S.00-56

• ?

SIMON FRASER UNIVERSITY

OFFICE OF THE VICE-PRESIDENT, ACADEMIC

?

MEMORANDUM

To: ?

Senate

From:

?

J. Osborne, Acting Chair

Senate Committee on Academic Planning

Subject: ?

Establishment of the Department of Statistics and Actuarial Science

(SCAP Reference: SCAP 00-15)

Date: ?

May 18, 2000

Action undertaken the Senate Committee on Academic Planning gives rise to the

following motion

"that Senate approve and recommend to the Board of Governors, the formation of

a new Department of Statistics and Actuarial Science within the Faculty of

Science effective 1 September 2001 and that the renaming of the remaining

department as the Department of Mathematics, as set forth in S.00-56 ."

---

SCAt'

00 -15

SIMON FRASER UNIVERSITY

?

MEMORANDUM

To:

?

Dr. J. Munro

?

From: W.

Davidson, Dean

VP Academic ?

Faculty of Science

Subject:

Formation of a Department of

?

Date:

April 3, 2000

Statistics & Actuarial Science

One of the recommendations that arose from the 1998 external review of the

Department of Mathematics and Statistics was that: "The University should consider

establishing a separate Department of Statistics within the Faculty of Science.' The

Statistics Group have taken this to heart and unanimously agreed that the time was

right to act on this recommendation. They submitted a proposal that had the blessing

of the rest of the Department of Mathematics and Statistics. It is fair to say that some

faculty members will be sorry to see the Statisticians leave, but all recognize that this

will ultimately benefit the teaching of Statistics and the discipline itself at Simon Fraser

University. The proposal was brought to a meeting of the Faculty of Science Council

last week and it passed with only a single vote against.

I am attaching a copy of the proposal that was endorsed by the Faculty of

Science. You will see that the new Department will include Statistics and Actuarial

Science. The rationale for this Department is sound. It is implied that the remainder of

the present Department of Mathematics and Statistics will become the Department of

Mathematics. The ongoing cost is minimal when one considers the overall benefits.

The question of resources should not be used to prevent this plan from proceeding. It

will not alter the hiring plan that is currently in place.

I ask that you personally support this initiative that has come from the faculty

members and take it to the appropriate committees for University approval.

O/L,

4

90--45&--

W. Davidson

c: ?

C. Dean

L. Berggren

Motion:

"That SCAP approve and recommend approval to Senate

and the Board of Governors, the formation of a

?

I ?

c

I ?V

new Department of Statistics and Actuarial Science

?

I

within the Faculty of Science effective 1 September 2001."\

FA

---

Dr. Charmaine Dean presented the following motion at the Faculty of

Science Meeting, March 28, 2000.

To

approve the formation of a

Department of

Statistics as outlined in

the attached paper.

MOVED: C. Dean

SECONDED: L. Berggren

MOTION CARRIED (with one against)

.

I]

2..

---

Simon Fraser University

MEMORANDUM

.

To:

Willie Davidson, Dean

?

From:

Charmaine Dean

Faculty of Science

?

Director of Statistics

Subject:

Proposal for the Formation of a

Date:

March 14, 2000

Department of Statistics

& Actuarial Science

Please find attached a proposal for the formation of a Department of Statistics & Actuarial

Science. I am requesting that this item be included on the agenda of the next meeting of

the Faculty of Science for consideration and approval.

1

?

3.

---

Proposal for the Formation of a

Department of Statistics and Actuarial Science

?

at Simon Fraser University

1. Introduction

The discipline of Statistics has expanded well beyond the strong mathematical focus that

was prevalent when Simon Fraser University was founded. Actuaries and statisticians are now

engaged much more in cross-disciplinary work. This fundamental change is well-represented

in the faculty complement at Simon Fraser University. The creation of a new Department of

Statistics and Actuarial Science would foster the further development of this dynamic field.

Specifically, the Department would promote the uniting of various faculty teaching statistics

across the campus. It could promote cross-disciplinary cooperation more effectively than the

existing administrative arrangement. It would provide more immediate credibility with po-

tential research collaborators, co-op employers, and other external contacts. It would create a

more secure environment for recruiting young faculty were they to know that career evaluations

would be conducted primarily by a committee of their peers. Finally, the priorities for Statistics

and Actuarial Science are now quite different from those for Mathematics. The Department

would foster the development of facilities truly geared to the unique needs of statistics and ac-

tuarial science (e.g. undergraduate laboratories for statistical and actuarial computing and for

providing direct experience with running sample surveys and designing experiments). It would

S ?

create an opportunity for Simon Fraser University to become a leading center for statistical

and actuarial research and education in western Canada.

The development of this initiative to form a department of Statistics and Actuarial Science

has a history at Simon Fraser University, and indeed, has by no means been hasty. The

Report of the External Review Committee for the Department of Mathematics and Statistics,

submitted April, 1998, listed as one of its eight recommendations that

'The University should consider establishing a separate Department of Statistics

within the Faculty of Science.'

This was the second consecutive time that external reviewers recommended further autonomy

for Statistics. Following the previous review, the statistics group developed a proposal for

an Institute for Statistics. This was approved by the Faculty of Science on February 25, 1991

(Paper FSC 3-91). Furthermore, immediately prior to the vote, a spokesperson for the statistics

group made it clear that the Institute was to be viewed as a first step toward the creation of a

fully autonomous Department of Statistics. Hence, the Faculty of Science has already expressed

approval in principle for this development. Unfortunately, subsequent financial exigencies led to

the indefinite postponement of this initiative. The external reviewers have, in effect, reminded

us all that it is time to complete the task, and stimulated the development of this proposal.

In the following section we identify and describe the faculty affiliated with the proposed

new Department. Section 3 identifies our immediate goals and initiatives. Section 4 discusses

S

structural requirements. It should be emphasized that no new programs are being proposed.

M.Sc. and Ph.D. programs already exist and are flourishing. We have recently created under-

graduate minor, majors and honors programs, at the request of our students, because of the

---

benefit they receive from an official designation on their academic record as having completed

these specialized programs. Our proposal is for an administrative reorganization which will

lead to a clear distinction between the new Department and that of Mathematics. Only a

modest administrative expenditure will be required to implement this proposal. The proposal

ends with a summary of the goals of the discipline of statistics and actuarial science. A sep-

arate document, describing the teaching programs of the proposed Department of Statistics

and Actuarial Science, and our teaching philosophy, is attached. The mathematicians in the

Department support the creation of the Department of Statistics and Actuarial Science, and

this move to create two new units has not been controversial. We anticipate that the good

relationship which currently exists between Mathematics and Statistics and Actuarial Science

should continue after the Department of Statistics and Actuarial Science is established.

2.

Faculty Affiliated with the

Proposed Department of Statistics and Actuarial

Science.

The statisticians and actuarial scientist affiliated with the proposed new department form

an active research group with a variety of interests, including fisheries and resource manage-

ment, model checking, survey sampling, applied probability modeling, statistical computing,

actuarial science, foundations of statistics, Bayesian methods, design of experiments, industrial

statistics, biostatistics, genetics, epidemiology and data analysis. We have carefully nurtured

links with governmental agencies, industry and medical practitioners through helpful contacts

and selective hiring. Although these interactions have been fostered by individual faculty mem-

bers, we have created a focus for such development through the Statistical Consulting Service.

As a by-product of these endeavours we have secured the first graduate cooperative program in

the university and apparently even now it is the only thriving one. This strong applied focus

to the Masters graduate Statistics program distinguishes it from the more theoretical bent that

UBC has taken, thus providing a complementary program in the province.

The ten regular faculty operate as a cohesive unit and have been heavily involved in work for

national and international Statistical and Actuarial Societies and for NSERC - two Presidents

of the Statistical Society of Canada (SSC); President, SSC Survey Methods Section; President,

SSC Biostatistics Section; Chair, SSC Elections Committee; Chair, SSC Committee on Women

in Statistics; Chair, SSC Education Committee; Chair, SSC Awards Committee; SSC Research

Committee; SSC Programme Chair and Secretary; SSC Board Member; Chair, Joint Meetings

Advisory Committee (a joint committee of five North American statistical societies); Advisory

Committee on Statistical Methods for Statistics Canada; NSERC Selection Committees and

Review Panels (five of the eight faculty who are Associate or Full Professors have served on

these); Review Panels for the Research Programs Directorate of Health Canada, and previously

Health and Welfare Canada; FRBC Selection Committee; Science Council of B.C.; International

Biometric Society Regional Advisory Board; Advisory Panel to the U.S. Energy Information

Administration; Associate Editor, Canadian J. Statistics; Associate Editor, Statistica Sinica;

Associate Editor, Biometrics; Associate Editor, Survey Methodology; Editor and Associate

Editor, Liaison; Associate Editor, Journal of Statistical Education; Statistical Appraiser for

the American Association of Public Health; Associate Editor, Statistics in Medicine, Special

Edition on Disease Mapping.

2 ?

(a.

---

The statisticians have been involved in several research initiatives locally: Drs. Routledge

and Schwarz at the Pacific Salmon Commission and the Department of Fisheries and Oceans;

• Dr. Dean at the B.C. Research Institute for Child and Family Health, St. Paul's Hospital

and the B.C. Ministry of Health; Dr. Sitter at the Pulp and Paper Research Institute of

Canada, MacMillan Bloedel and Pacifica Paper; Dr. Weldon at Environment Canada, MacMil-

lan Bloedel Research and the B.C. Ministry of Forests; Dr. Schwarz at B.C. Hydro, the Pacific

Urchin Harvesters Association and Kiahoose Aquaculture. The discipline is rapidly evolving

toward a more interactive role with the sciences and technology. Our group is well-positioned

to take advantage of these newly evolving opportunities. In this regard we are particularly for-

tunate in our recent recruitments and in the continuing, active involvement of our two emeritus

professors.

