Wednesday, August 15, 2007

Scary Numbers


One of my biggest peeves is misleading statistics and factoids. The other day on the Achenblog, Joel Achenbach quoted eminent contrarian scientist David Pimentel. Dr. Pimintel is famous for claiming ethanol costs more energy to make than it produces, and he may have a point. In this case he was talking about how much death and misery is attributable to pollution and throws out this nugget:

Of the world population of about 6.5 billion, 57 percent is malnourished, compared with 20 percent of a world population of 2.5 billion in 1950.

I did the math and 57% of the world is about 3.7 billion people, which is about the combined populations of China, India, Africa, Indonesia, Brazil, Pakistan, and Bangladesh. I called “bull-hockey” and decided to find out if this was accurate. I may not know much about world hunger, but I have mad Googlin’ skilz and ran across some stats that were below his number by a factor of four.

From CARE (an organization that raises money to fight hunger, so they ought to know):

More than 840 million people in the world are malnourished — 799 million of them live in the developing world.

From Bread for the World, another hunger group:

854 million people across the world are hungry, up from 852 million a year ago.

(Never mind that 75 million people were added to the world in that same time frame.)

From the Food and Agriculture Organization (FAO) of the United Nations (which is where most of these other groups get their stats):
In 2001-03, FAO estimates there were still 854 million undernourished people world wide: 820 million in the developing countries, 25 million in the transition countries and 9 million in the industrialized countries.

Wikipedia had this very nice map sourced from the United Nations World Food Programme:
(click to enlarge)

According to an accompanying table, the hunger rate in China, the world’s most populous country, was 12%. The most desparate countries were concentrated in war-torn parts of Africa. I asked Joel to follow up on this magic multi-billion number and he got this reply from Pimintel:

The 850 million people who are malnourished refers only to the people who are protein/calorie malnourished and ignores the people who are iron, iodine, and several vitamins malnourished. WHO reports there are 2 billion who are iron malnourished and the number of deaths from iron malnourishment equals the number of deaths from protein/calorie malnourishment. WHO references for 3.7 billion malnourished (57%) are:
WHO 2004 World Health Report.
WHO 2004 World Health Report.

Heck, by that definition I may be malnourished. I haven’t taken my One-A-Days in years. I searched all through his links which had some very distrubing mortality rate tables, but I just couldn’t find that 3.7 billion number.

Giving him the benefit of the doubt that he didn’t just pull that number out of his posterior, the second part of his alarmist statement is an even more blatant falsehood. He claims world hunger has gotten worse since 1950. He is either comparing apples with oranges or his numbers come from an alternate universe. Remember, in 1950, your parents were being told to eat their vegetables because there were starving children in post-war Europe. In the 60s, between 20 and 43 million people died of famine in China’s Great Leap Forward. According to this FAO table, the rate of malnutrition in the developing world has been cut in half since 1970.



In China today, the biggest problem isn’t starvation, it’s pollution from the hyper-rapid economic expansion. Vietnam, which also had famine in recent memory, now exports rice to Africa. Most food security issues are distribution problems, not production issues. The highest rates of hunger are in war-torn regions where food is used as a weapon.

Hunger is a serious problem and having nearly a billion people without enough to eat is a epic tragedy, but we don’t need to go spreading half-truths to make it sound even worse. One of the most insidious aspects of the internet is that once bad information is out there, it stays and spreads. I tried to find malnutrition information for the 1950s but I kept coming across Pimintel's unsourced quote. The real problem with this sort of fear-mongering is that it makes your other numbers suspect. You need to guard your credibility for when the real wolf arrives.

BlatantCommentWhoring™: What sort of absurd fear-inducing numbers have you seen?

Update (8/16/07):

David Pimintel was kind enough to reply to me by e-mail and this is what he had to say:

The difficulty with understanding malnourishment figures is that few studies present the data on how they were measured. For the 1950 data on malnourishment, no information was presented on how malnourishment was measured.
Concerning the current situation getting better, I note that the previous figure of 800 million for protein/calories malnourishment has increased to 850 million.
There are many reasons for the increase in malnourishment. First, per capita cereal grain production has declined and grains make up 80% of the world food. FAO reports that per capita grain production has been CONTINUOUSLY declining for the past 22 years. Supporting this decline in per capita food are the following facts for the past DECADE:
1) Cropland declined 20%.
2) Irrigation declined 10%.
3) Fertilizer use declined 17%.
I hope that these data are of help to you.

I’m still skeptical. All previous Malthusian predictions have proven wrong. The area with the largest increases in food instability have been in sub-Saharan Africa where genocidal campaigns have disrupted the already tenuous food production system.

Please check the comments for some other great remarks.

13 comments:

Anonymous said...

Pimentel's bigger point was that pollution is causing many deaths - that it affects malnourished people more. Whether it's 1 billion or 3.7 billion, it's still a lot of people, and probably depends on the definition of malnourished. I imagine the numbers are estimates, anyway.

The Republican party regularly comes up with fear-inducing numbers - number of small businesses that will fail if the minimum wage is raised, number of illegal immigrants, etc. I take them all with a grain of salt (just one).

mostlylurking

Mooselet said...

The problem when you start throwing numbers out there is that most people, when they figure out that the numbers have been manipulated, assume that the actual figures are manipulated as well and therefore underestimate the problem and then cease to care very much. Which is pretty much what you said.

I can't recall any fear-inducing numbers off the top of my head, but I do remember something I learned in statistics class - numbers can always be manipulated, so never believe the statement "numbers don't lie".

yellojkt said...

mostlylurking,
As you should, but he implies the problem is getting worse when it isn't.

Anonymous said...

