 00:45:45.440
average of two adults
and a little child.

00:45:45.440 --> 00:45:46.430
The next one, and so on.

00:45:46.430 --> 00:45:48.160
So he said, you can do
a kind of a formula,

00:45:48.160 --> 00:45:49.330
which is like this.

00:45:49.330 --> 00:45:52.760
Pretend the two parents are
each 30 and you're born.

00:45:52.760 --> 00:45:55.050
We'll call your family
intellectual stimulating

00:45:55.050 --> 00:45:56.110
environment, 30.

00:45:56.110 --> 00:45:57.890
Because we'll add up the ages
and divide by the number of

00:45:57.890 --> 00:45:59.210
people in the family.

00:45:59.210 --> 00:46:00.670
Does that make sense?

00:46:00.670 --> 00:46:04.260
This is not meant to be
a real formula, just

00:46:04.260 --> 00:46:06.470
a conceptual plan.

00:46:06.470 --> 00:46:08.570
And then he said, now you have
the two parents and you have a

00:46:08.570 --> 00:46:11.550
four-year-old sibling, because
you're the second born.

00:46:11.550 --> 00:46:12.980
So now there's four of you.

00:46:12.980 --> 00:46:13.840
We divide by 4.

00:46:13.840 --> 00:46:15.930
And we'll say the average
intellectual thing in the

00:46:15.930 --> 00:46:17.770
family is 16.

00:46:17.770 --> 00:46:19.950
By the time the next child comes
along, a seven-year-old

00:46:19.950 --> 00:46:22.810
sibling, three-year-old,
divide by five, and the

00:46:22.810 --> 00:46:25.190
average intellectual
environment is 14.

00:46:25.190 --> 00:46:27.060
Here's the idea.

00:46:27.060 --> 00:46:30.730
On average, it's better for
a child to have lots of

00:46:30.730 --> 00:46:33.210
interaction with a parent,
that with a little kid.

00:46:33.210 --> 00:46:35.190
Because a parent will stimulate
you more than a

00:46:35.190 --> 00:46:38.460
little kid, although a little
sibling might be a lot of fun.

00:46:38.460 --> 00:46:42.430
But purely IQ land, two parents
talking to you about

00:46:42.430 --> 00:46:47.150
stuff, is a more educational
environment than two parents

00:46:47.150 --> 00:46:51.130
plus a sibling who is two
years older than you.

00:46:51.130 --> 00:46:52.560
Because the two-year-old
can't help.

00:46:52.560 --> 00:46:55.360
No matter how smart he or she
is, they can't talk with you

00:46:55.360 --> 00:46:56.270
like an adult.

00:46:56.270 --> 00:46:58.680
This is the thought.

00:46:58.680 --> 00:47:02.460
And now, here's the really
fun observation.

00:47:02.460 --> 00:47:04.860
And again, we don't really know
if this is exactly right.

00:47:04.860 --> 00:47:08.990
But every year, if you're around
issues of education,

00:47:08.990 --> 00:47:11.380
there's test scores that
different states have.

00:47:11.380 --> 00:47:13.480
And the test scores go up.

00:47:13.480 --> 00:47:17.690
And you have taken high-tests
tests right, in high schools

00:47:17.690 --> 00:47:19.620
and grade schools and stuff.

00:47:19.620 --> 00:47:23.140
If the test scores goes up,
everybody goes, oh, reelect us

00:47:23.140 --> 00:47:24.860
because we're awesome
administration and our test

00:47:24.860 --> 00:47:25.530
scores are going up.

00:47:25.530 --> 00:47:26.830
We're doing everything right.

00:47:26.830 --> 00:47:30.020
If the test scores go down, they
go oh, throw the bums out

00:47:30.020 --> 00:47:30.950
of the administration.

00:47:30.950 --> 00:47:33.410
Their kids are watching too
much MTV, turn off that

00:47:33.410 --> 00:47:34.710
computer, and so on.

00:47:34.710 --> 00:47:35.910
Everybody gets the
test scores.

00:47:35.910 --> 00:47:39.100
And they figure out who's good
and who's bad, parents,

00:47:39.100 --> 00:47:42.200
teachers, school systems,
politicians, whatever.

00:47:42.200 --> 00:47:44.030
Zajonc says, and this is kind
of an interesting thought.

00:47:44.030 --> 00:47:48.750
He says no, he can predict
whether scores will go up or

00:47:48.750 --> 00:47:54.480
down simply by on average in a
given year, whether there are

00:47:54.480 --> 00:47:56.950
more first borns, or second
borns, or third borns.

00:47:56.950 --> 00:47:59.460
That if he just takes that
population average of every

00:47:59.460 --> 00:48:04.260
kid, the moment they're born,
he can predict years later

00:48:04.260 --> 00:48:05.800
where the scores go.

00:48:05.800 --> 00:48:08.970
And so here's the
scores in Iowa.

00:48:08.970 --> 00:48:10.860
Of the actual scores, they're
going up and down.

00:48:10.860 --> 00:48:12.410
So you can imagine, scores
are going down.

00:48:12.410 --> 00:48:15.810
Everybody's really upset about
kids who don't study, and

00:48:15.810 --> 00:48:18.350
teachers who don't teach, and
parents who don't parent.

00:48:18.350 --> 00:48:19.200
Scores are going up.

00:48:19.200 --> 00:48:20.920
And the kids are awesome, the
teachers are awesome, the

00:48:20.920 --> 00:48:21.920
parents are awesome.

00:48:21.920 --> 00:48:23.670
Because that's who
gets the credit.

00:48:23.670 --> 00:48:25.990
And he's saying, uh-uh.

00:48:25.990 --> 00:48:32.370
This is my prediction here,
based simply on for a given

00:48:32.370 --> 00:48:35.220
birth year, the average
number of older

00:48:35.220 --> 00:48:37.800
children in your family.

00:48:37.800 --> 00:48:38.550
That's it.

00:48:38.550 --> 00:48:39.720
And look how well that tracks.

00:48:39.720 --> 00:48:41.460
It's unbelievable.

00:48:41.460 --> 00:48:44.680
So it just makes you rethink
that so often when we get

00:48:44.680 --> 00:48:49.310
information about society of
what's going on, this may be

00:48:49.310 --> 00:48:50.440
right or wrong.

00:48:50.440 --> 00:48:52.460
But we don't really know
what's going on.

00:48:52.460 --> 00:48:55.190
There's all kinds of huge things
out there in the world

00:48:55.190 --> 00:49:00.220
that aren't the easiest things
to blame or give credit to.

