All posts by Matthew Zefferman

About Matthew Zefferman

I am a biologist, engineer, and social scientist who uses mathematical modeling techniques from evolutionary biology to tackle questions related to the origins and operation of complex large-scale economic and political institutions.

Status quo bias and the Scottish ballot

Scottish Ballot
Source.

Above is a photo of the ballot for Scottish independence.

The ballot wording seems incredibly straightforward (which is the point of this short Slate article). However, way it is worded might be biasing the vote towards independence.

People seem to have a large number of biases and heuristics they use when making decisions.  One of these is called the “status-quo bias.”  Samuelson and Zeckhouser, in their seminal work,  demonstrated this bias by asking people to make decisions in hypothetical scenarios. In some treatments, one of the options was said to be the status quo and in other treatments none of the options were. Here is an example from the questionnaire in the appendix to the paper linked above:

teaching_job

Individuals given the first scenario are more likely to choose A than individuals given the second scenario. This demonstrates a bias because individuals are choosing between the same four choices, but in the second option the only difference is that individuals are already at College A. (It could be argued that the decision-makers are taking moving cost into account in their decision, but the status quo bias has been demonstrated in many scenarios where this type of cost is not a factor.)

The status quo bias is pretty robust – it has been demonstrated in questionnaires, field experiments and observational data. A famous example is the observation that countries where organ donation in the default option on the donation form have drastically higher organ donor rates than countries where not being an organ donor is the default. This holds even for cases where the same country changes from an opt-out to an opt-in form.

Taking the status-quo bias in mind, suppose that the Scottish ballot were worded:

“Should Scotland become an independent country?”

This slight change in wording would, I predict, highlight the status quo and result in a somewhat lower vote for independence. (Granted, this is a pretty safe prediction to make since no one can test it.) But would this difference be enough to sway the election?  I hear it will be pretty close…

 

 

 

 

 

 

 

 

 

 

 

 

 

al-Baghdadi is almost certainly descended from Muhammad

al Baghdadi source

 When news sites discuss the attempt by the leader of ISIS, al-Baghdadi, to form a new caliphate, they often mention his claim to be a “direct descendant of Muhammad.” This claim is often met with skepticism because, as  this article in the New Republic puts it:

...because no one knows much about Baghdadi - certainly not enough to trace his lineage back 1,400 years to a preliterate society a thousand miles away - it’s hard (and in the Islamic State, probably fatal) to suggest he’s lying.

I am not at all skeptical of al-Baghdadi’s claim.  In fact, if Muhammad has any surviving descendants, I would be extremely surprised if al-Baghdadi were not one of them. I would also be extremely surprised if I were not one of them. I would also bet a great deal, dear reader, that you are also a descendant of Muhammad, especially if you are of Eurasian decent.

These counter-intuitive notions follow from the findings of coalescent theory.  The logic is fairly simple, as we look backwards on our family tree the number of our ancestors grows exponentially.  We have two parents, four grandparents, eight great-grandparents, sixteen great-great-grandparents, etc.  The thing about exponentials is that the increase, well, exponentially. Exponentials are relentless in their increasing. Also, as you look backwards in time, there are fewer people so number of people who could be your ancestor decreases exponentially. Eventually, you reach a time where not only are you are descended from every person who has an ancestor today, but a time where everyone alive today shares the same set of ancestors.1

A 2002 Atlantic article by Steve Olson describes work by Joe Chang who found that this point for all Europeans was only around 1400 (Muhammad died much earlier in 632).2 This analysis has been backed-up by a more recent genetic analysis as described by Carl Zimmer.

Chang’s paper assumed random mating, so there may be concern about isolated groups. A more recent paper looked at a model that assumed pretty conservative migration rates between areas of the world (they actually did two models – one simple and once complex and had similar answers). Both models support the notion that Muhammad is the ancestor of most humans alive today.

