Alumni Experiences

If I get a Graduate degree in Statistics/or at OSU, what kind of work does this prepare me for?

To answer this question, we asked our graduate alumni to write us a paragraph describing their job and how statistics enters into it.  Here are some of their replies (in random order):

Manuela Huso, M.S.  Statistics, 1989 
I work as a consulting statistician in the Department of Forest Science at Oregon State University.  My job has three primary components.  First, I am available to all FS faculty and their graduate students for consultation on study design and subsequent analysis.  It is still an uphill battle to assure that each graduate student has some sound statistical guidance at the design phase, before execution and analysis, but it's getting better.  A second component is collaboration with other faculty on a research topic from its inception through publication in the peer-reviewed literature.  The third component is teaching.  I teach short courses on use of statistical software packages and co-teach a graduate level course entitled Natural Resources Data Analysis.

I have applied statistics to such diverse topics as phtyochemistry of dead Alaska yellow cedar, effects of grazing on native and exotic plant density, and most recently, estimation of bird and bat fatality in wind power generation facilities. Approaches I have used include generalized mixed models, hierarchical models, simulation analysis, as well as simple ANOVAs. I very much appreciate the autonomy and flexibility I have in my work place. And the work itself is always challenging and ever-changing, and most importantly, never boring.

Stephen Jensen, M.S. Statistics, 2004
After receiving my MS in statistics at OSU, I began working for CleverSet, a small, Corvallis-based software research  company founded by a former OSU computer science professor.  At CleverSet I helped research and develop applications of statistical relational learning, specifically relational Bayesian networks, to various domains, such as satellite defense, ecological studies, sensored smart-homes, and e-commerce shopping behavior.  CleverSet's expertise in understanding online shopper behavior led to the development of CleverSet's product recommendation service for online retailers.  CleverSet was purchased in early 2008 by ATG, a large e-commerce software provider, and incorporated as ATG Recommendations.  Today I work for ATG as part of a team that develops the statistical relational models that drive ATG Recommendations.

Ken Fujii, M.S. Operations Research, 1998
Currently, my title is Associate Consultant.  My role with BARRA is to help our clients (institutional managers such as JP Morgan, Goldman Sachs, Fidelity, etc.) use our analytical software and models to help evaluate their portfolios.  Barra's software helps the portfolio manager target where their risk or return is coming from.  For example, was holding large cap assets in the semiconductor industry a good idea?

The concepts I learned from my MS degree in Operations Research that I use for my job are:

1. We use a multi-factor model (i.e., regression model) to forecast risk;
2. We use quadratic programming to find the optimal portfolio.

Having a quantitative background is very important since all of BARRA's models are based on statistical and operations research concepts.

Ken Newman, M.S. Statistics, 1982 (Currently Associate Prof., U. of Idaho)  
My current projects are primarily statistical problems in salmon science and management. 

After finishing my M.S. at OSU , I worked from 1983 to 1988 as a biometrician for native American fisheries in western Washington.  Having a Ph.D. was not essential to developing useful solutions to problems encountered there, and I believe that well-trained MS- level statisticians can provide valuable know-how in the areas of fisheries and wildlife.  One key to success and influence is to learn a particular subject matter area well enough for the statistician to serve as a co-principal investigator, not just as a statistical consultant.

One problem I worked on was to estimate the catch of salmon by sports fishermen in Puget Sound….the two sampling courses I took at OSU, sampling methods and sampling theory, were invaluable in terms of thinking about alternative ways to construct the sampling frame and then thinking about alternative estimation procedures….the results of this work, which took 4 years, led to differing adjustments of the voluntary-sample catch estimates in different regions and had a real impact on catch allocation issues between Indian and non-Indian fishermen.

John Van Sickle, M.S. Statistics, 1992  
In 1991 I began working as a statistician at the US Environmental Protection Agency research laboratory here in Corvallis.  Since then, I have been involved in a variety of statistical modeling projects which have focused on detecting and estimating the effects of human activities on aquatic ecosystems, against a background of high natural variation in those systems.