The research expertise of our group and descriptions of current research programs are out-

lined in Appendix A. We highlight here that Drs. Stephens and Lockhart have established

a highly successful long-term collaboration on goodness-of-fit testing. Drs. Routledge and

Schwarz are heavily involved in fisheries management. Dr. Swartz works in the topical area of

statistical computing and algorithms for computationally intensive problems. Dr. Sitter works

in survey sampling and experimental design, with much of this research directly applicable

to industrial experimentation. Dr. Weldon's focus is on foundations of statistics and applied

probability and statistics. Dr. Dean works in health statistics, in particular the assessment

of spatial patterns in disease incidence and mortality. Dr. Graham is a statistical geneticist,

while Dr. McNeney works in epidemiologic study design. Dr. Wirch, our professor of actuarial

science, develops and assesses risk measures for financial risk management.

.

3. Future Development: Goals and Initiatives

The Department's long-term and on-going goals will be management of the Statistics and

Actuarial Science undergraduate and graduate degree programs, liaison with other departments

and groups which we service, creation of a centre for excellence in research, and outreach in

statistics and actuarial science. We will focus our efforts on areas where we perceive there to

be a strong demand for highly qualified personnel. The immediate short-term initiatives are

described below.

3.1 Biostatistics and SFU's Health Initiative

We plan to develop a graduate focus in Biostatistics, with possible linkages with the Univer-

sity's Health Initiative. The proposal to develop a graduate focus in Biostatistics will permit a

broadening of offerings in that field and in epidemiology and statistical genetics. Students will

also have training in this area through their work for the Consulting Service, which we propose

to have linkages to Children's Hospital (see section 3.2), likely leading to a greater number of

theses in that field. We also envision cross-disciplinary training of graduate students entering

the biostatistical field with supervisory committee members encompassing a variety of fields

. ?

including basic sciences, epidemiology, economics, statistics. In this era when cross-disciplinary

training is so important, we expect that our graduates from such a novel Biostatistical program

3 ?

1.

---

will be heavily recruited. There will also be enhanced recruitment to our graduate school. There

are few programs in Biostatistics in Canada, and none in the west. Here is a potential to build

something unique to Western Canada. The program will also be highly beneficial to Canada

as a whole, by reducing the flow of Canadian students to the US for training in this important

area. We will in addition experiment with offering courses in biostatistics at medical institutes

in the Vancouver region to draw as an audience professional biostatisticians working at those

institutes. There have been requests for such courses and we have responded with interest; in

particular, we are currently consulting with Vancouver General Hospital in this regard. We

intend to offer such courses jointly for our own graduates students and the external students.

This will foster interactions between our graduate students and practicing biostatisticians. Our

single course in survival analysis has already provided employment for many of our students.

We have recently made a biostatistical adjunct appointment, Dr. Rob Baishaw, whose main

'home' of employment is a Health Consulting Firm, Synectics Health Corporation, in Vancou-

ver. Another adjunct appointment has been made for a researcher at the BC Cancer Agency,

Dr. John Spinelli, whose MRC funding has supported co-op work-terms for our students in the

past, and with whom we have had on-going research collaborations. Our new NSERC UFA

appointment, Dr. Graham, is a statistical geneticist with an epidemiological flavour. With her

appointment, and Dr. McNeney's, a statistician whose main research focus is statistical issues

in epidemiology, particularly study design, we envision further links to the medical community.

3.2 Industrial Outreach: The Statistical Consulting Service

In general, we have two main goals for this service: i) to bring an understanding of everyday

concepts in statistics to as wide an audience as possible, to forge closer ties with the internal

and external community of users of statistical methodology, to use modern statistical methods

to improve quality and productivity and thus contribute to provincial economic development;

and ii) to provide a place where our students may receive valuable professional training in

consulting. We have aimed at producing 'top flight' students and have been successful, through

the introduction of co-op and the training of students in the Consulting Service. Appendix B

provides a brief description of the service and examples of recent projects completed by the

service.

In the last 2 years the SCS has evolved into a partially self-funding service by charging faculty

and graduate students from other disciplines for services rendered, and providing contract

services to industry and government. Already the SCS has had contracts with such diverse

clients as: The Office of National Statistics for the United Kingdom, The Pulp and Paper

Research Institute of Canada, B.C. Hydro, MacMillan Bloedel, The Department of Fisheries

and Oceans, the Pacific Urchin Harvesters Assoc. and Klahoose Aquaculture. This activity has

many benefits to the Department: raises the profile both nationally and internationally; exposes

our graduate students to current applied problems and possible future employers; funds our

graduate students involved in particular contracts; exposes our faculty to currently emerging

applied problems; provides sound statistical advice to other disciplines at a subsidized rate;

and partially funds the position of Director of the SCS who also teaches two consulting courses

to our graduate students as part of his duties.

Our first goal is secure funding for the SCS so that we can make the position of Director

permanent. At present, it is a renewable term position with funding coming from the De-

partment of Mathematics and Statistics ($20K/yr) in part for the teaching of Stat 811 and

Stat 812. (Each of these 2 hour/week graduate courses in Statistical Consulting is offered two

4

---

terms per year; the two courses are required in sequence

for

all M.Sc. students in Statistics.)

The remainder of the SCS funding has come from consulting fees charged to off-campus clients

• (see above) and to on-campus clients at a reduced rate ($25K/yr after SFU overhead charges).

Note that about $5.5K per year has also been generated as support for graduate students from

off-campus clients.

We have made some progress toward our goal of stable funding for the SCS. The ALRP of

the department has agreed to let us use the remainder of a statistics faculty retirement position

(part of it was used for budget cuts - approximately 820K is left) and the current support

given to the service, originally from the Dean's office ($20K) as a base for securing permanent

funding for the Directorship. Children's Hospital has expressed a strong interest in a joint

venture to provide the remaining funding needed (approx. $25K). It is envisioned that the

Director of the SCS will spend part of his working week at Children's Hospital. This will have

the great benefit both for Children's and SFU that statistics graduate students become involved

in research projects emanating from Children's, since graduate training for our students is an

important feature of the Consulting service. Graduate students will obtain experience working

with medical investigators and may be able to use studies emanating from such collaborations

as bases for their M-SC. projects or theses. Such in-depth attention to studies will benefit

both the students and the investigators at Children's. Children's will also benefit from the

provision of biostatistical expertise. This arrangement will also permit closer collaboration

between faculty at the two institutions. Such a venture can be fostered by the University's

Health Initiative.

Once permanent funding is established, we envision using funds generated from external

contracts to regularly support chosen graduate students, perhaps in conjunction with our grad-

uate co-op program, to work in the SCS handling the administration and consulting for the

multitude of small on-campus projects and smaller off-campus projects. This has been done

for a few specific projects in the past 2 years: Derek Bingham was funded for one term for

his work on a project with Macmillan Bloedel; Changbao Wu was funded for one term for his

work on a project with the Office of National Statistics in the UK, Peter Lui, Chuck Paltiel,

Jason Sutherland and Ruihua Yin were funded to work on smaller projects, including two from

Children's Hospital. We hope that by hiring appropriately chosen statistics graduate students,

we can free-up a portion of the Director's time to pursue, jointly with faculty, larger long-term

projects which have substantial research components.

Recent activity at the SCS has had a major effect in the arena of public advocacy. In

response to the 1992 Pearse-Larkin Report, Dr. Routledge was asked by the Pacific Salmon

Commission to help assess the potential for errors in their estimates of fish passage at Mission.

This project formed the basis of a Ph.D. student's dissertation, and led to Dr. Routledge

being asked to serve on the Fraser River Sockeye Public Review Board in 1994. One of the

Review Board's key recommendations was for the creation of a Pacific Fisheries Conservation

Council to provide independent assessments of the health of our fish stocks and fish habitat.

Dr. Routledge was seconded to serve on this council for a three-year term starting September,

1998.

3.3 Actuarial Research and Education

We are working to create an endowed Chair in Actuarial Science with strong links to the

.

?

Faculty of Business. SFU is the only university in British Columbia to offer a program of stud-

ies in Actuarial Science. Students entering the actuarial program are typically very talented

q.

5

---

and interested in business careers. They also have an interest in the mathematical sciences and

show good statistical and computing skills. They are generally employed by financial institu-

tions, pension and employee benefits consulting firms, and employee benefits departments of

large corporations or government. They are expected to write the professional examinations

set by the Society of Actuaries. Our graduates rank among those who have received the highest

scores in the world on the first of these international examinations. Last year, five actuarial

firms came to campus to recruit our graduating students. Our multidisciplinary approach re-

quires students to build a foundation of statistics, business, economics, computing, and literary

skills. Our new actuarial appointment, Dr. Julia Wirch, with bachelor's degrees in commerce

and in mathematics and a masters and doctorate in Actuarial Science, and with considerable

experience in the actuarial industry, will bring both academic and practical strength to the

program. Our Certificate program in Actuarial Science provides a convenient conduit for stu-

dents who have completed degrees in other areas, such as business, economics or computing,

to pursue careers in the actuarial field.

We are exploring an opportunity to secure sufficient endowment to hire a first-rate Economic

Actuary and to use this hiring to build bridges to Economic Policy and to the Business faculty in

general. We have been working on this initiative with the SFU Development Office. With such

a position in place there will be opportunities for the development of professional courses for

practicing actuaries and for specialized courses in employee benefits and pension and retirement

plans.

