OVERVIEW
I enjoy collaborating across a broad spectrum of
business, science, engineering, and manufacturing, to design and
improve processes and products to better meet customer needs. I believe in a hands-on approach to problem solving, and
spend a substantial amount of time "on the floor" getting
to know the people and the processes involved.
I am results focused- a statistician with an engineering, as
much as a mathematical, bent- and while capable of sophistication, I
look for simple where simple will do.
I enjoy teaching others the skills and tools that enable
sound decisions in the face of uncertainty, and I continue to learn
and leverage new methodologies to that same end.
I have consulted as a statistician for several years in
industry and academia, on a broad range of subject areas, making use
of a background in science and engineering.
My current focus is helping businesses improve through
implementation of Six Sigma (a process improvement methodology made
famous by Motorola and GE). I
worked for GE as a quality consultant (Master Black Belt) supporting
process improvement and process redesign teams in their Six Sigma
quality initiative. At Eli Lilly I worked as a Senior Statistician and was
involved with project strategy, design and analysis of clinical
trials, and presentations to the FDA.
I also managed the R&D statistics department for an
agricultural feed products company, involved with testing feed
supplements, and process/product development and improvement.
My interest in design of experiments led to work with GM
Research Labs and INTEL Corporation, to develop expert systems for
experimental design, and I have conducted several workshops and
courses on the design and analysis of experiments.
EDUCATION
|
9/88 - 5/98
|
|
Ph.D.
Statistics, University of Manitoba. Representing
Restrictions on Randomization in the Linear Model
|
|
9/86 - 8/88
|
|
M.S.
Statistics, Brigham Young University (BYU). An Automated and Expert
System for Experimental Design Based on the General Linear
Model.
|
|
9/79 - 4/80
9/82 - 8/86
|
|
B.S.
Computer Science/Statistics, BYU.
|
EMPLOYMENT
|
6/01 – current
|
|
PROMAX
Consulting Services, Inc. Independent Consultant,
Robert Johnston & Associates Inc., Wake Forest, NC.
Consult
and teach in the areas of quality, process improvement, and
statisitcs:
·
Six Sigma (business
process improvement)
·
Design of experiments
& advanced data analysis
·
Clinical Trial design,
analysis, and reporting
·
SAS programming
|
|
11/99 – 6/01
|
|
Quality Consultant (Master Black Belt), GE Mortgage
Insurance Corp., Raleigh,
NC.
Leadership role in GE’s Six Sigma quality
initiative:
·
Direct teams in process
improvement (DMAIC) and redesign (DMADV).
·
Direct teams in Web
enabling of key processes.
Focus on radical redesign of key processes prior to web
enabling.
·
Manage the process
control system, links all core processes to top level business
drivers.
|
|
1/97 – 11/99
|
|
Senior
Statistician, Eli Lilly and Co.,
Indianapolis, IN.
Design and analysis of
Phase III clinical trials related to neuroscience.
·
Served as lead
statistician on an NDA submission for Weekly
ProzacÔ.
·
Represented Lilly at the
FDA on various occasions.
·
Developed protocols for
acute and long-term treatment, that included dose escalation,
re-randomization, and blinded rescue therapy.
·
Employed statistical
methods such as survival analysis (Kaplan-Meier, Proportional
Hazards Regression), equivalency/non-inferiority testing,
repeated measures models with censored data, and dose-response
relationships.
·
Worked with large
cross-functional teams. Creatively supported the team by
recognizing the need for, and developing, tools to aid trial
implementation, specifically, an automated visit scheduler for
on-site use and programs to help track enrolment and dropout
rates in time sensitive trials with interim analyses.
Adapted drug supply process to more efficiently utilize
limited supplies of clinical trial material.
·
Instructed the annual
PERI (industry-wide) course for new clinical statisticians.
|
|
5/93 - 8/96
|
|
Manager
of Research Statistics, Moorman's Inc., Quincy, IL.
