字幕列表 影片播放
"New drug may cure cancer."
譯者: Ann Chen 審譯者: Carol Wang
"Aspirin may reduce risk of heart attacks."
[你能點出這些標題 要如何改進嗎?]
"Eating breakfast can help you lose weight."
「新藥可能可以治癌症」
Health headlines like these flood the news,
「阿司匹林可能會 降低心臟病發作的風險」
often contradicting each other.
「吃早餐可以幫助減重」
So how can you figure out what's a genuine health concern
像這樣充斥新聞的醫藥標題,
or a truly promising remedy,
往往相互矛盾。
and what's less conclusive?
那麼,你如何辨別 哪些是真正的健康議題
In medicine,
或真正有希望的療法,
there's often a disconnect between news headlines
以及哪些是較沒有確證的呢?
and the scientific research they cover.
醫學上,
That's because a headline is designed to catch attention—
新聞標題與其報導的 科學研究之間常不一致。
it's most effective when it makes a big claim.
這是因為標題旨在吸引注意──
By contrast,
當新聞做出重大宣佈時是最有效的。
many scientific studies produce meaningful results
相對地,
when they focus on a narrow, specific question.
許多科學研究專注於 限定的特定問題時,
The best way to bridge this gap
才會產生有意義的結果。
is to look at the original research behind a headline.
彌補這一差距的最佳方法
We've come up with a simplified research scenario
是去查看標題背後的原始研究。
for each of these three headlines to test your skills.
針對下面三個標題, 我們分別給出了簡單的研究情境,
Keep watching for the explanation of the first study;
來測試你的技巧。
then pause at the headline to figure out the flaw.
接下來看第一項研究的解說,
Assume all the information you need to spot the flaw is included.
然後在標題處暫停,找出問題所在。
Let's start with this hypothetical scenario:
假定你點出問題所需要的信息 都已涵蓋在內了。
a study using mice to test a new cancer drug.
我們先從這個假設場景開始:
The study includes two groups of mice,
一項使用老鼠 測試新抗癌藥物的研究。
one treated with the drug, the other with a placebo.
該研究包括兩組老鼠,
At the end of the trial,
一組用藥物治療,另一組用安慰劑。
the mice that receive the drug are cured,
在研究結束時,
while those that received the placebo are not.
服用藥物的老鼠痊癒了,
Can you spot the problem with this headline:
而服用安慰劑的老鼠則沒有。
"Study shows new drug could cure cancer"
你能點出
Since the subjects of the study were mice,
「研究顯示新藥可能會治癒癌症」
we can't draw conclusions about human disease based on this research.
這個標題的問題所在嗎?
In real life, early research on new drugs and therapies is not conducted on humans.
因為研究對象是老鼠,
If the early results are promising,
我們無法根據這項研究 來對人類疾病下定論。
clinical trials follow to determine if they hold up in humans.
實際上,新藥及治療的早期研究 並不在人體上執行。
Now that you've warmed up,
如果早期結果是可行的,
let's try a trickier example:
隨後的臨床實驗 會確認在人體是否也可行。
a study about the impact of aspirin on heart attack risk.
現在你已小試牛刀,
The study randomly divides a pool of men into two groups.
我們來試一個更棘手的例子:
The members of one group take aspirin daily,
「阿司匹林對心臟病發作 風險影響的研究」。
while the others take a daily placebo.
研究將一群男性隨機分為兩組,
By the end of the trial,
一組成員每天服用阿司匹林,
the control group suffered significantly more heart attacks
而另一組每天服用安慰劑。
than the group that took aspirin.
實驗結束時,
Based on this situation, what's wrong with the headline:
對照組的心臟病發作
"Aspirin may reduce risk of heart attacks"
明顯多於服用阿司匹林組。
In this case, the study shows evidence that aspirin reduces heart attacks in men,
根據這情況,
because all the participants were men.
「阿司匹林可能降低心臟病發作 風險」的標題有什麼錯誤呢?
But the conclusion “aspirin reduces risk of heart attacks” is too broad;
此例中,研究證據顯示 阿司匹林減少男性心臟病發作,
we can't assume that results found in men would also apply to women.
因為所有參與者都是男性。
Studies often limit participants based on geographic location, age, gender,
但「阿司匹林降低心臟病發作風險」 這樣的結論涵蓋太廣;
or many other factors.
我們不能假定男性的結果 也適用於女性。
Before these findings can be generalized,
研究常根據地理位置、年齡、性別
similar studies need to be run on other groups.
或許多其他因素來限制參與者。
If a headline makes a general claim,
在這些結果能被普遍適用前,
it should draw its evidence from a diverse body of research, not one study.
需要對其他群體進行類似的研究。
Can you take your skills from the first two questions to the next level?
如果標題作了普遍適用的聲明,
Try this example about the impact of eating breakfast on weight loss.
它應該從多種研究群體獲得證據, 而不是單一研究。
Researchers recruit a group of people who had always skipped breakfast
你能把前兩題的技巧再提升一級嗎?
and ask them to start eating breakfast everyday.
試試這個關於 吃早餐對減重影響的例子。
The participants include men and women of a range of ages and backgrounds.
研究人員招募了一群不吃早餐的人,
Over a year-long period,
要求他們開始每天吃早餐。
participants lose an average of five pounds.
參與者包括不同年齡和背景的男女。
So what's wrong with the headline:
經過一年,
"Eating breakfast can help you lose weight"
參與者平均減了 5 磅。
The people in the study started eating breakfast and lost weight—
「吃早餐有助於減重」的標題
but we don't know that they lost weight because they started eating breakfast;
有什麼問題呢?
perhaps having their weight tracked
實驗對象開始吃早餐後減重──
inspired them to change their eating habits in other ways.
但我們不知道他們減重 是否是開始吃早餐的緣故;
To rule out the possibility that some other factor caused weight loss,
也許是體重被追蹤,
we would need to compare these participants
促使他們改變了飲食習慣。
to a group who didn't eat breakfast before the study
為了排除某些其他因素 導致減重的可能性,
and continued to skip it during the study.
我們需要將這些參與者
A headline certainly shouldn't claim the results of this research
與另一組在研究前及研究中
are generally applicable.
都不吃早餐的人進行比較。
And if the study itself made such a claim without a comparison group,
標題真的不應該聲稱這研究成果
then you should question its credibility.
是普遍適用的。
Now that you've battle-tested your skills
如果研究本身 沒有對照組就做出斷言,
on these hypothetical studies and headlines,
那麼你應該懷疑其可信度。
you can test them on real-world news.
在虛擬研究和標題中實戰後,
Even when full papers aren't available without a fee,
你不妨在真實新聞中試試身手。
you can often find summaries of experimental design and results
雖然完整的研究報告是付費的,
in freely available abstracts,
實驗設計和結果的概要
or even within the text of a news article.
還是可以在免費的摘要、
Individual studies have results
甚至新聞報導中找到。
that don't necessarily correspond to a grabby headline.
個別研究的結果
Big conclusions for human health issues
不一定非要寫出聳動的標題。
require lots of evidence accumulated over time.
人類健康議題的重大結論,
But in the meantime,
需要時間來累積大量證據。
we can keep on top of the science, by reading past the headlines.
但與此同時,