WEBVTT

00:00:00.111 --> 00:00:02.271
- [Colin] We have a saying 
in neuroscience sometimes:

00:00:02.271 --> 00:00:04.383
"Don't ask the person,
ask the brain,"

00:00:04.383 --> 00:00:06.484
because the brain activity
may be something

00:00:06.484 --> 00:00:09.256
that's less than fully conscious.

00:00:09.678 --> 00:00:11.795
♪ [music] ♪

00:00:18.815 --> 00:00:20.850
My research is about
behavioral economics

00:00:20.850 --> 00:00:22.105
and neuroeconomics.

00:00:22.105 --> 00:00:23.490
And behavioral economics

00:00:23.490 --> 00:00:26.828
is using ideas from psychology
and other social sciences

00:00:26.828 --> 00:00:28.799
to make economics
a little bit more lifelike

00:00:28.799 --> 00:00:31.310
and fit human behavior.

00:00:31.310 --> 00:00:33.761
The neuroeconomics part 
is that we actually try to see

00:00:33.761 --> 00:00:35.211
what's happening in the brain

00:00:35.211 --> 00:00:37.300
when people are making 
economic decisions.

00:00:37.300 --> 00:00:40.270
Hypothetical bias is a term 
for when you ask somebody

00:00:40.270 --> 00:00:41.748
whether they're going
to do something,

00:00:41.748 --> 00:00:43.147
but there's no actual consequences,

00:00:43.147 --> 00:00:45.042
like a lot of surveys:
"Are you going to vote?"

00:00:45.042 --> 00:00:47.061
"Will you buy 
this new product we have?"

00:00:47.621 --> 00:00:50.397
You tend to get a kind 
of upward "yes" bias.

00:00:50.397 --> 00:00:52.793
People are more likely to say,
"Yeah, I'd think I'd buy it"

00:00:52.793 --> 00:00:54.239
or "Oh, yeah,
I'm planning to vote."

00:00:54.239 --> 00:00:56.131
The hypothetical bias
can be pretty high,

00:00:56.131 --> 00:00:57.881
and it can be also retrospective.

00:00:57.881 --> 00:00:59.529
So if you ask people did they vote,

00:00:59.529 --> 00:01:02.862
70% say yes and really 
the answer was 45%.

00:01:03.206 --> 00:01:05.097
One application is 
in things like marketing.

00:01:05.097 --> 00:01:07.913
A lot of new products fail,
and one reason they fail

00:01:07.913 --> 00:01:09.631
is because when 
they test marketed,

00:01:09.631 --> 00:01:11.414
a lot of people said, 
"Yes, I'd buy it,"

00:01:11.414 --> 00:01:12.630
who weren't going to buy it.

00:01:12.630 --> 00:01:14.247
One thing that people 
have been chasing

00:01:14.247 --> 00:01:16.861
in different fields in economics
and psychology is

00:01:16.861 --> 00:01:20.303
how could we measure the size 
of this bias and adjust for it?

00:01:20.303 --> 00:01:23.102
So that if 70% of the people say
they're going to buy a new product,

00:01:23.102 --> 00:01:25.683
we know that 
the real number is 45%.

00:01:29.619 --> 00:01:32.742
We did a couple of studies
using brain imaging to say

00:01:32.742 --> 00:01:34.773
is there a kind
of signature in the brain

00:01:34.773 --> 00:01:37.026
of when somebody says, 
"Yes, I would buy it,"

00:01:37.026 --> 00:01:39.461
but when they really have 
to choose, they say no.

00:01:39.461 --> 00:01:41.442
So we showed them pictures
of different goods,

00:01:41.442 --> 00:01:43.601
and the first part 
of the experiment, we asked them,

00:01:43.601 --> 00:01:46.655
"Would you pay $27 
for this backpack, yes or no?" --

00:01:46.655 --> 00:01:47.908
that's the hypothetical part.

00:01:47.908 --> 00:01:50.671
And then we kind of surprised them,
when they come out of the scanner,

00:01:50.671 --> 00:01:52.071
and we say to them,
"Oh, by the way,

00:01:52.071 --> 00:01:54.816
now we're going to actually
have you decide to spend money,

00:01:54.816 --> 00:01:56.746
so we're going to give you $50."

00:01:56.746 --> 00:01:59.118
If you want to buy
the backpack for $27,

00:01:59.118 --> 00:02:00.693
we're going to take it
out of your 50,

00:02:00.693 --> 00:02:02.560
so that now they have
to make real decisions.

