Earlier this month, George Krause posted the following questions to the
PolMeth Listserv:
(1) How many basic or introductory type methods courses (e.g.,
Scope & Methods, introductory statistics and probability, and
introductory regression) does your PhD program require students
take? What is the nature of these courses in terms of the
material/topics covered? What are the objectives relating to the
basic methods courses for PhD training?
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time
Series, Game Theory) does your department offer? Are any of these
courses required of all PhD students in your graduate program? How
many "advanced" courses must be taken by PhD students for them to
count Methods as a 2nd field as well as a 3rd field? Are there any
other Methods field requirements?
This posting reports summaries of graduate methods training, that were
offered in response to these questions.
Scott Demarchi: Duke University
George: We just redid our methods curriculum -- answers below. I've just
finished our document for grad students on the new curriculum, so if you
would like a peek at it in addition to the answers let me know and I'll send
it.
Best.
Scott
(1) How many basic or introductory type methods courses (e.g., Scope &
Methods, introductory statistics and probability, and introductory
regression) does your PhD program require students take? What is the nature
of these courses in terms of the material/topics covered? What are the
objectives relating to the basic methods courses for PhD training?
We have two research design courses. I taught the first course in the
sequence last year with Bob Keohane, and this year with Steven Wilkinson (we
try to balance a quantiative type w/ a more substantive person). The focus
is epistemology (kuhn, lakatos, etc.), historiography (e..g, we read
Novick), and modern mathematical modeling. The second research design
course is focused on qualitative methods (historiography, field work, etc.).
The second course is optional; almost all grad students take the first
course (except for political phiosophy).
Our goal in both courses is to force students to do research designs that
aren't rotten. Our general belief is that many candidates fresh out of grad
school aren't all that critical about their designs.
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time Series,
Game Theory) does your department offer? Are any of these courses required
of all PhD students in your graduate program? How many "advanced" courses
must be taken by PhD students for them to count Methods as a 2nd field as
well as a 3rd field? Are there any other Methods field requirements?
We have few methods courses required of everyone. Instead of taking a
language, students can substitute two intro stats courses taught in the
dept. But that's it for requirements.
Many students do methods as a second field, fewer as a first. In the dept.,
I teach a computational course and a MLE / non-parametric course. Munger
teaches pooled-cross-sectional time series (mainly for comparativists).
Aldrich, Soskice, and Niou teach intro game theory, social choice theory,
and more advanced workshops on formal models.
For all students taking methods as a field (primary or secondary), we force
them to do some graduate work in stats, econ, or computer science. Most take
some subset of the micro sequence, or econometrics, or Bayesian courses in
our stats dept.
Best.
Scott
Damon Cann: University of Georgia
George,
A note by Chirs Zorn in The Political Methodologist summarizes some of the
issues surrounding graduate methodology sequences. It has some interesting
insights. The link is below.
http://polmeth.wustl.edu/tpm/tpm_v11_n2.pdf
(PS: The Zorn article is located on pages 16-20)
-Damon
==========================
Damon M. Cann
Assistant Professor
Dept. of Political Science
104 Baldwin Hall
University of Georgia
Athens, GA 30602-1615
[log in to unmask]
========
Chris Achen: Princeton University
Hi, George.
In our current sequence here, we teach two tracks. The "reading track"
is just a stats course and an intro formal theory course, both fairly
informal. Students can take either one or both.
The "research track" starts with a semester of math, then three semesters of
stats. The first two stats semesters are standard probability theory and
basic regression, followed by econometrics. The third semester is more
variable--MLE or Bayes or something interesting to the prof and more
advanced.
The formal theory research track also starts with the same math course.
Then we have four additional courses, the last two being political economy.
We don't require anything. To count "formal and quant" as a first or second
field, grad students need to pass an exam, weighted toward one side or the
other. The third field can be done with three courses chosen from stats and
formal. (The math course doesn't count.)
Chris
Jeff Gill: UC-Davis
George:
Anyway, your questions.... We just revamped our methods requirements for the
Ph.D. students, so this is very recent.
(1) How many basic or introductory type methods courses (e.g.,
Scope & Methods, introductory statistics and probability, and
introductory regression) does your PhD program require students
take? What is the nature of these courses in terms of the
material/topics covered? What are the objectives relating to the
basic methods courses for PhD training?