3.4

Experimentation in Teaching

We plan to expand the role of the Statistics Workshop. We have been utilizing a concept

called 'Activity-Based Statistics' in a few of our courses. This requires experimentation, and

has been an extremely successful teaching tool. In particular, those students whose home de-

partments are in the other sciences readily appreciate the focus on concepts, the experience

with real experiments and surveys, and the proficiency they develop in computer-assisted data

analysis. However, the space occupied by the Statistics Workshop is not suited to experimenta-

tion, and, in addition, we would like to expand the role of the workshop to accommodate such

experimentation in many of our courses. To this end, we will request support in the future for

expanding the workshop space and facilities. We require an additional space about the same

size as the existing workshop to run the experiments.

We have been offering every semester since 97-3, a completely online version of our basic

service course STAT 101. Recent developments include the use of a JAVA-based suite of

interactive modules for learning statistics. Study groups of students are involved through use

of online conferencing, and interaction with a tutor on an almost continuous basis is achieved.

Our statistics workshop is not currently used for this initiative but with some upgrading could

do so.

3.5 Course Development

Course development is an on-going process and priority. Currently, we plan to develop

a new first-year undergraduate course in Statistics, aimed at potential majors in Statistics.

The course would stress the broad applicability of concepts in probability and data analysis,

and would include the use of statistical software for simulation experiments and data analysis

our

projects.

introductory

In the past,

course

the

(STAT

need for

270)

calculus

be at

in

the

our

200-level.

introductory

However,

statistics

the advent

course

of

required

statistical

that

?

is

6 ?

10.

---

software has made in-depth coverage of probability and data-analytic ideas accessible without

calculus. Moreover, because of changes to the school mathematics curriculum, more statistical

ideas are being taught at high school, and students are ready for these ideas in first year. The

development of this course will impact the current core courses in our program.

Because of changes in the high school curriculum to provide greater exposure to statistical

ideas, we will also be exploring, with the Faculty of Education, the creation of a course for

high-school educators on the teaching of statistical concepts and experimentation.

In addition, we propose to replace several introductory courses in Statistics for non-majors

by a new course. In this course, the technical expertise developed will be similar to that

developed in current introductory statistics courses. Students from a variety of disciplines will

receive three hours of lectures per week jointly. However, there will be scheduled tutorials

which are subject-specific, each with a maximum of thirty students. In these tutorials, specific

examples taken from the students' home departments will be explored. The tutorials may

include experimentation. The content of the tutorials will be developed in consultation with

home departments. Assignments will also be discipline specific. In conjunction with this, the

role of the senior lecturer in coordinating and overseeing these courses will be reviewed.

Finally, we are revising our offerings in Actuarial Science to correspond with the new syllabus

for the actuarial examinations. These revisions will include proposals for the introduction of

some graduate level Actuarial Science courses.

8.6 Bridges with other Departments

We envision creating several cross-appointments to tie together those working at SFU in

the discipline of Statistics. Adjunct appointments in statistics and actuarial science will also

. form solid links with our industrial contacts. These will allow them an opportunity to share

in the development of the disciplines and to help shape the graduates that emerge. There is

significant potential for interdisciplinary and external partnerships here. With the creation of

cross-appointments we plan to explore further program development through team-teaching

and other flexible arrangements with other departments. For example, we have already been

exploring the use of web-based and other innovational instructional approaches with some

success, and would like to further this development. With an autonomous department, we shall

be able to present more credibly our applied nature to other departments, to researchers both

on and off campus, and to our external contacts.

3.7 Linkages with UBC

We have begun experimenting with offering our courses to graduate students at UBC, since

some of our courses are unique in the province because of our special focus in applied statistics.

Graduate students from UBC attended STAT 890 Biometrics in 98-2 and will do so again in

00-2. We. held all lectures and tutorials on a single day every week to facilitate attendance

by UBC students, as they requested. They have also expressed an interest in attending our

Longitudinal and Life-History Data Analysis course and we have agreed to try to schedule it

at a convenient time for them in the future. Conversely, we encourage our students to take

advantage of courses available at UBC whenever these can be used to complement our own

offerings.

I

It is worth highlighting the areas we have recently helped to develop and wish to focus on:

(i) environmental statistics, through the environmental science program, and (ii) biostatistics,

though our specialized courses, co-op positions, adjunct appointments and our vision for our

7 ?

ii.

---

graduate program. These are areas which NSERC itself targets as ones in which development

in statistics should focus. There were four emerging areas identified as important by NSERC:

statistical methodology for massive data sets and information technology, genetics and medical

science, stochastic modeling, and statistics and probability in environmental science. We have

focussed on two of these, blending our interests with those of other members of the Faculty of

Science.

3.8 Training and the Job Market

We will promote the growth of graduate and undergraduate co-op opportunities to their

full potential. Actuaries and statisticians find employment in a wide variety of fields, and the

relatively few graduates from Canadian statistical and actuarial programs are in high demand.

Our own graduates have been highly successful in gaining employment, not only because of

the state of the market and of the profession, but also because we provide opportunities for

training and experience in industry and government. We have one of the few thriving graduate

co-op Masters' programs at SFU and one of only two in Statistics in Canada. We aim to

become nationally and internationally recognized as a centre for excellence in our Applied

Co-op Masters' Program. We believe that we are well underway. Recently, admission into

our Statistics Masters' program has been highly competitive. Our graduates typically find

employment before they actually complete their studies. Appendix C provides a list of our

recent graduates and their current employment. Appendix D lists graduate co-op placements

since the program's inception in 1995. Our Statistics Masters' degree is similar in nature to a

professional accreditation and the co-op experience is vital for the students. There are 5 or more

placements per year from our Masters' student body, which is usually comprised of about 15

students. We wish to build upon our ability to offer students good career prospects. To this end,

we are aggressive in seeking co-op placements for our graduate and undergraduate students. In

this work, we find that employers sometimes state that they prefer students from departments

or programs of Statistics. See Appendix E for four letters from recent doctoral graduates in this

regard. Though we are well-known in the statistical community as an emerging group, housing

our students under an official Department will enhance their employment opportunities.

4. Structural Requirements

.

1. Physical Space

It would be desirable that the proposed Department of Statistics and Actuarial Science be

given new space. However, contingency plans which separate the current departmental space

into space for the two units are developed in Appendix F.

Creating physically separate Departments of Statistics and Actuarial Science and of Math-

ematics from the existing Department of Mathematics and Statistics will be challenging. The

main reasons for this are: (a) the need to create two separate departmental offices; and (b) the

fact that the current Department of Mathematics and Statistics is already under-spaced. In

light of these two points we have tried to identify the space currently being used solely for the

Statistics and Actuarial Science programs within the present department and the space which

8 ?

II-.

---

is currently shared between mathematics, statistics and actuarial science. In Appendix F, we

- ?

then propose how the space could be reallocated in such a way that the new Department of

• Statistics and Actuarial Science shoulders some of the current space burden. In addition we

have tried to build into our proposal a net gain in usable space for both departments. We finish

with an evaluation of additional space that the Department of Statistics and Actuarial Science

will need as soon as it can be obtained.

4

.

Administrative Staff

We shall need one full-time, and one half-time secretary and one Departmental Assistant.

This seems to reflect current practice for units the size of the proposed department.

.3 Technical Support

A half-time UNIX/Mac/PC technician. The technician manages computing support re-

quired for teaching and for research endeavours.

4

.

4

Budget

A general budget description is provided in Appendix G. However, no specific details are

listed there regarding the Teaching Assistant and Operating Budget allocations for Statistics

and Actuarial Science. These allocations, and that for Mathematics, would follow the normal

criteria for the Faculty of Science.

Note that no additional resources for the library will be required for the formation of the

. Department since no new programs are being planned. The incremental cost for creating a

Department of Statistics and Actuarial Science and one of Mathematics is the salary for a

departmental assistant and a half-time secretary, and minor renovation and equipment non-

recurring costs.

5. The Vision of the Department of Statistics and Actuarial Science

As the 20th century comes to a close, Statistics and Actuarial Science has completed its

development into a dynamic, highly interactive discipline. Simon Fraser University is well-

positioned to move forward on this front. The core of the proposed department has a strong

focus on applied research with special expertise in health, genetics, industrial experimenta-

tion, sample surveys, resource management, and actuarial science. This expertise is enhanced

through o

-

ur Statistical Consulting Service, adjunct professors in biomedical research, and in-

formal interactions with faculty in other departments at the university.

We see the new Department of Statistics and Actuarial Science at Simon Fraser University

emerging as a high-profile, internationally recognized centre for applied research, and a grad-

uate and undergraduate training centre with graduates placed throughout local industry and

government and further afield.

The following page provides a declaration of internal support by all the current members of

our group for the formation of the Department of Statistics and Actuarial Science.

9 ?

13.

---

Department of Statistics and Actuarial Science

?

Declaration of Internal Support

We, the undersigned', support the creation of a Department of Statistics

and

Actuarial

Science at Simon Fraser University, and wish to be a part of this new venture.

,J1

?

-

Charmaine Dean, Associate Professor

and Director of Statistics

4

1

7/ZA)

MJ4M/1

Jir4/o Graham, Ais)ant Professor

(72,42.