I make heavy use of multiple regression, multivariate methods, classification tools, and graphical tools to explore large, high-dimensional sample data sets collected across broad geographic regions.  Since we are interested in spatial patterns, I work with spatial correlation models and have learned to work with Geographic Information System (GIS) data sets and software.

I design and implement my own research projects, but also work on research teams and do one-on-one statistical consulting.  There is no question that communication skills are at least as important as statistical knowledge for me to work effectively in my research environment.

 Jeannie Sifneos, M.S. Statistics, 1986
I provide statistical support primarily for EPA funded research projects as a faculty research assistant in the Geosciences Department at OSU.  I work with a multidisciplinary team of scientists on the design and implementation of projects and use statistical analyses to describe relationships. For the last few years, I've been working on wetlands data, specifically looking at various aspects of rapid assessment methods used to determine wetland condition. Over the years, I've worked with many different types of data (e.g., fish and vegetation community data, biodiversity of terrestrial invertebrates data, landscape data, wetland 404 permit data) to answer environmental research questions using statistical tools.  I write up research results, publish in peer-reviewed journals, and occasionally get to go out in the field.

 Mike Cavanaugh, M.S. Statistics, 2000
My current position is instructor in statistics at Col Poly, San Luis Obispo (CA).  Over the summer I taught a stats for business students class, similar in content to ST 351 at OSU.  This fall I will teach two sections of the above class in addition to a section of stats for life sciences, which is probably similar to ST 411/511. 

When I decided to pursue a career in statistics it was not with the idea of becoming an instructor.  I was always geared toward some kind of industrial situation.  But I am glad that I decided to try teaching.  I have had 8 to 5 industrial "real jobs" before and I am much happier in my current position.

 Warren Aney, M.A. Statistics, 1972
I entered the OSU Department of Statistics in 1969 under a biometrics grant program aimed at making biometricians out of mathematicians and biologists.  I already had 10 years experience as a wildlife biologist and was one of the few biologists who entered this biometrics program.  Most of my course work was in math and statistics with a few graduate level courses in ecology.  My math cohorts took a lot more biology.

Over the almost 30 years since receiving my M.A., I first worked 6 years as head of the Oregon Department of Fish and Wildlife's Statistical Services Section.  My work during this time was primarily design and analysis for field and mail sample surveys, plus some early work on using satellite spectral data for habitat classification.  I spent the rest of my career with ODFW (15 years) in middle management.  Then for the past 7 years I've worked as a consulting wildlife ecologist.  Some of this work has involved statistical applications, usually rather simple in nature

(t-tests, chi-square, confidence limits).  However, my statistical training has made me a more astute ecologist with considerable appreciation for the value of information obtained from well designed, well conducted, and well evaluated data collection.

Len Deaton, Ph.D.  Statistics, 1974 
I am currently teaching both mathematics and statistics on a part time basis, at four different universities or colleges.  However, until 3 years ago, I  worked for 18 years for the Navy, as a civilian.  This was a very interesting and challenging job.  I started out designing, maintaining, updating, and using my own data base on all Navy air launched missile firings.  I produced quarterly reports and annual trend reports on the firing results which had statistics on the success rates and failure causes for various types of firings.  In addition to the regular reports, I provided responses to questions at a rate of about 4 per week.  I was often asked by high-ranking officers and civilians whether the results were statistically significant.  This caused me to develop a computer program for an exact test of a hypothesis on a Bernoulli population (sample sizes were often very small).  Results of one of my reports changed entirely the way the Navy trained missile and aircraft maintenance crews stationed in the Far East.  Even though all data was classified, some of my tables and words were quoted in Aviation Week on comparison of captive-carry failures of missiles.  My database was cited in the Congressional Record as being far superior to any such database held by the Air Force or the Army.

Vikki Maurer, M.S. Statistics, 1990
Currently I teach in the math department at Linn-Benton Community College.  I have also taught at Western Oregon State College.  When I went to graduate school, teaching was always something I was interested in.  I wanted options when I was done with school, which is why I went to graduate school to study statistics.  I did not want to be locked into teaching.  I wanted to be able to work as a statistician.