(1) Required:
I. Scope and Methods (i.e. a KKV type course)
II. Data Analysis (manipulating data in R)
III. The Linear Model
IV. MLE/GLM
V. Basic Game Theory
(note: normative theorists are exempted from MLE/GLM).
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time
Series, Game Theory) does your department offer? Are any of these
courses required of all PhD students in your graduate program?
How many "advanced" courses must be taken by PhD students for them
to count Methods as a 2nd field as well as a 3rd field? Are there
any other Methods field requirements?
(2) Advanced courses we've offered
I. Statistical Computing for Social Scientists
II. Bayesian Methods
III. Advanced Game Theory
IV. Qualitative Methods
V. Social Networks
A first or second field methods student needs three courses
beyond the core. These can be the above or taken "elsewhere".
The "elsewhere" has recently been:
I. Durations Models in the Soc. Dept.
II. TSCS at Berkeley (Henry Brady)
III. The Microeconomic series (3 quarters)
IV. Qualitative Choice Models in the Econ. Dept.
A first field methods student is expected to write a methods
dissertation. Since we're just now developing our first
students of this type, its unclear what that means for the
moment.
I hope this helps. Cheers.
-Jeff
Christian Grose: Vanderbilt University
George:
Hi there. We're working on some changes to our own methods curriculum here
at Vanderbilt too. Right now, I think three courses are required of all
students (I'm new so I'm still learning). These may change though as they
are revisited by some of the new folks that have been hired over time (we
are currently working on getting the methods field together).
One course on scope/method/research design that leads into probability and
statistics; a second course picking up where that one left off dealing w/
multivariate statistics, regression, logit, probit, etc.; and the third
dealing with event count, non-linear models, time-series, etc.)
Then there is an additional, non-required course, which I am teaching which
will deal with Bayesian methods, empirical tests of game theoretic and
formal models, and other advanced topics).
Hope this helps--and again this is what I think we do, though things might
change. We have 7 new faculty in the last 2 years, so the graduate
curriculum may be revised given the change.
Christian
John Jackson: University of Michigan
George,
I will try to answer your questions for Michigan, though our structure is
different than yours appears to be.
First, we include modeling, research design, qualitative methods and
empirical methods in one subfield. This means that our course offerings and
requirements encompass all these topics. But, let me mostly concentrate on
the empirical side, with only small references to the others. And, I will
answer in reverse order as I think that then makes the course offerings a
little more coherent.
Students are required to complete the work in a major field, a first minor,
and a second minor in order to proceed to the dissertation. There are no
"required" courses except for ones major and minor fields. This means that
someone could escape methods, or political philosophy, all together.
The requirements for the second minor in methods is a basic grad level
course in three of the four sub-sub-fields. For the empirical work this
means a fairly rigorous course in the linear model, including matrix
algebra, generalized least squares and a brief intro to MLE. We also teach
a course previous to this that does basic statistics, the linear model
without matrix and without the formalities. The first course does NOT count
for the minor subfield.
A first minor in methods requires what is needed for the second minor plus
advanced work in one of the sub-sub-fields. In empirical methods this means
our advanced multivariate course that covers structural equation models,
MLE, the range of limited dependent variables, selection bias, multilevel
modeling, hazard models, and the like. We also teach a second advanced
course where the content depends more on who is teaching it. it could be
rigorous time series, bayesian, quasi-experiments, etc. For the methods
major the additional course work would be taken outside the dept, say in
statistics and the person is expected to write a research paper so they
demonstrate the potential for original research, not just effective
applications of current methods.
The course offerings in the empirical part then are:
PS599 -- The intro course
PS699 -- The basic linear models course
PS787 -- The multivariate course
PS789 -- The specialized course, depends on instructor PS790 -- Directed
study.
Hope this helps.
John Jackson
Beth Leech: Rutgers University
Dear George,
I include the info for Rutgers below. When you've finished compiling this,
would you mind sharing the information with me? We're trying to revise and
upgrade our methods sequence and could use some ammunition. I started to
compile this information for a few top schools last year but stopped because
schools like Yale and University of Chicago tended to have no methods
requirements whatsoever (requirements were up to the field and/or the
advisor), which didn't help my argument.
(1) How many basic or introductory type methods courses (e.g., Scope &
Methods, introductory statistics and probability, and introductory
regression) does your PhD program require students take? What is the nature
of these courses in terms of the material/topics covered? What are the
objectives relating to the basic methods courses for PhD training?
