Scientists are seeing increasing demands imposed upon them by journals, funding agencies, and an extremely competitive environment for highly refined statistical applications in implementing the scientific method. Industries are finding their competitiveness and survival dependent upon efficient experimental designs, high quality processes, and reliable products. Further, the best analytical techniques for research and industry, and the computer software used to implement them, evolve rapidly. These demands on researchers and industry will not diminish. In an age dominated by computers and information, the ability to assimilate and process large quantities of information is crucial. It is increasingly obvious the need for complex statistical analysis in solving real world problems in very tight time schedules. Our faculty plans to meet the statistical needs of the State of Louisiana, Louisiana State University, and the Louisiana State University Agricultural Center, and to be a national leader in the development and use of state of the art statistical technology.
Our vision is to be a leading department in applied
statistics, advancing the discipline through scholarship, teaching, practice,
and enhancing the quality of research in the University community through
interdisciplinary research, statistical support and statistical education. Fulfilling our dual roles as educators and
researchers in the College of Agriculture and researchers and statistical
consultants for the Louisiana Agricultural Experiment Station makes the
department a natural site for the development of innovative interdisciplinary
statistical techniques and teaching methodologies.
The
mission of the Department of Experimental Statistics is the study, research,
development, practice, and dissemination of the knowledge and applications of
the discipline of statistics.
· To be the source for research, teaching, and service in statistics at the University.
· To further the development of statistics as an academic discipline and profession by providing an environment that promotes faculty scholarship.
· To maintain a quality curriculum for the development of applied statisticians and to provide quality service courses in statistics for other curricula.
· To promote quality research throughout the University community and the State of Louisiana through statistical support and collaborative research.
Actions:
1. Expand total faculty size by at least 3 additional faculty members.
2. Convert Instructor positions as they become vacant into tenure-track assistant professor positions.
3. Establish a full-time consulting laboratory director position (program review recommendation).
Outcomes:
· Experimental Statistics will be recognized as a leader in statistical science and applied statistics for the region.
· Statistical research through collaboration with Federal and State agencies, industry, and the University community will enhance economic development for the State.
· Quality and quantity of research at LSU will be enhanced due to increased statistical consulting support and collaborative research. Statistical support is a necessary and integral part of a National Flagship Agenda. Practically all other major research universities that LSU has selected as peers have statistics departments with faculty sizes 2 to 4 times that of Experimental Statistics.
· The addition of tenure-track faculty will add continuity and diversity to the existing program, and provide manpower for additional consulting, teaching, and research requirements.
· The consulting activities of the Department will operate more efficiently, and service to clients can be enhanced through the addition of a consulting laboratory director position (see Zhang, J. 2003. Statistics consulting in an academic institution. The Statistical Consultant 20(1), 3-4).
Actions:
1. Establish a PhD program in Applied Statistics (program review recommendation).
2. Have nationally competitive assistantships for all MAPST and PhD students.
3. Implement a thesis option for the Master of Applied Statistics degree (program review recommendation).
4. Develop concentrations such as biostatistics, data mining, bioinformatics, industrial statistics, and/or statistical computing as concentrations within the Master of Applied Statistics degree program (program review recommendation).
5. Implement coursework/activities to integrate statistical theory and statistical methods (program review recommendation).
6. Implement a thesis/dissertation review system for appropriate statistical design, analysis, and reporting of scientific research in conjunction with the Office of Research and Graduate Studies.
Outcomes:
· Research productivity will increase through the PhD program and thesis option, and through the statistical consulting support that these additional students will provide.
· Students will be better prepared upon graduation, and will be more attractive in the job market.
· Quality and retention of graduate students entering our program will improve due to competitiveness and availability of stipends.
· Quality of research at LSU will be enhanced through proper statistical design, analysis, and reporting, leading to the recognition of research excellence at LSU.
Actions:
1. Teach all introductory statistics courses (undergraduate and graduate) at the University (program review recommendation).
2. Increase awareness of the undergraduate minor in Applied Statistics, and integrate the minor as a concentration within other degree program curricula.
Outcomes:
· Students will be better consumers and users of statistics, satisfying key learning outcomes of applying mathematical concepts and problem solving techniques along with the use of appropriate computing technology.
· Graduates will be more competitive nationally, and more attractive to graduate programs.
· Louisiana graduates will be better prepared to pursue graduate studies in Applied Statistics.
Actions:
1. Establish a funded statistics help desk for statistics tutoring and consulting (program review recommendation).
3. Establish a funding mechanism for replacement of laboratory computers on a regular basis (life-cycle funding), or establish the requirement of student-owned laptops for laboratory use with the Department supplying Internet access points.