Pachitjanut Siripanich, Ph.D, 1988
I am teaching in the School of Applied Statistics, The National Institute of Development Administration (Bangkok, Thailand).  Besides teaching graduate level statistics, I also do some consulting in statistics on the outside, in both the private and government sectors.  I do a lot of survey research and some public opinion surveys on various topics.  In analyzing data of this type I use simple descriptive statistics, chi-square tests and hypothesis tests.  Once in a while I use some multivariate techniques such as factor analysis.

I've found that it is easy to abuse statistics if your background is not strong enough.  Starting from the first step in data collecting, nonprobability sampling is what most people do along with inappropriate methods for analyzing data.  Bias occurs in every step, when there is no sampling frame and improper design of sampling. 

Statistical computer packages play a big role in my work.  Training programs for analyzing data with SPSS are always oversubscribed, and other programs training users of statistics packages are in great demand here in Bangkok.

Bryan Wright, M.S. Statistics, 1999  
Following graduation, I spent one year working as a statistician for a government contractor to the USGS in Missouri.  I was involved in a pilot program to monitor the status and trends of freshwater fish health in large rivers across the US.  I used a variety of descriptive and inferential statistical tools to summarize and compare fish health indicators among sampling locations. 

I left that position to return to Corvallis last summer to work as a wildlife biometrician for the Oregon Department of Fish and Wildlife.  Specifically, I'm working in the marine mammal program where I recently designed a survey sample to estimate the number of salmon being consumed by harbor seals in a coastal estuary.  Currently I'm modeling harbor seal population growth as well as consulting on a deer survival study.  The great thing about this job is that I actually get to help biologists with their field work, which recently included capturing and marking over 100 harbor seals!

Yadolah Dodge, Ph.D., 1974
Dr. Dodge is Professor of Statistics and Operations Research and Head of the Department of Statistics at the University of Neuchatel in Switzerland..  His research interests are in computational statistics, regression analysis, optimization techniques in statistics, design and analysis of experiments, and statistical analysis based on the L₁-norm.

Dale Olsen, Ph.D., 1973
He joined the Johns Hopkins U. Applied Physics Laboratory in 1973 and developed techniques for evaluating both the reliability and accuracy of submarine launched ballistic missiles.  He developed a series of methodologies, implemented them in software, and successfully applied them on projects including those for modeling reliability, for validating the weapon system accuracy models at various levels, and for building detailed accuracy models using the output of massive Kalman filters.

In 1989 he developed an algorithm for the automatic detection of epileptic seizures using EEG data.  This research led to the development of algorithms for evaluating polygraph data and to a software system called PolyScore (tm ) for the analysis of the data.  He has served as a section supervisor, assistant group supervisor, and program manager, and is currently a Program Development Director.

Scott Seggerman, M.S. Statistics, 1995
Upon graduating, I returned to the Defense Manpower Data Center (near Monterey, CA), an agency that maintains various databases relating to the military.  My official occupation series is GS-1529 - Mathematical Statistician.  My job is basically as a SAS programmer to run ad hoc reports, recurring reports, and other analysis pertaining to many files located at DMDC.  My 7+ years at DMDC have included:

• Running analyses to determine which military personnel are eligible for WIC;
• Maintaining and utilizing our drug testing database;
• Supervising the security clearance file;
• Using random sampling to Desert Shield/Storm surveys.

The experience in consulting, multi-tasking, and people skills I learned through being a TA at OSU have improved my standing at DMDC as now I am in a position of team leader, in charge of 6 other junior analysts.

George Fegan, Ph.D., 1973  
Positions held have included:

• Asst. Prof., School of Engineering, Univ. of Portland;
• Consultant to Corporate Planning Dept. of Portland General Electric;
• Branch Chief of Utility Applications for the national Solar Energy Research Institute;
• Research Investigator for Energy Management Associates;
• Research Director for ATAC, a Silicon Valley software and consulting firm;
• Associate Prof. and Chair of the Applied Math. Dept. at Santa Clara University.