00:49:00.220 --> 00:49:03.160
So they're relatively small
differences, but they seem to

00:49:03.160 --> 00:49:06.090
track interestingly in the
number of siblings you have

00:49:06.090 --> 00:49:08.842
who are older than you.

00:49:08.842 --> 00:49:11.490
And at the same time,
this idea of

00:49:11.490 --> 00:49:13.650
parental exposure to language.

00:49:13.650 --> 00:49:17.980
There's this widely cited study
from Hart and Risley.

00:49:17.980 --> 00:49:21.890
So they recorded for each month,
for 2 and 1/2 years, a

00:49:21.890 --> 00:49:25.240
heroic recording, one full hour
of every word spoken at

00:49:25.240 --> 00:49:28.550
home between parents and
children in just 42 families.

00:49:28.550 --> 00:49:32.220
But they categorized them
as high, medium, or low

00:49:32.220 --> 00:49:33.130
socioeconomically.

00:49:33.130 --> 00:49:34.830
So in terms of income
and education.

00:49:34.830 --> 00:49:35.570
OK.

00:49:35.570 --> 00:49:37.730
And they're recording
lots and lots of

00:49:37.730 --> 00:49:39.420
conversations verbatim.

00:49:39.420 --> 00:49:42.870
And they code and analyzed
every utterance in 1,300

00:49:42.870 --> 00:49:46.140
transcripts of 30,000 pages.

00:49:46.140 --> 00:49:48.680
The reason why this is not
done a lot, it's not very

00:49:48.680 --> 00:49:50.350
glamorous work to do.

00:49:50.350 --> 00:49:51.440
You can imagine sitting
down there.

00:49:51.440 --> 00:49:55.010
But they're recording lots of
conversations at home in

00:49:55.010 --> 00:49:58.880
families who are better off,
medium off, or poor.

00:49:58.880 --> 00:50:02.280
And they're looking at what is
the conversation like, for an

00:50:02.280 --> 00:50:03.640
hour a day.

00:50:03.640 --> 00:50:05.120
And here's what they find.

00:50:05.120 --> 00:50:07.460
That variation in children's
IQ and language abilities

00:50:07.460 --> 00:50:09.020
related to the amount
that the parents

00:50:09.020 --> 00:50:10.780
speak to their children.

00:50:10.780 --> 00:50:14.000
And that the academic success by
age 8 or 9 or 10 is related

00:50:14.000 --> 00:50:16.860
a lot to speech at 0 to 3.

00:50:16.860 --> 00:50:19.080
Now again, you think causally.

00:50:19.080 --> 00:50:21.780
Because you think, parents are
talking to you at 0 and 3.

00:50:21.780 --> 00:50:22.840
And 9 and 10, you're doing

00:50:22.840 --> 00:50:25.330
differential equations, causally.

00:50:25.330 --> 00:50:25.630
No.

00:50:25.630 --> 00:50:28.030
The parents that are talking
to you a lot, 0 to 3, are

00:50:28.030 --> 00:50:30.870
talking to you at 4, 5, and
6 a lot, most of the time.

00:50:30.870 --> 00:50:33.420
So if these things are not like,
you get an injection,

00:50:33.420 --> 00:50:34.740
you're done.

00:50:34.740 --> 00:50:36.960
Many of the factors that we
can measure, continue on

00:50:36.960 --> 00:50:40.570
average, for most people
in their environment.

00:50:40.570 --> 00:50:43.960
But it's very impressive how
the environment altogether

00:50:43.960 --> 00:50:44.990
influences things.

00:50:44.990 --> 00:50:47.880
By age 3, the cumulative
vocabulary is 1,100 words in a

00:50:47.880 --> 00:50:51.480
professional family as a child,
750 in a working class,

00:50:51.480 --> 00:50:53.400
and 500 in a welfare family.

00:50:53.400 --> 00:50:56.870
And welfare families, 600 words
per hour, working class

00:50:56.870 --> 00:50:59.800
750, professionals 2,153.

00:50:59.800 --> 00:51:03.580
By age 3, the vocabulary used
by children in the high SES

00:51:03.580 --> 00:51:08.270
homes, is larger than the
vocabulary used by parents in

00:51:08.270 --> 00:51:09.080
the low SES.

00:51:09.080 --> 00:51:11.210
By age 3, larger vocabulary.

00:51:11.210 --> 00:51:15.600
This is just telling you the
incredible cumulative power of

00:51:15.600 --> 00:51:18.120
strong environmental
influences.

00:51:18.120 --> 00:51:20.160
Because it's unbelievably
multiplicative.

00:51:20.160 --> 00:51:21.635
It's every day, all day long.

00:51:21.635 --> 00:51:24.170
And if you're in an environment
that's supportive

00:51:24.170 --> 00:51:26.420
or an environment that's not.

00:51:26.420 --> 00:51:27.630
It's not one shot.

00:51:27.630 --> 00:51:31.840
It's every day, multiply,
multiply, multiply.

00:51:31.840 --> 00:51:33.940
300 words more per hour
for professional,

00:51:33.940 --> 00:51:34.770
than the welfare family.

00:51:34.770 --> 00:51:35.460
Extrapolate to a year.

00:51:35.460 --> 00:51:37.870
A child in a professional family
hears 11 million words,

00:51:37.870 --> 00:51:40.250
versus 3 million in
a welfare family.

00:51:40.250 --> 00:51:43.110
And strong correlation with
IQ scores by age 9.

00:51:43.110 --> 00:51:46.360
So very huge differences
in what children are

00:51:46.360 --> 00:51:47.480
exposed to at home.

00:51:47.480 --> 00:51:49.610
And they correlate highly
with these IQ things.

00:51:52.210 --> 00:51:56.440
So the literature is that on
average, on average, the IQ

00:51:56.440 --> 00:51:58.830
scores that children have around
9 or 10, correlate

00:51:58.830 --> 00:52:00.160
pretty well with ones
they'll have for the

00:52:00.160 --> 00:52:01.750
rest of their lives.

00:52:01.750 --> 00:52:04.560
Before that, it extremely
jumps around.

00:52:04.560 --> 00:52:06.090
We don't know why.

00:52:06.090 --> 00:52:07.720
But that's they're finding.

00:52:07.720 --> 00:52:11.360
By the time in 9 or 10, whatever
you have, on average

00:52:11.360 --> 00:52:12.640
tends to hang in there.

00:52:12.640 --> 00:52:15.950
If you're 100 at 10, you'll
tend to be 100 at 20.