The paper uses the example of Tasmania, which was geographically isolated from Australia for thousands of years before European colonization in the 19th century.  Although Tasmania was reproductively isolated well before the birth of Muhammad, there is no one alive today of pure Tasmanian descent. So if Muhammad has any descendants, it is extremely likely that all of today’s Tasmanians are all descendants of Muhammad.

In summary, if being a descendant of Muhammad is a requirement for being Caliph, it is either a very low bar or a very high bar. Either Muhammad has no descendants and, therefore, no one can be Caliph.  Or almost all seven billion of us are Muhammad’s decedent and the requirement is trivial. In any case, the press should not treat al-Baghdadi’s claims with skepticism, but indifference.

1- I should probably point out that these are “genealogical ancestors,” the ancestors that you would find on your family tree. You would have many fewer “genetic ancestors,” that is individuals who you share genes with through descent. Because you have a finite amount of genetic material and it is transmitted in discrete chunks, not all of your genealogical ancestors have contributed to your genome.

2 – This also explains one of the major scientific problems with the plot of the da Vinci Code.

 

Tit for Tat, It Ain’t All That – Part 1: The Prisoner’s Dilemma, except repeated.

I have been traveling to conferences and getting out publications, so have not been posting my planned “Tit for Tat, It Ain’t All That” series as quickly as I intended.  But here is Part 1:

Many readers of this blog will be familiar with the Prisoner’s Dilemma.  In case you are looking for a refresher, Zach Weinersmith has an SMBC comic with a pretty good summary of the usual set up.

In this series of posts I am going to use the formulation of the 2-player Prisoner’s Dilemma commonly used by biologists. One advantage of the biologists’ formulation is that the choices and payoffs are easier to remember (at least for me)  than the list of years a prisoner might get for confessing or not.  Another is that it only has two terms, which makes it more easy to work with mathematically. The disadvantage is that it is not as general as the standard economists formulation (by this I mean that there are 2-player Prisoner’s Dilemmas that cannot be expressed in the biologists’ formulation), but for our purposes in these blog posts we will be able to capture the complexity we need.

In the biologists’ formulation, you (let’s pretend that you are playing the game) are paired with another player and can give that player a gift.  The gift gives them a benefit of size b. However, giving the gift costs you.  The amount it costs you is c. Meanwhile and simultaneously, the other player is deciding whether to give you a gift that will benefit you by b and cost them c. We will call giving a gift “cooperation” and not giving a gift “defection.”

The below matrix shows the payoffs to the row player. As in Zach’s comic, your highest payoff is to defect, no matter what the other player does. In addition, the other player’s highest payoff is also to defect. However, as long as b > c, you would have both been better off had you both cooperated. Hence, the dilemma.

The 2-player Prisoner’s Dilemma. The payoff is to the row player.

The “rational solution” to the 2-player Prisoner’s Dilemma is well known.  Both players tragically defect.

But what if we add a twist where both players might meet again? If players play the game repeatedly, the thought is, perhaps cooperating in earlier interactions would encourage the other player to cooperate in later interactions, in the long-run, both players will  “escape” the dilemma by harvesting the rewards of repeated cooperation. This is a fairly seductive idea and is, in my experience, the thought to be the primary mechanism for escaping the Prisoner’s Dilemma in economics and, especially, political science.

However, in this series of blog posts, I plan to show that repeated interactions are not the panacea for solving the problems of cooperation that many think it is. The real story is a little more complicated and, frankly, a ton more interesting!

 

Tit-for-Tat. It ain’t all that.

The Iterated Prisoner’s Dilemma (IPD) is a deceptively simple game that has sparked a lot of research in many different fields. However, despite all of this attention, the IPD’s lessons are still, surprisingly, widely misunderstood.

Over the next few weeks I plan to post an eight-part series that will hopefully be helpful in bringing interested readers up-to-speed with the latest thinking about this important and fascinating game.

Here are the parts I am planning.  I will add links as I add posts.