Richard Lockhart, Professor

Brad McNeney, Assistant Professor

btI/

Rick Routledge, Professor

Randy 4(er,'Associate Professor

Tim SWartz. Ms*iate Professor

Larry Weldon, Associate Professor

Julia l

irch, Ass

tan

t 1rofessor

L

Ian Bercovitz, Directdr'of Statistical

Consulting Service

Rob Baishaw, Adjunct Professor

John SpinelJAdjunct Professor

Michael Stephens, Professor Emeritus

..t&&eo

P/

&^ ^

Cesareo Villegas, Professor Em ritus

.

This list represents the entire complement of SFU faculty

and

academic support staff associated with

statistics who are currently employed in the Department of Mathematics and Statistics at Simon Fraser

University. The venture has the unanimous support of this group.

?

0

---

Teaching Programs in the Proposed

Department of Statistics and Actuarial Science

?

at Simon Fraser University

1. Our Philosophy as a Department with regards to our Teaching Responsibilities

We have adopted an organizational management philosophy of continual quality improve-

ment as a means to reach our goals in our teaching responsibilities. We are committed to (i)

excellence, (ii) exceeding 'customer needs', (iii) the creating of a reputation for total quality,

(iv) working together as a team, and (v) providing members in all categories with opportunities

to improve relevant skills. Goals related to undergraduate and graduate teaching are itemized

below.

Our Undergraduate Service Course Goals are

• to provide a curriculum that reflects the needs of "customers": students, faculty teaching

those courses which depend upon this course as a prerequisite, future employers, and

Canadian society at large

• to increase the numeracy skills of students and foster an ability to think statistically

.

?

• to enable students to understand and evaluate statistical descriptions and the information

they contain

• to enable students to use statistical tools to help with problems encountered at work,

research, and everyday life

• to engender an appreciation for the usefulness and relevance of statistics

• to teach students new and interesting facts about the world via the use of real data (e.g.

through Chance News and newspaper and magazine articles)

Goals of the Undergraduate Major Program are

• to provide a curriculum that gives a basic training for entry level positions as Statisticians

or entry to a graduate program at other institutions

• to provide an emphasis on applied statistics

• to provide attractive lower division courses that will stimulate capable students to pursue

our major program

• to extensively use real life data and experimentation in our courses

• to extensively use computer packages in our courses

Graduate Course Goals are

1

?

/

---

• to provide graduate instruction to develop high caliber applied statisticians

• to increase the written and oral communication skills of the graduate students

• to increase the computer competency of our students

• to enable M.Sc. graduates to write a research report based on a problem they have

analyzed.

• to enable Ph.D. graduates to start an academic career. Upon completion of Ph.D. a

student should have the potential of three papers in their thesis.

Delivery Goals

• that all staff involved in the course are committed to continuous improvement and the

development of both the course and their own skills in areas that impact the course

• that all students are treated with respect and that no student encounters unnecessary

barriers or problems in regard to their learning or experience of the course

• that all students are helped to meet the goals of the course and program as efficiently as

possible

• that the course be conducted in a manner that encourages everyone involved (students,

lecturers, tutors, markers and support staff) to feel that they are integral members of a

team, that they have both rights and responsibilities in ensuring its effectiveness and its

improvement

• graduate students should receive adequate financial support (as allowed by our resources)

through TA and RAs.

In general we try to ensure that quality problems are promptly investigated and appropriate

action taken, as far as possible, to prevent recurrences. We also share material related to

courses and course development and provide an environment where the learning experience

and particularly the successes of the course team in improving the quality of this course be

documented as far as possible so they can be made available to assist future providers of the

course and others wanting to improve the quality of their own courses. We conduct regular

"one minute papers" in our courses to receive feedback from students on what is going right

and wrong in courses. Mid-semester post-mortems are held at Statistics group meetings to

discuss course progress. Each service course is assigned a steering committee consisting of a

statistician and members of client departments. These steering committees are responsible for

development and modifications to the course outlines, and for assessing student progress and

feedback. -

2.

Undergraduate

Programs

The department offers honors, major and minor degrees in Actuarial Science and in Statis-

tics. The Actuarial Science degree is offered within the Faculty of Science while the statistics

2

---

degrees can be held within the Faculty of Science or the Faculty of Arts. We also offer interdisci-

plinary degrees with Statistical content in Mathematics and Computing Science, Environmental

• Science and Management and Systems Science, and a certificate program in Actuarial Science.

Both our Statistics and Actuarial Science degree programs have only recently been formally

approved. This is of substantial importance to our students who will benefit from an official

designation as having completed these specialized programs. Our interdisciplinary programs

offer students flexibility and encourage them to extend their focus of knowledge. It should be

noted that we offer a large number of courses for servicing the statistical requirements of other

programs. In the sections which follow are descriptions of our programs and their sizes, how

we operate our workshop and our mechanism for advising students.

2.1 Core Programs

The Actuarial programs are designed to give students a strong foundation in actuarial

science and to prepare them to apply their actuarial skills in a business context to problems

concerning the evaluation of financial risk and financial risk management, particularly as it

relates to the insurance industry. In particular, the core courses have been chosen to prepare

students for the examinations of the Society of Actuaries/ Canadian Institute of Actuaries and

the Casuality Actuary Society.

The Statistics programs have a distinctly applied flavour. They emphasize, along with the

necessary foundations and theory, specific techniques that will make our graduates attractive to

potential employers. Many students take advantage of valuable co-op education opportunities

in business, industry, and government agencies. Our statistics major and honors programs also

provide a solid basis for post-graduate work. Graduates have been accepted at such major

. schools as Waterloo and Carnegie-Mellon. Statistical methodology plays a fundamental role in

such key fields as quality control, industrial experimentation, environmental health, and natural

resource management. One of the requirements of the program is a minor in another field to

provide the student with a good background in an area of application. We are continually

working to ensure that our undergraduate programs are relevant to important potential fields

of application.

Appendix I lists the requirements of these programs.

2.2 Joint Programs

The department administers two quantitative streams in the university's Environmental

Science undergraduate program: Environmetrics and Quantitative Techniques for Resource

Management. This is a recently established program. It is distinguished by its heavy emphasis

on core science disciplines. All students must take, e.g., at least 2 semesters of calculus and

3 semesters of statistics. The quantitative streams provide in-depth training in statistics and

quantitative modelling. The aim of it is to prepare students for such tasks as designing a sample

survey of a contaminated site or studying the dynamics of a highly unpredictable fish stock.

The Management and Systems Science Program (MSSC) is an interdisciplinary program

involving the Faculty of Business, the School of Computing Science, the Department of Eco-

nomics, and the Department of Mathematics and Statistics. It is administered through Mathe-

matics and Statistics. The Program is designed to produce graduates with solid problem solving

and computational skills for the increasing information technology spreading through the busi-

ness world. The courses in Business and Economics in the program build a general knowledge

3 ?

(.

---

Table 1: Average Enrolment in Undergraduate Courses between 94/95 and 97/98 by Course

Type and Level

?

0

Course Year

?

1- ?

2- ?

3- 4-

ACMA 1994/95 ?

14.0

1995/96 ?

10.5

1996/97 ?

15.3

1997/98 ?

15.8

MACM 1994/95 ?

51.3

1995/96 ?

78.0 ?

69.3

1996/97 ?

105.0 ?

55.3

1997/98 120.5 178.5 ?

97.8

MATH 1994/95 111.5 82.4 42.2 15.7

1995/96 117.8 ?

68.4 ?

38.7 14.5

1996/97 123.8

?

88.2 ?

40.4 11.1

1997/98 144.1 108.5

?

45.9 15.1

STAT ?

1994/95 147.7 98.5 50.2 12.5

1995/96 188.0 ?

85.6 ?

65.7 16.4

1996/97 164.5 136.0

?

57.3 15.8

1997/98 234.0 120.0 105.8 17.8

in these two areas. The courses in Computing are oriented towards software and algorithms.

The courses from Mathematics and Statistics emphasize discrete mathematics, optimization

and statistics. The Program has been well received by the business community. The students

in the co-op program find placements easily and graduates of the program have been quick to

find employment.

The department offers, jointly with the School of Computing Science, an honors program

in Mathematics and Computing Science (MACM). This program is designed for students inter-

ested in the mathematical aspects of computing or the computational aspects of mathematics.

It establishes a solid base in both disciplines, and then allows the choice of a specialty in a wide

variety of areas, including statistics.

We intend to continue our participation in these valuable programs, and will initiate dis-

cussions with other stakeholders to address whether significant changes are required to the

programs resulting from the formation of a Department of Statistics and Actuarial Science.

2.3 Class .Sizes

Details on enrolment for our various course types, and those for mathematics, for compari-

son, are given in Table 1. We are experiencing larger enrolments in many of our courses, and

even some of our 300 level courses have enrolments of over 100 students. Large courses include

STAT 203, with enrolments of about 100 students, STAT 270 (Introduction to Probability

and Statistics), offered every term with typical enrolments of over 150 students, STAT 301

(Statistics for the Life Sciences), with typical enrollment of over 300 students.

4 ?

19.

---

24 The Statistics Workshop

The department runs scheduled tutorials for some upper-year courses. In addition, the

Statistics Workshop provides instruction to students studying introductory level material. It is

open usually five days a week and is staffed with teaching assistants who provide instruction on a

one-to-one basis. The operation of the workshop is coordinated by a permanent faculty member,

the Laboratory Instructor or Workshop Coordinator, who also provides assistance to students

coming to the workshops. Marking of homework assignments, midterms and examinations is

also organized through the workshop. The Statistics Workshop and tutorials offer students

different styles of teaching by providing access to various teaching assistants in addition to

faculty. They also provide an opportunity for students to meet others from their classes, to

work in teams and study and learn together. The laboratory instructor and teaching assistants

encourage and organize cooperative learning. The laboratory instructor will often organize

review sessions, discussion groups and other activities to help students succeed in their course

or to enrich their learning experience. The courses covered in the Statistics Workshop are listed

in the table below.