Outcomes:
· Increase access to statistical support services by students, faculty, staff, and administration.
· Increase the quality of the learning environment, and provide students with training on computing resources commonly encountered in the work environment.
Actions:
1. Develop a regularly scheduled series of statistics workshops for industry and academics directed at State, national, and international audiences.
2. Develop a comprehensive examination system to track students’ progress through our degree programs, that can provide feedback to the Department to enhance the curricula.
Outcomes:
· The Department will increase the number of supportive constituencies at the local, State, and national levels.
· Students will obtain regular feedback on their progress through the degree program, helping them to identify areas of their strengths and weaknesses.
· The Department will obtain feedback to help monitor effects of curriculum modifications, recruiting, etc. on student performance, and monitor progress toward the achievement of program learning outcomes.
Actions:
1. Establish a consulting unit to interface with industry and State agencies to provide fee-based statistical consulting services (program review recommendation).
2. Increase federal, state, and private dollars to support the missions of the Department.
3. Establish at least 1 endowed chair/professorship in Applied Statistics.
Outcomes:
· Faculty recruitment will be nationally competitive and can be more targeted.
· More graduate students will be supported on assistantship, increasing the support for research and teaching, and enhancing graduate student recruitment, resulting in an increase in the diversity of the student body.
· Additional faculty and staff, such as Research Associates, Visiting Professorships, and Post-docs, will be available to support and enhance research and teaching functions.
· State-of-the-art computing equipment can be obtained and maintained in support of the Department’s missions.
Implement a thesis option into the Master of Applied Statistics degree (program review recommendation; national flagship action agenda 1 and 2).
Develop concentrations such as biostatistics, data mining, bioinformatics, industrial statistics, and/or statistical computing as concentrations within the Master of Applied Statistics degree program (program review recommendation; national flagship action agenda 1 and 2).
Submit a proposal to the Board Of Regents for approval of a PhD program in Applied Statistics (program review recommendation; national flagship action agenda 1, 2, and 4).
Submit proposals for Board Of Regents Fellowships for MAPST students (program review recommendation; national flagship action agenda 2 and 6).
Implement a student activity-based course or capstone to reinforce coursework training in Applied Statistics, and the integration of statistical methods and theory (addresses a program review recommendation; faculty discussions concerning written and final oral examinations; national flagship action agenda 2).
1. MAPST Learning Outcomes and
Performance Indicators
2. Resource Needs
In 1974 the Department submitted its application to the Board of Regents to offer the Master of Applied Statistics degree. The degree was formed with the following objectives:
· Equip the student with a sound statistical background.
· Provide the student with a sound background in a research area.
· Utilize the tremendous research problem resources available to give the student a thorough grounding in the interaction between statistics and research through consulting experiences.
Further, the Department committed to the following principles:
· Maximum opportunity must be given for interaction between the students and faculty members. Each student must have a major professor and graduate committee to provide input into the program and training of the student. The student would serve as an apprentice under the major professor, as the major professor would involve the student in consulting efforts (EXST 7083, 7084) and application of statistics to a particular area of research (EXST 7085).
· Emphasis will be given to the application of statistics to the various research areas. Each student should have a minor area of study with a minor professor. The minor professor along with the major professor would direct the student in developing the application of statistics to the minor field.
· The requirements for entering the program should not be so stringent as to discourage students. However, the quality of the program must be maintained and students must be required to make up any deficiencies in their undergraduate training. For example, to successfully complete the program the student must have a working knowledge of calculus through differentiation and integration. If the student does not have this, he may enter the program and make up the mathematics on a non-credit basis.
· Since there is a wide diversity of interests among students and since the specific statistical techniques used vary from one research area to another, students must be given as much flexibility as possible in planning their degree program. The student should be allowed to plan a degree program that is consistent with his interest and one that will provide the best training for consulting and joint research efforts in the minor area of study.
The
Master of Applied Statistics program is organized around four groups of
courses:
· Core courses – All students take these, and they include two theory courses (EXST 7060, 7061) and two methods courses (one of EXST 7003, 7004, or 7005, and one of EXST 7013, 7014, or 7015).
· Professional courses – Each student is required to have 4 credit hours of consulting practicum (EXST 7083, 7084). Each will also be required to take the advanced seminar (EXST 7086) and special problem (EXST 7085).
· Advanced courses – Each student is required to have nine hours of advanced graduate coursework in EXST courses. The student will select these courses depending on his goals.
· Minor courses – The student will take nine hours of selected minor course work from other departments to strengthen the knowledge in a research area.
This results in a program with a total coursework requirement of 38 hours. Further, the student must pass a comprehensive written examination covering his entire program. The final step is an oral examination administered by the student’s advisory committee.