Tit-for-Tat. It ain’t all that.
Part 1: The Prisoner’s Dilemma, except repeated.
Part 2: When Tit-for-Tat was all that.
Part 3: The population is important.
Part 4: When everything was all that.
Part 5: When nothing was all that (for long).
Part 6: Whither uncertainty?
Part 7: Selecting this from that.
Part 8: Learning to play the game.

 

 

Some fortnights of links

It has been a long time since I posted any links.  Here is the backlog.

Soccer, a Beautiful Game of Chance –  John Tierney. One thing about soccer is that, because it is so low-scoring, the outcomes of individual games are very stochastic. John Tierney argues that, because there is such a large skew in skill level between teams, this randomness is a good thing for the sport. It is the only way that the less-skilled teams have any chance to win and keeps the matches exciting.  Sort of like slot machines, I guess.

Developmental Psychology’s Weird Problem – Jane Hu. I’ve posted about the issues with doing experiments mostly with people from WEIRD (Western Educated Industrialized Rich and Democratic) societies before. This problem even extend to developmental psychology.

How economics became a science – Giorgos Kallis. This is a brief article about some non-conformist economists in the 1960s (including Herb Gintis and Sam Bowles) challenging the economics establishment.

How Not to Be Misled by the Jobs Report –  Neil Irwin and Kevin Quealy.  Why we simulate data.

A list of 26 Species “Concepts” – John Wilkins. It is difficult to define a species. I’ve found that this topic is good fodder for particularly nerdy conversation on long car trips. You mileage might (literally!) vary.

Avoiding “Sagan Syndrome.” Why Astronomers and Journalists should pay heed to Biologists about ET – Nathan Taylor.

Insanity: genes ‘versus’ environment as causes – Ken Weiss. Seriously, we need to come up with a sensible way to talk about this stuff.

Developmental Plasticity and the “Hard-Wired” Problem – Patrick Clarkin with a long and thoughtful post on a similar problem.

How Political Science Makes Politics Make Us Less Stupid – John Patty at the Math of Politics blog.

Nicolas Wade recently published a book on human population genetics called “A Troublesome Inheritance.” Various science bloggers and scientists who study aspects of population genetics explain why his science is all mixed up. Jeremy Yoder, Eric Michael Johnson, H. Allen Orr, Massimo Pigliucci, Michael Eisen, Jennifer Raff.

How to get an ecology PhD in four years (a 12-step program) – Rosemary Hartman at the UC Davis Ecology blog.  As a bonus, you may-or-may-not see a photo of Biased Transmission’s author rocking a pink short skirt.
Entertaining (in that internet sorta way):

Cookie Monster has really strong priors (about cookies).

Controversial Jeopardy! champ Arthur Chu tells his story – Marah Eakin. This is from back in February.  The summary is that Chu used some game theory to do better at Jeopardy!, but it was not as nice for regular viewers of the show.

Old School Language – a PNAS paper inspires an old school rap song by Zach Sherwin.

John Conway sort-of hates the game of life –  Maybe someday I will be so famous for some scientific idea that I will hate it.

A Rough Guide to Spotting Bad Science – a poster in pdf for download.

Lisa Goldberg explains the Monty Hall problem. I try using the “millions of doors” example when explaining the solution with mixed success, but the video does a better job than I do. I think another problem is that  Let’s Make a Deal stopped airing in 1976.  I, in my mid-30s, barely remember it (in syndication) from my childhood. I think I probably don’t do as good a job explaining the set-up to people unfamiliar with the show as they do in the video.

Best Books Read in 2013

This is round-up of the best books I read (or listened to) in 2013. They are ordered, roughly, from less-to-more academic.

Warning: In theory, if you click a link and buy a book, I get some sort of kickback. But please don’t buy a book just to see if I get a kickback.

Trial of the Clone: An Interactive Adventure! by Zach Wienersmith. In December 2012 my partner and I had placed two suspiciously similar packages for each other under the tree next to the couch. We opened them simultaneously and discovered that both gifts were an interactive adventure by the husband of parasite ecologist Kelly Wienersmith. We spent an entire  plane ride to Omaha, and much of the next day, working through our own adventures. I was an engineer. Emily was a fighter.  Both of us died.  Many. Times.

