Philosophers suffer the same cognitive biases as everyone else

Behavioural economics is essentially founded on the principle that decision-makers are affected by a range of cognitive biases. In his book "Thinking Fast and Slow" Daniel Kahneman distinguishes between two systems of thought: 'System 1' is fast, instinctive, emotional and subject to many of the observed cognitive biases; and  'System 2' is slower, logical and more deliberative and able to avoid at least some of the biases that System 1 is subject to. The obvious implication is that if you could train people to slow down their thinking and apply more logical reasoning and be more deliberative, you could help them avoid many of the common cognitive biases.

Which brings me to this recent paper (ungated version here) by Eric Schwitzgebel (University of California at Riverside) and Fiery Cushman (Harvard). In the paper, the authors test whether academic philosophers are subject to some common cognitive biases to the same extent as similarly-educated non-philosophers. They use two common experiments: (1) the trolley problem (a common ethics problem); and (2) the 'Asian disease' problem described by Kahneman and Tversky (which is explained here and here, and which I use in my ECON110 class each year).

Essentially the authors were looking for two cognitive biases: (1) Ordering effects, where the order that scenarios are presented affects how they are evaluated; and (2) Framing effects, where the way that different options are framed (or presented) affects how they are evaluated. They find:

...substantial order effects on participants’ judgments about the Switch version of trolley problem, substantial order effects on their judgments about making risky choices in loss-aversion-type scenarios, and substantial framing effects on their judgments about making risky choices in loss-aversion-type scenarios.

Moreover, we could find no level of philosophical expertise that reduced the size of the order effects or the framing effects on judgments of specific cases. Across the board, professional philosophers (94% with PhD’s) showed about the same size order and framing effects as similarly educated non-philosophers. Nor were order effects and framing effects reduced by assignment to a condition enforcing a delay before responding and encouraging participants to reflect on “different variants of the scenario or different ways of describing the case”. Nor were order effects any smaller for the majority of philosopher participants reporting antecedent familiarity with the issues. Nor were order effects any smaller for the minority of philosopher participants reporting expertise on the very issues under investigation. Nor were order effects any smaller for the minority of philosopher participants reporting that before participating in our experiment they had stable views about the issues under investigation.

In other words, even academic philosophers are subject to the same cognitive biases as non-philosophers, and even when they are familiar with the problems they are being asked to evaluate. Scary stuff, particularly as the authors conclude:

Our results cast doubt on some commonsense approaches to bias reduction in scenario evaluation: training in logical reasoning, encouraging deliberative thought, exposure to information both about the specific biases in question and about the specific scenarios in which those biases manifest.

All of which suggests that nudging may be one of the few solutions to cognitive bias.

[HT: Marginal Revolution]

Lions are dying to be saved

You would have to have been living under a rock not to have heard about the furore set off by the killing of Cecil the lion last month (Wikipedia summarises the reactions to the killing here). From my reading though, the best report is this one by Norimitsu Onishi in the New York Times:

Despite intensifying calls to ban or restrict trophy hunting in Africa after the killing of a lion named Cecil in Zimbabwe, most conservation groups, wildlife management experts and African governments support the practice as a way to maintain wildlife. Hunting, they contend, is part of a complex economy that has so far proven to be the most effective method of conservation, not only in Africa but around the world as well.

While hunting is banned in government parks here in South Africa, animals inside their boundaries are routinely sold to game ranches when their populations are considered excessive, generating money to maintain habitats and fight poachers.

And because trophy hunting is legal in private game reserves, the animals end up fetching higher prices than they would in being killed for food or other reasons, conservationists contend.

In other words, hunting helps to save threatened species like lions. How does that work? There are two ways to explain this argument.

First, the trouble with most threatened or endangered species (like lions) is that they are a common resource - rival and non-excludable. Rival goods are those where one person's use of the good reduces the amount available to everyone else, i.e. in this case killing one lion reduces the number of lions available to everyone. Non-excludable goods are those where you cannot easily prevent a person from obtaining the benefit from them, i.e. in this case it is difficult to stop the hunters from killing lions. In contrast, if only hunters with an appropriate licence are allowed to hunt lions, that makes lions a private good - rival but excludable (because only licensed hunters can kill them). Making threatened species a private good creates incentives for local communities (or farmers) to conserve the species - it is valuable to sell the rights to hunt to hunters, and this value can only be realised in the long run if the species is managed sustainably.

Second, consider the opportunity costs of killing a lion. If there are hunting licences and the licence to kill a lion can be sold for between $24,000 and $71,000 (according to the Onishi article linked above), then the opportunity cost of killing a lion is the tens of thousands of dollars in income foregone. This provides incentives for locals not to kill lions, since they are very valuable as hunting trophies. On the other hand, if there are no licences to kill lions that can be sold to hunters, then the value of a lion falls to the value of the meat (less than 60 cents per pound according to the article, or about $300 for a mature male lion). The opportunity cost of locals killing lions for food is much lower, so more lions may be killed (a simple result of the Law of Demand - the 'price' (opportunity cost) of killing lions is lower, so more lions are likely to be killed).

Either way, hunting is a solution to the problem of threatened species, in a similar way that farming could also be a solution. So, many of the arguments against hunting threatened species may well be misplaced.

More on endangered species:

• Can technology save the rhinos?
• Because crushing ivory is so much better than burning it
• Poisoning rhino horns to save the rhinos
• This couldn't backfire, could it?... Ivory burning edition

The drug dealers' dilemma

Probably the most famous example in game theory is the prisoners' dilemma. The general story goes something like this (with lots of variants; this is the version I use in ECON100):

Bonnie and Clyde are two criminals who have been captured by police. The police have enough evidence to convict both Bonnie and Clyde of the minor offence of carrying an unregistered gun. This would result in a sentence of one year in jail for each of them.