00:52:15.950 --> 00:52:18.090
If you're 120 at 10, you'll
tend to be 120 at 20.

00:52:18.090 --> 00:52:20.860
Does that make sense?

00:52:20.860 --> 00:52:22.620
So there's been a lot of
interest in evidence about

00:52:22.620 --> 00:52:26.850
whether you can change
your IQ as an adult.

00:52:26.850 --> 00:52:29.230
So here's one study that got a
lot of attention, because it

00:52:29.230 --> 00:52:31.130
was kind of interesting
in that way.

00:52:31.130 --> 00:52:34.530
So they did a complicated
task like this.

00:52:34.530 --> 00:52:35.750
This was an n-back task.

00:52:35.750 --> 00:52:38.560
Simultaneously you hear letters
and you see little

00:52:38.560 --> 00:52:40.220
spatial locations like this.

00:52:40.220 --> 00:52:44.040
And every time something matched
two items ago, you

00:52:44.040 --> 00:52:44.730
would push a button.

00:52:44.730 --> 00:52:45.920
So you're hearing them
at the same time.

00:52:45.920 --> 00:52:47.170
And if you hear a letter
that was two letters

00:52:47.170 --> 00:52:47.920
ago, you push a button.

00:52:47.920 --> 00:52:49.480
Or same spatial location,
you push one.

00:52:49.480 --> 00:52:50.710
It's pretty hard.

00:52:50.710 --> 00:52:52.610
They took 34 adults.

00:52:52.610 --> 00:52:55.590
They trained them from
eight to 19 days,

00:52:55.590 --> 00:52:56.480
so less than a month.

00:52:56.480 --> 00:52:57.550
Skipped weekends.

00:52:57.550 --> 00:52:58.450
25 minutes a day.

00:52:58.450 --> 00:53:00.470
It's only half an hour a day.

00:53:00.470 --> 00:53:01.780
They train on this.

00:53:01.780 --> 00:53:05.280
And then they tested them on
some of these widely used IQ

00:53:05.280 --> 00:53:07.080
types of measures.

00:53:07.080 --> 00:53:09.940
Here's their performance on
the task itself, getting

00:53:09.940 --> 00:53:10.770
better and better.

00:53:10.770 --> 00:53:12.280
OK, that's not surprising.

00:53:12.280 --> 00:53:14.120
And the more they practiced,
the higher they got.

00:53:14.120 --> 00:53:16.305
So if you practice really hard
at something, you get better

00:53:16.305 --> 00:53:17.980
at doing it.

00:53:17.980 --> 00:53:20.690
That's not the thing they
were interested in.

00:53:20.690 --> 00:53:22.380
They were interested in this.

00:53:22.380 --> 00:53:24.170
Here's the IQ scores before
they were trained.

00:53:24.170 --> 00:53:25.020
The IQ scores.

00:53:25.020 --> 00:53:26.180
They didn't practice
the IQ tests.

00:53:26.180 --> 00:53:27.930
And here's the IQ
ones afterwards.

00:53:27.930 --> 00:53:29.670
This is the control group,
who did nothing.

00:53:29.670 --> 00:53:31.320
This is the group that did
the mental practice.

00:53:31.320 --> 00:53:32.960
And you can see, they pushed
up their IQ scores.

00:53:32.960 --> 00:53:33.850
Not a lot.

00:53:33.850 --> 00:53:38.010
But it was only 20 minutes a
day for less than a month.

00:53:38.010 --> 00:53:40.010
So these kinds of studies
really intrigued

00:53:40.010 --> 00:53:41.810
people in two ways.

00:53:41.810 --> 00:53:43.920
And the more training you
got, the more your

00:53:43.920 --> 00:53:44.596
IQ score went up.

00:53:44.596 --> 00:53:46.410
The more weeks you did this.

00:53:46.410 --> 00:53:49.330
Because a), maybe we could push
up IQ scores, if that's

00:53:49.330 --> 00:53:50.480
important to push up.

00:53:50.480 --> 00:53:52.240
This is a big decision.

00:53:52.240 --> 00:53:53.790
Even when you're an adult.

00:53:53.790 --> 00:53:56.650
What if you did more than 20
minutes a day for three weeks.

00:53:56.650 --> 00:54:00.400
And b), maybe we could push up
everybody's IQ score, who

00:54:00.400 --> 00:54:02.090
needed help in that, with
the right kind of

00:54:02.090 --> 00:54:04.070
training, at any age.

00:54:04.070 --> 00:54:06.880
So a lot of interest in that.

00:54:06.880 --> 00:54:10.720
So let me take a step now to
sort of harder topics.

00:54:10.720 --> 00:54:14.730
Now, one of the reasons that
intelligence and IQ topics

00:54:14.730 --> 00:54:18.860
have gotten appropriately lots
of concern is a terrible

00:54:18.860 --> 00:54:22.930
history of racism and sexism
associated with IQ tests and

00:54:22.930 --> 00:54:26.790
the way people use that to argue
basically inferiority of

00:54:26.790 --> 00:54:29.250
groups, other than themselves.

00:54:29.250 --> 00:54:32.490
So the US Immigration Act of
1924, the official government

00:54:32.490 --> 00:54:34.800
talked about biologically
weak stocks.

00:54:34.800 --> 00:54:38.060
At that time, they were focused
or prejudiced against

00:54:38.060 --> 00:54:40.070
Italians and Jews from Europe.

00:54:40.070 --> 00:54:42.540
Different societies, different
times, as you know,

00:54:42.540 --> 00:54:47.020
tragically, had biases against
different groups in the world.

00:54:47.020 --> 00:54:49.520
Unfortunately, it's part of
the world we live in.

00:54:49.520 --> 00:54:52.930
And even though we know now with
certainty that we're all

00:54:52.930 --> 00:54:55.660
almost identical genetically in
any practical sense across

00:54:55.660 --> 00:55:01.720
racial groups, race remains a
powerful social category.

00:55:01.720 --> 00:55:03.645
It's a very complicated topic.

00:55:03.645 --> 00:55:06.530
The experience you've had, is
the experience you've had.

00:55:06.530 --> 00:55:09.680
But we know it's a
social category.

00:55:09.680 --> 00:55:12.380
People identify to a certain
extent with that.

00:55:12.380 --> 00:55:15.060
They recognize other people
in that context.

00:55:15.060 --> 00:55:16.780
We'll come back to that
later in the course.