Courses: STAT 101 Introduction to Statistics

STAT 203 Introduction to Statistics for the Social Sciences

STAT 270 Introduction to Probability and Statistics

STAT 301 Statistics for the Life Sciences

STAT 302 Analysis of Experimental and Observational Data

The courses covered in the Statistics Workshop provide an introduction to statistical meth-

ods for students enrolled in both the sciences and the social sciences. In any one semester it is

normal to service three of the above courses. Occasionally, only two courses may be serviced

depending upon course offerings.

Each of the courses serviced, except Statistics 270, makes quite extensive use of the statistical

software which is available on each of the workshop's fifteen computers. The courses also

require that the Workshop operate as a laboratory since students are expected to carry out

real experimentation; this is a major challenge for the tutors because of the large numbers

of students requiring help in the planning and design of their experiments. Such 'hands-on'

studies illustrate important points of applied statistics and statistical reasoning. They provide

the student with exposure to sampling, design, data collection, variability, model development,

selection of factors, outliers, interactions, the need for randomization and many other statistical

activities and concepts. The students tend to enjoy such work and through it, gain a reasonably

solid understanding of these statistical concepts and an appreciation of statistics.

Two examples of this occur in Statistics 301 where students are required to gather and

analyze both survey data and experimental data:

Survey- Students are required to gather and analyze data that they obtain about the types

of automobiles that park in 'B lot' at the University.

Experimental- Students conduct an experiment and analyze the data obtained on the effect

of wing design on the distance that paper planes will fly.

2.5 Directed Studies Courses

• The department offers directed studies courses to undergraduates on an individual basis.

These courses would not be counted as part of the faculty member's regular contribution to

teaching, and are purely voluntary. Faculty members are often sympathetic to offering courses

5

?

R.

---

to students who require them for completion of their program in a timely fashion, or offering

specialized courses to highly motivated students.

2.6 Advising

?

S

Much procedural advice on registering is available from staff in our general office. The

department also provides advice through faculty advisors who provide academic advice to

students taking our courses, including providing career and graduate school information and

approving programs. There are two faculty advisors in Statistics, and one for Actuarial Science.

We encourage our majors and honors students to speak with a departmental advisor at least

once per term. Usually advisors speak quite regularly with students, sometimes on an informal

basis, and are a vital part of their support structure.

2.7 Endowments

The department has a joint mathematics and statistics endowment fund. It is used for

funding undergraduate course awards at the lower level, student travel awards to attend the

Canadian Undergraduate Mathematics Conference, Putman awards, and it has been proposed

that some funds are used for awards to our majors. We propose that one-third of the endowment

be used for creating an endowment fund for the Statistics and Actuarial Science Department.

3. Graduate Programs

Statistics

We currently

?

offer programs in statistics leading to an M.Sc. or a Ph.D.

S

The M.Sc. degree has a strong applied focus and requires a project of statistical analysis

and 30 semester hours of course work including a minimum of 2 semesters participation in the

Statistical Consulting Service through the courses STAT 811-2 and 812-2. The program has an

optional co-op component which has been very successful, providing solid practical training to

participants. There are currently two sets of rules under which a statistics student within the

Department of Mathematics and Statistics can get an M.Sc. degree. One set of rules is common

to the entire department and requires a thesis and 20 units of course work of which at least 12

must be at the graduate level. The other set of rules is based on a project and more extensive

course work. We have not had any students graduate under the thesis rules in recent years and

propose not to maintain such a set of rules in our new department. However, we wish to make

sure it is still possible for a student to substitute a thesis for the project when the department

agrees that this would be in the student's best interest. We therefore propose to make some

small changes to the rules to make explicit the possibility of using the thesis.

At the Ph.D. level we propose no changes to the rules currently in place for statistics

students. We believe that these rules will serve well those Ph.D. students wishing to write

dissertations in areas on the Actuarial Science/ Statistics interface.

Actuarial Science

Students interested in Actuarial Science can currently do either an M.Sc. or a Ph.D. in the

Department of Mathematics and Statistics following either the department's general rules or

6 ?

)D.

---

the special statistics rules. We propose to modify the statistics rules to make it clear that the

project or thesis could focus on actuarial science. We further propose to create reading and

•

?

special topics numbers in Actuarial Science to give us the flexibility to accommodate students.

Here are the basic ingredients of our current programs together with the proposed modifi-

cations.

The M.Sc. Program

• We require 30 semester hours of credit. Of these at least 24 must be at the graduate level.

• We require completion of STAT 801-4 (Mathematical Statistics).

• We require completion of our two Statistical Consulting Courses: STAT 811-2 and STAT

812-2.

• We require completion of a project; our proposal is to modify the wording to permit the

substitution of a thesis for the project. A project is essentially a work of data analysis;

a thesis would require either a research component or a substantial survey of a method-

ological area.

• Students are able to participate in the Co-op program. This has been very successful; see

Appendix D of the proposal for the formation of a Department of Statistics and Actuarial

Science for details of recent placements.

• ?

The Ph.D. Program

• We require 30 semester hours of credit beyond the B.Sc. level. Of these at least 24 must

be at the graduate level. We currently give credit for 22 of these hours to students with

an M.Sc. Students with an M.Sc. then require 8 semester hours of credit at least 4 of

which are at the graduate level.

• We require successful completion of a general examination covering senior undergraduate

statistics material. We propose to make explicit our commitment to modify the syllabus

for Actuarial Science students.

• We will modify the language describing the thesis to include the possibilities of contribu-

tions to Actuarial Science rather than simply statistics.

Appendix II contains a proposed graduate statistics calendar entry which has received ap-

proval at the Faculty committee level. The proposed entry is only slightly changed from the

1999/200L entry. It also contains the current graduate statistics calendar entry.

Funding for Students

The department has always provided financial support for all its graduate students without

external funding. Some support is provided through research assistantships from the research

grants of individual faculty members. We have about 20 students at any one time, with 3

to 6 of these being doctoral students. In conjunction with our colleagues in Mathematics we

maintain an active recruiting program; we intend to continue that collaboration through PIMS.

7 ?

Al.

---

We anticipate no significant change in the size of the program. We seek to maintain current

annual teaching assistantship and Graduate Fellowship funding levels for our students.

Student Employment

?

fl

Graduates of our program have been very successful in the job market. In Appendix C of

the proposal for the formation of a Department we show recent graduates and their current

employment. We are particularly proud of the success rate of our Ph.D. graduates in gaining

faculty positions.

.

is

8

---

Appendix I

.

?

Undergraduate Program Requirements

.

0

---

University

calendar

?

J'ZsIsIa

Statistics Program

applied in a broad range of fields. To fully appreciate their application, it is important that students

also gain advanced training in an area of potential application. To this end, students pursuing a

major or honors in statistics are required to complete a minor in a field other than mathematics and

.

statistics. In keeping with the almost universal applicability of statistical methodology, there are no

other restrictions on the selection of a minor. Students are encouraged to discuss the selection of a

minor with an advisor early in their program.

Prerequisite Grade Requirement

Students must have obtained a grade of C- or better in prerequisites for courses offered by the

Department of Mathematics and Statistics.

Faculty of Science Requirements

Students must satisfy the Faculty of Science upper division credit, breadth and grade point average

requirements.

Major Program

Students will also be required by the Department of Mathematics and Statistics to obtain credit for

the following courses.

a) Lower Division Requirements

Mathematics

Students must complete one of

• MATH 151-3 Calculus I

• MATH

154-3

Calculus I for the Biological Sciences

• MATH

157-3

Calculus for the Social Sciences I

plus one of

• MATH

152-3

Calculus II

• MATH

155-3

Calculus II for the Biological Sciences

• MATH

158-3

Calculus for the Social Sciences II

plus both of

• MATH 232-3 Elementary Linear Algebra

• MATH 251-3 Calculus III

Statistics

Students must complete both of

• STAT 270-3 Introduction to Probability and Statistics

• STAT 280-3 Applied Probability Models

Computing Science

(

.1'.

---

Students must complete one of

• CMPT 100-3 Software Packages and Programming

• CMPT 101-4 Introduction to Computer Programming

• CMPT 102-3 Introduction to

Scientific

Computer Programming

b)

Upper Division Requirements

Mathematics and Computing Science

Students must complete

• MACM 316-3 Numerical Analysis I

Statistics

Students must complete all of

• STAT 330-3 Introduction to Statistical Inference

• STAT 350-3 Linear Models in Applied Statistics

• STAT 402-3 Generalized Linear and Nonlinear Modelling

• STAT 410-3 Statistical Analysis of Sample Surveys

• STAT 430-3 Statistical Design and Analysis of Experiments

• STAT 450-3 Statistical Theory

c)

Students are also required to complete a minor in some discipline other than mathematics or

statistics. The certificate in actuarial mathematics may also be used to fulfil this requirement.

d)

Faculty of Science requirements stipulate that at least three other upper division courses be taken

in mathematics, statistics, actuarial mathematics, or mathematics and computing science. Students

should consult a departmental advisor before selecting these courses. STAT 340,420 and 460 are

recommended. Neither STAT 301 nor STAT 302 may be used to fulfil this requirement.

Honors Program

A bachelor of science with honors in statistics requires 132 credit hours. See general regulations in

the

Faculty of Science

section for further breadth, upper division credit, and other requirements.

