Eating chocolate reduces stress in students.

一種簡單的維他命就可以 減少你得到心臟病的風險。

New drug prolongs lives of patients with rare disease.

學生吃巧克力可以減少壓力。

Health headlines like these are published every day,

新藥品延長罕見疾病病人的壽命。

sometimes making opposite claims from each other.

每天都可以看到 像這類的健康標題被刊出，

There can be a disconnect between broad,

有時，這些標題的主張還彼此相反。

attention-grabbing headlines and the often specific,

這些廣泛、吸引注意力的標題

incremental results of the medical research they cover.

和它們所報導的那些通常很明確、

So how can you avoid being misled by grabby headlines?

漸進式的醫療研究結果 之間會有斷層。

The best way to assess a headline's credibility

所以，我們要如何避免 被聳動的標題誤導？

is to look at the original research it reports on.

評估標題可信性的最佳辦法

We've come up with a hypothetical research scenario

就是去看它報導的原始研究。

for each of these three headlines.

我們針對前述這三則標題分別

Keep watching for the explanation of the first example;

提出了一個假設性的研究情境。

then pause at the headline to answer the question.

繼續看第一個例子的說明；

These are simplified scenarios.

接著在標題出現後 按暫停，回答問題。

A real study would detail many more factors and how it accounted for them,

這些都是簡化的情境。

but for the purposes of this exercise,

真實的研究會詳細說明更多因子 以及它們與研究的關係，

assume all the information you need is included.

但為了做這項練習，

Let's start by considering the cardiovascular effects

就假設所需的資訊已全被納入。

of a certain vitamin, Healthium.

咱們先來思考一下 一種維他命 Healthium

The study finds that participants taking Healthium

對心血管的影響。

had a higher level of healthy cholesterol than those taking a placebo.

該研究發現吃 Healthium 的受試者的

Their levels became similar to those of people with naturally high levels

高密度脂蛋白膽固醇（HDL 好膽固醇）

of this kind of cholesterol.

比吃安慰劑的要來的高。

Previous research has shown that people with naturally high levels

前者的膽固醇值很接近那些

of healthy cholesterol have lower rates of heart disease.

天生就具有高水平好膽固醇的人。

So what makes this headline misleading:

先前的研究顯示

"Healthium reduces risk of heart disease."

天生就具有高水平好膽固醇的人

The problem with this headline is that the research didn't actually investigate

比較不會罹患心臟疾病。

whether Healthium reduces heart disease.

那麼，這標題是怎麼誤導的：

It only measured Healthium's impact

「Healthium 降低心臟疾病的風險」？

on levels of a particular kind of cholesterol.

這個標題的問題是， 研究並沒有真正去調查

The fact that people with naturally high levels of that cholesterol

Healthium 是否會減少心臟疾病。

have lower risk of heart attacks

它只是測量了 Healthium

doesn't mean that the same will be true of people

對於特定種類膽固醇水平的影響。

who elevate their cholesterol levels using Healthium.

那種膽固醇天生就很高的人

Now that you've cracked the case of Healthium,

比較少有心臟病，

try your hand at a particularly alluring mystery:

這並不一定表示

the relationship between eating chocolate and stress.

用 Healthium 來提高 膽固醇水平的人亦會如此。

This hypothetical study recruits ten students.

現在，你破解了 Healthium 的案例，

Half begin consuming a daily dose of chocolate,

試試看這個特別誘人的謎：

while half abstain.

吃巧克力和壓力之間的關係。

As classmates, they all follow the same schedule.

這項假設性研究招募了十名學生。

By the end of the study, the chocolate eaters are less stressed

其中一半學生開始 每天吃一定劑量的巧克力，

than their chocolate-free counterparts.

另一半則避吃巧克力。

What's wrong with this headline:

他們都是同學，所以課表一樣。

"Eating chocolate reduces stress in students"

在研究尾聲，吃巧克力的人

It's a stretch to draw a conclusion about students in general from a sample of ten.