Learning outcomes proposed by the American Statistical
Association for certification of consulting statisticians includes items listed
below (taken from ASA Committee on Certification: Expanded Proposal: 28 May
1993: Section 3.2, American Statistical Association, www.amstat.org). The vast majority of these
items are covered specifically by the basic statistical methods sequences (EXST
700x, 701x), the mathematical statistics sequence (EXST 7060, 7061), and the
professional courses (EXST 7083, 7084, 7085, 7086). General awareness of
statistical methods and theory is given throughout all courses offered by the
Department.
1. Fundamentals of Probability and Inference. The students should show an understanding in the following areas, including relationships among items where appropriate.
· Basic probability calculations
· Conditional probability and independence
· Random variables and transformations of variables
· Specific probability distributions including binomial, multinomial, Poisson, normal, F, t, chi-square, exponential, and uniform
· Central limit theorem
· Population versus sample
· Likelihood methods
· Bias and mean square error
· Least squares estimation
· Differences in broad terms between other methods of estimation (weighted least squares, maximum likelihood, moments, and Bayesian)
· Confidence intervals (confidence level, coverage, pivotals)
· Hypothesis testing (Neyman-Pearson framework)
· Meaning of a P-value
· Types of errors in hypothesis testing
· Power of a test
· Linear and non-linear models
2. Basic Methods Widely Used in Applied Statistics. Students are to know the strengths and weaknesses of common methods used to summarize data, for performing statistical tests, of common general-purpose statistical software, and be able to use and interpret output from statistical packages for the analysis of data. This will include:
· Mean, median, mode, variance, standard deviation, standard error, coefficient of variation
· Graphical methods to display data (stem-and-leaf, box plots, scatter plots, probability plots)
· Sampling from a normal distribution
· T-tests, paired and unpaired
· Basic nonparametrics (Wilcoxon/Mann-Whitney equivalents to t-tests; sign test)
· Linear regression (simple, multiple, multi-source, diagnostics)
· Correlation (simple, multiple, partial)
· Binomial data; estimation, confidence intervals for p, tests of significance, comparison of two populations (chi-squared and Fisher’s exact test)
· R x C tables (contingency tables, chi-squared tests)
· Analysis of variance: one and two-way; fixed and random effects; analysis of unbalanced designs; relationship to regression; contrasts and multiple comparisons
· Principles of experimental design including: randomization, pairing, blocking, Latin squares, factorial experiments, split-plot (factors within factors), response surfaces, sample size calculations, design of data collection forms
· Principles of survey design including: population, types of sampling, questionnaire design
3. Awareness of the Broad Field of Statistics. Students should have an awareness of concepts such as:
· Bayesian analysis
· Computer intensive methods (bootstrap, jackknife, randomization tests, smoothing)
· Categorical data analysis (logistic and loglinear models)
· Generalized linear model
· Multivariate analysis (principal components, factor analysis, MANOVA, structural equations, etc.)
· Spatial statistics
· Observational studies
· Survey research methods
· Reliability and survival analysis
· Quality control including statistical standards1
· Time series analysis[1], Markov and semi-Markov processes1, queuing theory1
In addition, students should be aware of the many areas within which statistics is frequently applied: Biometry, Biopharmaceuticals, Business, Computing, Education, Environment, Epidemiology, Government, Graphics, Health Sciences, Marketing, Quality, Physical and Engineering Sciences, Productivity, Social Statistics, and Sports.
4. Awareness of the Ethical Aspects of Statistical Consulting.
5. Ability to Interact Constructively with Non-statisticians.
6. Demonstrated Skills in the Application of Statistics.
Course work including consulting practica, special problem, and the written comprehensive and final oral examinations are the primary means for evaluating the success of the program objectives. Each candidate for a MAPST degree is required to pass a comprehensive final examination that consists of two parts, a written comprehensive examination and a final oral examination. The two parts need not be taken in the same semester, but a student must receive a passing grade on the written comprehensive examination before the date of the final oral examination. In addition, students are required to submit a special problem project (EXST 7085) approved by the student’s committee to the Department prior to graduation.
The final written exam is compiled and graded by a
Departmental Committee that is assigned by the Department Head on a
semester-by-semester basis. The exam is designed to evaluate a student's
command of that body of material that each and every Master of Applied
Statistics student should possess. As such, a wide range of topics is covered.
