Newsy Numbers

EriDay writes "The Wall Street Journal has a new feature called The Numbers Guy about "the way numbers and statistics are used - and abused - in the news, business and politics". The first installment lets us know that somewhere between 0 and 1 Billion (or more) people will be killed by Asian bird flu."

Comments

Statistical Lies...

[Allen Zadr]

First published in 1954: How to Lie With Statistics

Good book, recommended reading, if you like the above article.

Re:Statistical Lies...

[tgrigsby]

Or if you want the Reader's Digest version, there's a quick and easy explanation of how to use critical thinking when you hear statistics here: http://www.wsu.edu:8080/~taflinge/evistats.html

Re:Statistical Lies...

[hackstraw]

I've heard of the book, but have not actually read it personally. I remember reading something in 4th grade or thereabouts that talked about how advertisers used numbers to make their product seem better with things like 3 out of 4 doctors surveyed, etc.

Also, there is a popular quote that goes "Lies, damn lies, and statistics".

Similar to the article, there is a strange number game that was done a while back when the SARS "epidemic" hit the world. A total of about 850 people died from the thing, yet annually 10,000 or so people die from influenza. SARS is an epidemic, influenza not.

However, people have heard of influenza and not SARS, so I guess it makes for better headlines.

Re:Statistical Lies...

[Rude Turnip]

Hell, 70% of people know that!

My personal favorite

[TildeMan]

During one of the 2004 presidential debates: "We increased federal wetlands by 3 million!" -- GWB

I'm still not really sure what that means.

So...

[FireballX301]

The first installment lets us know that somewhere between 0 and 1 Billion (or more)

Excellent, it's nice to know that a negative number of people won't die.

Re:So...

[fred fleenblat]

However, he does not rule out fractional numbers of dead people.

When I'm in the mood to tweak, I'll bring up the idea that deaths should be scaled by life expectancy. An extreme example would be that maybe the death of a 90 year-old guy with cancer should only count about 1% as much as the death of a healthy college kid.

This is at first a bit horrifying, but it changes the perception of health risk a bit. A car accident can strike at any time no matter your health or life-expectancy, but the flu is far harsher on the very young and the very old. Heart attacks quite rare for the under-30 crowd, become very common in the 50's and 60's and start tapering off since people who are susceptible have already had them. Various other ailments have other relationships to life-expectancy, both for susceptibility and for impact.

The logical conclusion I always get to is that we should focus a lot more health resources on the very young, i.e. pre-natal and neo-natal care, free vaccinations, healthy childhood diets and exercise, lifelong sunscreen habits, semi-intentional exposure to a variety of colds and flus in the teens and 20's, and moderation of alcohol and fatty foods after that.

It's all common sense stuff and would pay off 100:1 compared to after-the-event treatments for things like heart attacks and cancer.

Did you know...

[se2schul]

...83% off all statistics quoted are made up on the spot!

Pi

[savagedome]

Also, in his diary, the following excerpt was found:

11:15, restate my assumptions: 1. Mathematics is the language of nature. 2. Everything around us can be represented and understood through numbers. 3. If you graph these numbers, patterns emerge. Therefore: There are patterns everywhere in nature.

aarch... panic !

[selderrr]

...somewhere between 0 and 1 Billion (or more) people...

Is that zero , or zero billion ?

[...head explodes...]

Funny Statistic

[Lord Kano]

1/3 of all auto accidents involve people who test positive for marijuana use.

This means that 2/3 of all auto accidents are cause by people who are not high.

We sober people are KILLING each other while the stoners are not.

LK

Re:Funny Statistic

[Shadow Wrought]

I remember a statistic in College (way to long ago to remember the cite) about the disparity between males and females when it comes to causing car accidents. Males were far more more likely to be responsible as the cause of an accident- unless alcohol or drugs were not involved. Then women were far more likely to be the cause;-)

Re:Funny Statistic

[i41Overlord]

I do hope that less than 1/3 of the population uses marijuana, as it's illegal.

Legality is not morality.

People used to drink also. They drank before prohibition, they drank during prohibition, and they drank after prohibition. The law didn't really change much, other than the fact that the same people went from being law abiding citizens, to criminals who supported the Kaiser (the same old Communist/Terrorist enemy tactic used forever) back to law abiding citizens.