Back when I was about eight, my younger sister and I enrolled in a summer reading program at the library where we could earn a sticker for every five books we read. So we blew through 50-or-so Choose Your Own Adventure books in a week and earned sooo many stickers. So the library changed its policy to explicitly exclude CYOA books from the program. This is not that kind of book. Zach built in some role-playing-game style stochasticity and (unless you cheat) you are also likely to die. A lot.  But it is worth it. Zach is the cartoonist behind SMBC comics, and if you enjoy its nerdy and irreverent sense of humor, you are likely to enjoy this book. (If not, you might still like it, but probably not). Also, if you use technology, there is an Android app of the book for your phone (voiced by Will “Wesley Crusher” Wheaton).  And there appears to be a sequel in the works.

Nudge: Improving Decisions About Health, Wealth, and Happiness by Richard H. Thaler and Cass R. Sunstein. Thaler and Sunstein advance “libertarian paternalism.” The basic idea is that people should have many choices, but, using tricks of psychology, they can be “nudged” to make the “right” choices. The “right” choice is defined by what is good for the person or good for society. The hope is that what is good for the person or good for the societies is somewhat correlated with the person doing the nudging so these powers are not used for evil. In the first part of the book, the authors go through a series of cognitive biases that people Western college students seem to have. One, the status quo bias, is where, when people are given a choice, they tend to go with the status quo option. For example, in countries where people have to check a box to be an organ donor, organ donation rates are much lower than in countries where people have to check a box to opt-out of being an organ donor. If a policy-maker thinks that organ donation is the “right” choice then they might want to make it an opt-out option. There are many examples like this.

The second part of the book delved into specific policy perscriptions, but the authors sort-of lost me here. In one example, they advocate for privatizing marraige (i.e., getting the government out of it completely). While this may be a good policy idea (it certainly is libertarian). It seemed only loosely related to the cognitive biases described in the first half of the book. There are many examples like this. I highly recommend the first part of the book, but if you decided to skip the second part, I wouldn’t blame you.

Thinking, Fast and Slow by Daniel Kahneman. This book hit the mark. I would recommend it to anyone interested in decision science. It is similar to Nudge in some of the ground it covers (the authors feature prominently in each other’s books), but spends more time on the science and less on the policy prescriptions. The book mainly discusses the series of experiments Kahneman and his late collaborator, Amos Tversky, conducted into the biases humans use in assessing risk and decision-making – experiments for which Kahneman won the economics version of the Noble prize. Here are a couple examples from their research.

Here is another. Take these two questions:

Would you accept a bet that offers a 90% chance to lose $5 and a 10% chance to win $95? 

Would you pay $5 to participate in a lottery that offers a 10% chance to win $100?

It turns out that many more people will take the second offer and than the first – even though the outcomes are the same for both. This is a problem for utility theory – a theory of decision-making often used by economists. Utility theory assumes agents have preferences over outcomes, independent of their previous state. Kahneman and Tversky’s “prospect theory”  describes how actual humans make decisions in reference to earlier states. That is, we make decisions, in part, on how much we stand to gain or lose relative to where we are now – our current state matters. This is one of those findings that sounds like “common sense” to most people, but has made people rethink the common simplifying assumptions of an entire discipline.

I heard a lot about these last two books at the Summer Institute on Bounded Rationality a few years ago. The main gripe against Nudge from some of the participants is that paternal libertarianism seemed too, well, “paternal.” The main gripe with Thinking, Fast and Slow, was that Kahneman seemed too focused on the negative aspects of decision-making biases and heuristics and not enough on the positive aspects. The argument was that most of the time these biases serve us well, which is why we use them. Kahneman, the argument goes, focuses too much on where they lead us astray missing the forest for the trees. If you are interested in this different perspective Gerd Gigerenzer’s  Simple Heuristics that Make Us Smart is a pretty good start.

