The method that can "prove" almost anything - James A. Smith

816,530 views

2021-08-05 ใƒป TED-Ed


New videos

The method that can "prove" almost anything - James A. Smith

816,530 views ใƒป 2021-08-05

TED-Ed


์•„๋ž˜ ์˜๋ฌธ์ž๋ง‰์„ ๋”๋ธ”ํด๋ฆญํ•˜์‹œ๋ฉด ์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค.

๋ฒˆ์—ญ: Soohyun Park ๊ฒ€ํ† : JY Kang
00:06
In 2011, a group of researchers conducted a scientific study
0
6371
4167
2011๋…„ ํ•œ ์—ฐ๊ตฌ์ง„์€
๊ณผํ•™์ ์ธ ์—ฐ๊ตฌ๋ฅผ ํ†ตํ•ด ๋ง๋„ ์•ˆ๋˜๋Š” ๊ฒฐ๋ก ์„ ๋„์ถœํ–ˆ์Šต๋‹ˆ๋‹ค.
00:10
to find an impossible result:
1
10538
2125
00:12
that listening to certain songs can make you younger.
2
12663
3500
ํŠน์ • ๋…ธ๋ž˜๋ฅผ ๋“ค์œผ๋ฉด ์ Š์–ด์ง„๋‹ค๋Š” ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค.
00:16
Their study involved real people, truthfully reported data,
3
16663
3625
์ด๋“ค์˜ ์—ฐ๊ตฌ๋Š” ์‹ค์ œ ์‚ฌ๋žŒ, ์žˆ๋Š” ๊ทธ๋Œ€๋กœ ์ˆ˜์ง‘๋œ ๋ฐ์ดํ„ฐ,
00:20
and commonplace statistical analyses.
4
20288
3000
์ผ๋ฐ˜์ ์ธ ํ†ต๊ณ„ ๋ถ„์„์œผ๋กœ ์ด๋ฃจ์–ด์กŒ์Šต๋‹ˆ๋‹ค.
00:23
So how did they do it?
5
23288
1416
์–ด๋–ป๊ฒŒ ๊ทธ๋Ÿฐ ๊ฒฐ๋ก ์„ ์–ป์€ ๊ฒƒ์ผ๊นŒ์š”?
00:24
The answer lies in a statistical method scientists often use
6
24704
4125
๋‹ต์€ ๊ณผํ•™์ž๋“ค์ด ์‚ฌ์šฉํ•œ ํ†ต๊ณ„์  ๋ฐฉ๋ฒ•์— ์žˆ์Šต๋‹ˆ๋‹ค.
00:28
to try to figure out whether their results mean something or if theyโ€™re random noise.
7
28829
4875
๊ฒฐ๊ณผ๊ฐ€ ์œ ์˜๋ฏธํ•œ ๊ฒƒ์ธ์ง€ ๋ถˆ๊ทœ์น™์ ์ธ ๋…ธ์ด์ฆˆ์ธ์ง€ ๋ถ„์„ํ•˜๋Š” ๊ฒƒ์ด์ฃ .
00:33
In fact, the whole point of the music study
8
33704
2625
์‚ฌ์‹ค ์Œ์•… ์‹คํ—˜์˜ ํ•ต์‹ฌ์€
00:36
was to point out ways this method can be misused.
9
36329
3917
ํ†ต๊ณ„์  ๋ฐฉ๋ฒ•์ด ์ž˜๋ชป ์‚ฌ์šฉ๋  ์ˆ˜ ์žˆ์Œ์„ ์ง€์ ํ•˜๋Š” ๊ฒƒ์ด์—ˆ์Šต๋‹ˆ๋‹ค.
00:40
A famous thought experiment explains the method:
10
40746
2791
์ด ๋ฐฉ๋ฒ•์„ ์„ค๋ช…ํ•ด์ฃผ๋Š” ์œ ๋ช…ํ•œ ์‚ฌ๊ณ  ์‹คํ—˜์ด ์žˆ์Šต๋‹ˆ๋‹ค.
