WEBVTT

1
00:00:00.303 --> 00:00:02.730
<v ->If you have a sneaky suspicion that liberals</v>

2
00:00:02.730 --> 00:00:04.640
are prone to violence, you're more likely to pay

3
00:00:04.640 --> 00:00:07.729
attention to raucous leftist demonstrations at Berkeley

4
00:00:07.729 --> 00:00:09.652
while down playing acts of violence

5
00:00:09.652 --> 00:00:12.333
at Donald Trump's own campaign rallies.

6
00:00:12.333 --> 00:00:14.492
<v ->They'd be carried out on a stretcher folks.</v>

7
00:00:14.492 --> 00:00:16.017
<v ->And liberals tend to do the opposite.</v>

8
00:00:16.017 --> 00:00:17.503
That's confirmation bias.

9
00:00:17.503 --> 00:00:19.866
The tendency we all have to interpret information

10
00:00:19.866 --> 00:00:22.295
in a way that supports what we already believe.

11
00:00:22.295 --> 00:00:23.484
And it's what we're talking about

12
00:00:23.484 --> 00:00:25.242
in this episode of Argument Clinic,

13
00:00:25.242 --> 00:00:28.162
a guide to spotting bad arguments on the internet.

14
00:00:28.162 --> 00:00:30.589
We are all inherently biased.

15
00:00:30.589 --> 00:00:32.347 line:15% 
We tend to weigh evidence more heavily if it

16
00:00:32.347 --> 00:00:35.625 line:15% 
confirms our belief and discount evidence that doesn't.

17
00:00:35.625 --> 00:00:38.239
And the internet, especially social media,

18
00:00:38.239 --> 00:00:41.179
is a giant confirmation biased machine.

19
00:00:41.179 --> 00:00:42.780
It makes it easy to find evidence to support

20
00:00:42.780 --> 00:00:44.725
just about any argument you want to make

21
00:00:44.725 --> 00:00:46.506
and ignore all the other stuff.

22
00:00:46.506 --> 00:00:48.482
For instance, let's say you wanna prove

23
00:00:48.482 --> 00:00:50.015
that Trump supporters are violent.

24
00:00:50.015 --> 00:00:52.321
So you Google 'violent Trump supporters.'

25
00:00:52.321 --> 00:00:54.997
You'll find a long list of ugly incidents.

26
00:00:54.997 --> 00:00:56.202
Of course, there may be millions of

27
00:00:56.202 --> 00:00:57.834
peaceful Trump supporters out there

28
00:00:57.834 --> 00:00:59.221
but you didn't Google that.

29
00:00:59.221 --> 00:01:02.214
Social media like Facebook presents the same problem.

30
00:01:02.214 --> 00:01:04.445
We tend to like information that we agree with,

31
00:01:04.445 --> 00:01:05.629
which means we see more of it,

32
00:01:05.629 --> 00:01:07.437
which means we think there's more of it,

33
00:01:07.437 --> 00:01:08.941
which means we must be right.

34
00:01:08.941 --> 00:01:09.774
Right?

35
00:01:09.774 --> 00:01:12.124
That's why we have stuff like the scientific method,

36
00:01:12.124 --> 00:01:14.822
which uses control groups and placebos

37
00:01:14.822 --> 00:01:16.362
and peer review to ensure that our

38
00:01:16.362 --> 00:01:19.677
preexisting ideas don't effect our findings.

39
00:01:19.677 --> 00:01:21.641
It's also why lawyers try to screen out

40
00:01:21.641 --> 00:01:23.777
biased jury members who's experiences

41
00:01:23.777 --> 00:01:25.474
and beliefs might prevent them

42
00:01:25.474 --> 00:01:27.921
from weighing evidence and arguments objectively.

43
00:01:27.921 --> 00:01:30.271
Now unfortunately, most of the information online

44
00:01:30.271 --> 00:01:31.954
doesn't meet those standards.

45
00:01:31.954 --> 00:01:33.418
Just the opposite in fact.

46
00:01:33.418 --> 00:01:36.089
Most of it is engineered to create an emotional response,

47
00:01:36.089 --> 00:01:38.134
not a rational one, because emotions are

48
00:01:38.134 --> 00:01:39.900
what get us to click and share.

49
00:01:39.900 --> 00:01:41.880
And there's no better way to stir people's emotions

50
00:01:41.880 --> 00:01:44.686
than by convincing them that they are right about everything

51
00:01:44.686 --> 00:01:48.317
and that anyone who disagrees is a moron.

52
00:01:48.317 --> 00:01:50.783
We'll never be able to eradicate confirmation bias,

53
00:01:50.783 --> 00:01:53.465
but we can at least try to limit it in our own lives.

54
00:01:53.465 --> 00:01:56.162
So just remember, anecdotes are not data.

55
00:01:56.162 --> 00:01:58.371
Be open to arguments that challenge your beliefs

56
00:01:58.371 --> 00:02:01.117
and be modest and skeptical of everything you read,

57
00:02:01.117 --> 00:02:04.539
even if, no, especially if, it seems to prove

58
00:02:04.539 --> 00:02:07.162
what you already thought in the first place.

59
00:02:07.162 --> 00:02:08.623
[jolly music]