WEBVTT
Kind: captions
Language: en

1
00:00:00.250 --> 00:00:04.880
Imagine that a researcher asks you to write
down your most important goal.

2
00:00:04.880 --> 00:00:06.549
Then, once you’ve done that,

3
00:00:06.549 --> 00:00:09.580
they give you a story about trees and
ask you to read it to yourself.

4
00:00:09.580 --> 00:00:11.300
"The way trees produce leaves

5
00:00:11.300 --> 00:00:16.010
"is one of the many examples of the
orderly patterns created by nature," it says.

6
00:00:16.010 --> 00:00:19.780
Subconsciously, the thought of trees sprouting
leaves in an orderly fashion

7
00:00:19.780 --> 00:00:21.450
fills you with determination.

8
00:00:21.450 --> 00:00:25.430
And then the researchers ask you some
questions about your most important goal:

9
00:00:25.430 --> 00:00:28.220
how likely are you to pursue it?

10
00:00:28.220 --> 00:00:30.039
This was a real study done in 2014,

11
00:00:30.039 --> 00:00:33.390
in which researchers found that being exposed
to a story about order

12
00:00:33.390 --> 00:00:36.950
would be more motivating than a similar story
about randomness.

13
00:00:36.950 --> 00:00:38.940
And that’s based in the concept of priming,

14
00:00:38.940 --> 00:00:42.760
the idea that exposure to a stimulus,
so a word or a concept,

15
00:00:42.760 --> 00:00:47.240
can cause you to change your response to a
completely different stimulus.

16
00:00:47.240 --> 00:00:48.620
So for this study,

17
00:00:48.620 --> 00:00:51.620
researchers switched out the word "order"
or "randomness" in the story,

18
00:00:51.620 --> 00:00:55.539
and they found that exposure to the word "order"
would prompt a person to say

19
00:00:55.540 --> 00:00:57.600
they were more motivated to pursue a goal.

20
00:00:58.750 --> 00:01:02.960
If that sounds a little too simple to be true,
well, it probably is.

21
00:01:02.969 --> 00:01:04.980
Many Labs, a group of 186 researchers,

22
00:01:04.980 --> 00:01:07.880
have been trying to recreate a lot of
psychology studies like that

23
00:01:07.880 --> 00:01:12.360
using over 60 different labs around the world
and using much larger sample sizes.

24
00:01:12.360 --> 00:01:14.650
14 of the 28 studies that they tested,

25
00:01:14.650 --> 00:01:16.659
including that ‘order’ and ‘randomness’ study,

26
00:01:16.659 --> 00:01:19.720
couldn’t be replicated with the same significance.

27
00:01:19.720 --> 00:01:22.909
Another of the Many Labs studies tested the
impact of a word scramble

28
00:01:22.909 --> 00:01:26.840
with either priming words conveying
"heat" or "cold," or a neutral control,

29
00:01:26.840 --> 00:01:28.800
on people's perception of global warming.

30
00:01:28.800 --> 00:01:32.270
And while the original study found that those
who solved heat-related word-scrambles

31
00:01:32.270 --> 00:01:35.659
were more likely to then say they were worried
about the climate,

32
00:01:35.660 --> 00:01:37.500
Many Labs couldn’t replicate that.

33
00:01:37.500 --> 00:01:40.520
Now, that doesn’t mean
priming doesn’t exist.

34
00:01:40.520 --> 00:01:44.060
Semantic priming has been observed in several
well-respected studies.

35
00:01:44.060 --> 00:01:48.480
Researchers have tracked eye movement and
gaze-length and found that people

36
00:01:48.490 --> 00:01:52.030
are quicker to read word groups
that are semantically related.

37
00:01:52.030 --> 00:01:57.020
So, "gold and silver" is read more quickly
than "gold and horse."

38
00:01:57.020 --> 00:02:00.049
Another study showed that people are faster
to recall pairs of words

39
00:02:00.049 --> 00:02:02.100
if they’re semantically related
to each other.

40
00:02:02.100 --> 00:02:05.010
The researchers followed-up on their own studies,
trying to replicate their results,

41
00:02:05.010 --> 00:02:08.310
and, yeah, they found that the
priming effects did keep happening,

42
00:02:08.310 --> 00:02:11.180
although not always in the same exact way.