00:43
there are eight cups of tea,
11
43746
1750
์—ฌ๊ธฐ์— 8์ž”์˜ ์ฐจ๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.
00:45
four with the milk added first, and four with the tea added first.
12
45496
4416
4๊ฐœ๋Š” ์šฐ์œ ๋ฅผ ๋จผ์ € ๋„ฃ๊ณ  ๋‹ค๋ฅธ 4๊ฐœ์—๋Š” ์ฐจ๋ฅผ ๋จผ์ € ๋„ฃ์Šต๋‹ˆ๋‹ค.
00:50
A participant must determine which are which according to taste.
13
50162
3625
์ฐธ๊ฐ€์ž๋Š” ๋ง›์œผ๋กœ ์ด๋“ค์„ ๊ตฌ๋ถ„ํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.
00:53
There are 70 different ways the cups can be sorted into two groups of four,
14
53871
4583
8์ž”์„ 4๊ฐœ์”ฉ 2๊ฐœ ๋ฌถ์Œ์œผ๋กœ ๋‚˜๋ˆ„๋Š” ๋ฐ๋Š” 70๊ฐ€์ง€์˜ ๋ฐฉ๋ฒ•์ด ์žˆ๊ณ 
00:58
and only one is correct.
15
58454
2000
๊ทธ ์ค‘์˜ ํ•˜๋‚˜๋งŒ์ด ์ •๋‹ต์ž…๋‹ˆ๋‹ค.
01:00
So, can she taste the difference?
16
60662
2584
์ฐธ๊ฐ€์ž๋Š” ๋ง›์˜ ์ฐจ์ด๋ฅผ ๋Š๋‚„ ์ˆ˜ ์žˆ์„๊นŒ์š”?
01:03
Thatโ€™s our research question.
17
63246
1625
์ด๊ฒƒ์ด ์šฐ๋ฆฌ์˜ ์—ฐ๊ตฌ ๊ณผ์ œ์ž…๋‹ˆ๋‹ค.
01:04
To analyze her choices, we define whatโ€™s called a null hypothesis:
18
64871
4625
์ฐธ๊ฐ€์ž์˜ ์„ ํƒ์„ ๋ถ„์„ํ•˜๊ธฐ ์œ„ํ•ด ๋‹ค์Œ๊ณผ ๊ฐ™์€ ๊ท€๋ฌด๊ฐ€์„ค์„ ์ •์˜ํ•ฉ๋‹ˆ๋‹ค.
01:09
that she canโ€™t distinguish the teas.
19
69496
2167
๊ฐ€์„ค : ์ฐธ๊ฐ€์ž๋Š” ๋ง›์„ ๊ตฌ๋ถ„ํ•  ์ˆ˜ ์—†๋‹ค.
01:11
If she canโ€™t distinguish the teas,
20
71871
2042
๋งŒ์•ฝ ๋ง›์„ ๊ตฌ๋ถ„ํ•  ์ˆ˜ ์—†๋‹ค๊ณ  ํ•˜๋”๋ผ๋„
01:13
sheโ€™ll still get the right answer 1 in 70 times by chance.
21
73913
5166
์ฐธ๊ฐ€์ž๋Š” 70๋ฒˆ์— 1๋ฒˆ ๊ผด๋กœ ์šฐ์—ฐํžˆ ์ •๋‹ต์„ ๋งž์ถœ ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
01:19
1 in 70 is roughly .014.
22
79079
3334
70๋ฒˆ์— 1๋ฒˆ์ด๋ฉด ์•ฝ 0.014 ์ด๊ณ ,
01:22
That single number is called a p-value.
23
82746
3292
์ด ์ˆซ์ž๋ฅผ p-๊ฐ’์ด๋ผ๊ณ  ํ•ฉ๋‹ˆ๋‹ค.
01:26
In many fields, a p-value of .05 or below is considered statistically significant,
24
86038
6916
๋งŽ์€ ๋ถ„์•ผ์—์„œ 0.05 ์ดํ•˜์˜ p-๊ฐ’์€ ํ†ต๊ณ„์ ์œผ๋กœ ์œ ์˜๋ฏธํ•œ ๊ฒƒ์œผ๋กœ ๊ฐ„์ฃผ๋˜๋ฉฐ
01:32
meaning thereโ€™s enough evidence to reject the null hypothesis.
