I develop statistical methods to study evolution and extinction in the fossil record.
Mass extinctions and the Signor-Lipps effect. When a collection of taxa (e.g., species of dinosaurs or ammonites) are killed off simultaneously in a mass extinction, their last occurrences in the fossil record may nonetheless give the impression of a gradual extinction, due to the incompleteness of the fossil record (the Signor-Lipps effect). I develop methods for testing whether a mass (sudden) extincton has occurred, and for estimating the position and duration of such an extinction. I also study the related topic of placing confidence intervals on stratigraphic ranges, i.e., estimating the time of extinction of a taxon based on its fossil record.
Evolutionary trends. Large-scale trends in the history of life (such as increase in
body size, a.k.a. "Cope's Rule") may be a result of driven
mechanisms (e.g., selection) or passive mechanisms (e.g., diffusion away
from a lower bound). Several tests have been developed to categorize
a trend as either driven or passive, but real trends are
likely to be a combination of both types.
I have developed a new method, the Analysis of Skewness, to quantify the extent to which a trend
is a combination of driven and passive trends.
With Jon Payne of Stanford and members of his lab,
I also study trends in animal body sizes over time; see our recent paper in
Macroevolution. Using statistical analysis and paleontological databases, I study large-scale questions about the history of life. Sample questions include: (1) Are mass extinctions merely larger manifestations of processes responsible for background extinction, or are they a fundamentally different phenomenon?
(2) How many kinds of dinosaurs ever lived, and were dinosaurs already declining before an impact at the end of the Cretaceous Period caused their ultimate extinction? (3) Are observed extinction rates in mass extinctions a reliable guide to the severity of the triggering mechanism? That is, does a higher extinction rate indicate a more severe cause, or can factors such as Simpson's paradox make such an inference misleading?
The tempo of the Cambrian explosion.
The Cambrian explosion was a key event in the history of life, in which representatives of most phyla (animal body plans) first appear in the fossil record. With
Susannah Porter and
John Moore and
of UC Santa Barbara and
Adam Maloof of Princeton,
I am working on developing a statistically rigorous timeline of Early Cambrian originations, accounting for a number of possible biases such as dating error and incomplete fossil sampling. This problem is analogous to the Signor-Lipps effect in mass extinctions, but in the reverse direction.
Dynamics of extinction in food webs.
The end-Permian mass extinction was the most severe in the history of life, but its causes remain enigmatic.
Peter Roopnarine of the California Academy of Sciences and
Ken Angielczyk of the Field Museum,
I study the dynamics of extinction in terrestrial ecosystems during the Permian.
Using models of Permian food webs, we explore how the extinction of primary producers (e.g. green plants) could have led to the collapse of terrestrial ecosystems. More generally, we study factors leading to stability or instability in ecosystems.
Lab members and collaborators
Other areas of interest:
Statistical graphics and visualization.
I am interested in how to effectively visualize high-dimensional datasets
using graphical methods such as parallel axis plots and data image plots, and more generally
in principles of effective graphical representation and communication.
Statistical methods in baseball research,
also known as sabermetrics.
I am interested in the analysis of
managerial strategies (featured in the
New York Times and the
Philadelphia Inquirer; also see
for additional media coverage).
My MS thesis was on a Markov Chain model
for baseball lineups and player evaluation, with advisor
Here is a
graphic I did with RIch Wicentowski
of the Swarthmore Computer Science department on the history of the Yankees,
interview on the growing use of statistical analysis in baseball.
Opportunities for students:
I am planning to have
two openings for undergraduate student researchers for each of the next two upcoming summers (2016-2017). These are paid positions funded by an NSF grant.
I am also open to supervising projects on statistical paleontology during the school year for course credit (independent study/directed reading).
Please email me if you are interested. Due to funding restrictions, eligibility is usually limited to Swarthmore students.
The preferred prerequisites are
Stat 61 and
CS 21 (or equivalent), but some combination of Stat 11, Stat 31, Bio 2, or other experience may suffice as well.
No paleontology background is necessary.
Learn more about projects and publications by previous students.
Courses at Swarthmore College
Stat 1: Statistical Thinking
Stat 11: Statistical Methods I
Stat 21: Statistical Methods II (new course for Fall 2016)
Stat 31: Data Analysis and Visualization
old Stat 21: Quantitative Paleobiology (note: not the same as the new Stat 21)
Stat 61: Probability and Mathematical Statistics I
Stat 111: Probability and Mathematical Statistics II seminar
View information on my
previous courses and course evaluations at Stanford, Harvard, Williams, and The University of Chicago.
I was awarded the 2008
Waller Education Award from the
American Statistical Association
for innovation in the instruction of elementary statistics.
video of my talk at
Requesting a letter of recommendation from me? Please read
this page first.
Professor of Statistics, 2016-; Associate Professor 2008-2016; Assistant Professor 2002-2008:
Mathematics and Statistics,
Blaustein Visiting Associate Professor, 2009-2010:
Geological and Environmental Sciences,
Department of Statistics,
Visiting Assistant Professor, 1998-1999:
Department of Mathematics,
The University of Chicago,
Department of Statistics,
Data Desk statistical software,
Data Description, Inc.,
The University of Chicago,
Department of Statistics,
Statistics and Biometry,
Trivia: What do Stanford, Swarthmore, Harvard, Chicago, and Cornell have in common? (Hint: think team nicknames.)
I participated in academic quiz competitions while
I was a student at Cornell and Chicago.
I captained the University of Chicago
team, the NAQT 1997 International
Champions. I also played for the winning team
(a three-man Cornell team) at the first ever
Questions on the Crum tournament at Swarthmore in 1992.
I was a contestant on
Jeopardy! in 1996.
I was an alternate on
Who Wants to be a Millionaire in 1999 but didn't appear on the show.
I did get to meet Regis and John Carpenter (the first million-dollar
winner, who was on my taping). Look for me in the audience!
One of my hobbies is music composition and songwriting. Some musicians/composers I like are
(read an interview I did with Sharon Corr:
Some representative pieces of mine include a
piece for piano and mallet percussion, a
piece for rock band, harp, and cello inspired by Copland, and a
piece inspired by the 2012 Newtown school shootings.
New: My Sextet for Percussion (Part 3) was performed at Swarthmore
in May 2016 (photo at left). The performers from L to R are Madeline Conca '17 (marimba),
Andrew Hauze '04 (piano), Rebecca Norling-Ruggles '19 (vibraphone), Deborah How '89 (piano), Ben Schreiber '16 (xylophone),
and Helen Thum '16 (xylophone).
Spring 2016 Office Hours
Mondays and Wednesdays, 3:30-5:00
or by appointment, or just drop in any time my door is open
Science Center 159
s c w a n g [at] s w a r t h m o r e [dot] e d u
Department of Mathematics and Statistics
500 College Ave
voice: (610) 690-5769
fax: (610) 690-6854
Last update: February 20, 2016