Material for each semester’s exam is drawn from a large pool of questions compiled
by the faculty and taken from a variety of sources. This exam is designed to
accomplish several objectives simultaneously:
A final oral examination by the student’s advisory committee is given following successful completion of the written comprehensive exam and all other course work requirements (some courses may be in progress during the semester within which the student is taking the final exam). This exam follows Graduate School regulations on committee size and graduate faculty composition and has the same objectives as the written exam. In addition, the oral exam permits faculty to observe and assess verbal communication skills [speaking, listening] and the ability of a student to discuss and address statistics topics [problem comprehension, use of “statistical language”].
The special problem project (EXST 7085) undertaken with a faculty advisor gives the advisor, and the student’s committee, insight into the student’s level of achievement and success with research, and the student’s writing and organizational skills. Further, the special problem requires that the student demonstrate skills in the application of statistics.
Tenure-track faculty positions. The highest
priority that the Department has is the acquisition of new tenure-track faculty
lines with specialties from broad areas of applied statistics including such
areas as data mining, biostatistics, bioinformatics, statistical genetics,
statistical computing, linear and non-linear modeling, Monte Carlo methods,
experimental design, and survey sampling. These positions could be joint
appointments between the College of Agriculture and the Louisiana Agricultural
Experiment Station, and would have responsibilities to perform research in
Applied Statistics, collaborate and consult with faculty throughout the
University community, teach courses in statistics, secure grants and contracts
to support Departmental activities and research, and to work with Master
(MAPST) and Ph.D. students of Applied Statistics.
Median 9-month Starting Salary Assistant Professor: $60,500
(2002-2003 Salary Report of Academic Statisticians, American Statistical
Association, http://www.amstat.org/news/salaryanalysis.html)
Assistantship Resources. To support a viable PhD program in Applied Statistics, the Department will require additional assistantships. Currently the Department’s assistantship rate is $12,500/year with most Teaching Assistantships restricted to 9-month appointments. In 2003, the rate needs to be increased to at least $14,500/year.
Consulting Center Director. In order to operate a University-wide consulting center and an external consulting service center for State government and industry, a full-time director will be required. See Zhang (2003) [Zhang, J. 2003. Statistics consulting in an academic institution. The Statistical Consultant 20(1), 3-4.] for additional comments regarding this need. The position would need to be carefully described and evaluated to avoid difficulties with promotion and tenure.
Computing Support Technician. Currently the Department uses a single graduate assistant for these duties that it shares with the College of Agriculture. The College provides the support funds for this position. A full-time technician is required for the day-to-day operation of 4 computing laboratories, several administrative offices, faculty computers, and the associated network, printers, and servers.
Statistical Packages. The Department needs access to a variety of statistical computing environments in support of its research, teaching, and consulting missions. Software such as SAS, SAS Data Miner, SAS/INTRNET, SPSS, Splus, Minitab, Stat-Exact, PV-Wave, Mathematica, MAPLE, ArcView and ArcGIS are needed and require annual maintenance.
Operating Systems. Departmental servers need to be upgraded to Windows 2003 server, and laboratory and office machines need to be upgraded accordingly.
Vehicle. The Department has struggled for years with frequent repairs to unreliable vehicles. Essentially all out-of-Baton Rouge travel occurs in personal vehicles due to concerns of reliability and safety of the State vehicle. Presently we have a vehicle that has high mileage and will need to be replaced soon. Approximate costs: $12,000-$16,000 (sedan to station wagon).
Laboratory Computers and Printers. Laboratory computers and printers need replacing on a 3- to 4-year life-cycle. The Department has managed over many years to replace computers in these labs through a variety of funding outlets, though the computers tend to get very outdated by the time sufficient funds are obtained. Permanent life-cycle funding as a part of the budget is needed for this equipment. An alternative would be the installation of wireless access points in the labs along with a requirement that students supply their own laptop with the appropriate software for use in the labs.
Faculty, Staff, and Student Office Computers and Printers. Again, life-cycle funding is required to insure that these groups are working on modern and reliable computing hardware with access to the latest software.
Networking Upgrades. As the University upgrades to 1-gigabit speeds, the Department will need to upgrade its existing networking hardware to take advantage of these increased networking speeds.
Computer Projectors. A lot of instruction of statistical methodologies and statistical principles takes place in our computing laboratories. Computer projectors would greatly enhance these classes and enhance their use for statistics workshops. Approximate costs: 3 labs x $2,000/projector = $6,000.
Security Systems. Install video security system to monitor hallways at all times, and security systems for computer laboratories in Ag Admin 11 and 44. Security systems would be integrated building-wide. EXST Camera hallway system: $12,725. Laboratory security system (rooms 11 & 44): $8,820.
Office Space for New Faculty and PhD Students. Current Department space can be used to accommodate a few additional faculty members, but current space is not sufficient to house a fully functioning PhD program.