25
92954
3792
์ด๋Š” ๊ท€๋ฌด๊ฐ€์„ค์ด ํ‹€๋ ธ์Œ์„ ์ž…์ฆํ•˜๊ธฐ์— ์ถฉ๋ถ„ํ•˜๋‹ค๋Š” ๊ฑธ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.
01:36
Based on a p-value of .014,
26
96996
3375
์šฐ๋ฆฌ๋Š” 0.014์˜ p-๊ฐ’์— ๊ทผ๊ฑฐํ•ด
01:40
theyโ€™d rule out the null hypothesis that she canโ€™t distinguish the teas.
27
100371
4125
์ฐธ๊ฐ€์ž๊ฐ€ ๋ง›์„ ๊ตฌ๋ถ„ํ•˜์ง€ ๋ชป ํ•œ๋‹ค๋Š” ๊ท€๋ฌด๊ฐ€์„ค์ด ํ‹€๋ ธ๋‹ค๊ณ  ํ•  ์ˆ˜ ์žˆ๋Š” ๊ฑฐ์ฃ .
01:44
Though p-values are commonly used by both researchers and journals
28
104913
3916
์—ฐ๊ตฌ์ž์™€ ํ•™์ˆ ์ง€ ๋ชจ๋‘ ๊ณผํ•™์  ๊ฒฐ๊ณผ๋ฅผ ํ‰๊ฐ€ํ•˜๊ธฐ ์œ„ํ•ด
01:48
to evaluate scientific results,
29
108829
2084
p-๊ฐ’์„ ํ”ํ•˜๊ฒŒ ์ด์šฉํ•˜์ง€๋งŒ
01:50
theyโ€™re really confusing, even for many scientists.
30
110913
2958
๋งŽ์€ ๊ณผํ•™์ž๋“ค์—๊ฒŒ๋„ p-๊ฐ’์€ ๋งค์šฐ ํ—ท๊ฐˆ๋ฆฌ๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.
01:54
Thatโ€™s partly because all a p-value actually tells us
31
114329
4042
p-๊ฐ’์ด ์‹ค์ œ๋กœ ์šฐ๋ฆฌ์—๊ฒŒ ์•Œ๋ ค์ฃผ๋Š” ๊ฒƒ์€
01:58
is the probability of getting a certain result,
32
118371
3000
๊ท€๋ฌด๊ฐ€์„ค์ด ์ฐธ์ด๋ผ๊ณ  ๊ฐ€์ •ํ•  ๋•Œ
02:01
assuming the null hypothesis is true.
33
121371
2917
ํŠน์ • ๊ฒฐ๊ณผ๋ฅผ ์–ป์„ ํ™•๋ฅ ์ด๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
02:04
So if she correctly sorts the teas,
34
124663
2791
๋”ฐ๋ผ์„œ ๋งŒ์•ฝ์— ์ฐธ๊ฐ€์ž๊ฐ€ ์ฐจ๋ฅผ ์ •ํ™•ํ•˜๊ฒŒ ๋ถ„๋ฅ˜ํ•œ๋‹ค๋ฉด,
02:07
the p-value is the probability of her doing so
35
127454
3417
p-๊ฐ’์€ ์ฐธ๊ฐ€์ž๊ฐ€ ๋ง›์„ ๊ตฌ๋ถ„ํ•  ์ˆ˜ ์—†๋‹ค๋Š” ๊ฐ€์ •ํ•˜์—์„œ
02:10
assuming she canโ€™t tell the difference.
36
130871
2458
๋ง›์„ ๊ตฌ๋ถ„ํ•  ํ™•๋ฅ ์„ ์˜๋ฏธํ•ฉ๋‹ˆ๋‹ค.
02:13
But the reverse isnโ€™t true:
37
133329
2459
ํ•˜์ง€๋งŒ ์ด๊ฒƒ์˜ ๋ฐ˜๋Œ€๋Š” ์ฐธ์ด ์•„๋‹™๋‹ˆ๋‹ค.