In addition to the requirements a), b) and c) for a major, candidates for an honors degree in statistics

will be required to obtain credit for the following.

d) Additional Mathematics Requirements

Students must complete all of

• MATH 242-3 Introduction to Analysis

• MATH 320-3 Advanced Calculus of One Variable

• MATH 322-3 Complex Variables

plus one of

?

0

---

• MATH 332-3 Introduction to Applied Algebraic Systems

• MATH 339-3 Groups and Symmetry

• MATH 438-3 Linear Algebra

0

e) Additional Statistics Requirements

?

Students must complete both of

• STAT 420-3 Non-Parametric Statistics

• STAT 460-3 Decision Analysis and Bayesian Inference

f) Faculty of Science requirements stipulate that at least three other upper division courses be taken

in statistics, mathematics, actuarial mathematics or mathematics and computing science. These

courses should be chosen with the assistance of a statistics advisor. STAT 340 is recommended.

Neither STAT 301 nor STAT 302 may be used to fulfil this requirement.

Minor Program

Candidates for a minor in statistics are subject to the general regulations of the faculty in which

they are registered. In addition, students will be required by the Department of Mathematics and

Statistics to obtain credit for the following courses.

Mathematics Requirements

Students must complete one of

. ?

• MATH 151-3 Calculus I

• MATH

154-3

Calculus I for the Biological Sciences

• MATH 157-3 Calculus for the Social Sciences I

plus one of

• MATH 152-3 Calculus II

• MATH 155-3 Calculus II for the Biological Sciences

• MATH 158-3 Calculus for the Social Sciences II

plus both of

• MATH 232-3 Elementary Linear Algebra

• MATH 251-3 Calculus III

Statistics Requirements

Students must complete

• STAT 270-3 Introduction to Probability and Statistics

and at least five of the following courses.

.

?

• ACMA 315-3 Credibility Theory and Loss Distributions

• ACMA 320-3 Actuarial Mathematics I

PRI

---

• ACMA

335-3

Risk Theory

• ACMA

445-3

Survival Models

• STAT 330-3 Introduction to Statistical Inference*

• STAT 340-3 Statistical Quality Control

• STAT

350-3

Linear Models in Applied Statistics*

• STAT 380-3 Introduction to Stochastic Processes

• STAT 402-3 Generalized Linear and Nonlinear Modelling

• STAT 410-3 Statistical Analysis of Sample Surveys

• STAT 420-3 Non-Parametric Statistics

• STAT 430-3 Statistical Design and Analysis of Experiments

• STAT 450-3 Statistical Theory

• STAT 460-3 Decision Analysis and Bayesian Inference

• *these core courses are recommended

STAT Courses

Back to the Undergraduate Studies section of the 1999/2000 Simon Fraser University Calendar

Copyright(c) 1996 Registrar Publications. All rights reserved.

Technical questions/comments about this site? Contact: Byron_Henrysfu.ca

r

.

g.

---

Appendix II

Graduate Program Requirements

Graduate Statistics: Proposed 2000/01 Calendar Entry

This calendar entry has been approved by the Faculty of Science graduate studies committee. It

is anticipated that these regulations will be in place by September 2000. The current calendar

entry is also attached.

M.Sc. Program Requirements:

The program is intended to give students instruction on a wide range of statistical techniques

and also to provide experience in the practical application of statistics. The program should

be of interest to students who wish to acquire statistical expertise in preparation for a career

in either theoretical or applied statistics.

Students in the program will be required to

• complete at least 30 credit hours of course work in Statistics and related fields beyond

courses taken for the bachelor's degree. Of these 30 hours, at least 24 are to be in graduate

courses or graduate seminars, and the remaining six may be chosen from graduate courses

or those 400 level undergraduate courses which may be taken for credit for the BSc in

Statistics. Normally these courses will include STAT 801, 811 and 812 and at least four

of STAT 802, 803, 804, 805, 806, 870, 890, 891.

• submit and defend successfully a project (as outlined in the Graduate General Regula-

tions) based on some problem of statistical analysis. This problem will often arise out of

the statistical consulting service.

Students with a good undergraduate background in statistics will normally complete the course

work in 4 semesters. The project, including the defence, is expected to require 2 semesters or

less.

Students with backgrounds in other disciplines, or with an inadequate background in statistics,

may be required to take certain undergraduate courses in the department in addition to the

above requirements.

Co-op education

Students in the M.Sc. program or the Ph.D. program have the opportunity to obtain work ex-

perience during their degrees by participating in the co-operative education program. Employ-

ment lasting one or two semesters with government agencies, companies or other organizations

employing statisticians is arranged for qualified students. Such employment often provides the

problem which forms the basis of the M.Sc. project.

Ph.D. Program Requirements

A candidate will generally obtain at least 30 credits beyond courses taken for the bachelor's

degree. Of these, at least 22 will be graduate courses or seminars and the remaining 8 may be

from graduate courses or those 400 level undergraduate courses which may be taken for credit

for the BSc in Statistics. Students who hold an M.Sc. in statistics are deemed to have earned

18 of the 22 graduate hours and 4 of the 8 undergraduate or graduate hours required.

10 ?

14.

---

Candidates will normally pass a general examination covering a broad range of senior under-

graduate material in statistics. A candidate ordinarily will not be allowed to take the general

examination more than twice. The general examination must normally be completed within

four full time semesters of initial enrollment in the Ph.D. program.

Students will be required to submit and successfully defend a thesis which will embody a

significant contribution to statistical knowledge.

For further information and regulations, see Graduate General Regulations.

.

S

11

---

Graduate

Courses

in Statistics

I

.

STAT 602-3

Generalized Linear and Non-linear Modelling

STAT 650-5

Quantitative Analysis in Resource Management and Field Biology

STAT 801-4

Mathematical Statistics

STAT 802-4

Multivariate Analysis

STAT 803-4

Data Analysis

STAT 804-4

Time Series Analysis

STAT 805-4

Non-parametric Statistics and Discrete Data Analysis

STAT 806-4

Lifetime Data Analysis

STAT 811-2

Statistical Consulting I

STAT 812-2

Statistical Consulting II

STAT 870-4

Applied Probability Models

STAT 880-0

Practicum I

STAT 881-0

Practicum II

STAT 882-0

Practicum III

STAT 883-0

Practicum IV

STAT 890-4

Statistics: Selected Topics

STAT 891-2

Seminar

STAT 894-2

Reading

STAT 895-4

Reading

STAT 898-0

MSc Thesis/Project

STAT 899-0

PhD Thesis

Notes

• STAT 602 and 650 are service courses not for students in our department. We will

maintain these courses.

• STAT 880-883 are numberings for semesters of participation in Co-op education.

.

12

?

I.

---

iE ?

University

calendar ?

11I•I'

Statistics Program

TLX10512 Shrum Science Centre,

(604) 291-3331

Tel,

(604) 291-4947

the test of English as a foreign language results.

Applicants with degrees in areas other than statistics are encouraged to apply provided they have

• some formal training in statistical theory and practice.

MSc Program Requirements

The program is intended to give students instruction on a wide range of statistical techniques and

also to provide experience in the practical application of statistics. The program should be of

interest to students who wish to acquire statistical expertise in preparation for a career in either

theoretical or applied statistics.

Students in the program will be required to

• complete at least 30 credit hours of course work in Statistics and related fields beyond courses

taken for the bachelor's degree. Of these 30 hours, at least 24 are to be in graduate courses or

graduate seminars, and the remaining six may be chosen from graduate or undergraduate

seminars or 400 level undergraduate courses. Normally these courses will include STAT 801,

811 and 812 and at least four of STAT 802, 803, 804, 805, 806, 870, 890, 891.

• complete satisfactorily STAT 811 and 812

• submit and defend successfully a project (as outlined in the

Graduate General Regulations)

based on some problem of statistical analysis. This problem will ordinarily arise out of the

statistical consulting service.

Students with a good undergraduate background in statistics will normally complete the course

work in four semesters. The project, including the defence, is expected to require two semesters or

less.

Students with backgrounds in other disciplines, or with an inadequate background in statistics, may

be required to take certain undergraduate courses in the department in addition to the above

requirements.

Co-operative Education

Students in the MSc or PhD program may obtain work experience during their graduate studies by

participating in the co-operative education program. Employment lasting one or two semesters with

government agencies, companies or other organizations employing statisticians is arranged for

qualified students. Such employment often provides the problem which forms the basis of the MSc

project.

PhD Program Requirements

A candidate will generally obtain at least 28 credits beyond courses taken for the bachelor's degree.

Of these, at least 16 will be graduate courses or seminars and the remaining 12 may be from

graduate courses or seminars or 400 level undergraduate courses. Students who hold an MSc in

statistics are deemed to have earned 12 of the 16 graduate hours and eight of the 12 undergraduate

or graduate hours required. The course work in all cases will involve study in at least four different

areas of statistics and probability.

Candidates normally pass a general examination covering a broad range of senior undergraduate

statistics material. A candidate ordinarily is not allowed to take the general examination more than

33.

---

twice. Thegeneral examination is normally completed within four full time semesters of initial

enrolment in the PhD program.

Students submit and successfully defend a thesis which will embody a significant contribution to

statistical knowledge. ?

0

For further information and regulations, see

Graduate General Regulations.

STAT Graduate Courses

Back to the Graduate Studies section of the 1999/2000 Simon Fraser University Calendar

Copyright(c) 1996 Registrar Publications. All rights reserved.

Technical questions/comments about this site? Contact: Byron_Henrysfu.ca

.

3Lf

---

Appendix A

Research Profile of Faculty in the proposed

Department of Statistics and Actuarial Science

Dr. Rob Baishaw,

an adjunct professor, is a Senior Biostatistician with Synectics Health

Corporation, a medical research group based on campus at the University of British Columbia.

