Behavioral biases in charitable giving, installment #1637

by on September 25, 2013 at 6:28 am in Medicine, Philosophy, Science | Permalink

People pay more attention to the number of people killed in a natural disaster than to the number of survivors when deciding how much money to donate to disaster relief efforts, according to new research published in Psychological Science, a journal of the Association for Psychological Science.

…Their model estimated that about $9,300 was donated per person killed in a given disaster. The number of people affected in the disasters, on the other hand, appeared to have no influence on the amount donated to relief efforts.

The summary article is here, and the gated published version is here.  I do not see an ungated copy.  Here is a related paper (pdf) on how disasters drive aid decisions.

For the pointer I thank Bill Benzon.

Jonathan September 25, 2013 at 6:57 am

Actually, from the summary article, the behavior seems perfectly rational. People reject giving money to the number of people “affected” to the number of people killed because “affected” is a weaselly media word to make things look big. Death is a more reliable indicator than “affected.”
When the researchers swapped the term “affected” with the much less ambiguous term “homeless,” participants believed that money should be allocated according to the number of homeless people following a disaster.”
So need trumps death, as you’d hope..

8 September 25, 2013 at 7:54 am

I forget which hunger campaign it was, but there was one in the U.S. a few years back that said something like 1 in 5 children go hungry in America. It was absurd on its face based on the obesity rates, but I looked for their source for this number and found that hunger was defined as missing 1 meal in 1 month. In other words, if a child was punished with “no dinner tonight, straight to bed” or a child was merely petulant and refused to eat a meal, that qualified as hunger.

John Thacker September 25, 2013 at 9:03 am

Well, most statistics like that tend to be exaggerated by playing fast and loose with definitions. But I think that the general point, that some disasters can be worse or need more relief even if not as many people died, is quite valid.

Rahul September 25, 2013 at 12:31 pm

I think definitions are a key factor here. Death is traditionally more immune to fast and loose definitions.

Perhaps, that is why people pay more attention to the number of people killed.

Cliff September 25, 2013 at 10:55 am

I thought they usually used “food insecurity” which means “in danger of missing a meal”, not actually even missing one!

sort_of_knowledgable September 25, 2013 at 11:44 am

I expect that the involuntary modifier was part of the definition. Also 30 percent of the population being obese doesn’t preclude another 20% going hungry. Also one can be consuming large amount calories and still be food insecure. Michael Oher presumably missed meals during his periods of homelessness but was able to get enough to become NFL lineman.

Axa September 25, 2013 at 7:26 am

The article says that the number of survivors and the relief effort needed is NOT directly proportional to the number of deaths. “Attention should be diverted from the number of fatalities to the number of survivors in need.” Hope journalism catches up with this idea soon.

Andrew' September 25, 2013 at 8:00 am

Maybe why we should focus on the spending rather than the marketing.

prior_approval September 25, 2013 at 8:55 am

And here just might be a good test case –

‘A powerful earthquake killed at least 270 people in Pakistan’s remote south-west province of Balochistan.

The 7.7-magnitude quake struck on Tuesday afternoon at a depth of 20km (13 miles) north-east of Awaran, the US Geological Survey said.

Many houses were flattened and thousands of people have spent the night in the open.

After the quake, a small island appeared off the coast near the port of Gwadar.

People gathered on the beach to see the new island, which is reported to be about 200m (656ft) long, 100m wide and 20m high.’

John Thacker September 25, 2013 at 9:04 am

The appearance of a new island is fascinating, but I’m not sure exactly why that’s a particularly good test case. Are you sure you didn’t just want to share that interesting story with us?

Cliff September 25, 2013 at 10:58 am

He hid the lede- 300,000 people affected, 90% of houses destroyed

JJ September 25, 2013 at 11:06 am

The simple explanation: that number of deaths is a proxy for how bad a disaster was and therefore how badly in need survivors are, doesn’t seem so bad a judgment to me, given the alternatives. The number of survivors doesn’t tell you how much money they need at all, but the number of deaths probably contains some information in that direction.

Marie September 25, 2013 at 11:44 am

Yup, I stubbed my toe the other day and there were, what, over 6 billion survivors of that incident?
It’s not perfect, but it gives you an idea of how serious an event is, death toll. We do it with wars, too, even though injuries might be a better gauge in some cases of how much suffering has been caused in any one attack.

mkt September 25, 2013 at 2:43 pm

Excellent. And the problem of measuring how badly the survivors are affected is immense. The survivors of the earthquake in Haiti were in many cases living fairly miserably even before the quake. Are we supposed to apportion aid based on how poor their living standards are now, or based on how much their living standards were reduced?

An earthquake in the northeast US which caused say a week-long power blackout to tens of millions of people would be somewhat of a big thing (not as big as something which destroyed houses and injured people, but not trivial either). But there are many places in the world which don’t have electricity period.

Rahul September 25, 2013 at 12:55 pm

I collected some quick data and the Affected-to-killed ratio varies a lot 18 to 400. Of course, a lot depends on definitions and whose numbers you trust.

1991 Bangladesh cyclone
138,000 killed / 10 million homeless.

2010 Haiti earthquake
159,000 killed / 3 million affected

2004 Indian Ocean earthquake and tsunami
250,000 / 10 million affected

2008 Sichuan earthquake
90,000 / 4.8 million people homeles

Hurricane Katrina
1,833 / 700,000 homeless

FreeTheFiles September 25, 2013 at 1:56 pm

Here’s the paper.

Brendan Perrine September 25, 2013 at 4:55 pm

Also sometimes in Southern California they talk about number of homes destroyed but never campgrounds. % houses and bussnesses destroyed can be a good measure in an earthquake same with water and power supply being improtant to repair infracstuructre.

Go Kings, Go! September 25, 2013 at 9:17 pm

If you’re taking requests, can we have a post on “Behavioral biases in Blogging Tags”, with maybe the starting point on why this post was tagged: “Medicine, Philosophy, Science”?

Comments on this entry are closed.

Previous post:

Next post: