The Affect Heuristic

The af­fect heuris­tic is when sub­jec­tive im­pres­sions of good­ness/​bad­ness act as a heuris­tic—a source of fast, per­cep­tual judg­ments. Pleas­ant and un­pleas­ant feel­ings are cen­tral to hu­man rea­son­ing, and the af­fect heuris­tic comes with lovely bi­ases—some of my fa­vorites.

Let’s start with one of the rel­a­tively less crazy bi­ases. You’re about to move to a new city, and you have to ship an an­tique grand­father clock. In the first case, the grand­father clock was a gift from your grand­par­ents on your fifth birth­day. In the sec­ond case, the clock was a gift from a re­mote rel­a­tive and you have no spe­cial feel­ings for it. How much would you pay for an in­surance policy that paid out $100 if the clock were lost in ship­ping? Ac­cord­ing to Hsee and Kun­reuther, sub­jects stated will­ing­ness to pay more than twice as much in the first con­di­tion.1 This may sound ra­tio­nal—why not pay more to pro­tect the more valuable ob­ject?—un­til you re­al­ize that the in­surance doesn’t pro­tect the clock, it just pays if the clock is lost, and pays ex­actly the same amount for ei­ther clock. (And yes, it was stated that the in­surance was with an out­side com­pany, so it gives no spe­cial mo­tive to the movers.)

All right, but that doesn’t sound too in­sane. Maybe you could get away with claiming the sub­jects were in­sur­ing af­fec­tive out­comes, not fi­nan­cial out­comes—pur­chase of con­so­la­tion.

Then how about this? Ya­m­ag­ishi showed that sub­jects judged a dis­ease as more dan­ger­ous when it was de­scribed as kil­ling 1,286 peo­ple out of ev­ery 10,000, ver­sus a dis­ease that was 24.14% likely to be fatal.2 Ap­par­ently the men­tal image of a thou­sand dead bod­ies is much more alarm­ing, com­pared to a sin­gle per­son who’s more likely to sur­vive than not.

But wait, it gets worse.

Sup­pose an air­port must de­cide whether to spend money to pur­chase some new equip­ment, while crit­ics ar­gue that the money should be spent on other as­pects of air­port safety. Slovic et al. pre­sented two groups of sub­jects with the ar­gu­ments for and against pur­chas­ing the equip­ment, with a re­sponse scale rang­ing from 0 (would not sup­port at all) to 20 (very strong sup­port).3 One group saw the mea­sure de­scribed as sav­ing 150 lives. The other group saw the mea­sure de­scribed as sav­ing 98% of 150 lives. The hy­poth­e­sis mo­ti­vat­ing the ex­per­i­ment was that sav­ing 150 lives sounds vaguely good—is that a lot? a lit­tle?—while sav­ing 98% of some­thing is clearly very good be­cause 98% is so close to the up­per bound of the per­centage scale. Lo and be­hold, sav­ing 150 lives had mean sup­port of 10.4, while sav­ing 98% of 150 lives had mean sup­port of 13.6.

Or con­sider the re­port of Denes-Raj and Ep­stein: sub­jects who were offered an op­por­tu­nity to win $1 each time they ran­domly drew a red jelly bean from a bowl of­ten preferred to draw from a bowl with more red beans and a smaller pro­por­tion of red beans.4 E.g., 7 in 100 was preferred to 1 in 10.

Ac­cord­ing to Denes-Raj and Ep­stein, these sub­jects re­ported af­ter­ward that even though they knew the prob­a­bil­ities were against them, they felt they had a bet­ter chance when there were more red beans. This may sound crazy to you, oh Statis­ti­cally So­phis­ti­cated Reader, but if you think more care­fully you’ll re­al­ize that it makes perfect sense. A 7% prob­a­bil­ity ver­sus 10% prob­a­bil­ity may be bad news, but it’s more than made up for by the in­creased num­ber of red beans. It’s a worse prob­a­bil­ity, yes, but you’re still more likely to win, you see. You should med­i­tate upon this thought un­til you at­tain en­light­en­ment as to how the rest of the planet thinks about prob­a­bil­ity.

As I dis­cussed in “The Scales of Jus­tice, the Note­book of Ra­tion­al­ity,” Finu­cane et al. found that for nu­clear re­ac­tors, nat­u­ral gas, and food preser­va­tives, pre­sent­ing in­for­ma­tion about high benefits made peo­ple per­ceive lower risks; pre­sent­ing in­for­ma­tion about higher risks made peo­ple per­ceive lower benefits; and so on across the quad­rants.5 Peo­ple con­flate their judg­ments about par­tic­u­lar good/​bad as­pects of some­thing into an over­all good or bad feel­ing about that thing.

Finu­cane et al. also found that time pres­sure greatly in­creased the in­verse re­la­tion­ship be­tween per­ceived risk and per­ceived benefit, con­sis­tent with the gen­eral find­ing that time pres­sure, poor in­for­ma­tion, or dis­trac­tion all in­crease the dom­i­nance of per­cep­tual heuris­tics over an­a­lytic de­liber­a­tion.

Gan­zach found the same effect in the realm of fi­nance.6 Ac­cord­ing to or­di­nary eco­nomic the­ory, re­turn and risk should cor­re­late pos­i­tively—or to put it an­other way, peo­ple pay a pre­mium price for safe in­vest­ments, which low­ers the re­turn; stocks de­liver higher re­turns than bonds, but have cor­re­spond­ingly greater risk. When judg­ing fa­mil­iar stocks, an­a­lysts’ judg­ments of risks and re­turns were pos­i­tively cor­re­lated, as con­ven­tion­ally pre­dicted. But when judg­ing un­fa­mil­iar stocks, an­a­lysts tended to judge the stocks as if they were gen­er­ally good or gen­er­ally bad—low risk and high re­turns, or high risk and low re­turns.

For fur­ther read­ing I recom­mend Slovic’s fine sum­mary ar­ti­cle, “Ra­tional Ac­tors or Ra­tional Fools: Im­pli­ca­tions of the Affect Heuris­tic for Be­hav­ioral Eco­nomics.”

1Christo­pher K. Hsee and Howard C. Kun­reuther, “The Affec­tion Effect in In­surance De­ci­sions,” Jour­nal of Risk and Uncer­tainty 20 (2 2000): 141–159.

2Kimihiko Ya­m­ag­ishi, “When a 12.86% Mor­tal­ity Is More Danger­ous than 24.14%: Im­pli­ca­tions for Risk Com­mu­ni­ca­tion,” Ap­plied Cog­ni­tive Psy­chol­ogy 11 (6 1997): 461–554.

3Paul Slovic et al., “Ra­tional Ac­tors or Ra­tional Fools: Im­pli­ca­tions of the Affect Heuris­tic for Be­hav­ioral Eco­nomics,” Jour­nal of So­cio-Eco­nomics 31, no. 4 (2002): 329–342.

4Veronika Denes-Raj and Sey­mour Ep­stein, “Con­flict be­tween In­tu­itive and Ra­tional Pro­cess­ing: When Peo­ple Be­have against Their Bet­ter Judg­ment,” Jour­nal of Per­son­al­ity and So­cial Psy­chol­ogy 66 (5 1994): 819–829.

5Finu­cane et al., “The Affect Heuris­tic in Judg­ments of Risks and Benefits.”

6Yoav Gan­zach, “Judg­ing Risk and Re­turn of Fi­nan­cial As­sets,” Or­ga­ni­za­tional Be­hav­ior and Hu­man De­ci­sion Pro­cesses 83, no. 2 (2000): 353–370.