02:15
the p-value doesnโ€™t tell us the probability
38
135788
2416
p-๊ฐ’์€ ์šฐ๋ฆฌ๊ฐ€ ์•Œ๊ณ ์ž ํ•˜๋Š” ๊ฒƒ
02:18
that she can taste the difference,
39
138204
1625
์ฆ‰, ์ฐธ๊ฐ€์ž๊ฐ€ ๋ง›์„ ๊ตฌ๋ถ„ํ•  ํ™•๋ฅ ์„ ์•Œ๋ ค์ฃผ์ง€ ์•Š์Šต๋‹ˆ๋‹ค.
02:19
which is what weโ€™re trying to find out.
40
139829
2084
02:22
So if a p-value doesnโ€™t answer the research question,
41
142329
3250
p-๊ฐ’์ด ์—ฐ๊ตฌ ๋ฌธ์ œ์— ๋‹ต์„ ์ฃผ์ง€ ๋ชป ํ•œ๋‹ค๋ฉด
02:25
why does the scientific community use it?
42
145579
2292
๊ณผํ•™๊ณ„๋Š” ์™œ p-๊ฐ’์„ ์‚ฌ์šฉํ•˜๋Š” ๊ฑธ๊นŒ์š”?
02:28
Well, because even though a p-value doesnโ€™t directly state the probability
43
148329
4709
๊ฒฐ๊ณผ๊ฐ€ ๋ฌด์ž‘์œ„ํ•œ ์šฐ์—ฐ์— ์˜ํ•œ ๊ฒƒ์ผ ํ™•๋ฅ ์„
02:33
that the results are due to random chance,
44
153038
2500
p-๊ฐ’์ด ์ง์ ‘์ ์œผ๋กœ ๋‚˜ํƒ€๋‚ด์ง€๋Š” ์•Š๋”๋ผ๋„
02:35
it usually gives a pretty reliable indication.
45
155538
3333
p-๊ฐ’์€ ์ผ๋ฐ˜์ ์œผ๋กœ ๊ฝค ์‹ ๋ขฐํ• ๋งŒํ•œ ์ง€ํ‘œ๊ฐ€ ๋ฉ๋‹ˆ๋‹ค.
02:39
At least, it does when used correctly.
46
159204
2792
์ ์–ด๋„ ์˜ฌ๋ฐ”๋ฅด๊ฒŒ ์‚ฌ์šฉ๋  ๋•Œ์—๋Š” ๋ง์ด์ฃ .
02:41
And thatโ€™s where many researchers, and even whole fields,
47
161996
3917
๊ทธ๋ฆฌ๊ณ  ์ด ์ง€์ ์—์„œ ๋ชจ๋“  ๋ถ„์•ผ์˜ ๋งŽ์€ ์—ฐ๊ตฌ์ž๋“ค์€ ๋ฌธ์ œ์— ๋ด‰์ฐฉํ•ฉ๋‹ˆ๋‹ค.
02:45
have run into trouble.
48
165913
1458
02:47
Most real studies are more complex than the tea experiment.
49
167538
3458
๋Œ€๋ถ€๋ถ„์˜ ์‹ค์ œ ์—ฐ๊ตฌ๋Š” ์ฐจ ์‹คํ—˜๋ณด๋‹ค ํ›จ์”ฌ ๋ณต์žกํ•ฉ๋‹ˆ๋‹ค.
02:51
Scientists can test their research question in multiple ways,
50
171288
3375
๊ณผํ•™์ž๋“ค์€ ์—ฐ๊ตฌ ๋ฌธ์ œ๋ฅผ ์—ฌ๋Ÿฌ ๋ฐฉ๋ฒ•์œผ๋กœ ์‹คํ—˜ํ•ด ๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
02:54
and some of these tests might produce a statistically significant result,
51
174663
4375
์ด๋“ค ์ค‘ ์ผ๋ถ€๋งŒ ํ†ต๊ณ„์ ์œผ๋กœ ์œ ์˜๋ฏธํ•œ ๊ฒฐ๊ณผ๋ฅผ ๋ณด์ด๊ณ 
02:59
while others donโ€™t.
52
179038
1208
๋‚˜๋จธ์ง€๋Š” ์˜๋ฏธ ์—†์„ ์ˆ˜๋„ ์žˆ์ฃ .
03:00
It might seem like a good idea to test every possibility.
53
180454
3167
๋ชจ๋“  ๊ฐ€๋Šฅ์„ฑ์„ ์‹คํ—˜ํ•ด ๋ณด๋Š” ๊ฒƒ์ด ์ข‹์€ ์ƒ๊ฐ์ฒ˜๋Ÿผ ๋ณด์ผ ์ˆ˜๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
03:03
But itโ€™s not, because with each additional test,
54
183913
3083
ํ•˜์ง€๋งŒ ๊ทธ๋ ‡์ง€ ์•Š์Šต๋‹ˆ๋‹ค.
์ถ”๊ฐ€์ ์ธ ์‹คํ—˜์„ ํ•  ๋•Œ๋งˆ๋‹ค ๊ฑฐ์ง“ ์–‘์„ฑ์˜ ๊ฐ€๋Šฅ์„ฑ์ด ์ปค์ง€๊ธฐ ๋•Œ๋ฌธ์ด์ฃ .
03:07
the chance of a false positive increases.
55
187163
3208
03:10
Searching for a low p-value, and then presenting only that analysis,
56
190996
4500
๋‚ฎ์€ p-๊ฐ’์„ ์ฐพ๊ณ  ํ•ด๋‹น ๋ถ„์„๋งŒ์„ ์ œ์‹œํ•˜๋Š” ๊ฒƒ์„
03:15
is often called p-hacking.
57
195496
2750
ํ”ํžˆ p-ํ•ดํ‚น์ด๋ผ๊ณ  ํ•ฉ๋‹ˆ๋‹ค.
03:18
Itโ€™s like throwing darts until you hit a bullseye
58
198246
2750
์ด๊ฒƒ์€ ๋งˆ์น˜ ๊ณผ๋…์˜ ์ •์ค‘์•™์„ ๋งžํž ๋•Œ๊นŒ์ง€ ๋‹คํŠธ๋ฅผ ๋˜์ง€๊ณ ๋Š”
03:20
and then saying you only threw the dart that hit the bullโ€™s eye.
59
200996
3333
๊ทธ ๋‹คํŠธ ํ•˜๋‚˜๋งŒ์„ ๋˜์กŒ๋‹ค๊ณ  ๋งํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค.
03:24
This is exactly what the music researchers did.
60
204746
3208
์ด๊ฒƒ์ด ๋ฐ”๋กœ ์Œ์•… ์‹คํ—˜์˜ ์—ฐ๊ตฌ์ž๋“ค์ด ํ•œ ์ผ์ž…๋‹ˆ๋‹ค.
03:28
They played three groups of participants each a different song
61
208079
3709
๊ทธ๋“ค์€ ์„ธ ๊ทธ๋ฃน์˜ ์ฐธ๊ฐ€์ž๋“ค์—๊ฒŒ ๊ฐ๊ฐ ๋‹ค๋ฅธ ์Œ์•…์„ ๋“ค๋ ค์ฃผ์—ˆ๊ณ 
03:31
and collected lots of information about them.
62
211788
2500
์ฐธ๊ฐ€์ž๋“ค์— ๋Œ€ํ•œ ๋งŽ์€ ์ •๋ณด๋ฅผ ์ˆ˜์ง‘ํ–ˆ์Šต๋‹ˆ๋‹ค.
03:34
The analysis they published included only two out of the three groups.
