Blog

Gist

When peo­ple do not know with cer­tainty if an event A will occur or not, and if they have to make a deci­sion based on event A occur­ring, then they may post­pone their deci­sion until they know with cer­tainty if A has occurred, even if they would have made the same deci­sion whether or not A had occurred. As clearly put:

“We call the above pat­tern a dis­junc­tion effect. A dis­junc­tion effect occurs when peo­ple pre­fer x over y when they know that event A obtains, and they also pre­fer x over y when they know that A does not obtain, but they pre­fer y over x when it is unknown whether A obtains or not.”¹

Dis­junc­tion effects are at play in many decision-making sit­u­a­tions and can influ­ence how long peo­ples are tak­ing to decide, and even the deci­sions themselves.

Origins

To under­stand intu­itively what a dis­junc­tion effect is, let’s look at the fol­low­ing example:

“For another exam­ple of non­con­se­quen­tial rea­son­ing, imag­ine that you have agreed to bet on a toss of a coin in which you have equal chance to win $200 or to lose $100. Sup­pose that the coin has been tossed, but that you do nit know whether you have won or lost. Would you now want to play this gam­ble for a sec­ond time? Alter­na­tively, how would you feel about accept­ing the sec­ond gam­ble if you knew that you lost $100 on the first gam­ble? And finally, would you play the sec­ond gam­ble hav­ing dis­cov­ered that you won $200 on the first gam­ble? We have shown that, con­trary to STP, a major­ity of respon­dents accepted the sec­ond gam­ble both after hav­ing won as well as hav­ing lost the first, but a major­ity rejected the sec­ond gam­ble when the out­come of the first was not known. This pattern—accept when win, accept when lose, but reject when do not know—was the sin­gle most fre­quent pat­tern of pref­er­ences exhib­ited by our sub­jects.”¹

Another exam­ple stud­ied by Shafir and Tver­sky is a mod­i­fied ver­sion of the prisoner’s dilemma where the sub­ject is play­ing one-round-only PD with an unknown oppo­nent. The twist is that he is some­times given the deci­sion of the oppo­nent (whether (s)he com­petes or coop­er­ates). What is observed is that, when the choice of the oppo­nent is known, the sub­ject chooses “com­pete” the vast major­ity of the time, regard­less of whether his oppo­nent has cho­sen “com­pete” or “cooperate”.

When he does not know, how­ever, the pro­por­tion of sub­jects choos­ing to “coop­er­ate” is ris­ing sig­nif­i­cantly. Shafir and Tver­sky attribute this to an effort by the sub­ject to make his best to bring a mutu­ally ben­e­fi­cial sit­u­a­tion. When the sit­u­a­tion is played out how­ever, this argu­ment is no longer valid and the choice is made purely from an indi­vid­ual per­spec­tive. This is the same mech­a­nism that makes us vote, even though we know our impact won’t change the result, but make us totally unin­ter­ested in vot­ing as soon as the out­come is known.

Another expla­na­tion put for­ward is the following:

“Alter­na­tively, peo­ple may con­sider all the rel­e­vant out­comes but, due to the pres­ence of uncer­tainty, may not see their own pref­er­ences very clearly.”¹

At Stake

Strat­egy

Strate­gists must take care of Dis­junc­tion Effects and not delay­ing deci­sions just in the face of uncer­tainty around what is con­sid­ered a key event. Often times, even though the event is indeed key, the deci­sion taken is the same even after dif­fer­ent out­comes. They would in effect delay a deci­sion in order to wait for an event to be real­ized or not, even though this does not affect their final deci­sion. Put it another way:

“One result of non­con­se­quen­tial deci­sion mak­ing is that peo­ple will some­times seek infor­ma­tion that has no impact on their decision.”

Sub­jects are in effect will­ing to pay for infor­ma­tion that is not going to change their choice, but merely going to clar­ify their rea­sons for choosing.

Mar­ket­ing

If you put your clients in front of a choice where the uncer­tainty on the out­come of an event is lead­ing to a dis­junc­tive choice, you risk delay­ing your client’s choice, and thus risk los­ing him or her entirely if con­di­tions change dur­ing this delay. Price struc­tures in par­tic­u­lar can lead to such effects.

Cor­po­rate

Some man­agers can fall prey to this effect and choose to delay deci­sions before what is con­sid­ered a deci­sive event, even though exter­nal con­di­tions and fixed con­straints mean that heir choice will be the same regard­less of the out­come of the deci­sive event.

Actionable ?

Tver­sky and Shafir show that “once peo­ple are made aware of their pref­er­ences given each pos­si­ble out­come, the sure-thing prin­ci­ple is no longer vio­lated.”² Thus explicit analy­sis and descrip­tion of the dif­fer­ent pos­si­ble sce­nar­ios along with the dif­fer­ent course of actions decided in each case would tend to, at least, reduce the risk to fall into this trap.

More details

Dis­junc­tion effects go against the sure-thing prin­ci­ple (STP). “The sure-thing prin­ci­ple says that if we pre­fer x to y given any pos­si­ble state of the world, then we should pre­fer x to y even when the exact state of the world is not known. […] STP is an impor­tant impli­ca­tion of the con­se­quen­tial­ist view. It cap­tures a fun­da­men­tal intu­ition of what it means for a deci­sion to be deter­mined by the antic­i­pated con­se­quences. It is a cor­ner­stone of Expected Util­ity The­ory.”¹

References

1. (1992) Shafir, E., & Tver­sky, A. Think­ing through uncer­tainty: Non­con­se­quen­tial rea­son­ing and choice. Cog­ni­tive Psy­chol­ogy, 24, 4, 449–474. [↩] [↩] [↩] [↩]
2. (1992) Tver­sky, A., & Shafir, E. The dis­junc­tion effect in choice under uncer­tainty. Psy­cho­log­i­cal Sci­ence, 3, 5, 305– 309. [↩]