00:55:16.780 --> 00:55:19.320
But it's a sort of fact of life
that we often think of

00:55:19.320 --> 00:55:21.980
ourselves as belonging to one
racial group or another and

00:55:21.980 --> 00:55:26.070
somebody else as belonging to
one racial group or another.

00:55:26.070 --> 00:55:28.240
So I'm going to focus on one
thing that's both troubled

00:55:28.240 --> 00:55:31.910
people and led to bad things,
but has some interesting,

00:55:31.910 --> 00:55:34.070
important aspects
to think about.

00:55:34.070 --> 00:55:38.210
So in the US, when people
compare African American and

00:55:38.210 --> 00:55:41.840
European American scores,
there's a gap of scores that

00:55:41.840 --> 00:55:42.630
grows over time.

00:55:42.630 --> 00:55:45.130
So in infancy, to the extent
you can test these kinds of

00:55:45.130 --> 00:55:46.020
things, there's no difference.

00:55:46.020 --> 00:55:50.740
By age 4, on average, people who
are European American will

00:55:50.740 --> 00:55:52.160
tend to score about four
or five points

00:55:52.160 --> 00:55:53.540
higher on IQ tests.

00:55:53.540 --> 00:55:55.930
And as the goes up, it
continues to go up.

00:55:55.930 --> 00:55:57.530
There's a gap on average.

00:55:57.530 --> 00:55:58.320
This is a huge average.

00:55:58.320 --> 00:56:04.370
It's across many people across
time, as people get

00:56:04.370 --> 00:56:05.560
older in the US.

00:56:05.560 --> 00:56:10.140
Now, we've strongly believed
for many, many reasons

00:56:10.140 --> 00:56:12.400
empirically that this is an
environmental influence.

00:56:12.400 --> 00:56:15.040
And here's one example of
many pieces of evidence.

00:56:15.040 --> 00:56:18.340
So that children fathered, for
example, by Black America GIs

00:56:18.340 --> 00:56:21.260
in Germany, brought up by German
mothers, have exactly

00:56:21.260 --> 00:56:23.870
the same IQs as white
American GI

00:56:23.870 --> 00:56:24.740
children and German mothers.

00:56:24.740 --> 00:56:28.620
That is, in countries where
stereotypes were applied in

00:56:28.620 --> 00:56:32.120
general, partly because there
wasn't a history of African

00:56:32.120 --> 00:56:34.650
Americans in Germany at
that time, the IQ

00:56:34.650 --> 00:56:35.370
scores are the same.

00:56:35.370 --> 00:56:38.300
So we think it's overwhelmingly
evident that

00:56:38.300 --> 00:56:39.480
it's an environmental effect.

00:56:39.480 --> 00:56:43.610
But in the US, there's a lot
of relationship between, on

00:56:43.610 --> 00:56:46.630
average, across the entire
country, on average, between

00:56:46.630 --> 00:56:48.710
what group you're in
and average income.

00:56:48.710 --> 00:56:51.160
So here's the average income a
few years ago, depending on

00:56:51.160 --> 00:56:54.160
which group you identified
yourself as belonging to.

00:56:54.160 --> 00:56:59.130
So we talked earlier about the
relationship between economic

00:56:59.130 --> 00:57:01.880
comfort or stability or support
and things like the

00:57:01.880 --> 00:57:05.600
number of words a parent can
speak to or it's time to speak

00:57:05.600 --> 00:57:06.310
to their child.

00:57:06.310 --> 00:57:08.310
And the habit seems to correlate
with all kinds of

00:57:08.310 --> 00:57:09.730
advantages.

00:57:09.730 --> 00:57:14.270
So here's an incredibly powerful
and important idea.

00:57:14.270 --> 00:57:16.865
Claude Steele is a researcher
at Stanford, who I knew, who

00:57:16.865 --> 00:57:19.100
has played a big part in this.

00:57:19.100 --> 00:57:20.310
And here's the idea.

00:57:20.310 --> 00:57:22.390
So it's called stereotype
threat.

00:57:22.390 --> 00:57:23.980
It applies to all of us.

00:57:23.980 --> 00:57:27.270
But in some ways depending on
the world we live in, some of

00:57:27.270 --> 00:57:29.970
us can be more vulnerable than
others in certain situations.

00:57:29.970 --> 00:57:31.040
Here's the idea.

00:57:31.040 --> 00:57:32.350
So let me tell you the
experiment and then the

00:57:32.350 --> 00:57:33.070
interpretation.

00:57:33.070 --> 00:57:35.490
The experiment is that he'll
bring people in and give,

00:57:35.490 --> 00:57:39.390
Claude or his students, black
and white participants,

00:57:39.390 --> 00:57:40.400
college students.

00:57:40.400 --> 00:57:41.940
And they say, it's a laboratory
experiment.

00:57:41.940 --> 00:57:44.760
And they give them something
like an SAT test.

00:57:44.760 --> 00:57:48.140
And the groups perform
just the same.

00:57:48.140 --> 00:57:50.920
Stanford students, black and
white, perform just the same.

00:57:50.920 --> 00:57:52.330
That's the one group.

00:57:52.330 --> 00:57:55.480
Here comes the next group of
black and white students, who

00:57:55.480 --> 00:57:59.670
are told we're going to give
you a test of intelligence.

00:57:59.670 --> 00:58:00.710
Now, look what happens.

00:58:00.710 --> 00:58:03.800
The scores of African
Americans go down.

00:58:03.800 --> 00:58:05.610
And it doesn't even have to
be the phrase "test of

00:58:05.610 --> 00:58:08.530
intelligence," you can have
people just tick off, on a

00:58:08.530 --> 00:58:12.480
front page, what racial group
they belong to and you get

00:58:12.480 --> 00:58:14.790
almost the same effect.

00:58:14.790 --> 00:58:15.530
So what is this?

00:58:15.530 --> 00:58:18.380
This is understood as stereotype
vulnerability.

00:58:18.380 --> 00:58:23.580
That is, tragically in the US
there's a stereotype that

00:58:23.580 --> 00:58:26.680
African Americans
are not as good

00:58:26.680 --> 00:58:29.890
academically as white Americans.

00:58:29.890 --> 00:58:31.700
It's a pernicious stereotype.

00:58:31.700 --> 00:58:34.400
It's one we can decry morally
and say it's a bad thing.

00:58:34.400 --> 00:58:35.990
But it's out there.

00:58:35.990 --> 00:58:38.360
There's another area where
African Americans are thought

00:58:38.360 --> 00:58:42.100
of as phenomenal by stereotype,
which is sports,

00:58:42.100 --> 00:58:42.760
in this country.