However, the police suspect Bonnie and Clyde of committing a bank robbery (but they have no evidence). The police question Bonnie and Clyde separately and offer them a deal: if they confess to the bank robbery they will get immunity (and be set free) but the other criminal would get a sentence of 20 years. However, if both criminals confess they would both receive a sentence of 8 years (since their testimonies would not be needed).

The outcome of the game is that both criminals have a dominant strategy to confess. Confessing results in a payoff that is always better than the alternative, no matter what the other criminal decides to do.

The New York Times reports on a new real-world take on this game:

When the sheriff in Franklin County, Ky., posted a flier on Facebook asking local drug dealers to snitch on their competition, the response was more than a little incredulous.

That is, until a tip sent to a phone number on the flier led to an investigation that helped the sheriff arrest a local drug dealer...

“Asking drug dealers to turn in other drug dealers,” Sheriff Melton said. “It’s comical, and it’s working.”

Here's the flier they used:

And here's why it works, for drug dealers who have short time horizons. For simplicity, let's assume that there are two drug dealers only (A and B). In this short-run non-repeated game each drug dealer has a simple choice: Snitch on their competition, or stay silent. If both drug dealers stay silent, then make low profits (because they have to compete with each other). If one stays silent and the other snitches, then the snitch makes high profits while the other goes to jail. If both snitch, then both go to jail. The game in normal (payoff table) form is as follows:

What happens in this game? Consider Drug Dealer A first. They have a weakly dominant strategy to snitch. This is because the payoff is never worse than snitching (and sometimes it is better). If Drug Dealer B stays silent, Drug Dealer A is better off snitching (because high profits are better than low profits). If Drug Dealer B snitches, either strategy is equally good for Drug Dealer A (since they go to jail either way). So Drug Dealer A shouldn't choose to stay silent, because the better option (at least some of the time) is to choose to snitch.

Now consider Drug Dealer B. They also have a weakly dominant strategy to snitch. This is because the payoff is never worse than snitching (and sometimes it is better). If Drug Dealer A stays silent, Drug Dealer B is better off snitching (because high profits are better than low profits). If Drug Dealer A snitches, either strategy is equally good for Drug Dealer B (since they go to jail either way). So Drug Dealer A shouldn't choose to stay silent, because the better option (at least some of the time) is to choose to snitch.

So, both drug dealers acting in their own best interests leads both to snitch. If you're working out the Nash equilibriums, there are three - the only outcome that isn't a Nash equilibrium is the outcome where both drug dealers remain silent (to see why, take any of those outcomes and consider whether either player would be willing to change strategy - since they wouldn't (or would be indifferent to the change in strategy) those three outcomes are Nash equilibriums).

Now, the game isn't quite that simple of course, because the game isn't played just once (a non-repeated game, which is what the payoff table shows), but is played many times. You could think of the two drug dealers as making the choice to snitch or stay silent every morning when they wake up. The payoffs to the repeated game are the sums of the profits (low or high) or jail time from every play of the game.

So what happens in a repeated game? The 'best' choice for each drug dealer may be to cooperate (i.e. to remain silent). If either drug dealer snitches, then that is likely to cause the other drug dealer to snitch the next time the game is played. Snitching only has a short-term payoff, but ultimately leads to a worse outcome for everyone. So, the threat of reciprocal snitching may be enough to ensure an uneasy alliance results for drug dealers who have long time horizons (those who recognise this as a repeated game). Which suggests that there is probably such a thing as 'honour among drug dealers', and this police strategy will not work for long.

[HT: Marginal Revolution]

Your smartphone knows if you will pass or fail this semester

Tyler Cowan at Marginal Revolution pointed me to this new paper (pdf) by Rui Wang (Dartmouth College), Gabrielle Harari (University of Texas at Austin), Peilin Hao, Zia Zhou and Andrew Campbell (all Dartmouth College). The authors used continuous sensing data (like location, audio) collected from students' smartphones over a semester (apparently over 53GB of data per student!) to infer the times students spent studying (smartphone stationary in a location like the library, with no audio), in conversations (smartphone stationary, with audio), partying (smartphone in a location like a fraternity, with loud audio), etc. They also collected data on the students' personality, stress, sleep, and a bunch of other things.

What did they find? Nothing too surprising, according to the paper:

Our prediction model indicates that students with better grades are more conscientious, study more, experience positive moods across the term but register a drop in positive affect after the midterm point, experience lower levels of stress as the term progresses, are less social in terms of conversations during the evening period, and experience change in their conversation duration patterns later in the term.

In this age of worries about the data that is collected from us, I wonder how they got their 30 students to agree to this data collection exercise (and they must have had a really good story for their institutional review board!). Two takeaways from the paper though:

1. Students' behaviours really do matter for their chances of passing - studying more makes a difference! Though apparently attendance in class does not - that might be an artefact of the small sample size and the homogeneity of the sample. Note that they do find that students who increase their attendance after the mid-point of the semester do perform better; and
2. Student smartphones potentially contain a wealth of unexploited data that could be used to develop study aids. How long before there is an app that prompts you to decrease your time spent partying (or encourages you to party more early in the semester, and less later in the semester)?

Finally, this suggests that perhaps universities could use this data to identify students at risk of failure, and intervene. Provided students let their university have access to continuous sensing data from their phones, of course!

Related reading:

• Could early identification of students at risk of failure be a bad thing?