The Sports Gene: Inside the Science of Extraordinary Athletic Performance by David Epstein.  I expected to really dislike this book.  I expected to be grumpy about 90% of the time when reading it listening to it in my car. However, despite its terrible title, I thought it was a fairly balanced look about the possible contributions of genes, training, and culture to sports performance. I was only grumpy about roughly 15% of the time. Overall, I think the book could better highlight the distinction between explaining variation in athletic performance at the individual level (e.g., two athletes undergoing the same training with different results) and variation in performance between groups (why Jamaica seems to have so many good sprinters relative to its population).  And the author is much quicker than I’d like to suggest that unexplained variation at both scales is due to genetic variation. But he does point out, for example, that sprinters have a lot of prestige in Jamaica and that someone as large and fast as Usain Bolt in the United States would likely have gone into American football and be earning a lot more money (or be sidelined with injuries). So overall a mixed bag, but much better than I expected.

Why Nations Fail: The Origins of Power, Prosperity, and Poverty by Daron Acemoglu and James Robinson.  The authors of this book present a series of case studies, arguing that differences in whether nations succeed or fail is due to differences in political and economics institutions.  The basic premise and arguments seem very plausible and I think they do a convincing job pointing out flaws in competing hypotheses, such as Jared Diamond’s geography-centric ideas. However, I am really interested in why there are such differences in political and economic institutions in the first place.  Is it just historical accident?  Is there something more to it? This book does not get into these questions as much as I’d like, but it was still engaging.

Brainwashed: The Seductive Appeal of Mindless Neuroscience by Sally Satel and Scott O. Lilienfeld.  This was an engaging and rather short read about reasons you might be skeptical about the types of claims that are sometimes made in neuroscience. The two main problems are these:

1) There is an over-reliance on null-hypothesis testing and multiple comparisons. There are a very large number of possible areas of the brain that are tested and some non-trivial number of these will come back positive by chance using p < 0.05 without any controls for multiple tests.  This leads to a very high false positive rate. This is the point of the (in)famous (Ig) Nobel-prize winning dead salmon study.

2) When you read something like “the scientists found that X part of the brain lit up, which makes sense because X part of the brain is associated with function Y,” you should be skeptical.  The reason is because every X part of the brain is associated with a very large number of function Y’s.  And most Y’s are associated with multiple X’s. So it is often pretty easy to find an association between X and Y that “makes sense” after the fact.

Paleofantasy: What Evolution Really Tells Us about Sex, Diet, and How We Live by Marlene Zuk. The basic story here is that the ways humans lived in the paleolithic was much more varied than most people, including a surprising number of scientists, assume. It is problematic to make arguments about modern humans from the ways that our “hunter-gatherer” ancestors lived, because hunter-gatherers likely had many different ways of living. Some of her main targets are evolutionary psychologists, who have a reputation for hypothesizing about how human’s psychology “would have evolved” in hunter-gatherers, without being too concerned about the evidence for how hunter-gatherers actually lived. Another of the book’s targets, and the one that seems to have gotten the most attention, is the Paleo diet and its derivatives. I do not have much of a dog in this fight, but her arguments seemed reasonable. However, she spent more time than I’d have liked responding to pro-Paleo diet blog comments relative to claims made by the original Paleo-diet books. Sometimes I wondered how much of this was fighting straw men.

Institutions, Institutional Change and Economic Performance by Douglass North. This is one of those books that I wish I would have read back when I got into institutions. It is a classic and everyone references it.  As I was reading, I kept giving North little air high fives. I really liked this book! I realize that most people are not going to have the same reaction to this book that I did (which is why it is so far down on the list) but if you are interested in institutions or institutional change, this is part of the foundation.

