From a marketing perspective, R is increasingly popular in a lot of the public opinion type/consulting jobs in the DC region. For more general data scientist/programmer jobs, I've heard that Python is a really good investment, but you'd definitely want R as well for statistical work.
Most government agencies are still SAS shops, and SAS training is likely to pay off for folks looking for government or contractor work at the federal or state level. Though we also use Stata for modeling and R for various tasks, SAS is still our go-to program for data manipulation. We also use SAS-callable SUDAAN for dealing with data from highly complex sample designs (as with many government establishment surveys), but that's something to pick up after you're comfortable with SAS.
Anna Maria
Anna Maria Ortiz, Ph.D.
Assistant Director, Statistics
Applied Research and Methods
US Government Accountability Office
202-512-2788
-----Original Message-----
From: Political Methodology Society [mailto:[log in to unmask]] On Behalf Of Philip A. Schrodt
Sent: Wednesday, June 06, 2012 1:14 PM
To: [log in to unmask]
Subject: Re: [POLMETH] Software training for non-academic jobs
I would second Stephen Haptonstahl's recommendation of the R and Python
combination. While there is certainly going to be continued idiosyncratic
demand for the various proprietary systems -- one of my students recently did
some U.S. government work where everything, including the statistical
estimation, had to be in Matlab -- no one alternative is going to stand out the
way R does, both in terms of capabilities and, of course, price. R also now has
quite a bit of cross-over into the machine-learning (a.k.a. "data mining")
fields, so it is becoming a general data processing system, not just classical
statistics.
It can be *really* handy to know a general purpose computer language for
situations where there is no obvious general solution, such as extracting data
from downloaded web pages. Python can usually get things done in a relatively
small number of lines of code, and seems to be gradually displacing perl and
Java for these sorts of problems.
On Wed, Jun 6, 2012 11:43 AM Stephen Haptonstahl <[log in to unmask]> wrote:
>
>Having just recently returned to the non-academic working world as a
>Data Scientist I can tell you that I see some explicit call for SPSS,
>some for STATA, some for SAS, a little more for Matlab, and a LOT for R
>and Python. The impression I get is that demand for S* is decreasing
>slowly while demand for R and Python is increasing.
>
>The job descriptions I've seen are appropriate for more technical
>political scientists, statisticians, computer scientists, and the like.
>I don't know what is being used in jobs a polisci undergrad might pursue.
>
>Regardless of the particular software, I'm sure having some analytical
>skills puts almost any job seeker at some advantage.
>
>Best,
>Steve
>
>Stephen Haptonstahl
>530-712-3946
>
>On 6/6/2012 7:46 AM, Eric Prier wrote:
>> Colleagues,
>> What an interesting discussion thread, and it has made me reconsider my use
of the traditional Babbie Methods book along with my own SPSS
assignments/exercises. However, I am most curious about questions relating to
employability skills and software - is there a sense in this community of the
software (SPSS, Stata, R, etc.) that parallels or is most adaptable to
obtainable jobs for most political science students? Perhaps the software is
different for varying levels (BA/BS, MS/MA, and PhD)?
>>
>> Fortunately, I have not been on the nonacademic job market in a while, but I
am curious if someone has done a recent study of job skills and/or workforce
software compatibility for political science majors? Maybe someone is currently
working in the private sector and has some insight into this question. Thanks
for considering my question...
>>
>> Best,
>> EW
>> DR. ERIC PRIER
>> Associate Professor of Political Science
>> Dorothy F. Schmidt College of Arts and Letter
>> Florida Atlantic University
>> 3200 College Ave
>> Davie, FL 33314-7714
>> tel: 954-236-1340
>> fax: 954-236-1150
>> [log in to unmask]
>> www.fau.edu
>>
>>
>> -----Original Message-----
>> From: Political Methodology Society [mailto:[log in to unmask]] On
Behalf Of Paul Gronke
>> Sent: Tuesday, June 05, 2012 1:06 PM
>> To: [log in to unmask]
>> Subject: [POLMETH] Summary: Pollack book and other responses
>>
>> Dear Colleagues,
>>
>> As always, the PolMeth community has been extremely responsive and helpful
regarding my recent query about a statistics and research design text and a
programming companion volume. At the request of a number of readers, I have
summarized the responses below.
>>
>> ======
>>
>> The Course and the Students:
>> ========================
>>
>> There is a tremendous diversity among colleges and universities, and within
comparably situated institutions, a diversity in the ways that political
science and social sciences are structured. Rather than try to assemble some
sort of comprehensive list, I have provided below a brief description of the
kinds of courses and kinds of students that colleagues described:
>>
>> 1) Math-phobic students: These may include
>>
>> - Students for whom this course may be the one and only statistics and
design course they take.
>> - Older students in community colleges or schools with a mission for adult
or non-traditional students.