00:58:42.760 --> 00:58:44.010
It's a weird thing.

00:58:49.260 --> 00:58:53.130
The idea is that knowledge of
these things, even if this is

00:58:53.130 --> 00:58:56.040
your real performance, once that
thought enters your mind

00:58:56.040 --> 00:58:59.380
that I'm doing something that
some people say I'm not going

00:58:59.380 --> 00:59:02.820
to do such a good job at,
is enough to make you

00:59:02.820 --> 00:59:06.550
underperform compared to
your true potential.

00:59:06.550 --> 00:59:09.410
So the other area where this
has been studied a lot is

00:59:09.410 --> 00:59:11.040
women and mathematics.

00:59:11.040 --> 00:59:11.980
I think that's disappearing.

00:59:11.980 --> 00:59:16.110
When I grew up, everybody was
worried that women weren't

00:59:16.110 --> 00:59:18.160
given as good an academic
support as men

00:59:18.160 --> 00:59:19.290
or girls and boys.

00:59:19.290 --> 00:59:21.530
OK, it's reversed.

00:59:21.530 --> 00:59:22.810
In my lifetime, it's reversed.

00:59:22.810 --> 00:59:24.490
Because now, by every
criteria, women are

00:59:24.490 --> 00:59:27.040
outperforming men in the US by
every criteria, at least

00:59:27.040 --> 00:59:29.470
societally, the number who
finish school, your beginning

00:59:29.470 --> 00:59:30.700
salaries, everything.

00:59:30.700 --> 00:59:32.950
So now they're worried
that us boys are

00:59:32.950 --> 00:59:34.740
really being left behind.

00:59:34.740 --> 00:59:38.280
So these things change because
they're cultural inventions.

00:59:38.280 --> 00:59:39.650
They're cultural inventions.

00:59:39.650 --> 00:59:42.980
And cultural morays change.

00:59:42.980 --> 00:59:46.470
And sometimes in good
ways, hopefully.

00:59:46.470 --> 00:59:50.680
So stereotype threat is that
other's judgments or one's own

00:59:50.680 --> 00:59:53.930
actions will confirm negative
stereotypes about one's group.

00:59:53.930 --> 00:59:56.130
Oh, you should do badly on
something or you should do

00:59:56.130 --> 00:59:57.590
well on something.

00:59:57.590 --> 01:00:00.390
So they studied on women
on that, old people.

01:00:00.390 --> 01:00:02.380
Athletic ability of course
runs the opposite.

01:00:02.380 --> 01:00:03.710
The experiment was
kind of fun.

01:00:03.710 --> 01:00:06.330
They had a very athletic African
American instructor

01:00:06.330 --> 01:00:08.560
come up and start
doing push-ups.

01:00:08.560 --> 01:00:11.420
And all the white guys started
to lapse in their push-ups.

01:00:11.420 --> 01:00:13.680
They got tired really fast,
compared to an equally

01:00:13.680 --> 01:00:17.510
athletic white trainer who
came up to the front.

01:00:17.510 --> 01:00:19.290
Does that make sense?

01:00:19.290 --> 01:00:20.360
Does that make sense?

01:00:20.360 --> 01:00:22.490
It's whatever is the stereotype
you have in your

01:00:22.490 --> 01:00:25.793
head, if you say OK, oh, oh, I'm
not supposed to do so good

01:00:25.793 --> 01:00:27.160
in this, you underperform.

01:00:27.160 --> 01:00:28.850
Not everybody, all the
time of course.

01:00:28.850 --> 01:00:31.260
But enough that we have to
be concerned about that.

01:00:31.260 --> 01:00:34.050
More for academic performance
probably than push-ups.

01:00:34.050 --> 01:00:36.020
But who knows.

01:00:36.020 --> 01:00:38.730
So here's this achievement
gap.

01:00:38.730 --> 01:00:45.540
And different scores between
again, at age 13, between

01:00:45.540 --> 01:00:47.830
whites and blacks in the United
States, which we know

01:00:47.830 --> 01:00:49.190
shouldn't be there.

01:00:49.190 --> 01:00:50.565
Many factors are in it.

01:00:50.565 --> 01:00:52.590
A huge variation among
people, of course.

01:00:52.590 --> 01:00:54.220
These are huge averages.

01:00:54.220 --> 01:00:57.320
But one that everybody wants
to get rid of, in a just

01:00:57.320 --> 01:01:01.070
society where everybody has
an equal opportunity.

01:01:01.070 --> 01:01:05.440
So here's a result that's so
unbelievable, that if it

01:01:05.440 --> 01:01:07.410
weren't for some subsequent
results, I wouldn't even show

01:01:07.410 --> 01:01:09.550
it to you, even though it's
a well-done experiment.

01:01:09.550 --> 01:01:11.800
But the message of it
seems so incredible.

01:01:11.800 --> 01:01:13.330
So here's what they did.

01:01:13.330 --> 01:01:15.200
In this well-controlled study
in Science, and there's been

01:01:15.200 --> 01:01:17.850
some follow-ups that have
exactly done this.

01:01:17.850 --> 01:01:22.230
So they took African American
students and white students.

01:01:22.230 --> 01:01:25.690
And they had them take
these various tests.

01:01:25.690 --> 01:01:27.788
But some of them, at the
beginning of the year, wrote

01:01:27.788 --> 01:01:28.540
an affirmation.

01:01:28.540 --> 01:01:30.570
They had to choose a value
from a list and write an

01:01:30.570 --> 01:01:33.050
in-class essay about why
the chosen value is

01:01:33.050 --> 01:01:34.470
important to you.

01:01:34.470 --> 01:01:36.790
The controls chose a value
from those same lists and

01:01:36.790 --> 01:01:38.630
wrote an in-class essay about
why the value might be

01:01:38.630 --> 01:01:40.340
important for somebody else.

01:01:40.340 --> 01:01:42.710
You're writing a one-time essay
about why something

01:01:42.710 --> 01:01:45.370
matters to you or matters
to somebody else.

01:01:45.370 --> 01:01:49.380
So obviously what matters to you
is more powerful for you.

01:01:49.380 --> 01:01:52.650
And by just doing that, there's
a 40% reduction in the

01:01:52.650 --> 01:01:54.480
African American students
in the study in

01:01:54.480 --> 01:01:55.370
their achievement gap.