Managerial Dilemmas: The Political Economy of Hierarchy by Gary J. Miller. My main research program is currently the origins of hierarchy and this is a great overview of issues surrounding the theme hierarchy in both an institutional economics and organizational behavior framework. It connects social choice theory with principle-agent problems and public goods problems in a nice tidy package. I recommend it to anyone interested in hierarchy in moderately complex organizations. Again, I realize this stuff is not everyone’s cup of tea, but it really helped get my thoughts about these topics organized.

I guess my only issues with these last two books is that I would have liked more in the way of toy mathematical models. But, again, I recognize I am often in a small minority in this regard.

Signal or noise at the Olympics?

Last week, Slate’s Ben Blatt argued that bobsled was “the worst designed sport of all time” because, essentially, the results have too much signal and not enough noise – virtually guaranteeing that the fastest team would win.

This week, Slate’s  Josh Voorhees argues that Olympic Hockey has too much noise in the signal. The first rounds are poor predictors for who actually wins.

So is there an optimal amount of signal-to-noise-itude? Signal appeals to our sense of fairness and noise makes the games more exciting.

Some Fortnights of Links

I compiled this post a couple of weeks ago, but it seems I never actually posted it.

In Saving A Species, You Might Accidentally Doom It by Ed Yong.

A Calm Look at the Most Hyped Concept in Neuroscience – Mirror Neurons – Christian Jarret

We Really Have no Idea Why Political Attitudes Change (or Not) a guest post by Bernard Winograd at Social Evolution Forum.

It’s a scandal drug trial results are still being withheld at The Guardian.

How Many Times Does “Don’t Promote Misogyny” Need to Be Discussed? at Life as an Extreme Sport.

Silent Technical Privilege by Philip Guo at Slate.

Important new paper on impact of having women as conveners on gender ratio of speakers at Tree of Life

introducing R to a non-programmer in one hour by Alyssa Frazee.

Statistics Done Wrong: The Woefully Complete Guide by Alex Reinhart.

I Had My DNA Picture Taken, With Varying Results – Kira Peikoff in the NYT.

Entertaining (in that internet sorta way):

Facial hair trends over time

History of English in 10 Minutes

Interpreting Your Students’ Quiz Answers at Math With Bad Drawings

How to Name a Baby at Wait But Why

A Horse Walks into a Bar comic at Darwin Eats Cake.

The World’s Biggest Cities Over Time Since 4000 BC from Ian Morris’ book “Why The West Rules — For Now.” (at Business Insider)

What Scientific Idea is Ready for Retirement? Edge 2014 Question Roundup.

Edge 2014 Annual Question: What Scientific Idea is Ready for Retirement: 174 contributors! Edge, I do not really have time for all this. But I can’t help myself.

Edge really mixes it up this year. Instead of “who’s science do you like” it is “who’s science do you not like” or “please  knock around your favorite strawman.”

While this question encourages combativeness, what I like about what Edge does, is that it gets everyone’s often-hidden biases and beliefs out on the table (and in writing). These are types of beliefs that come out in conversation or at a conference presentation, but are often obscured in professional publications by careful constructed collegiality and jargon.

Many of the 174 responses have to do with the normal topics of this blog: evolution and social evolution, especially of humans.  Here are a subset of that subset . Most have to do with sussing out the relative contributions to behavior of genes, environment, culture, maternal effects, epigenetics…  I agree to these to various extents.  There are at least two that I just don’t understand, even after a few re-readings. But I include them anyway to contrast with the others.

Grouped roughly by topic:

Peter Richerson – Human Nature
Alison Gopnik – Innateness
Alun Anderson – We are Stone Age Thinkers
Patricia S. Churchland – Modules

Kiley Hamlin – Moral Blank State-ism
Laura Betzig – Culture
John Tooby – Learning and Culture

Robert Sapolsky – Heights and Lengths and Areas of Rectangles*
Stephen Pinker – Behavior = Genes + Environment
Timo Hannay – Nature Versus Nurture

Jamil Zaki – The Altruism Hierarchy
Tor Nørretranders – Altruism
Sandy Pentland – The Rational Individual

Martin Nowak – Inclusive Fitness

* – Look for a future post on this one.