The best thing that ever happened to me was OSU.  The statistics program was extremely flexible.  I entered with a M.S. in Mathematics and wanted a degree in statistics with a mathematical bent.  This diverse background helped me in every position….In the energy fields once you got past the nomenclature of the physicists and engineers, everything seemed to be statistics or straightforward mathematics.  At PGE the financial people threw in a little stochastics.  Software development and consulting was not much different:  you listened to the client in much the same way you did in the OSU consulting lab and worked very hard to re-cast the clients' problem into the more familiar statistical-mathematical framework.  Language translation seems to be the key…..I would like to end with a tip of the hat for the role of mathematics in statistics.

Kathy Hall, Ph.D. , 1990
For the last 16 years, I have worked as a statistician in a manufacturing facility in the electronics industry [Hewlett-Packard].  I have consciously remained in this role so that I can keep at least a part of my job as technical, hands-on consulting.  In the consulting role, I try to help clients and project teams understand their data.  Graphical output is the best kind.  Since our group's goal is making money (and not producing research knowledge), our loss function is often money.  This leads to really interesting application of many standard techniques for messy data sets that violate the assumptions of the theory.  There is also a lot of model building, especially cost models.

The other part is more strategic -- determining what data to collect and how to collect it to maximize information for minimum cost so that we make the right decisions.  Once these points are determined, I again need to educate everyone as to why they are the right things to do, and then on to doing something with the data….It's never boring, but provides a great deal of autonomy and job satisfaction.

Nancy Lo, Ph.D. 1972
I am a mathematical statistician at the Southwest Fisheries Science Center, National Marine Fisheries Service.  The types of problems I work on are:

• Survey methods for estimating fish biomass:  ichthyoplankton survey; trawl survey; aerial survey and, lately, airborne laser survey for pelagic fishes.
• Quantifying characteristics of fish and fish populations, in particular their early life history. In statistical terms, estimating parameters (abundance of fish population, growth rate, mortality rates in each life stage, etc.).  My recent interesting projects include using an adaptive allocation method to survey Pacific sardine eggs with the aid of an egg pump, and using spatial statistical techniques to examine the spatial characteristics of sardine eggs.

My advice to a student interested in pursuing a career in statistics is to take courses on communication skills, e.g., writing and public speaking, and courses on subjects to which you would like to apply your statistical skills.

David Liou, M.S. Statistics, 1992
I have been teaching statistics, research methods, and computer software applications in the Department of Social Work, Shih Chien University, Taiwan.  The statistics background has given me some advantages in my teaching career.  It makes me more qualified to teach statistics in the university and gives me advantages in doing research.  Statisticians, especially in social science, have been treated as an endangered species!

Statistics is a  powerful tool in my work and has been useful to my career.

Michel Lejeune, Ph.D., 1974
I am now Professor of Statistics at the University of Grenoble in France.  After moving a lot, first within Switzerland and then to Paris, I am now settled for good with my family.  I try to keep up with some research, although I get more and more involved with administrative tasks.  I am mostly interested in sampling problems with applied orientations, basically surveys.

I greet all staff members and express again my gratitude to the statistics department for all the benefits I received from it.

 Rebecca Rodabough, M.S. Statistics, 1996 & Joe Larson, M.S. Statistics, 2001
 We work for the Women's Health Initiative at the Fred Hutchinson Cancer Research Center in Seattle, WA.  The WHI is a large multi-center randomized clinical trial and observational study.  We have about 161,000 women, aged 50-79 at baseline, participating in one of 3 clinical trials (Hormone Replacement Therapy, Diet Modification, and Calcium/Vitamin D) or the observational study. The majority of out time is spent as consultants/analysts for scientific manuscripts.  We have participated in producing papers related to diabetes, cardiovascular disease, nutrition, and many more topics using analyses ranging from simple linear regression to factor analysis.  The second largest portion of our time is spent producing reports for review by a Data Safety Monitoring Board (DSMB).  The DSMB is responsible for monitoring the study twice a year and could stop the study either for harm or benefit at any point.  These reports include analyses ranging from frequencies and means to Cox regression models and survival analysis.  The remaining portion of our time is spent on periodic and ad hoc reports/analyses.  With all projects, there is always a great deal of data management to be done prior to performing analyses.  We use SAS for almost all the data management and analyses, with some S-Plus for graphical purposes.