00:02:02.560 --> 00:02:05.034
And then we study 
in the brain imaging,

00:02:05.034 --> 00:02:07.926
could we tell what areas 
were saying yes,

00:02:07.926 --> 00:02:09.966
but actually would later say no,

00:02:09.966 --> 00:02:12.911
compared to the areas that said,
"Yes, I think I'll buy it,"

00:02:12.911 --> 00:02:14.452
and, yes, they really did buy it.

00:02:14.452 --> 00:02:17.219
And we found both
more activity in certain regions

00:02:17.219 --> 00:02:18.927
associated with valuation

00:02:18.927 --> 00:02:20.843
and then activity
in different regions,

00:02:20.843 --> 00:02:22.126
which were somewhat predictive

00:02:22.126 --> 00:02:25.290
of when a yes was going to turn
into, "Well, not really."

00:02:28.821 --> 00:02:30.868
In some other studies 
on hypothetical bias,

00:02:30.868 --> 00:02:33.068
we used eye tracking, 
which is a computerized way

00:02:33.068 --> 00:02:35.467
of seeing what you're looking at
for how long.

00:02:35.467 --> 00:02:37.249
It also measures pupil dilation.

00:02:37.249 --> 00:02:39.679
When you're aroused 
by something that you like

00:02:39.679 --> 00:02:42.546
or possibly you're scared,
the pupil dilates a little bit.

00:02:42.546 --> 00:02:44.107
So we used that method

00:02:44.107 --> 00:02:46.217
and then another method
using mouse movements.

00:02:46.217 --> 00:02:48.005
And so we found 
that the mouse tracking

00:02:48.005 --> 00:02:50.337
and the eye tracking
actually could give us an idea

00:02:50.337 --> 00:02:52.974
of when people would say, "Yes, 
I'm going to buy this product,"

00:02:52.974 --> 00:02:55.282
but then they really didn't
when they had skin in the game

00:02:55.282 --> 00:02:56.498
and had to buy it.

00:02:56.498 --> 00:03:00.933
The quicker a person moves a mouse
to a box to click on something,

00:03:00.933 --> 00:03:02.284
the more they like it.

00:03:02.284 --> 00:03:05.190
It's like a fast trajectory,
"I really like this."

00:03:05.190 --> 00:03:08.500
A slow meandering trajectory is, 
"Well, I don't know. I'm not sure."

00:03:08.500 --> 00:03:12.457
The motor activity in the mouse is
actually an index in economic value

00:03:12.457 --> 00:03:15.542
as well as some other things,
like indecision.

00:03:19.701 --> 00:03:21.650
We live in a kind of golden age
of social science

00:03:21.650 --> 00:03:24.303
in which we can measure things
in lots and lots of different ways.

00:03:24.303 --> 00:03:25.819
So one thing we've just begun

00:03:25.819 --> 00:03:28.181
that I think is going 
to be really, really fun

00:03:28.181 --> 00:03:32.728
is in order to study habit
and a bunch of other things,

00:03:32.728 --> 00:03:34.371
we bought a smart vending machine,

00:03:34.371 --> 00:03:35.831
and a smart vending machine

00:03:35.831 --> 00:03:37.620
is basically a vending
machine in the back

00:03:37.620 --> 00:03:39.217
and a giant iPad in the front.

00:03:39.217 --> 00:03:42.285
So you could program the iPad
to show whatever you want.

00:03:42.285 --> 00:03:43.998
For example, if somebody
buys habitually,

00:03:43.998 --> 00:03:46.451
and you raise the price
by a few percent,

00:03:46.451 --> 00:03:47.524
do they just ignore that

00:03:47.524 --> 00:03:49.381
because they're not 
even looking at the price?

00:03:49.381 --> 00:03:51.539
Eventually, we'd like also
to be able to use

00:03:51.539 --> 00:03:53.508
webcams or cameras to record,

00:03:53.508 --> 00:03:55.516
so when I say somebody
is not looking at the price,

00:03:55.516 --> 00:03:57.537
I really mean they are
not looking at the price

00:03:57.537 --> 00:04:00.285
because the camera looked
at where their eyes were looking.

00:04:00.285 --> 00:04:02.418
- [Narrator] Want to see 
more economists in the wild?

00:04:02.418 --> 00:04:04.176
Check out our playlist.

00:04:04.176 --> 00:04:05.228
Are you a teacher?

00:04:05.228 --> 00:04:07.868
Here's some related material 
for your classroom.

00:04:07.868 --> 00:04:09.641
♪ [music] ♪