63
214288
4250
์ด๋“ค์ด ๋ฐœํ‘œํ•œ ๋ถ„์„์—๋Š” ์„ธ ๊ทธ๋ฃน ์ค‘ ๋‘ ๊ทธ๋ฃน๋งŒ์ด ํฌํ•จ๋˜์–ด ์žˆ์—ˆ๊ณ 
03:38
Of all the information they collected,
64
218538
2208
์ˆ˜์ง‘ํ•œ ๋งŽ์€ ์ •๋ณด๋“ค ์ค‘์—์„œ
03:40
their analysis only used participantsโ€™ fathersโ€™ ageโ€”
65
220746
3542
์ฐธ๊ฐ€์ž๋“ค์˜ ์•„๋ฒ„์ง€ ๋‚˜์ด๋งŒ์„ ๋ถ„์„์— ์ ์šฉํ•จ์œผ๋กœ์จ
03:44
to โ€œcontrol for variation in baseline age across participants.โ€
66
224288
4541
์ฐธ๊ฐ€์ž ๊ฐ„ ๊ธฐ์ค€ ์—ฐ๋ น ํŽธ์ฐจ๋ฅผ ์กฐ์ ˆํ–ˆ์ฃ ..
03:49
They also paused their experiment after every ten participants,
67
229246
4208
๋˜ํ•œ 10๋ช…์˜ ์ฐธ๊ฐ€์ž๊ฐ€ ์ถ”๊ฐ€๋  ๋•Œ๋งˆ๋‹ค ์‹คํ—˜์„ ์ž ์‹œ ์ค‘์ง€ํ•œ ํ›„
03:53
and continued if the p-value was above .05,
68
233454
4459
p-๊ฐ’์ด 0.05๋ฅผ ๋„˜์œผ๋ฉด ์‹คํ—˜์„ ๊ณ„์†ํ–ˆ์ง€๋งŒ
03:57
but stopped when it dipped below .05.
69
237913
3291
p-๊ฐ’์ด 0.05 ์•„๋ž˜๋กœ ๋–จ์–ด์ง€๋ฉด ์‹คํ—˜์„ ์ค‘๋‹จํ–ˆ์Šต๋‹ˆ๋‹ค.
04:01
They found that participants who heard one song were 1.5 years younger
70
241746
5208
๊ทธ๋“ค์€ ํŠน์ • ์Œ์•…์„ ๋“ค์€ ์ฐธ๊ฐ€์ž๋“ค์ด
๋‹ค๋ฅธ ์Œ์•…์„ ๋“ค์€ ์ฐธ๊ฐ€์ž๋“ค๋ณด๋‹ค 1.5์‚ด ์–ด๋ฆฌ๋‹ค๋Š” ๊ฒƒ์„ ๋ฐœ๊ฒฌํ–ˆ๋Š”๋ฐ
04:06
than those who heard the other song, with a p-value of .04.
71
246954
4375
์ด๋•Œ์˜ p-๊ฐ’์€ 0.04์˜€์Šต๋‹ˆ๋‹ค.
04:12
Usually itโ€™s much tougher to spot p-hacking,
72
252163
2833
์ผ๋ฐ˜์ ์œผ๋กœ p-ํ•ดํ‚น์„ ์žก์•„๋‚ด๊ธฐ๋Š” ์ƒ๋‹นํžˆ ์–ด๋ ต์Šต๋‹ˆ๋‹ค.
04:14
because we donโ€™t know the results are impossible:
73
254996
2667
์šฐ๋ฆฌ๋Š” ๊ฒฐ๊ณผ๊ฐ€ ๋ถˆ๊ฐ€๋Šฅํ•œ ๊ฒƒ์ธ์ง€ ์•Œ ์ˆ˜ ์—†๊ธฐ ๋•Œ๋ฌธ์ž…๋‹ˆ๋‹ค.
04:17
the whole point of doing experiments is to learn something new.
74
257663
3416
์‹คํ—˜์„ ํ•˜๋Š” ์ด์œ ๋Š” ์ƒˆ๋กœ์šด ๊ฒƒ์„ ์•Œ๊ธฐ ์œ„ํ•ด์„œ์ž…๋‹ˆ๋‹ค.
04:21
Fortunately, thereโ€™s a simple way to make p-values more reliable:
75
261329
4209
๋‹คํ–‰ํžˆ p-๊ฐ’์˜ ์‹ ๋ขฐ๋„๋ฅผ ๋†’์ด๋Š” ๊ฐ„๋‹จํ•œ ๋ฐฉ๋ฒ•์ด ์žˆ์Šต๋‹ˆ๋‹ค.