Laws such as this are mainly academic in my opinion.

Re:Funny Statistic

[The Spoonman]

I do hope that less than 1/3 of the population uses marijuana, as it's illegal.

Yes, because everyone who has half a brain and can think critically knows how dangerous marijuana can be, and that the government would never make illegal something that wasn't dangerous. They're fully acquainted with what should be illegal and what shouldn't.

Is it me, or is HTML like the PERFECT language of sarcasm?? :)

Re:Alright! Another thread where we can bash Bush

[lucabrasi999]

This thread provides and excellent opportunity to bash Bush, America, Corporations and Fox News!

Yeah, those folks at the Wall Street Journal are nothing but a bunch of crazy liberals.

I see plenty of cautions

[Anonymous Coward]

The Numbers Guy says, "The 100 million figure was reported widely, including in the New York Times, The Wall Street Journal Online, CNN, Newsweek and the U.K.'s Observer; but without much caution about how arbitrary it is."

So I looked and I couldn't find a single article supporting his claim that it was reported as fact.

Maybe it's The Numbers Guy who abusing facts.

It happens every day

[DavidBrown]

I have read (sorry, cannot cite source) that the claim that 100,000 civilians have been killed in Iraq is based on a statistical survey that says somewhere between 5,000 and 100,000 civilians had been killed.

Similar

[kodelab]

Like Bush winning the 2004 election with 0-60,608,582 votes.

Hello PR Stunt!

[torinth]

This column is just an superficial attempt by the WSJ to combat the "news is junk" meme that's been building over the last few years. They're trying to make it look like: "hell, we've got people who write fricking columns about statistical manipulation!" so that you don't think the rest of their paper prints it.

But odds are that in todays super-competitive least-necessary-change news market the WSJ has done nothing substantial to improve the accuracy of their paper and instead just inserted a column to improve the image.

A Mathematician Reads the Newspaper

[markh1967]

The article reminded me of 'A Mathematician Reads the Newspaper' by John Allen Paulos. A great read for those interested in the mangling of science and numeracy by the media and politicians.

Re:argh! Statistical abuse!

[parkrrrr]

How about a different Mark Twain quote, from Life on the Mississippi:

In the space of one hundred and seventy-six years the Lower Mississippi has shortened itself two hundred and forty-two miles. That is an average of a trifle over one mile and a third per year. Therefore, any calm person, who is not blind or idiotic, can see that in the Old Oolitic Silurian Period,' just a million years ago next November, the Lower Mississippi River was upwards of one million three hundred thousand miles long, and stuck out over the Gulf of Mexico like a fishing-rod. And by the same token any person can see that seven hundred and forty-two years from now the Lower Mississippi will be only a mile and three-quarters long, and Cairo and New Orleans will have joined their streets together, and be plodding comfortably along under a single mayor and a mutual board of aldermen. There is something fascinating about science. One gets such wholesale returns of conjecture out of such a trifling investment of fact.

A joke from the past

[Maimun]

During communism (Bulgaria), we had this joke. An American and a Soviet athlete competed in an official event of importance. The American won. Next day, the newspapers wrote: "The Soviet athlete took the second place, while the American only got the penultimate one". :)

A BBC Radio series worth listening to..

[Angostura]

The BBC has an excellent radio series called More Or Less" that unpicks the numbers and statistics that are bandied about in the news. It is authoritative, interesting and a remarkably good listen (available on demand using Real Audio)

But a billion COULD die ...

[Anonymous Coward]

First rule of government: inconclusive data means no action.

The article is about H5N1, better known as "bird flu." Some important things to know about avian influenza: in the small number of cases we've seen of it, it has a 75% or higher mortality rate (as opposed to 2.5% for the Spanish Flu pandemic of 1918); it is remarkably difficult to create vaccines for it, because it kills the eggs used to create traditional influenza vaccines; the variants we see are amantadine/rimantadine resistant, limiting antiviral treatment options and suggesting significant exchange of genetic material with human influenza viruses; it is pantropic (capable of infecting tissue across the body) in some animals, and both pneumotropic (as all influenza are) and neurotropic in others; and H5N1 is epidemic in Asia amongst many different waterfowl.

