| Date: | April 23, 2007 / year-entry #141 |
| Tags: | non-computer |
| Orig Link: | https://blogs.msdn.microsoft.com/oldnewthing/20070423-01/?p=27153 |
| Comments: | 33 |
| Summary: | Shankar Vedantam wrote a Washington Post article and also appeared on NPR to discuss The Decoy Effect: Given a choice between two options, introducing a third, clearly inferior, option can influence your original decision. You won't pick the third option, but a clever choice of the bad third option can sway the decision toward either... |
Shankar Vedantam wrote a Washington Post article and also appeared on NPR to discuss The Decoy Effect: Given a choice between two options, introducing a third, clearly inferior, option can influence your original decision. You won't pick the third option, but a clever choice of the bad third option can sway the decision toward either of the other two. Rationally, an inferior third option should have no effect on your choice between the two other options, but psychologists (and marketing majors) have discovered that human beings are not rational decision makers. (Anyone who has interacted with a two-year-old child is already well aware of this.) |
Comments (33)
Comments are closed. |
Genuinely fascinating.
I can’t say I find this surprising. One of the first things we studied in my first semester of statistics is the way people bend the numbers to favor their point of view or word questions/statements in a certain way to achieve a desired result. This article certainly falls in that category.
Lies. Damned Lies. And Statistics.
JamesNT
The presidential elections that Perot, and Nader were in showed this on a national level.
In many European countries the presidential election is done in two steps.
The first step includes all the candidates.
If in the first round no candidate gets more than 50% of the votes, then a second round is organized, with the first two.
Such a system is more expensive, but eliminates the decoy effect.
Imagine two rounds in 2004: Bush/Gore/Nader, followed by round two Bush vs Gore.
Andy:
They don’t, unless you are arguing that Perot made more people vote Bush than would have otherwise, and that Nader made more people vote Gore than would have otherwise.
Human beings are also quite fond of crapping in their pants, as my interactions with a two-year-old child have demonstrated.
"The presidential elections that Perot, and Nader were in showed this on a national level"
I don’t think the Presidential elections are an example of the Decoy Effect. Ross Perot and Ralph Nader did not make people choose to vote for George Bush. People actually chose to vote for Perot and Nader. This is the fact that hurt Al Gore. Perot and Nader were stealing votes from Gore, not giving votes to Bush.
The Decoy Effect states that Perot and Nader would be inferior candidates. They were not. It also states that people would then be more likely to vote for Bush. They didn’t, they voted for Perot and Nader.
Fast food restaurants understand the decoy effect very well. In most cases the "large" size of a sandwich, drink etc. on a menu is simply there to push people towards choosing the more profitable "medium" size instead of the "small" that is more than sufficient for most people.
No, fast food is not an example of the decoy effect. It’s different–a large soda is similar to how electronics companies will have an incredibly expensive model in their product line. Many people shy away from buying top end items. So a ridiculously expensive/large top end item will make the next best item seem more reasonable. If the most expensive computer were $2,000, consumers would tend to buy $1,000 computers. But introduce a $3,000 model and $2,000 won’t look so bad by comparison.
It’s similar, but it’s not the decoy effect as described here.
uhh. You young folks need some history.
When Perot ran, it was the ELDER (George Herbert Walker) Bush who ran along with Bill Clinton. Perot took votes AWAY FROM BUSH, and that is why we had Clinton as president.
That would be the US system. If a candidate fails to collect 51% of the electoral vote, the runoff voted on by the house is of the top 2.
They are inferior, at least in terms of attracting votes. In our electoral system, a third party candidate can split the vote, thus allowing the second most favored candidate to win – that’s why the republicans funded nader in 2004: it drew votes from lurch (who has 3 purple hearts!).
I think he’s referring to a runoff general election, which we don’t do. Personally, I’d prefer instant runoff voting, but I don’t see that happening any time soon.
another way to influence a choice: a winform, two buttons:
1) Yes
2) Of course
and for decoy effet:
3) For sure!
easy :-))))
Ryan, the US system is most definitely not the equivalent of the democratic French system. The US system is deliberately not democratic in the electoral college, being biased in favor of lower population states and eliminating direct voting for candidates by the populace. The Founders didn’t have such a great opinion of popular democracy and it shows in the system they created.
OK, so this is *really* off-topic, but the *2001* election was Bush/Gore/Nader; the 2004 election was Bush/Kerry/Nader (as a distant also-ran). One could make a good case that Nader cost Gore the election in 2001, but so far as I can tell, Republican donors didn’t send donations to Nader in *2001*. They sent donations to Nader in *2004*, and, quite frankly, wasted their money, as Nader did much worse than he did in 2001.
This is constantly used in politics to advance new ideas. You will often see someone propose a rather radical idea A which naturally will be attacked as way too radical. You may even have the impression the proposer must out of his mind for coming up with stuff like that. Shortly thereafter, a less radical idea B comes along (incidentally by the same people, but that’s not always as obvious) which goes in the same direction but, since it’s not as radical, it has a good chance of being accepted. Here the first proposal just served to set the boundaries of debate so the later proposal has a chance. If the first proposal hadn’t been made, the decision would have been between "status quo" and "idea B". Now that idea B falls in the middle of the spectrum of debate, it appears like a compromise and both "status quo" and "idea A" as too extreme.
Naturally, this is also a common tactic used in negotiation. Pick what you want, ask for more and negotiate yourself back to what you wanted in the first place. Very effective and makes you appear like someone who seeks win-win situations and constructive solutions.
*cough*Vista Home Basic*cough*
Linux between Windows and Mac came to my mind.