04:25
pre-registering a detailed plan for the experiment and analysis
76
265913
4708
์‹คํ—˜๊ณผ ๋ถ„์„์— ๋Œ€ํ•œ ์ž์„ธํ•œ ๊ณ„ํš์„ ์‚ฌ์ „์— ๋“ฑ๋กํ•˜๊ฒŒ ํ•˜์—ฌ
04:30
beforehand that others can check,
77
270621
2458
๋‹ค๋ฅธ ์‚ฌ๋žŒ์ด ํ™•์ธํ•  ์ˆ˜ ์žˆ๋„๋ก ํ•˜๋ฉด
04:33
so researchers canโ€™t keep trying different analyses
78
273079
3417
์—ฐ๊ตฌ์ž๋“ค์€ ์œ ์˜๋ฏธํ•œ ๊ฒฐ๊ณผ๋ฅผ ๋ฐœ๊ฒฌํ•  ๋•Œ๊นŒ์ง€
04:36
until they find a significant result.
79
276496
2125
๊ณ„์†ํ•ด์„œ ๋‹ค๋ฅธ ๋ถ„์„์„ ์‹œ๋„ํ•˜๋Š” ํ–‰์œ„๋ฅผ ํ•  ์ˆ˜ ์—†๊ฒŒ ๋ฉ๋‹ˆ๋‹ค.
04:38
And, in the true spirit of scientific inquiry,
80
278788
2458
์ฐธ๋œ ๊ณผํ•™์  ์—ฐ๊ตฌ ์ •์‹ ์—๋Š”
04:41
thereโ€™s even a new field thatโ€™s basically science doing science on itself:
81
281246
5375
๊ทผ๋ณธ์ ์œผ๋กœ ๊ณผํ•™ ์Šค์Šค๋กœ ๊ณผํ•™์„ ํ–‰ํ•˜์—ฌ ๊ณผํ•™์  ๊ด€ํ–‰์„ ์—ฐ๊ตฌํ•˜๊ณ  ๊ฐœ์„ ์‹œํ‚ค๋Š”
04:46
studying scientific practices in order to improve them.
82
286621
3667
์ƒˆ๋กœ์šด ๋ถ„์•ผ๋„ ์žˆ์Šต๋‹ˆ๋‹ค.
์ด ์›น์‚ฌ์ดํŠธ ์ •๋ณด

์ด ์‚ฌ์ดํŠธ๋Š” ์˜์–ด ํ•™์Šต์— ์œ ์šฉํ•œ YouTube ๋™์˜์ƒ์„ ์†Œ๊ฐœํ•ฉ๋‹ˆ๋‹ค. ์ „ ์„ธ๊ณ„ ์ตœ๊ณ ์˜ ์„ ์ƒ๋‹˜๋“ค์ด ๊ฐ€๋ฅด์น˜๋Š” ์˜์–ด ์ˆ˜์—…์„ ๋ณด๊ฒŒ ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๊ฐ ๋™์˜์ƒ ํŽ˜์ด์ง€์— ํ‘œ์‹œ๋˜๋Š” ์˜์–ด ์ž๋ง‰์„ ๋”๋ธ” ํด๋ฆญํ•˜๋ฉด ๊ทธ๊ณณ์—์„œ ๋™์˜์ƒ์ด ์žฌ์ƒ๋ฉ๋‹ˆ๋‹ค. ๋น„๋””์˜ค ์žฌ์ƒ์— ๋งž์ถฐ ์ž๋ง‰์ด ์Šคํฌ๋กค๋ฉ๋‹ˆ๋‹ค. ์˜๊ฒฌ์ด๋‚˜ ์š”์ฒญ์ด ์žˆ๋Š” ๊ฒฝ์šฐ ์ด ๋ฌธ์˜ ์–‘์‹์„ ์‚ฌ์šฉํ•˜์—ฌ ๋ฌธ์˜ํ•˜์‹ญ์‹œ์˜ค.

https://forms.gle/WvT1wiN1qDtmnspy7