His work at Synectics has focussed on the design and analysis of observational, longitudinal

medical studies, and on the analysis of pharmacokinetic data arising in clinical trials settings.

Current work includes several large scale studies in psychiatric disorders, auto-immune dis-

ease, and oncology. Dr. Baishaw's research interests include methodologies for the analysis of

longitudinal count data and approaches to the analysis of observational data. Drs. Baishaw

and Dean are currently examining methods for the analysis of life history data. Recent work

includes an examination of the use of cubic splines for the estimation of the intensity function

in a panel data context.

Dr. Charmaine Dean

works with Children's Hospital and the Ministry of Health in their

development of models for investigating spatial and temporal patterns in disease incidence and

mortality over the province, including the development of innovative statistical methodology to

do so. Assessing the spatial-temporal distribution of various kinds of mortality throughout the

province is an important task at the ministry. Policy-makers use these distributions, presented

in the form of maps, for allocating health funding. The new methodology Dr. Dean and her

recent Ph.D. graduate, Dr. Ying MacNab, have developed focuses specifically on addressing the

problem of producing meaningful statistics at the local health area level for small population

• groups. Identifying the distribution of rates over a region is a prediction problem, where random

effects, representing area-specific effects, are to be estimated. These area-specific effects are

correlated both spatially and temporally. In B.C. the data arise in a nested structure and she

and Dr. MacNab have exploited this structure to develop inference for a proposed hierarchical

model for the random effects. Primary applications at present are cancer and infant mortality.

With a previous Spanish postdoctoral student, Dr. Ugarte, and her Spanish colleague Dr.

Militino, Dr. Dean is involved in a project funded by Spain, to produce a mortality atlas for

the province of Navarra in Spain.

Dr. Dean has also worked in the general area of the analysis of count data, and has produced

several papers dealing with methods for handling a common and troublesome phenomenon in

such analyses called 'overdispersion'. This arises when the empirical variance in the data

exceeds the nominal variance under some presumed model. Dr. Dean has developed some

simple yet effective methods for accommodating overdispersion in analyses of count data. Her

recent work focuses on testing for variance components in Poisson mixture models using a

modification of penalized quasi-likelihood techniques. She has investigated the loss in efficiency

in the estimation of treatment effects in the analysis of counts versus actual event times in

Poisson processes. This has implications for cost-effectiveness in the design of clinical trials.

Dr. Dean has also been involved in research projects through the Vancouver General Hos-

pital, and the B.C. Institute for Child and Family Health. She is currently involved in AIDS

research with Health Canada, and in the analysis of the results of a major provincial study on

survival after cardiac surgery.

Dr. Jinko

Graham's research is in the area of statistical genetics and genetic epidemi-

ology. With Professors Ake Lernmark and Norman Breslow at the University of Washington,

10 ?

3i.

---

she is a coinvestigator on a Swedish population-based case-control study of the genetic and

environmental basis of type 1 diabetes, a complex genetic disease. The research is a team effort

with the Swedish Childhood Diabetes Study Group, a well-known network of physicians and

diabetes researchers based in hospitals across Sweden.

Disease associations within populations are a focus of the ongoing research in type 1 diabetes,

and also of related work with Professor Elizabeth Thompson at the University of Washington,

developing new methods of

disequilibrium fine-mapping.

In disequilibrium mapping, population

associations are used to infer the genomic location of a disease mutation at finer resolution than

is possible with traditional genetic linkage studies based on family data. Fine-mapping reduces

the substantial costs associated with isolating and cloning a disease gene in a large candidate

interval of the genome. The idea is that, on average, marker loci closer to the disease locus will

have alleles which are more strongly associated with the disease than farther marker loci. Since

genealogical relationships within a population lead to such associations, stochastic models of

gene ancestry have proven useful.

Besides disequilibrium mapping, many other problems in population genetics can be phrased

in terms of a latent ancestry connecting genes. For example, in work with Professor Bruce

Weir at North Carolina State University and Dr. James Curran at the University of Waikato,

Dr. Graham is developing improved estimators of forensic match probabilities for modern

microsatellite DNA profiles. In forensic matching of the DNA of a suspect to a perpetrator, the

more related the suspect and perpetrator are, the more likely that there will be a match. The

microsatellite loci currently popular in forensic DNA profiling mutate in stepwise increments,

and so the allelic state provides some information on the underlying ancestry. However, current

estimators of match probabilities are based on a non-stepwise mutation model, and do not make

use of this important information on relatedness.

An ancestral perspective is also useful in a research project with Professor Françoise Seillier-

Moiseiwitsch at the University of North Carolina at Chapel Hill and Dr. Brad McNeney at

Simon Fraser University. The group is using data on the genetic diversity of HIV populations

within individuals to estimate the diversity of strains at the time of infection. The number of

transmitted HIV quasispecies is of interest to scientists developing vaccine strategies.

Dr. Richard Lockhart

is completing work with T. W. Anderson and M. A. Stephens on

goodness-of-fit testing for standard time series models like AR(1) and MA(1) based on com-

parison of the cumulative spectral distribution and the empirical spectral distribution function.

With Gemai Chen and M. A. Stephens he has recently developed asymptotic approximations

for the usual likelihood objects in the Box-Cox transformation analysis in a framework which

leads to asymptotic approximations which (unlike those of Bickel and Doksum) depend contin-

uously on the parameter values. They have developed goodness-of-fit tests for the (false) null

hypothesis of normal errors in these models, along the way trying to make clear the meaning

of testing such a hypothesis. They intend to see if this sort of small sigma large n asymptotics

applies usefully in other estimated transformation contexts.

One of his recent Ph.D. graduates, Ken Butler, developed goodness-of-link tests for logistic

regression in his Ph.D. thesis and Dr. Lockhart expects to continue work in this area generally.

Related work with John Spinelli and M. A. Stephens on assessing the Poisson assumption in

Poisson regression models is almost complete.

Another of his recent Ph.D. students, Chandanie Perera, developed small sigma asymptotic

analysis of a method (thermo-luminescence) of dating sedimentary deposits such as sand dunes.

The problem requires further work particularly on a question of the following form. If the

11 ?

3(0

---

regression of

Y

on X has a slope which does not depend on a further covariate

W

over a range

W

1

< W < W,

where the limits

W

1

and

W,

are unknown, how should this common slope

. be estimated and how do we attach a standard error to its estimate. The problem is further

complicated because in the thermo-luminescence context the data sets for

Y

versus X are

correlated at adjacent settings of

W

(and the regressions are not linear). In her thesis, Perera

developed tests for normality when you have many small samples which may not have common

means or variances. This work is incomplete and related to other work still underway with

Gemai Chen in which they try to deal with goodness-of-fit tests after fitting a large number of

parameters.

Dr. Lockhart is interested quite generally in the problem of model assessment, particularly

in models with many parameters. He seeks tests which control overall error rates according

to some clear protocol but which, having detected a model failure, can be decomposed into

components which can be used to identify the particular sort of model failure.

Dr. Brad

McNeney's research is in statistical issues related to public health. Much of his

current work is concerned with the design of epidemiologic and clinical studies, and associated

questions of data access such as security and privacy, extraction from large central registries,

and distributed computing, particularly in health services or health outcomes settings. In

collaboration with Professors Jon Wellner and Norman Breslow at the University of Washington,

he is developing statistical methods for the cost-effective design of such studies. Recently,

two-phase or response-selective design strategies have been proposed as economical. The case-

control study in which cases are oversampled relative to their overall population frequency may

be viewed as an example, and is known to be far more efficient than a cohort study when

the disease is rare. More generally, the response and a set of inexpensive covariates can be

. measured on all subjects. Then, in a second phase, those with rare responses and/or exposures

can be oversampled for a more informative and costly covariate, while a reduced fraction of

individuals with common responses and exposures can be sampled. Thus, a subset of subjects

will have data

missing by design

on the costly covariate. The work carries over naturally to more

general missing-data problems in the statistical analysis of clinical trial data. With Professor

A. (Butch) Tsiatis at North Carolina State University, Dr. McNeney is investigating efficient

estimation and testing in semiparametric models with data that are missing by chance.

Dr. McNeney also has research interests in statistical issues related to HIV vaccine devel-

opment. Given the scale of the public health problem HIV poses in the developing world, the

need for effective vaccines is clear. In a project with Professor Francoise Seillier-Moiseiwitsch

at the University of North Carolina, and Dr. Jinko Graham at Simon Fraser University, he is

developing statistical methods for assessment of the number and diversity of HIV quasispecies

transmitted to an individual based on HIV envelope sequences collected over time from newly

infected individuals. The larger the number of transmitted quasispecies, the more difficult the

task of developing an effective vaccine.

The goal of

Dr. Rick Routledge's

research program is to develop statistical methodology

of particular value to population biologists and resource managers. This work requires that he

uses, and make contributions to, newly evolving statistical theory, and that he keeps abreast

of modern developments in population biology and related instrumentation. Examples of the

technical methodology that Dr. Routledge works with include saddlepoint approximations,

•

?

exact p-value calculations, spline-fitting, and split-beam echosounding.

Dr. Routledge's work has been incorporated into undergraduate courses and texts in pop-

ulation biology, and has contributed to high-profile public debates on resource management

12

---

issues. Spin-off work in estimating human populations has also contributed to debates over

street prostitution, the issuance of local arrest warrants, needle exchanges for intravenous drug

users, and health care for aboriginal populations.