比不吃巧克力的人壓力小。

That's because the fewer participants are in a random sample,

這個標題有什麼問題：

the less likely it is that the sample will closely represent

「吃巧克力會減少學生壓力」

the target population as a whole.

樣本只有十個人，卻把結果套用到 所有學生，這就是過度擴大。

For example, if the broader population of students is half male and half female,

原因是，在隨機樣本中的 受試者人數越少，

the chance of drawing a sample of 10

該樣本就越無法代表

that's skewed 70% male and 30% is about 12%.

整個目標整體。

In a sample of 100 that would be less than a .0025% chance,

比如，如果廣大的學生母體中 有一半是男性，一半是女性，

and for a sample of 1000,

抽出十個人做為樣本，

the odds are less than 6 x 10^-36.

此樣本有 12% 的可能性 會偏向七成男性、三成女性。

Similarly, with fewer participants,

若是一百人的樣本，

each individual's outcome has a larger impact on the overall results—

此機率則不到 0.0025%，

and can therefore skew big-picture trends.

若樣本中有一千人，

Still, there are a lot of good reasons for scientists to run small studies.

此機率不到 6 x 10^-36。

By starting with a small sample,

同樣的，當受試的人數很少時，

they can evaluate whether the results are promising enough

個別受試者的結果 對整體結果的影響會比較大，

to run a more comprehensive, expensive study.

因此可以造成整體趨勢的偏差。

And some research requires very specific participants

不過，科學家還是有很多 好理由去進行小型研究。

that may be impossible to recruit in large numbers.

從小樣本開始，

The key is reproducibility—

他們可以評估結果是否夠理想，

if an article draws a conclusion from one small study,

再來進行更全面、廣泛的研究。

that conclusion may be suspect—

有些研究會需要非常特定的受試者，

but if it's based on many studies that have found similar results,

可能無法讓很多人參與。

it's more credible.

關鍵在於可重複性：

We've still got one more puzzle.

如果文章從一項小型研究得出結論，

In this scenario, a study tests a new drug for a rare, fatal disease.

那結論可能不可信；

In a sample of 2,000 patients,

但如果文章的根據是 很多項發現類似結果的研究，

the ones who start taking the drug upon diagnosis

就會比較可信。

live longer than those who take the placebo.

我們還有一個謎要解。

This time, the question is slightly different.

這個情境中的研究，是在測試 治療致命罕見疾病的新藥。

What's one more thing you'd like to know before deciding if the headline,

樣本是兩千名病人，

"New drug prolongs lives of patients with rare disease", is justified?

被診斷出此疾病後 就開始吃這種藥的人，

Before making this call,

比吃安慰劑的人活得更久。

you'd want to know how much the drug prolonged the patients' lives.

這次，問題有點不同。

Sometimes, a study can have results that,

你還需要知道哪一項資訊， 才能判斷這個標題是否合理：

while scientifically valid, don't have much bearing on real world outcomes.

「新藥延長了罕見疾病患者的生命」？

For example, one real-life clinical trial of a pancreatic cancer drug

在做決定之前，

found an increase in life expectancy— of ten days.

你需要先知道，這種藥 能延伸病人多長的壽命。

The next time you see a surprising medical headline,

有時研究的結果可能在科學上有效，

take a look at the science it's reporting on.

但對現實世界的結果影響不大。

Even when full papers aren't available without a fee,

比如，胰藏癌藥物的實際臨床試驗

you can often find summaries of experimental design

發現可以延長壽命「十天」。

and results in freely available abstracts,

下次看到驚人的醫療相關標題時，

or even within the text of a news article.

先看一下它所報導的科學。

It's exciting to see scientific research covered in the news,

即使需要付費才能 看到完整研究論文，

and important to understand the studies' findings.

你通常仍然能在免費取得的摘要中