I can't think of any specific fear-mongering numbers, but virtually any statistics that come up in the news are either thoroughly massaged to produce the desired outcome or, as you put, pulled from the posterior (10% is a favorite for this group).

But ethanol is a real boondoggle. Not only are there questions about energy output compared to input, but it simply can't meet demand. If all the arable land in the US was given over to ethanol it would meet something like 1/5 of US oil consumption. Not to mention what mere speculation has done to food prices; Mexico is a good example.

Anonymous said...

It's not so much the numbers for me as it is the way they're presented. Clearly there is some counting on the fact that people don't really think about what they're being told.

I was horrified when No Child Left Behind was first being discussed and it appeared that nobody was challenging the idea that every student in America could achieve above-average test scores.

I have a hard time with things being "a thousand times thinner/smaller/less" than other things, rather than 1/1000th the size.

At a community meeting a couple of months ago, someone told the group that the impact of moving a business' driveway was "between zero and none". I asked him where that falls on the number line, because by definition you've already had an impact. His reply: "That's what I mean." He was counting on faking me out by agreeing with me.

This sort of thing drives me crazy.

If you haven't already, check out a book called Innumeracy. It's a fascinating look at this sort of manipulation.

Anonymous said...

The most appalling statistic that crops up regularly is one you briefly mentioned yesterday: child kidnappings/missing children. It's horrible because it feeds on every parent's greatest irrational fear: that someone will snatch your child away. In fact, of course, it's very rare that a stranger kidnaps a child. What is much more common is that a non-custodial relative who is not on good terms with the custodial parent will fail to bring the child back after visitation. And that's something that many people don't need to worry about at all.

The bad use of statistics I hate most is when two numbers are compared, found to be correlated, and the conclusion is immediately that one must cause the other.

The only example I can think of right now is, I've seen it reported that people who attend church regularly tend to live longer than people who don't. That makes sense to me, but I'm sure going to church doesn't extend your life expectancy. It's just that people who go to church every week are the same kind of people who eat moderate, balanced meals, manage the stress in their lives, and have a stable network of family, friends and community.

The other problem with statitistics is that by putting a number on something, you create the illusion that it's a hard-edged fact. If you want to increase your credibility, just add some decimal places. If I say 10% of the people in prison have AIDS, that just doesn't sound as true as if I say 12.3 percent. but essentially, given the fuzziness of the information, it's the same statistic. And either number could be completely off; just having the number doesn't mean you are right (I just made those numbers up, for instance).

"67.3% of all statistics are made up on the spot."

But back to your original rant. Who cares how many people are starving; that point is, we are wallowing in wealth while billions of people suffer. We need to try harder to make the whole world a better place, not just our own backyard.

Anonymous said...

"Who cares how many people are hungry" isn't exactly what I meant to say. Try this: I don't care about the numbers. I care about the people.

yellojkt said...

d-x,
Burning perfectly edible food or converting farmland to biofuel growing seems even more ridiculous than sucking up all the dead dinosaurs that make such good plastics and other durable goods. Clearly we need something new and renewable.

kb,
Missing children scare numbers and Super Bowl Sunday spousal abuse stats are some of the phony numbers that rankle me. We should try to make the world a better place. We just shouldn't lie or mislead to make our point. I don't respect the leaders that have done that recently.

Claude,
Good luck with hitting those Lake Wobegon goals. Innumeracy is a book I have browsed but need to give a full read.

Anonymous said...

The California legislature actually passed a law earlier this year essentially requiring schools to get all of their students to above average test scores. *palm to face*

I am reminded, of course, of the line attributed to Disraeli, Twain, and others: There are three kinds of untruth:lies, damn lies, and statistics.

Jamy said...

It's silly to blame the statistics--blame the people who lie with them.

As far as getting all kids above average test scores, it's absurd. Now, something we could strive for is getting all kids under the median test score, closer to the MEDIAN. Then again, someone will have to explain what the median is...

As a trained demographer, I agree with yellowjkt's assessment: the Malthusians have always been wrong. And this guy is wrong too--he's just making stuff up! When all the reputable sources are this close on the number, it's probably as close as you can get to true.

And the difference between 1 billion and 3 billion is HUGE. It would require quite a different approach to the problem if nearly half the world's population were malnourished.

Also, people don't usually die of starvation--they die of the diseases they become more susceptible to as a result of being malnourished. Not to diminish the problem, but the mechanism is different, which means different interventions are required. Basically: more public health (clean water, clean food, clean air) PLUS more food is needed.

Elizabeth said...

The people who say that autism has increased 500% in the last 20 years is shameful. I understand what they are tyring to say, but when they throw these imposible seeming numbers at me my first reaction is "What are they trying to pull" instead of listening to their message.

Jeff and Charli Lee said...

I think people can be misled by such doomsday statistics as the alarming increase in certain cancer rates, when what we're sometimes really hearing about is the increase in earlier cancer detection.

With new technologies helping to skew the data every year, the benchmarks used to measure the old data don't always remain valid - thereby creating apples and oranges all over again.

Leave it to you to take on a subject like world hunger statistics. I wrote about Alanis Morrisette's 10 year old song the other day.

2fs said...

Karen: The logic you suggest for church-goers supposedly outliving non-church-goers sounds, uh, logical...until you compare things like the divorce and illegitimacy rates in "red" states (which tend to have higher numbers of church-goers) to those in "blue" states: there's more of that there divorcin' and cheatin' and baby-havin' happening redly than bluely.

My favorite bogus statistic: the average man thinks about sex every seven seconds. How the hell do you measure that? Once you ask the question - duh, the guy's thinking about sex. You pretty much can't monitor thinking about sex without thinking about sex.

Unless they hook up a little dingus-monitor or something. But then, you know, a lot of guys would be, huh, I'll pump that thing more than any other guy and show them...