01:01:55.370 --> 01:02:00.570
Their scores went up 40% because
they wrote one essay

01:02:00.570 --> 01:02:04.580
at the beginning of the year
about why values are important

01:02:04.580 --> 01:02:06.430
to them and related
to education.

01:02:06.430 --> 01:02:08.770
One essay did that.

01:02:08.770 --> 01:02:09.970
Unbelievable.

01:02:09.970 --> 01:02:13.870
And just a reminder about how
much one's attitude incredibly

01:02:13.870 --> 01:02:16.220
influences your performance.

01:02:16.220 --> 01:02:18.310
You have a range of things
you can do it.

01:02:18.310 --> 01:02:21.040
And what gets you to the top of
your range that's in you,

01:02:21.040 --> 01:02:23.060
or the bottom of your range,
has a lot to do with

01:02:23.060 --> 01:02:25.940
attitudes, that can be moved
around by a lot of things.

01:02:25.940 --> 01:02:29.450
One essay overcame a tremendous
amount of the

01:02:29.450 --> 01:02:32.680
concerns that on average,
African American students have

01:02:32.680 --> 01:02:35.520
about academic achievement.

01:02:35.520 --> 01:02:36.980
And it's not just that.

01:02:36.980 --> 01:02:39.630
So in this course, you only get
a couple of tips of things

01:02:39.630 --> 01:02:40.880
that are useful to
you for courses.

01:02:40.880 --> 01:02:41.550
Here's one.

01:02:41.550 --> 01:02:43.260
It's the same principle,
but it seems

01:02:43.260 --> 01:02:45.560
to apply for everybody.

01:02:45.560 --> 01:02:47.600
So here's the experiment, also
published in Science.

01:02:47.600 --> 01:02:49.050
It's really cool.

01:02:49.050 --> 01:02:51.850
So they brought in students.

01:02:51.850 --> 01:02:54.020
And they said, look students
often are worried about how

01:02:54.020 --> 01:02:56.000
they're going to do in a test,
especially if it's an area you

01:02:56.000 --> 01:02:58.440
think you're not so good in.

01:02:58.440 --> 01:03:02.200
So they had them take a test on
Gauss's modular arithmetic.

01:03:02.200 --> 01:03:03.030
Here's the pre-test.

01:03:03.030 --> 01:03:04.800
Everybody did about the same.

01:03:04.800 --> 01:03:07.580
Then they said, OK now, they
have to create pressure.

01:03:07.580 --> 01:03:10.240
Because of course if you take
a test in an experiment,

01:03:10.240 --> 01:03:12.190
that's not pressured, right?

01:03:12.190 --> 01:03:14.240
Pressure is the grade you're
going to get, whether it lets

01:03:14.240 --> 01:03:16.420
you into medical school, whether
you get to become the

01:03:16.420 --> 01:03:19.400
head of the American Medical
Association.

01:03:19.400 --> 01:03:20.590
That's pressure, right?

01:03:20.590 --> 01:03:22.860
So they say oh, here's what
we're going to do now.

01:03:22.860 --> 01:03:24.830
We're going to give you another
test, everybody

01:03:24.830 --> 01:03:26.000
scoring about the same.

01:03:26.000 --> 01:03:29.030
But you can make a fair bit of
money with a partner who's

01:03:29.030 --> 01:03:29.770
down the hall.

01:03:29.770 --> 01:03:31.420
We should tell you that partner
has already done a

01:03:31.420 --> 01:03:34.230
really good job on the second
version of the test.

01:03:34.230 --> 01:03:36.930
So if you don't do well, you
don't get money and the

01:03:36.930 --> 01:03:39.490
partner down the hall doesn't
get your money.

01:03:39.490 --> 01:03:41.985
If you do well in the second
test you're about to take now,

01:03:41.985 --> 01:03:42.900
then everybody wins.

01:03:42.900 --> 01:03:44.240
So they're trying to make
you feel pressured.

01:03:44.240 --> 01:03:45.750
On man, I want to
win the money.

01:03:45.750 --> 01:03:47.740
Plus there's a guy down the hall
who already did well and

01:03:47.740 --> 01:03:49.430
I'm going to let us down.

01:03:49.430 --> 01:03:50.750
They're trying to
make you feel as

01:03:50.750 --> 01:03:51.610
pressured as you might.

01:03:51.610 --> 01:03:54.370
As much as they can in an
experiment, like you do in a

01:03:54.370 --> 01:03:56.330
regular test, in a
regular course.

01:03:56.330 --> 01:03:58.750
So trying to create pressure
and anxiety.

01:03:58.750 --> 01:04:01.130
And then for 10 minutes, the
controls just sit there and

01:04:01.130 --> 01:04:01.850
think about it.

01:04:01.850 --> 01:04:03.480
Oh, I hope I don't
mess up now.

01:04:03.480 --> 01:04:04.270
The pressure is on.

01:04:04.270 --> 01:04:06.120
I hope I don't meet the person,
if I don't do a good

01:04:06.120 --> 01:04:07.845
job, down the hall.

01:04:07.845 --> 01:04:10.530
An expressive group wrote an
essay for 10 minutes about

01:04:10.530 --> 01:04:13.676
their thoughts and feelings.

01:04:13.676 --> 01:04:15.030
Oh, I wrote this.

01:04:15.030 --> 01:04:16.980
This is a huge mistake.

01:04:16.980 --> 01:04:21.240
This should be "better."
I'll fix it up.

01:04:21.240 --> 01:04:22.350
Better, better, better.

01:04:22.350 --> 01:04:23.860
I did it unrelated,
in writing.

01:04:23.860 --> 01:04:25.420
So here's what happens.

01:04:25.420 --> 01:04:28.920
The group that just sat
there did nothing.

01:04:28.920 --> 01:04:32.990
Their performance doesn't--
let me get this right.

01:04:32.990 --> 01:04:34.380
The performers that wrote the
expressive writing, their

01:04:34.380 --> 01:04:35.690
score goes up.

01:04:35.690 --> 01:04:37.990
Everybody else's
scores go down.

01:04:37.990 --> 01:04:40.330
Because here, here's where
they're beginning.

01:04:40.330 --> 01:04:42.810
Now, they get pressured.

01:04:42.810 --> 01:04:44.930
And they perform less well.

01:04:44.930 --> 01:04:47.670
They're choking.

01:04:47.670 --> 01:04:49.190
The group that wrote
the essay, their

01:04:49.190 --> 01:04:52.040
performance stays up.