<bashing follows>
@James Day: "The Founders didn’t have such a great opinion of popular democracy and it shows in the system they created."
James, you do understand that’s not the only, or necessarily even primary, reason for the EC, don’t you? It’s there to force candidates to not ignore lightly-populated states. That’s why proposals to abolish the EC are so wrong-headed: they will allow candidates to focus on a dozen or so major metropolitan areas and ignore the entire rest of the country.
I just don’t know why you’d need a two-year-old to demonstrate how people don’t make rational decisions. Virtually the entire advertising business is predicated on subverting what little logic there is that goes into purchasing. As Hanan Levin says: "People buy with emotion & justify with reason always."
The Electoral College has a few purposes, none of which have to do with "distrust" of the populace.
You have to also remember that States are entrusted to develop their own election systems. If Washington State suddenly decided to select a Presidental candidate by playing cow-plop bingo, we’re free to do so… the Electoral College makes this possible by abstracting the Federal election process from the State election process. In theory, Washington State could also become a dictatorship and still be a member of the United States, although in reality I doubt anybody in the Federal Government would stand for that.
As Rick states, it also has the effect of lessening the influence of densely populated states so that farmers in Iowa get the same (or at least similar) representation as movie producers in LA.
Mihai: It doesn’t have to be much more expensive.
People can rank candidates by preference on the first ballot, and the vote-counting process can throw out very low-scoring candidates until someone has 50% of the vote.
It’s called Instant Runoff, I think.
> That’s not the definition of inferior used in this post. As far as the decoy effect goes, inferior means “worse with respect to measures A and B”, not “less popular”.
In elections, inferior means ‘bad at getting votes’. The skills needed in order to get elected are different from the skills needed to do the job. Sound familiar?
The use of a supe-premium product to increase acceptance of a premium product is well-established, as per the fast food example, as well as Michelob, Cadillac, etc. But there are very few niches in which it makes sense to deliberately introduce an inferior product, so while the decoy effect may be seen in the marketplace, I don’t think it’s on the list of major marketing strategies.
For the decoy effect to work all the choices need to be desirable. In the wikepeia description both of the MP3 players were desirable. Adding a 3rd desirable player allowed you to more easily choose between the original two.
In politics the effect can be applied when choosing a candidate from within your party because the leading two candidates are both desirable choices when the third desirable candidate is added.
The problem is that for a Presidential election the effect breaks down because you are no longer choosing between two desirable candidates. Based on your political party one candidate will be desirable while the other is undesirable. When the 3rd choice is added he will again be either desirable or undesirable. In this case, if he is undesirable your vote does not change otherwise you now have two desirable candidates to choose from. This may be difficult since you don’t have a third desirable choice applying the Decoy Effect.
I guess it could still apply for those who cannot decide what party they are in and list themselves as undecided.
—"They sent donations to Nader in *2004*, and, quite frankly, wasted their money, as Nader did much worse than he did in 2001."—-
Could the two factors have been related?
"I just don’t know why you’d need a two-year-old to demonstrate how people don’t make rational decisions."
*That* people don’t make rational decisions is obvious. I think *how* they (we) are irrational, though, is quite interesting.
"They are inferior, at least in terms of attracting votes."
That’s not the definition of inferior used in this post. As far as the decoy effect goes, inferior means "worse with respect to measures A and B", not "less popular".
Let’s say, for the sake of argument, that you are a pretty liberal voter. Then two measures of a candidate’s worth might be liberalness (the more liberal, the better) and popularity (all things equal, you’d rather vote for a candidate who has a chance of winning). Nader was more liberal, but far less popular. Better at one measure, but worse at another. He wasn’t a strictly inferior alternative, and so whatever effect he had wasn’t the decoy effect.
> They sent donations to Nader in *2004*, and, quite frankly, wasted their money, as Nader did much worse than he did in 2001.
Did they? They drew votes from lurch, improving junior’s shot at the white house, and it worked.
"James, you do understand that’s not the only, or necessarily even primary, reason for the EC, don’t you? It’s there to force candidates to not ignore lightly-populated states. That’s why proposals to abolish the EC are so wrong-headed: they will allow candidates to focus on a dozen or so major metropolitan areas and ignore the entire rest of the country."
Instead, they ignore the states where they can’t possibly win and the states where they can’t possibly lose, and concentrate on the states that are "up in the air" and on the states with the most votes.
If the electoral votes were split according to the split of the popular vote in the state, as some states do, then this would be valid.
When the EC was set up in the 1700s, there was no national communications system. The candidates didn’t have the exposure they have now. Traveling across the country took a long time. The candidates couldn’t hope to visit every area of the country, so the average person would have no idea who they were voting for. Instead, they had this system where they voted for someone local whom they "trusted" to go look at the candidates and vote for the best person.
Okay, this is getting too far off topic, I’ll stop now.
A related phenom is giving people three buying choices: more than with one or two options, more people will choose one thing of the three to buy and the average spent per-person goes up
One of the things a teacher in some study-hall type class said on the topic of how to take multiple-choice tests was that if there are a pair of opposite answers, one of them is more likely to be correct than any of the other available answers.
This is just an example of the failure of Independence of Irrelevant Alternatives in Arrow’s theorem. If there is no perfect choice system, then there’s no surprise that the human brain doesn’t have one.
@Richard: Actually, since Arrow’s Theorem only applies in elections (i.e. two or more voters), the fact that the brain fails IIA says a lot about the physical process by which the brain internally works: the brain must have subcomponents, and the final outcome is determined by a vote.
Of course, neurologists already had a fairly good idea that that’s what’s going on under the hood. But it’s interesting to see independent confirmation of that.