Consultation with PhD
nutritionists on experimental design for swine, beef, and dairy
research, as well as with feed technology and manufacturing on
process/product improvement.
·
Managed the analysis and
reporting of experimental data for more than 100 trials-
parallel, crossover, split-plot, latin square and incomplete
block designs.
·
Conceived, and managed
the design and implementation of a relational database system
for research farms. Reduced
data cycle-time by more than 90% (from 4 months to less than 1
week).
·
Automated the data
cleaning, analysis, and reporting for experiments involving
blocking, repeated measures, covariates, and
treatment-by-covariate interactions.
Could accommodate twice the volume of experiments with
half the resources.
·
Worked with feed
technology on process/product design and improvement.
For example, experiments resulting in redesign of mixers
and calibration instruments.
·
Designed and implemented
an on-line factory system for statistical process control.
·
Consulted with other
areas as-needed (survey design, marketing and finance data
analysis, etc.)
|
|
9/88 - 4/93
|
|
Lecturer,
Department of Statistics, University of Manitoba, and University
of Winnipeg.
Courses taught:
5.333
Research Methods- Multiple regression, analysis of
variance and covariance, experimental design.
5.221
Introduction to Statistical Analysis- EDA, estimation and
testing, regression and correlation, experimental design.
Consultant,
Statistical Advisory Service, University of Manitoba (see section: Summary of University Consulting)
|
|
5/89 - 8/89
|
|
Researcher,
General Motors Research Labs, Warren, Michigan.
Design and development of an expert system for
experimental design. Focus
was on design creation based on user inputs, rather than on
design selection from a limited database. Included capabilities
for response surface, asymmetric fractional factorials, robust
designs, and restricted randomization.
Investigated methods for analyzing fractions of
asymmetric mixed models.
|
|
9/86 - 4/87
|
|
Consultant,
Department of Food Science and Nutrition, BYU (see section: Summary of
University Consulting)
|
|
5/87 - 8/87
|
|
Researcher,
Expert Systems Division, INTEL Corp., Chandler, AZ.
Reviewed and tested expert system shells.
Developed expert system technology for experimental
design applications. Consulted
with engineering staff on design and analysis of experiments
dealing with Integrated Circuit production.
|
|
5/85 - 8/88
|
|
Consultant,
Statistical Consulting Center, BYU (see section: Summary of
University Consulting)
|
|
1/83 - 4/85
|
|
Teaching assistant in Physics and Statistics
Departments, BYU.
|
SUMMARY
OF UNIVERSITY CONSULTING
The University of Manitoba and BYU operate statistical
consulting centers to assist the university community.
Clients are a mix of master's students, Ph.D. students, and
faculty, encompassing a wide array of disciplines.
A sample of areas worked on include:
|
Anthropology
|
|
Reviewed and developed methods for estimating
seasonality of death from growth rings in bones, teeth, or
shells.
|
|
Psychology
|
|
Evaluated Multivariate models for post-partum
depression in women.
|
|
Linguistics
|
|
Devised a graphical presentation for word group
pronunciations in Slavic languages.
|
|
Food Science
|
|
Consulted on several projects for the Food Science
and Nutrition department at BYU including taste panels and toxin
studies, that involved large
companies such as NASA and Norbest Foods.
Developed and taught a faculty short course on design of
experiments.
|
GRANTS/AWARDS
& SERVICE
|
2001
|
|
Nominated by senior management for GE Capitals
Pinnacle Award, for process redesign and digitization
(Mediterranean Cruise).
|
|
1991-1992
|
|
University of Manitoba Fellowship ($10,000 CAN).
|
|
1989-1991
|
|
Natural Science and Engineering Research Council
Scholarship ($30,000 CAN).
|
|
1987-1988
|
|
Partial funding of thesis work by INTEL Corp. ($8,000
U.S.).
|
|
1988
|
|
Best In Class award, for presentation at Physical
& Mathematical Sciences Symposium, BYU.