01:04:52.040 --> 01:04:56.260
And it's kind of ironic, because
a theme of comedy is

01:04:56.260 --> 01:04:59.610
frequently when they tell you,
whatever you do, don't tell

01:04:59.610 --> 01:05:00.830
somebody x.

01:05:00.830 --> 01:05:02.030
And then my mistake,
you tell x.

01:05:02.030 --> 01:05:04.590
Have you seen that
a comedy show?

01:05:04.590 --> 01:05:07.120
So you could have thought, well
writing an essay about my

01:05:07.120 --> 01:05:09.880
thoughts and feelings about how
much anxiety and pressure

01:05:09.880 --> 01:05:11.910
I feel, is like the worst
thing you can do.

01:05:11.910 --> 01:05:13.610
It's like, I don't
want face it.

01:05:13.610 --> 01:05:15.530
I'd rather clear my mind.

01:05:15.530 --> 01:05:18.300
But it turns out that when
people write this little essay

01:05:18.300 --> 01:05:21.030
about the thoughts and feelings
about their anxiety,

01:05:21.030 --> 01:05:24.100
their scores rise.

01:05:24.100 --> 01:05:26.920
And if they don't write that
kind of an essay, the scores

01:05:26.920 --> 01:05:29.030
fall, because you've introduced
here the stress and

01:05:29.030 --> 01:05:31.780
anxiety that we think is a
part of regular testing.

01:05:31.780 --> 01:05:33.280
Does that makes sense?

01:05:33.280 --> 01:05:33.990
I'll fix the slide.

01:05:33.990 --> 01:05:35.480
I'm sorry about the typo.

01:05:35.480 --> 01:05:36.730
Is that OK?

01:05:39.460 --> 01:05:41.130
I'm almost a little afraid to
give you advice, because

01:05:41.130 --> 01:05:43.200
somebody will write an essay,
do bad on the physics test,

01:05:43.200 --> 01:05:44.690
and say oh, this is terrible.

01:05:44.690 --> 01:05:46.090
I would have gotten a A.

01:05:46.090 --> 01:05:48.790
But it seems like writing a
little essay about your

01:05:48.790 --> 01:05:50.900
thoughts and feelings actually
pushes scores up.

01:05:50.900 --> 01:05:52.860
I mean that's the scientific
evidence.

01:05:52.860 --> 01:05:56.540
But again, it's just a reminder
of how powerful

01:05:56.540 --> 01:05:59.440
feelings and attitudes are,
whether they're in us or

01:05:59.440 --> 01:06:03.390
visited on us by the world, in
how well we perform, within

01:06:03.390 --> 01:06:08.260
the range of performance that
are within us to start with.

01:06:08.260 --> 01:06:09.890
I have two more points.

01:06:09.890 --> 01:06:12.440
So we talked a little bit about
entity versus growth.

01:06:12.440 --> 01:06:14.700
You know people who believe
they're like, I'm only so good

01:06:14.700 --> 01:06:15.560
at something.

01:06:15.560 --> 01:06:17.260
I'm only so good at basketball,
I'm only so good

01:06:17.260 --> 01:06:19.220
at math, I'm only so good at
programming, I'm only so good

01:06:19.220 --> 01:06:21.700
of writing, or whatever, versus
people who say, the

01:06:21.700 --> 01:06:24.400
harder I work, the
better I'll get.

01:06:24.400 --> 01:06:27.300
And that pretty much, this work
from Carol Dweck, that

01:06:27.300 --> 01:06:30.090
many people tend to go for
one thing or the other.

01:06:30.090 --> 01:06:31.720
Is it are we born with
talent and that's

01:06:31.720 --> 01:06:33.040
pretty much the deal?

01:06:33.040 --> 01:06:34.620
Or the harder we work,
do we get more

01:06:34.620 --> 01:06:36.500
talented at doing something?

01:06:36.500 --> 01:06:37.700
And so this is the kind
of questionnaire.

01:06:37.700 --> 01:06:39.280
And these are literally
questions, so she gives

01:06:39.280 --> 01:06:42.170
students, you have a certain
amount of intelligence and you

01:06:42.170 --> 01:06:43.360
really can't change it.

01:06:43.360 --> 01:06:44.840
I strongly agree.

01:06:44.840 --> 01:06:47.210
That's this trait model.

01:06:47.210 --> 01:06:48.140
Or I disagree.

01:06:48.140 --> 01:06:51.480
The harder I work, the
smarter I can be.

01:06:51.480 --> 01:06:54.170
And what she did for example, is
give that kind of question

01:06:54.170 --> 01:06:55.800
to students.

01:06:55.800 --> 01:06:58.450
And then follow them in their
performance on a math course.

01:06:58.450 --> 01:06:59.950
Here's the ones who believed
they could get better.

01:06:59.950 --> 01:07:02.140
Here's the ones who said they
only have so much math ability

01:07:02.140 --> 01:07:03.570
and intelligence.

01:07:03.570 --> 01:07:05.050
And look at the ones who thought
they could get better,

01:07:05.050 --> 01:07:06.500
got better.

01:07:06.500 --> 01:07:08.830
So that seems like a really
good thing to have, is not

01:07:08.830 --> 01:07:10.500
only the positive attitude.

01:07:10.500 --> 01:07:14.540
But a real belief that the
harder the work, the smarter

01:07:14.540 --> 01:07:15.790
you'll get.

01:07:18.890 --> 01:07:20.720
Last two slides.

01:07:20.720 --> 01:07:22.300
So finally, finally, finally.

01:07:22.300 --> 01:07:25.220
And we've had a couple of
them, but almost none, a

01:07:25.220 --> 01:07:27.800
longitudinal, random
assignment study.

01:07:27.800 --> 01:07:30.870
Which doesn't sound hugely
exciting, but that's the best

01:07:30.870 --> 01:07:32.400
science we can ever do.

01:07:32.400 --> 01:07:34.160
Randomly assign people
to things.

01:07:34.160 --> 01:07:37.200
And then follow them over time
to understand how life works.

01:07:37.200 --> 01:07:42.180
So one of the famous studies is,
this is Perry Preschool.

01:07:42.180 --> 01:07:46.030
So it followed into adulthood,
123 children from low SES

01:07:46.030 --> 01:07:49.200
environments in Michigan.

01:07:49.200 --> 01:07:51.740
And they were randomly divided,
randomly divided.

01:07:51.740 --> 01:07:54.540
And there's no randomness
right, in life, mostly.

01:07:54.540 --> 01:07:56.470
You're not randomly
given genes.

01:07:56.470 --> 01:07:58.650
You're not randomly put
into a household.