We have one year-long methods sequence that includes 10 weeks on research
design, 10 weeks on introductory statistics up to regression, and 10 weeks
on philosophy of science/epistomology. All students are required to take
this sequence. Its goal is to make sure everyone, including our many
students in political theory, have a basic understanding of the assumptions
and goals of empirical research and a rudimentary knowledge of data analysis
(we use SPSS and ANES data in the stats part of the course).
Students are required to take one other methods course, but it doesn't have
to be in quantitative methods (for instance, students in political theory
might decide to take a foreign language readings course).
Students in American politics are required to take an additional regression
course and strongly encouraged to take additional classes, attend ICPSR, or
take classes from other area schools in our consortium (Princeton, Columbia,
NYU, Stony Brook). Students in other empirical fields may or may not be
required to have additional quantitative training, depending on their
advisor. The regression course covers the topic in much more depth than the
introductory statistics class. A matrix approach is used and regression
diagnostics are a major focus. In the last few weeks of the course, a brief
introduction to MLE is given, with the focus on logit and probit. The aim
is to prepare students to conduct basic data analysis and prepare them for
more advanced work in the future.
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time Series,
Game Theory) does your department offer? Are any of these courses required
of all PhD students in your graduate program? How many "advanced" courses
must be taken by PhD students for them to count Methods as a 2nd field as
well as a 3rd field? Are there any other Methods field requirements?of all
PhD students in your graduate program? How many "advanced" courses must be
taken by PhD students for them to count Methods as a 2nd field as well as a
3rd field? Are there any other Methods field requirements?
We offer two courses in advanced methods: a game theory course and a
course on MLE (which also touches on times series methods). They are not
required. Students can't claim methods as their 2nd field -- the
department doesn't recognize methods as a distinct subfield and doesn't
offer qualifying exams in methods. The 3rd field, however, can be a
student-created field and there is no exam. Students could choose methods
as that 3rd field. The only requirements are that a student get
departmental permission and take three courses in the field.
Hope that is helpful and I look forward to seeing your results!
Beth Leech
Dave Peterson: Texas A&M University
George
We have a requirement that all students have to take four classes two
of which are specified for them. An intro research design course and an
intro regression (usually taught out of Gujarati) are required. There is an
expectation that they will take a second regression (Greene) class as well.
The fourth class is open. We offer a philosophy of science class that used
to be required but is now optional class (but given the incremental changes
it gets preferential treatment in the grad handbook). Students can
substitute basically any methods class for that fourth course.
We do not have nay other courses on the books but offer a bunch of
stuff as a variable topics course. We have (in my time here) offered
experimental methods, game theory, time series, MLE, Hierarchical modeling,
and a math/modeling course. For methods as a student's second field, they
have to take four additional courses (a total of 8). For methods as a
"third" (we call it a topical) field, they have to take three.
We are in the process of dong something similar. We just approved
methods as a field last year. Can I get the results of the informal survey?
Hope all is well,
Dave
David A. M. Peterson
Assistant Professor
Department of Political Science
Texas A&M University
(979)845-6783
Sean Gailmard: Northwestern University
Hi George,
I saw your message on POLMETH -- we have only two required courses of any
kind in the Ph.D. program here, Prob & Stat (which I am teaching
now) and Linear Models. All students in American, IR, Comparative, and of
course Methods have to take these. (Students in the Theory field do not have
any formal requirements.) We are on quarters, so that amounts to 20 weeks of
training as a bare minimum.
There is a third course that for historical reasons is called "Advanced
Methods" but really is discrete choice, IV's, panel methods, and a dose of
partial identification & causal inference. Of the students that take the
first two stats classes, probably 2/3 to 3/4 take this course (or are
planning to take it this year).
We also offer more advanced courses of varying topics depending on
instructor and student interest. These courses might get into MLE,
statistical computation, Bayesian methods, ecological inference, etc.
That's all on the stats side...we are not offering a game theory course in
the department this year but that is an aberration. We will next year offer
"Introduction to Formal Modeling in Political Science," which is a brief
review of highlights in social choice and a longer treatment of game theory.
That course would be accessible
(presumptively) to any student in the Ph.D. program. We are also launching a
joint Ph.D. program with Kellogg, which combines our stats with their more
technical offerings in decision theory, game theory, and applications in
political economy.
Sean
Matt Lebo: Stony Brook University
Hey George,
Here's the story at Stony Brook:
Incoming PhDs take the following required courses:
POL600 (first and second semester) A scope and methods type course where
each faculty member takes a turn presenting their research to the class.
The class also has readings and works -- in consultation with a
faculty member of their choice -- on a single paper that has data analysis
in it and maybe becomes a conference paper or (sometimes) even a
dissertation.
POL602 (first semester) Introduction to probability and statistics. I teach
this one. We cover the basics of probability theory, distributions,
hypothesis testing, random variables, and other topics. By the
end the students are ready for regression.
POL603 (second semester) Intro to Regression. Bahar Leventoglu teaches this
one. Using a matrix algebra approach the students work their way through
regression. Its pretty deep and doesn't get to Logit.
POL 604 (third semester) Maximum likelihood. Stanley Feldman teaches.
Begins with Logit and covers a lot of ground.
Then the students may take optional courses after their first 3 semesters
are over.
POL 606 Time Series. Me again. Taught every other spring. Begins with Box
Jenkins and covers most of the topics your class does. About 2 weeks on
PCSTS, also GARCH, ARFIMA, Cointegration and coming soon Dynamic Conditional
Correlations (Engle 2002) which I'm learning in a project I'm working on
wtih Jan Box-S.
POL 611 Game Theory -- Scott Basinger. Taught every other fall. I think
first years can take this one.
POL 610 Measurement. Stanley Feldman. Every other year. Advanced students
take this and say its great.
POL 6?? Bayesian. Hopefully we will hire someone this fall who can teach it.
We had a grad student, Jack Buckley, who taught himself and then taught a
class in it, but he is long gone.
All students are required to test in methods for their qualifying exams.
The exams generally don't test beyond the required course though a question
might pop on from an elective. So methods is a field for all. No other
requirements, though there is math camp at the beginning.
I hope thats helpful. I know SB is pretty intensive and our students
usually come here knowing that. But the required courses really set up the
students well even if they don't take any electives. Toning them down a bit
they could be squeezed into 3 or even two classes.
Let me know if I can be of any more help.
Matt
Chuck Shipan: University of Iowa
Hi, George.
(1) How many basic or introductory type methods courses (e.g., Scope &
Methods, introductory statistics and probability, and introductory
regression) does your PhD program require students take? What is the nature
of these courses in terms of the material/topics covered? What are the
objectives relating to the basic methods courses for PhD training?
First semester:
Scope and methods
Math
Intro to econometrics (e.g., basic regression)
Second semester:
Probability and statistics
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time Series,
Game Theory) does your department offer? Are any of these courses required
of all PhD students in your graduate program? How many "advanced" courses
must be taken by PhD students for them to count
Methods as a 2nd field as well as a 3rd field? Are there any other Methods
field requirements?
Offered yearly:
Game theory
Offered once every year or two:
Testing formal models in American politics Dynamic modelling in IR
Advanced methods:
Usually one per year, topics vary (e.g., MLE, time series, and special
topics have been the most recent)
Good luck!
Chuck
Nathan Kelly: University of Tennessee
Dear Professor Krause,
This is something that I have been thinking about with regard to the
Tennessee program as well. My guess is that we don't require/offer much in
the way of methods compared to others, but I would be curious about what you
learn from your query if you would be willing to share what you learn. I'll
include responses for our department below.
Required introductory methods oriented courses are: Scope and Methods,
Research Design, and Introductory Statistics II. Scope and Methods is you
basic intro to the discipline, approaches, and philosophy of science.
Research design is run as a sort of workshop providing students the
opportunity to get their feet wet in research by writing a design. The stats
course introduces students to regression (multivariate), some limited
dependent variable methods, and some time series methods. We also offer an
even lower level intro to stats that covers probability theory, what is a
variable, and univariate statistics. It isn't required, though it is de
facto required since it is a prerequisite for the second stat course. So,
unless students have this knowledge coming in and can get the prereq.
waived, they have to take it.
We offer no advanced methods courses of any kind. We rely on other
departments for this. Some of us are trying to put together some advanced
stat seminars. This is where it would be useful for me to have the same
information that you are after. Again, I would be appreciative if you would
be willing to share what you learn with me after you receive more responses.
Best,
Nate
--
*************************************************
Nathan J. Kelly, Ph.D.
Assistant Professor
Department of Political Science
University of Tennessee
*************************************************
Stephen Shellman: University of Georgia
Hi George.
I am at Georgia and we require 3 courses - scope and methods, baby stats,
and regression (with some logit/probit).
I prefer the Florida State model where I did my grad work.
4 courses:
I. Scope & Methods
II. Quantitative Analysis(Baby Stats - paired sample t-tests,
chi-square/contingency tables, ANOVA, basic multivariate regression) III.
Regression (with some simulataneous equation stuff & structural equation
modeling) IV. Advanced Quantitative Analysis (MLE/Time Series)
To get a minor, I believe you have to take 3 additional courses - one of
which may be at ICPSR. One of which may be "Game Theory"
They also offer advanced courses like: Time Series Analysis
So to get my Methods Minor, I took the 4-course sequence (required by all
Ph.D. Students), Game Theory, Simultaneous Equations at ICPSR, and Time
Series in my Dept.
They also allowed you to take advanced courses in the Econ or Stats depts
for credit if approved.
Hope this helps.....
Steve
Stephen M. Shellman
Director, Project Civil Strife
Director, Summer Workshop On Teaching about Terrorism
Assistant Professor
Department of International Affairs
University of Georgia
[log in to unmask]
http://www.arches.uga.edu/~smshel
Ole Forsberg (University of Tennessee)
George,
The University of Tennessee has a Scope and Methods (SAM) course and a
Research Methods (RM) course in addition to two basic quantitative methods
courses. The requirement for PhD students is currently SAM, RM, and the
first quant course. Starting in 2006, the PhD requirement will be SAM, RM,
and the second quant course. This is our first step to increasing the number
of offerings in quantitative methodology. We used to offer methods as a PhD
field, but it was dropped about a decade ago... unfortunately for me.
The SAM course covers the philosophy of science, epistemology, and different
different manners of studying political science. It was(is) taught by an old
behavioralist from Michigan State, which suited me with my natural science
background fine. The RM course teaches us how to actually write a
publishable piece and what the individual components of a research paper
are.
We have no 'advanced' quantitative methods courses yet. There is simply a
two-semester sequence covering up to, and including, time series analysis
and a little binary dependent variable regression (only logit). There is no
discussion of MLE in any of it. To go further, I took an independent study
and attend classes at ICPSR. Oh, and I also read.
I hope this helps you. If you have other questions, let me know.
Ole
Kosuke Imai (Princeton University)
Hi George,
You can take a look at
http://www.princeton.edu/politics/research/fields/formal/ for what we are
doing for graduate courses in formal and quant courses at Princeton. You can
also have a look at my webpage for syllabus of some of those courses.
http://imai.princeton.edu/teaching/index.html
Good luck,
Kosuke
-----------------------------------------------------
Kosuke Imai Office: Corwin Hall 041
Assistant Professor Phone: 609-258-6601
Department of Politics eFax: 973-556-1929
Princeton University Email: [log in to unmask]
Princeton, NJ 08544-1012 http://imai.princeton.edu
-----------------------------------------------------
Bob Bernstein (Auburn University)
We have minimal requirements at Auburn, where our Ph.D. is in Public Admin
and Public Policy. We have 2 required methods courses--one focusing on
quantitative, one on qualitative. We want students to be able to design
analyses of complex causation, but that mostly goes through regression (inc.
logistic and GLM) and path analysis, with a heavy focus on measurement.
More advanced topics are introduced so students know they exist and have
some idea of advantages and disadvantages. But if they want to take them,
they typically do them in other depts or via directed studies. They need 3
additional courses for a minor in methods.
Bob Bernstein
B. Dan Wood (Texas A&M University)
See below.
B. Dan Wood
(1) How many basic or introductory type methods courses (e.g., Scope &
Methods, introductory statistics and probability, and introductory
regression) does your PhD program require students take? What is the nature
of these courses in terms of the material/topics covered? What are the
objectives relating to the basic methods courses for PhD training?
Most students are required to take 4 methods courses to receive a PhD. These
are 1) Scope and methods, 2) Intro Stats, 3 Advanced Stats, and 4) At least
one advanced methods such as MLE, Time Series, theory of science, etc.
(2) How many advanced methods courses (e.g., MLE, Bayseian, Time
Series,Game Theory) does your department offer? Are any of these courses
required of all PhD students in your graduate program? How many "advanced"
courses must be taken by PhD students for them to count Methods as a 2nd
field as well as a 3rd field? Are there any other Methods field
requirements?
We have advanced classes in Time Series, Maximum Likelihood, HLM,
Experimentation, Game Theory, and Mathematical Modeling. We are considering
adding a Bayesian course and probably will by next year.
Students who take methods as a minor are required to take 4 methods courses
beyond the basic requirements listed above in 1).
Renee Johnson (University of Florida)
Hi George-
This is response to your question about graduate methods course offerings on
the methods listserve.
The University of Florida has developed a four-course core that is required
of all of our Ph.D. students irrespective of field. The titles of theses
four courses are:
POS 6736 Conduct of Inquiry
POS 6716 The Scope and Epistemologies of Political Science POS 6737
Political Data Analysis POT 6502 Politics and Theory
I think you would consider two of these courses to be part of a core methods
training.
"Conduct of Inquiry" is designed to introduce students to the different
types of research approaches that political scientists use in answering
questions of interest to them. Students in this class are exposed to
various works in political science that use both quantitative and
qualitative approaches to conducting empirical research including survey
research, participant observation, archival research, focus groups, etc.
Students must also actively engage in some of these efforts during the
course of the semester. Reading materials in this course include KKV and a
general research design textbook in addition to supplemental texts and
articles demonstrating the use of various different research strategies in
political science.
"Political Data Analysis" is an introduction to the use of statistical
techniques in political science. It begins with a discussion of levels of
measurement of data and continues with probability theory, basic statistical
tests and ends with a basic introduction to mulitivariate regression.
Students in this course complete a data-based project that requires the use
of techniques they learned in the class as well as two examinations and a
selection of homeworks.
These are the only methods courses required of all of our Ph.D. students.
We offer a major field in methodology, a minor field in methodology, and a
non-testing minor field in methodology. Students preparing for a major
field in methodology (in addition to the core requirements listed
above) must complete an intermediate statistics course (linear models) and a
maximum likelihood course in addition to four other areas of specialization.
Students preparing for a minor field in methodology must complete the
intermediate statistics and maximum likelihood course in addition to two
other areas of specialization. Students studying methodology as a
non-testing field must complete intermediate statistics, maximum likelihood,
and two other areas of specialization.
Students may take courses from our statistics department (or other
departments) with the approval of the methodology committee.
A selection of Advanced Courses that we currently offer in the department
include maximum likelihood, formal theory, fuzzy logic, and survey research.
Courses that students frequently take as advanced topics in the statistics
department include categorical data analysis and time series. We have just
recently hired two new methodologists so our internal offerings are likely
to expand in the coming years.
I hope this is helpful,
Renee
Laura Langbein (American University)
George,
Hope this answers your questions.
1) We require 3 courses: epistemology, stats, applied econometrics.
2) We have one advanced course in MLE. Program evaluation is arguably
another "advanced" course; it merges the applied econometrics with research
design, and gives students lots of information about how to properly use and
interpret econometrics in the "real" world.
3) Methods is an (optional) 3rd field....it requires 5 courses (2, in
addition to the required 3). Students select 2 from program evaluation;
qualitiative methods; or the MLE course. It is probably the modal choice of
a 3rd field. (All students take 3 fields; the 3rd field is a
"non-comprehensive" field, certified by courses and in the dissertation
proposal defense.)
4) We expect that all PhD grads from our program can teach undergrad
research methods, and can comfortably use standard quantitative methods in
their research.
Irfan Noorudin (The Ohio State University)
Dear George,
Here are the answers with respect to Ohio State, though if any of my senior
colleages do respond to you and their answers differ from mine, I'd go with
their interpretation of our rules and regs.
> (1) How many basic or introductory type methods courses (e.g.,
> Scope &
> Methods, introductory statistics and probability, and introductory
> regression) does your PhD program require students take? What is the
> nature of these courses in terms of the material/topics covered? What
> are the objectives relating to the basic methods courses for PhD
> training?
We have recently changed our requirements. All of these are spelled out in
gory detail in our graduate handbook which is on-line here:
http://psweb.sbs.ohio-state.edu/gradstudy/gradhandbook.pdf.
All our students are required to take the equivalent of 3 courses (recall we
are on the quarter system so this is a year's worth of courses) in "research
methods" including a course all of them must take on "Research Design." This
is typically taken in the first quarter of their first year. And it covers
the typical scope and methods territory of such courses, with some
discussion of philosophy of science, debates over qualitative and
quantitative and interpretivist methods etc.
After that course students have more flexibility as to how they fulfill the
remainder of the requirement, and this varies by sub-field. All the
Americanists must take the first two courses in statistical methods. These
cover probability and distribution theory, and regression analysis
respectively. The latter aims to get them through OLS, its assumptions and
the consequences of their violations, and a brief introduction to MLE for
binary dependent variables. These courses are taught in the first two
quarters of the year, so that in the third quarter students can "complete"
the stats sequence by taking an elective course on MLE. This also means that
most first-year students take both Research Design and Intro to Prob/Dist
Theory in their first quarter in addition to their field pro-seminar. My
sense is that most students who take the first two courses take the third
MLE course, especially in American politics.
Our comparative and IR students have more flexibility. My sense here is that
most are still taking the stats sequence described above but that, as demand
rises and courses are created, more will start to take real sequences in
qualitative methods.
For the stats courses, we haven't really had a discussion since I've been
here of the objectives of these courses with respect to the PhD program in
general. But I think it's fair to say that they are general service courses,
and not "weed out" courses. To enable most of our students to survive and
learn in them, we now also have a three week "Math boot camp" that runs from
right after APSA till the start of the fall quarter in the third week of
September. This isn't required, but I think the norm is that everyone ends
up taking it.
> (2) How many advanced methods courses (e.g., MLE, Bayseian,
> Time Series,
> Game Theory) does your department offer? Are any of these courses
> requiredof all PhD students in your graduate program? How many
> "advanced" courses must be taken by PhD students for them to count
> Methods as a 2nd field as well as a 3rd field? Are there any other
> Methods field requirements?
We offer quite a few advanced methods courses. There is now a full sequence
in formal theory, though none of these are required for all students, though
one is for those declaring Political Economy as a second field. On the stats
side, our access to the ITV program, of which Wisconsin, Illinois, and
Minnesota, are also a part, allows us to offer a wide range of advanced
methods courses. A listing of methods courses taught under the ITV auspices
is available here: http://psweb.sbs.ohio-state.edu/prism/methodscourses.pdf.
For more info on this program, you should contact Jan Box-Steffensmeier.
If students declare methods as a second field, a student must take 3 courses
beyond MLE. So, in theory, a student would take the probability/distribution
theory, regression, MLE, and 3 more courses. Methods can be declared a
second major (it can't be the only major field), in which case students must
take 6 courses beyond MLE.
I hope this helps answer your questions. FYI, archived syllabi for our
methods courses (though not totally up-to-date) are available here:
http://psweb.sbs.ohio-state.edu/courses/syllabus%20project/method.htm
Please let me know if you have other questions.
Cheers, Irfan
Mark Hallerberg (Emory University)
George,
To answer your questions, Emory has a required two-semester sequence. There
is also a research design course that I am almost sure is required that
Micheal Giles regularly staffs. So, we seem to require only one additional
methods course of all students than Pitt does (although i should add that
students deal with topics like endogeneity and intro time series in the
second semester stats, which is further along than Jon usually goes in his
course). We offer time series (which I am doing now) and MLE in alternate
years at the more advanced levels. On the game theory side, there are
presently two courses. We hope after this year's hires that we can regularly
offer a third course at the graduate level. Finally, we also have a
qualitative methods course. I was at first skeptical (what are students
getting beyond a standard research design course?), but students seem to
like it and the quality of some qualitative dissertation proposals is
somewhat better. if one has to do a case study, at least do it as well as
possible.
In terms of the methods exam, I am doing them for the first time this fall.
Students must take two of three subfields if they do Methods as a
field--quantitative, formal, or qualitative. My experience is that students
taking the quant side have had the four stats courses in the department and
at least one summer in Michigan. I presume that the qualitative types have
much less, although to be fair two of the four stats courses are required of
everyone and the qualitative types now have their own "Michigan" they attend
(at ASU), so the differences may not be as great as I presume.
I hope this helps!
Mark
George A. Krause
Professor
Department of Political Science
University of Pittsburgh
4442 Wesley W. Posvar Hall
230 South Bouquet Street
Pittsburgh, PA 15260 [USA]
412.648.7278 (Office Phone)
412.648.7277 (Department Fax)
412.648.7250 (Department Phone)
[log in to unmask] (E-Mail Address)
**********************************************************
Political Methodology E-Mail List
Editor: Karen Long Jusko <[log in to unmask]>
**********************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
http://polmeth.wustl.edu/polmeth.php
**********************************************************
|