|
|
1987-1988
|
|
Russell Richards Memorial Scholarship.
|
|
1998 & 1999
|
|
Instructed PERI (industry-wide) courses for new
clinical statisticians.
|
|
1996
|
|
Instructed workshop on experimental design for ASQ.
|
|
1995-1996
|
|
Program chair for local ASQ.
|
|
1990-1991
|
|
Served on Internal Review Committee for the Dept. of
Statistics, University of Manitoba.
|
|
1986
|
|
Data analysis and report compilation for Internal
Review for Re-accreditation, BYU.
|
SAMPLE
OF SKILLS/SUBJECT AREAS
Statistics
|
Design
of Experiments
·
split-plots
·
fractional factorials
·
response surface
·
mixture designs
·
restricted randomization
Linear
models
·
repeated measures
·
heterogeneous covariates
·
missing data
·
randomization tests
Quality
Control
·
EWMA
·
MIL Standards
|
Multivariate
·
MANOVA
·
factor analysis
·
classification
Categorical
·
loglinear and GSK
Survival
·
Kaplan-Meier
·
Proportional Hazards
Regression
Sampling
·
stratified
·
cluster
·
varying probability
|
Languages
SAS (Base, Stat, Graph, Macro, IML), S-PLUS, JMP,
Minitab, Gauss
Fortran, C, Pascal, Prolog, Smalltalk, MAPLEV
Science/Engineering
Chemistry (3 courses)
Elec. Eng.
Physics (3 courses)
Chem. Eng. (2 courses) Civil
Eng.
Biology
Foreign Language
Portuguese (spent 2 years in Brazil)
PRESENTATIONS,
PUBLICATIONS, AND TECHNICAL REPORTS
Johnston,
R.S. (1998, 1999).
PERI Course for New Clinical Statisticians.
Washington D.C. and Fort Lauderdale, FL.
Johnston,
R.S. (1996). Experimental
Design. A workshop
presented for the ASQ, Quincy, Illinois.
Monks,
G., Johnston, R. (1993). Estimating
season of death from growth increment data: a critical review.
Archaeozoologia, Vol.
V/2, pp. 17-40.
Johnston,
R.S. (1991). A
Generalized and Unified Approach to Representing Restriction Errors in
the Linear Model. Technical
report, University of Manitoba, Winnipeg, Manitoba.
Johnston,
R.S., Balshaw, R. (1991). Principal
Components and Factor Analysis by Example.
Technical report, University of Manitoba, Winnipeg, Manitoba.
Johnston,
R.S. (1991). Expected
Mean Squares in Fractional Designs Represented by Mixed Models. Technical report, University of Manitoba, Winnipeg, Manitoba.
Johnston,
R.S. (1991). A Matrix
Based Approach to Restrictions on Randomization.
Presented at the ASA annual meeting, Atlanta, Georgia.
Johnston,
R.S. (1990). Expected
Mean Squares in Fractions of Unbalanced Mixed Models.
Presented at the ASA annual meeting, Anaheim, California.
Johnston,
R.S. (1989). Expert
Systems in Design and Analysis of Experiments.
Presented at the annual meeting of the Statistical Association
of Manitoba, Winnipeg.
Monks,
G., Johnston, R. (1989). Measuring
Incremental Growth Structures. Presented
at the 54th Annual Meeting of the Society for American Anthropology,
Atlanta, GA.
Johnston,
R.S., Scott, D., Bryce, G., Hilton, C. (1988).
An Automated and Expert System for Experimental Design Based on
the Linear Model. Proceedings
of the American Statistical Association, New Orleans.
Johnston,
R. (1987). Building and
Using General Linear Models. A
workshop presented to the Department of Food Science and Nutrition,
BYU, Provo, Utah.
Scott,
D.T, Peto, G., Johnston, R., Hendrix, L. (1987).
An Expert System for Building a General Linear Model.
Proceedings of the American Statistical Association, San
Francisco.
|
