Mom, I looked better than her in that dress... right?
Decoy effects and politics - a simplified discussion through the lens of the 2012 Republican primary
Take a look at this field and what do you see?
Ok… maybe that’s a bit of a loaded question – let’s come back to that.
I’m a self-described “election junkie.” No really – like I really geek out on elections. I’ve been listening to FiveThirtyEight’s politics podcast since high school. I’m the type of person that sets up with two screens in front of the TV on election night tracking polling, live blogs, and watching whatever flavor of this guy I’m feeling describe to me how whomever candidate is outperforming their early vote predictions. That said, I’m not here to write about election forecasting (a topic which I could write on and on about…) but rather voter psychology. At the end of the day elections are about irrational actors making supposedly rational choices – a classic problem for decision theory.
Let’s say we can boil down the success of a candidate in an election to a few variables:
Candidate Quality
Incumbency
Electability
Demographic makeup of the electorate (i.e., who is actually voting)
Distance from the median voter on policy (i.e., how far are you from the “center”)
There is countless research devoted to understanding each one of these factors and how they play into understanding who wins an election, but instead of thinking about the importance of each one of these factors let’s think for a moment about how a candidate is perceived by the voter along each one of these factors. Apart from incumbency and the demographics of the electorate, each of the remaining factors is relative. Where relativity exists, subjectivity exists which means a voter’s evaluation of these factors can change based on who’s running against said candidate. The more candidates in a field, the more relativity is introduced into voters’ perceptions of said candidates, opening the door to heuristics.
One of these is what’s called the decoy effect - let’s think about this through a business example:
A decoy effect is characterized as a dominated choice: when an individual’s preference for one option over another changes as a result of the addition of a third, less preferable, option. For example, if someone prefers beer to wine, we’d expect that placing a horrible wine next to an amazing wine and an average beer would nudge their preference toward the good wine as it “looks better” than the alternatives.
Decoy effects are rampant in marketing and in studies of consumer psychology. They’re also probably what your off-the-shelf McKinsey consultant tells you about when you invite them over to analyze your business. Evidence of decoy effects have changed the way many social scientists have understood how individuals perceive “indifference” when confronted with two choices or how to incentivize a particular choice by introducing a decoy.
Let’s think through this with a classic example from Dan Ariely’s Predictably Irrational1:
You’re a college student looking to subscribe to The Economist. The Economist, having hired “value-maximizing” 22-year-old management consultants to provide a pricing strategy for their new student plan have come up with the following options:
Option 1: Economist.com subscription - $59. A one-year subscription to Economist.com. Includes online access to all articles from The Economist since 1997
Option 2: Print subscription - $125. A one-year subscription to the print edition of The Economist
Option 3: Print & web subscription - $125. A one-year subscription to the print edition of The Economist and online access to all articles from The Economist since 1997
Being the savvy student you are, you notice the following: Option 3 provides everything option 2 does and more all for the same price!
Regardless of the fact that you’ll never actually read the paper version – Option 3 looks most attractive of this set given that it “strictly dominates” Option 2 (i.e., it provides more benefits than Option 2 at the same cost)
In this real-life example:
16% of students chose Option 1
0% chose Option 2
84% chose Option 3.
Yes, you read that right. Nobody chose Option 2. That is the power of the decoy effect.
That said, what’s more interesting is what happens when a group of students were only given options 1 and 3 – in this case, 68% of students chose the web-only option with the rest opting for the combined print and internet offering.
With the presence of the decoy, more customers shifted to the higher-priced offering simply because it strictly dominated another choice. You can see countless examples of this pricing strategy in business (Apple’s iPhone line-up being a prime example).
Ok, so we see this effect in marketing – but how does it relate to politics?