43
00:02:11.180 --> 00:02:15.840
It’s safe to say that ‘semantic priming’
does happen when you’re reading and listening:

44
00:02:15.840 --> 00:02:18.900
your brain understands what’s going on
a lot faster

45
00:02:18.900 --> 00:02:18.900
if you hear the words and parts of speech
that you expect to.

46
00:02:18.900 --> 00:02:22.220
if you hear the words and parts of speech
that you jellyfish.

47
00:02:23.840 --> 00:02:26.160
...your brain understands what’s going on
a lot faster

48
00:02:29.390 --> 00:02:32.890
That questionable idea that
priming influences behaviour?

49
00:02:32.890 --> 00:02:36.980
Well, that’s part of a larger issue called
the Replication Crisis.

50
00:02:36.980 --> 00:02:40.930
My own Masters thesis, many years ago, was
an accidental example of this.

51
00:02:40.930 --> 00:02:43.660
In my undergraduate degree, I found a possible
priming effect

52
00:02:43.660 --> 00:02:47.489
where teachers could be influenced to give
higher grades to essays

53
00:02:47.489 --> 00:02:49.440
if they were ‘primed’
by words in those essays,

54
00:02:49.440 --> 00:02:52.910
if the student talked about
success rather than failure.

55
00:02:52.910 --> 00:02:55.800
This was a fascinating result, so for my Masters,

56
00:02:55.800 --> 00:02:59.580
I spent a long time refining, testing, and
making a larger study

57
00:02:59.580 --> 00:03:02.400
that was as scientifically
rigorous as I could make it.

58
00:03:03.180 --> 00:03:04.820
And I found no effect.

59
00:03:04.830 --> 00:03:05.830
Nothing at all.

60
00:03:05.830 --> 00:03:09.420
And, like, sure, I still got my degree,
proving the null hypothesis is still a success,

61
00:03:09.420 --> 00:03:12.780
but… it still hurt to know I was wrong.

62
00:03:12.780 --> 00:03:14.870
The Replication Crisis is the realisation,

63
00:03:14.870 --> 00:03:18.700
across a lot of branches of medicine, life,
behavioral and linguistic studies,

64
00:03:18.700 --> 00:03:23.629
that results we thought were statistically
significant… might not be.

65
00:03:23.629 --> 00:03:25.200
Maybe the sample size was too small,

66
00:03:25.200 --> 00:03:27.150
or biases in sampling
weren’t accounted for:

67
00:03:27.150 --> 00:03:29.500
like they only tested US college students.

68
00:03:29.500 --> 00:03:32.440
Maybe there wasn’t enough control for the
impact of the testing environment,

69
00:03:32.440 --> 00:03:34.560
or there was human error in the test design,

70
00:03:34.560 --> 00:03:38.940
or maybe it's just the sheer complexity
of humans as subjects to begin with.

71
00:03:38.940 --> 00:03:41.840
It turns out that researchers are not always

72
00:03:41.840 --> 00:03:44.680
properly and openly accounting
for all those factors;

73
00:03:44.680 --> 00:03:47.220
they, or their university’s PR department,

74
00:03:47.220 --> 00:03:50.680
present their findings without talking about
the limitations.

75
00:03:50.680 --> 00:03:53.650
Or maybe, if the results are negative or
unimpressive,

76
00:03:53.650 --> 00:03:56.170
they just get filed away and never published.

77
00:03:56.170 --> 00:03:59.370
It doesn’t mean that every statistical study
you see is bunk.

78
00:03:59.370 --> 00:04:00.580
Absolutely not.

79
00:04:00.580 --> 00:04:05.450
But when you see bold claims about how your
brain can be influenced by the language you read,

80
00:04:05.450 --> 00:04:07.140
and they seem a little too good to be true?

81
00:04:07.140 --> 00:04:09.650
It’s right to be skeptical.

82
00:04:09.650 --> 00:04:13.400
Researchers have to take the Replication Crisis
and learn from it, adapt their methods,

83
00:04:13.400 --> 00:04:16.320
and keep trying to understand ourselves.

84
00:04:16.320 --> 00:04:19.400
The script for this video was put together
by a team of writers, including

85
00:04:19.400 --> 00:04:22.920
Gretchen McCulloch, whose podcast
Lingthusiasm is both wonderful

86
00:04:22.920 --> 00:04:24.380
and linked in the description below.