>> - Students in non-quantitatively oriented public administration programs
>>
>> 2) Students who have a higher capacity for mathematics, or for whom the
course is as much a math or quantitative reasoning as it is a design course.
>>
>> 3) Students whose curriculum requires a statistics course AND a research
design course
>>
>> 4) Students in colleges which require student theses, capstone projects, or
independent research projects
>>
>> Comments:
>>
>> Groups (2) and (3) seem to have it easiest. These colleagues as a group
generally opted for more challenging texts such as Agresti, Gujuarti, Levin and
Fox. The courses are usually proof-based with lighter coverage of design
issues.
>>
>> Group (1) probably has it hardest, since some mentioned that even the
Pollack is too challenging for some of their students. My read of the emails
is that this is probably the least well-served group (any aspiring writers take
note!). One mentioned a book by Healey as "almost exclusively verbal."
>>
>> Group (4) is distinguished by those schools that are also in group (3)
(design + stats) and those that are not. For the latter, most expressed
frustration at balancing design, data management, and stats in a single course.
>>
>> The Books:
>> =========
>>
>> These comments focus on Pollack of necessity since that is what I asked
about, but there is a clear competitor--Kellstedt and Whitten--and the two
books appear to be complementary, overlapping in some respects but appealing to
different markets.
>>
>> Pollack: Received generally positive remarks, except for those who found
it *too* simple.
>>
>> Pluses: Intuitive explanations, good for math phobic students: "nice,
intuitive explanations of regression and statistical significance that really
get through to students that are intimidated by the math". Good as a
supplement because of the design section: "I have used it without the stata
companion volume just for my elections course - exactly because it is good
reading for them on social science methods."
>>
>> Minuses: Too easy "The one problem is that it makes things seem too easy."
A number of people thought it had a misleading or flat our incorrect discussion
of t-tests: " I'm not that thrilled with its treatment of tailedness of tests,"
"I found the discussion of foundational statistical concepts in later chapters
to be too loose, or even misleading, in crucial places."
>>
>> Kellstedt and Whitten: The clear competitor among the email respondents.
>>
>> Pluses: Good coverage of design issues, and links design to statistics:
"The first about half focuses on research design and the second introduces
stats as a way of fulfilling part of the research design strategy laid out
earlier in the text. " Better discussion of foundational inferential issues: "
I will likely use other books, such as Kellstedt& Whitten, to accompany the
discussion of more foundational, non-statistical issues."
>>
>> Minuses: None mentioned, but I did not ask for a review of this book.
>>
>>
>> The Stata Companion
>> =================
>>
>> This volume received a huge thumbs up from almost every respondent.
>>
>> Most lauded it's lucid approach, the number of problems at the end of each
chapter. One respondent said that if you require students to fill out the end
of each chapter, they will end up being very adept at Stata: "They have the
book. I give them no instructions other than a five minute "Here's the
program, you click open file here, and then you start." Students are required
to complete all of the assignments at the end of the chapters. I tell them
that by the end it seems like busywork, but one of the goals is to have them
work on the stats package enough that it seems like busywork. By the end, they
know their way around the programs. "
>>
>> Many said they successfully use the supplement without the main text. A
number noted that they liked the SPSS guide as well.
>>
>> One respondent noted that there is a shorter Stata PDF guide available on
Cambridge's "resources" page for Kellstedt and Whitten.
>>
>> THE most common request was the Pollack or someone else write a similarly
designed supplementary text for R.
>>
>>
>> ---
>> Paul Gronke Ph: 503-517-7393
>> Fax: 503-661-0601
>>
>> Professor, Reed College
>> 3203 SE Woodstock Blvd
>> Portland OR 97202
>>
>> **********************************************************
>> Political Methodology E-Mail List
>> Editors: Diana O'Brien<[log in to unmask]>
>> Jon C. Rogowski<[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
>>
>> **********************************************************
>>
>> **********************************************************
>> Political Methodology E-Mail List
>> Editors: Diana O'Brien<[log in to unmask]>
>> Jon C. Rogowski<[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
>>
>> **********************************************************
>
>**********************************************************
> Political Methodology E-Mail List
> Editors: Diana O'Brien <[log in to unmask]>
> Jon C. Rogowski <[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
>
>**********************************************************
>
>
===================================
Philip A. Schrodt
Dept of Political Science
Pennsylvania State University
University Park, PA 16802
Phone: +1-814-863-8978
January - June 2012:
Peace Research Institute, Oslo
Hausmanns gate 7 NO-0186 Oslo
NORWAY
Phone: (+47) 22 54 77 00
Email: [log in to unmask]
Event Data Project: http://eventdata.psu.edu
===================================
**********************************************************
Political Methodology E-Mail List
Editors: Diana O'Brien <[log in to unmask]>
Jon C. Rogowski <[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
**********************************************************
|