Back to this picture:
For those unfamiliar this is a picture of the 2012 Republican primary field – arguably one of the most accomplished and ideologically diverse primary fields to date. From left to right we have:
Rick Santorum (Senator, Pennsylvania)
Michelle Bachman (Representative, Minnesota)
Newt Gingrich (Former House Speaker, Georgia)
Mitt Romney (Governor, Massachusetts)
Ron Paul (Representative, Texas)
Tim Pawlenty (Governor, Minnesota)
Herman Cain (Former CEO, Godfather’s Pizza)
Now it took a while for delegates to coalesce around Romney, the eventual nominee among this star-studded cast. In retrospect, though, its not hard to see how decoy effects may have played a role in primary voters’ choices. Let’s zoom in for a second on the top 4 candidates in the race – I’ll list the dates they withdrew as well to really hone in on the apparent decoy effect present:
Mitt Romney (1st place, mathematically secured nomination on May 29th)
Rick Santorum (2nd place, withdrew on April 10th)
Ron Paul (3rd place, never withdrew and lost during convention August 30th)
Newt Gingrich (4th place, withdrew May 2nd)
If we take a look at the withdrawal periods we see one thing – Gingrich and Santorum (who together had more delegates than Paul) withdrew when it was clear they weren’t able to compete with Romney while Paul stuck in the race. Both Gingrich and Santorum eventually endorsed Romney, which is where we can start to see the effect of perception across this candidate field:
You have Ron Paul who’s an unabashed libertarian – unafraid to stick it to the establishment on the Iraq war, bailouts, and a plethora of other issues
You’ve also got Gingrich, a known commodity who made a name for himself under the presidency of Bill Clinton. He’s been out of the limelight for a while but is chasing the nomination
Then there’s Santorum, another senator who was a surprise in this race but is quite similar in policy positioning to both Gingrich and Romney – though clearly more right-wing on social issues
For those unfamiliar with Romney – and although I express sincere admiration for him – I’d best describe him as a “middle of the road flip-flopper.” At his heart, he’s an establishment Republican but his record in Massachusetts painted him more as a Blue Dog Democrat than anything else. That said, he was the party’s “chosen son” after waiting his turn behind McCain in 2008
It's easier, perhaps, to separate this field into 3 groups:
The outsider, in Paul
The establishment, in Gingrich and Romney
The upstart, in Santorum
Here, Santorum, Gingrich, and Romney have more similarities than differences – though it’s quite clear for those who remember that Romney provided all of the right messaging on social issues without the “extreme positions’ of Santorum and all of the right experience on economic issues without the perceived “politicking” of Gingrich.
If we take this back to the pricing example, we can see three buckets form:
Your internet only plan (aka the outsider plan) in Paul
Your paper only plan (aka the “traditionalists” with some flaws) in Santorum and Gingrich
Your paper + online plan in Romney
Are there other dynamics at play? Sure, but the premise of this (though rather simplistic analysis) is to understand that voters stack candidates against each other and are compelled by things like decoy effects in primaries. In any 3 candidate field, the candidate that is most different from the 3 is often left behind, while the “non-decoy” from the remaining pair is most often the winner. We can see this trend occur generally across the past few presidential primaries:
2008 Republican Primary:
Mccain (Winner)
Paul (Outsider)
Romney/Huckabee (Decoy)
2016 Democratic Primary:
Clinton (Winner)
Sanders (Outsider)
O’Malley (Decoy)
2020 Democratic Primary
Biden (Winner)
Sanders (Outsider)
Harris/Warren/Bloomberg (Decoys)
Regardless of the efficacy of this observation in forecasting elections – it better helps us understand candidate positioning in elections. As we look to the Republican primary in 2024, which may very well be contested between Donald Trump and Ron DeSantis (two, let’s say “similar” candidates) understanding who voters might perceive as the “decoy” may provide better insight into whose campaign will perform better.
Next piece we’ll flip back to education.
I am aware about the recent controversies surrounding Dan and his use of potentially manipulated data, among other things. That said, I believe this effect has been replicated through real-world RCTs (The Economist example being one of them) and is credible.