01:07:58.650 --> 01:08:01.550
It's always things are super
correlated with each other.

01:08:01.550 --> 01:08:02.890
And it's very hard to
tease them apart.

01:08:02.890 --> 01:08:05.910
But here they took kids who were
from poor families and

01:08:05.910 --> 01:08:07.990
put them randomly
into two groups.

01:08:07.990 --> 01:08:10.830
One group is a program group,
who received a high-quality

01:08:10.830 --> 01:08:14.080
active learning preschool
program and home visits, to be

01:08:14.080 --> 01:08:14.900
supportive as well.

01:08:14.900 --> 01:08:19.130
So they got help at home
and help in preschool.

01:08:19.130 --> 01:08:22.010
And a no-program group, who
didn't get any of these.

01:08:22.010 --> 01:08:24.960
I can tell you that when you do
random assignment studies

01:08:24.960 --> 01:08:29.100
in real life like this, some
people argue, I don't agreed

01:08:29.100 --> 01:08:30.399
with this, that it's unethical,
because you should

01:08:30.399 --> 01:08:33.450
help everybody all the time.

01:08:33.450 --> 01:08:34.810
But if you don't do this,
you never figure

01:08:34.810 --> 01:08:35.630
out what helps people.

01:08:35.630 --> 01:08:38.520
OK It's a little bit
of a circular--

01:08:38.520 --> 01:08:41.520
you can't figure out what
really helps people.

01:08:41.520 --> 01:08:43.439
And then in a really rare
thing-- so it's random

01:08:43.439 --> 01:08:44.140
assignment.

01:08:44.140 --> 01:08:48.760
They followed them ages 3 to 11,
14 to 15, 19, 27, and 40.

01:08:48.760 --> 01:08:50.290
They follow you through
most of your

01:08:50.290 --> 01:08:52.600
childhood and adult life.

01:08:52.600 --> 01:08:55.250
So what happens?

01:08:55.250 --> 01:08:57.600
Here's their IQ scores?

01:08:57.600 --> 01:09:00.160
Here's the program people
in red and the

01:09:00.160 --> 01:09:02.000
no-program people in blue.

01:09:02.000 --> 01:09:05.290
And you can see that for a few
years, they separate by IQ.

01:09:05.290 --> 01:09:08.340
But by age 7, they're pretty
much the same.

01:09:08.340 --> 01:09:11.880
And when these kind of data came
out, because of course

01:09:11.880 --> 01:09:13.630
after the preschool, they
go back to their

01:09:13.630 --> 01:09:15.330
regular local schools.

01:09:15.330 --> 01:09:17.350
They're pretty much
the same schools.

01:09:17.350 --> 01:09:20.660
So people, after that they said,
well Head Start is a

01:09:20.660 --> 01:09:24.210
kind of a program to help kids
who are poor, get a head start

01:09:24.210 --> 01:09:25.500
in education and support.

01:09:25.500 --> 01:09:26.810
It's no good.

01:09:26.810 --> 01:09:28.770
It doesn't do anything.

01:09:28.770 --> 01:09:31.000
Because once they get back into
schools that not very

01:09:31.000 --> 01:09:33.740
supportive, the benefits of
the program disappear.

01:09:33.740 --> 01:09:37.120
You didn't do anything for the
kid, because unless you fix

01:09:37.120 --> 01:09:40.350
everything in their lives, that
one year just doesn't

01:09:40.350 --> 01:09:42.470
make a difference
in the long run.

01:09:42.470 --> 01:09:46.080
So everybody said, Head Start
programs are just, sadly, a

01:09:46.080 --> 01:09:49.870
waste of resources, because
they don't do anything for

01:09:49.870 --> 01:09:51.689
these children over
the longer term.

01:09:51.689 --> 01:09:55.350
But here's the remarkable
thing.

01:09:55.350 --> 01:09:57.100
They kept following
these kids.

01:09:57.100 --> 01:09:58.040
And here's what they found.

01:09:58.040 --> 01:10:02.640
For example, in the number of
arrests, 36 had five or more

01:10:02.640 --> 01:10:06.120
arrests in the program group.

01:10:06.120 --> 01:10:08.580
But all the way up to 55, for
the no-program group.

01:10:08.580 --> 01:10:10.920
That's a big difference, how
often you're arrested.

01:10:10.920 --> 01:10:13.410
That's a big outcome in life.

01:10:13.410 --> 01:10:17.110
Earning $20,000 or more at age
40, program group 60%;

01:10:17.110 --> 01:10:19.180
no-program group, 40%.

01:10:19.180 --> 01:10:21.650
Completed high school,
45% to 65%.

01:10:21.650 --> 01:10:26.850
20% increase in completion of
high school, from one year of

01:10:26.850 --> 01:10:28.385
help before kindergarten.

01:10:30.980 --> 01:10:34.230
Basic achievement at
age 14, way up.

01:10:34.230 --> 01:10:37.680
Willingness to do homework,
way up.

01:10:37.680 --> 01:10:39.940
So what's that's telling
us is, IQ is a

01:10:39.940 --> 01:10:40.940
pretty interesting thing.

01:10:40.940 --> 01:10:42.100
And we've talked
a lot about it.

01:10:42.100 --> 01:10:44.410
It is an interesting thing.

01:10:44.410 --> 01:10:47.890
But there's something really
deep in the ways you can help

01:10:47.890 --> 01:10:51.360
people, that make them
succeed in life.

01:10:51.360 --> 01:10:55.390
Not get in trouble, avoid bad
things, make money, finish

01:10:55.390 --> 01:10:57.610
school, do your work.

01:10:57.610 --> 01:11:01.750
That one year of school changed
these kids for the

01:11:01.750 --> 01:11:04.390
rest of their lives,
on average.

01:11:04.390 --> 01:11:08.260
And imagine if you had done
two years or five years?

01:11:08.260 --> 01:11:10.730
And IQ in this case doesn't
capture it at all.

01:11:10.730 --> 01:11:13.310
There's something else in them
that changed, that made them

01:11:13.310 --> 01:11:16.730
more resilient in life and
successful in the long run.

01:11:16.730 --> 01:11:20.380
So there's a lot of pieces to
who does well and not well in

01:11:20.380 --> 01:11:21.850
ways we wish for everybody.

01:11:21.850 --> 01:11:23.430
And IQ is a piece
of that story.

01:11:23.430 --> 01:11:25.450
But there's a lot of pieces.

01:11:25.450 --> 01:11:26.700
So thanks very much.