So, what we know is that if an H5N1 variant emerges that is human-infectuous and easily transmissible, the chances are very, very high that the resultant pandemic would burn through populations like a wildfire. Furthermore, the chances of this happening are greater than either the appearance of or the damages from various high-profile, high-budget "homeland security" scenarios, such as smallpox (unlikely to occur) or a dirty bomb (more panic than damage).

So, what are the right risk factors? That's hard to say, since it depends on the right mutations being hit. But what we do know is that H5N1 represents at least as dangerous a threat as al-Qaeda.

Here's a classic: about marriages and divorce

[jwd-oh]

The often quoted statistic:
In the US, 1/2 of all marriages end in divorce.

The correct statistic:
In the US, the annual divorce rate is 1/2 the annual wedding rate.

These are extremely different.

Bad statistics jokes

[fbform]

A statistician discovered that the probability of a bomb being on board a given aircraft was alarmingly high. But he realized that the probability of two bombs being on board the same aircraft was reassuringly low.

So these days, whenever he flies, he carries a bomb with him.

---- ____ ----

A university surveyed its graduate students, and found that the male students averaged 1.8 children each, while he female students averaged 1.4 children each. Therefore men have more kids than women.

Re:I hope

[SharpFang]

But of course! Iraq owns the oil! US only rebuilds several $bln worth of destroyed country's infrastructure (which they have destroyed themselves), and they will have the operation of rebuilding the country paid in oil.
A perfectly legal transaction. Like a doctor breaks your leg and then charges you for putting it back together...

SARS

[KMSelf]

Why was SARS so significant?

• Novel disease with (initially) unknown etiology.
• Novel disease agent (SARS virus) of unknown origin.
• Extremely high mortality rate (as noted by other responses). The ultimate mortality rate appears to be around 9%, though localized rates in excess of 20% exist, possibly due to variance in the infectious agent (more below). Moreover, as I was following stats at the time, the mortality rate was grossly underreported for several weeks as the epidemic unfolded. I wrote several nastygrams myself to The Economist which was quoting a much lower (3-8% IIRC) mortality. Mortality varied greatly with age, from Wikepedia: "below 1% for people aged 24 or younger, 6% for those 25 to 44, 15% in those 45 to 64 and more than 50% for those over 65." The article has a wealth of information.
• Extremely high transmission rate. SARS was passed between victims based on very casual contact, including apparently nothing more than sharing a confined room for a brief period of time.
• Poor response to therapy. Once ill, a victim's prognosis was largely independent of treatment. Viruses are difficult to treat in any event, and the few nominally useful antiviral treatments which do exist were largely ineffectual.
• Rapid mutation and/or wide variance among viral strains. Based on my after-the-fact recollection of SARS mortality rates. China had among the highest mortality, rates were far lower elsewhere. This may have been due to differences in treatment or more strains of the virus present in China (where SARS originated) than elsewhere.
• Suppression of initial information. China's government and health authorities initially responded to the SARS outbreak by supressing information. This confounded responses be making unavailable useful information and generating rumors and speculation.
• High morbidity and mortality among healthcare workers. Among the hardest hit communities were the doctors, nurses, and researchers initially responding to SARS. Among the victims were several of those who first identified, treated, and isolated the disease. In a broader outbreak, healthcare workers would likely have suffered significantly. A friend's wife, staff at one of the few US hospitals to encounter SARS (Belvue, NYC) was very concerned.

So you've got a new, disease with unknown agent, few treatments, high mortality, and a large impact on healthcare infrastructure. Not a good sign.

The extent to which cases and deaths due to SARS were minimized is not an indication that the disease was overblown, but that the response to it was highly effective. Remember that there was a massive quarantine effort made. Again from Wikipedia:

Attempts were made to control further SARS infection through the use of quarantine. Over 1200 were under quarantine in Hong Kong, while in Singapore and Taiwan, 977 and 1147 were quarantined respectively. Canada also put thousands of people under quarantine. [ 12] In Singapore, schools were closed for 10 days and in Hong Kong they are closed until April 21 to contain the spread of SARS.

SARS was a very close call, and a big wakeup alert.