A major focus of his recent research has been in fish management. This began with a

collaborative project with the Pacific Salmon Commission, in which he and his graduate student

developed geometric probability arguments to improve the process for estimating the numbers

of adult sockeye salmon migrating up the Fraser River. This work led to his appointment to the

Fraser River Sockeye Public Review Board, a ministerial inquiry into so-called missing fish in

the Fraser River in 1994. He is now co-supervising (with J. O'Hara-Hines at the University of

Waterloo) a Ph.D. student who is investigating estimation formulas for use with newly evolving

split-beam echosounding technology.

Fisheries data typically contain large measurement errors. One of his M.Sc. students recently

looked at applying the SIMEX method for generating bias adjustment to a nonlinear regression

model commonly used in fish stock recruitment. Furthermore, the escalating conservation crisis

for coho salmon has prompted Dr. Routledge and a senior coho biologist in the Department of

Fisheries and Oceans to study models for predicting extinction probabilities for small, isolated

fish stocks.

Other components of his research program include studying the properties of saddlepoint

approximations and the mid-p-value, developing new change-point methodology for specific

problems arising in physiological threshold phenomena, developing new diversity estimation

techniques for use in implementing forest practices codes, and refining mark-recapture tech -

niques for fish and wildlife estimation problems.

Dr. Carl

Schwarz studies the dynamics of animal populations and a variety of ecological

problems. His research is motivated by real problems encountered by ecologists. Much of his

research involves capture-recapture experiments. In these experiments animals are captured,

tagged, released and then subsequently recaptured. The pattern of captures and recaptures

enables the researcher to estimate survival rates and the number of animals present in the

population.

One of his immediate research objectives is to extend a method called the stratified-Petersen

estimator which is often used in salmon escapement studies to take account of certain assump-

tions which are not valid in practice. The estimate of escapement in the Nass river in northern

B.C. uses fish wheels to capture, tag, and release fish. Recaptures take place upriver at the

Meziadin Dam where all returning salmon must pass through a single fish way. Unfortunately,

due to the large volume, it is not possible to examine all the fish, and sub-samples are taken-

but, it is not possible to read all of the individual tags. Dr. Schwarz is developing methods to

deal with the observed, but unread tag numbers in order to obtain estimates of the number of

salmon that originally passed the fish wheels.

He is also investigating the effect of tag reading errors. In a study on Sable island, seals were

individually branded as pups and subsequent surveys have used spotting scopes to record the

brand numbers of adult seals that have returned to breed. Based upon the observed capture

histories, it appears that a substantial number of the brand readings are in error. Surprisingly,

no previous work has been done on the effects of tag misread errors upon estimates of survival,

catch-ability, and abundance in capture-recapture studies.

Students under Dr. Schwarz's supervision have worked in a number of areas motivated by

real world problems. Some recent projects include methods for creel surveys when gear loses

efficiency; estimating the number of razor clams on beaches in Haida Gwaii; estimating oyster

13

?

39.

---

densities in the Kiahoose First Nation's waters; estimating breeding probabilites for grey seals;

and comparing methods of estimating timber volumes for the Ministry of Forests.

Dr. Randy

Sitter's research contributions have been primarily in two rather unrelated

areas: analysis of complex survey data; and the design of experiments with a specific focus on

industrial applications.

In complex surveys, often the sampling design induces a structure to the data which is not

independent and identically distributed. Though techniques for variance estimation and confi-

dence intervals do exist, they are often cumbersome to implement, or do not extend to complex

designs. It is desirable to have resampling methods which re-utilize the existing estimation sys-

tem repeatedly, using computing power to avoid theoretical work, that can be applied to such

data. Dr. Sitter has made a number of contributions to the development of such resampling

methods in this setting. At present he and one of his recently graduated Ph.D. students (C.

Wu, now a faculty member at U. Waterloo) are considering development of such methods for

cross-classified sampling as is used to estimate the Retail Price Index in the United Kingdom,

as well as other countries. This work was motivated by a long-term consulting project which

the SCS together with Dr. Sitter have had with the Office of National Statistics in the U.K.

More recently, Dr. Sitter has developed (with J. Chen and with C. Wu) empirical likelihood

methodology to incorporate complicated auxiliary information into complex survey designs. It

is shown that, under stratified multi-stage and other complex designs and that when estimating

the total of y when the auxiliary variable x has known mean, the method is asymptotically

equivalent to the commonly used generalized regression estimator. Initial investigations suggest

better small sample performance and robustness to deviations from a linear relation.

Dr. Sitter has also made contributions to optimal design and combinatorical design. Mo-

tivated by applications in the pharmaceutical industry, he developed (with C.F.J. Wu and a

M.Sc. student, B. Forbes) a simple two-stage approach to the optimal design of binary response

studies. Simulations suggest inferences based on such a procedure are better than a one-stage

approach. In the process, many useful theoretical results were established in both one- and

two-stage optimal designs in this and more complicated non-linear modeling situations. In

many industrial settings, unreplicated fractional factorial (FF) designs are used to investigate

the possible importance (in terms of quality) of a large number of factors using only a small

number of experimental runs. A common way to choose the best such design is by ranking the

design using aberration as a measure, where smaller aberration essentially implies a design is

more able to consider interaction effects. Dr. Sitter has extended the notion of aberration to

blocked FF designs and obtained minimum aberrations (MA) designs for this common experi-

mental setting (with J. Chen and M. Feder). Together with another of his recently graduated

Ph.D. students (D. Bingham, now a faculty member at U. Michigan) he has obtained theoret-

ical results and developed algorithms for obtaining such MA designs in the case of split plot

structures and robust parameter experiments.

Dr. John

Spinelli is a Senior Biostatistician at the British Columbia Cancer Agency

and a Research Advisor to the British Columbia Cardiac Registries. He has been involved in

design and analysis of a multitude of research studies at the BC Cancer Agency, St. Paul's

Hospital, Vancouver General Hospital and the Ministry of Health. His current projects include

risk factors for childhood leukemia, sun exposure as a risk factor for lymphoma, assessing the

• outcome after coronary interventions (surgery/angioplasty), mortality while waiting for cardiac

surgery, early discharge after narcotic overdose and estimating the cost of smoking in British

Columbia.

3,".

14

---

With Dr. Stephens and Dr. Lockhart he is working on problems in goodness-of-fit, par-

ticularly for regression models. Current work involves tests of fit for the binomial distribution

and binomial regression. A recent project involves the development of tests of fit for grouped

continuous

With Dr.

data.Dean

?

and Melody Ghahramani, he is working on the development of statistical

0

models for simultaneous modeling of short and long-term mortality after cardiac surgery.

Dr. Michael Stephens,

Professor Emeritus, remains very active in research receiving

numerous invitations to speak and collaborate both nationally and internationally. He contin-

ues to supervise students. He is currently working on a number of projects in goodness-of-fit

including: testing time series models with T.W. Anderson and Richard Lockhart, testing Pois-

son regression models, tests for specific distributions with V. Choulakian and P. Puig, general

results for functions of expected values of order statistics with his recent Ph.D. student H.

Coronel-Brizio and general consistency results for the Shapiro-Wilk goodness-of-fit statistic.

Dr. Tim Swartz's

research interests lie in the field of statistical computing. Most of his

work attempts to take advantage of the power of modern computing machinery by developing

computationally intensive algorithms to solve statistical problems. A particular area where he

has devoted a lot of attention is the integration problem which arises in Bayesian applications

as well as other fields in the physical sciences. With M. Evans of the University of Toronto, he

is near completion of a research level text on integration that will be published in the Oxford

Statistical Science Series.

Dr. Swartz also has interests in certain areas of applied statistics. He has written a series

of papers that develop empirical voting indices and is near completion of some modelling work

concerning final offer arbitration. He also has written papers involving statistics in sport.

Dr. Larry Weldon's

primary area of research is the foundations of statistics - a re-

evaluation of the important tools and concepts that are necessary for statisticians and others

to use statistical theory effectively. His years of experience with consulting in the health,

social and life sciences has broadened his view of the discipline of statistics. He has written

two textbooks in statistics which reflect this research, and emphasize logic and conceptual

understanding rather than the traditional mathematics of statistics. He is currently writing

a book addressing the new role of probability modeling. He has worked for several years on

how to revise statistical education in the light of modern technology, and his experience with a

statistical education project in Indonesia has been a model for developing his ideas. Dr. Weldon

is also interested in the use of applied probability models for studying complex systems, and

graphical methods for data analysis and for describing simulation outcomes.

Dr. Julia Wirch's

research is in Actuarial Science. Her current work investigates cap-

ital adequacy risk measures for insurance asset portfolios. The development of quantitative,

objective methods for the determination of the risk factor distributions and risk measurement

techniques used in solvency testing and margin calculations is of increasing importance as the

risks of insurance companies become more complex.

With Dr. Mary Hardy at the University of Waterloo, Dr. Wirch examined risk measures

currently used for insurance regulation, and investigated distortion risk measures which led her

to construct the coherent beta risk measure. The coherent beta risk measure satisfies the notion

of second order stochastic dominance, an important economic ordering property pertaining to

risk aversion, which is not satisfied by the risk measures used in practice.

15 ?

qo.

---

The significance of this research is in its application to risk management. In financial risk

management, the use of coherent risk measures with consistent ordering properties will improve

• the information available to managers, regulators and shareholders of financial institutions. Two

issues adding to the risk of insurers are the increased used of embedded options in insurance

contracts and the increased use of derivative securities in insurance fund management. These

instruments are often more volatile than traditional insurance products with respect to risk

factors. This research has shown how the coherent beta risk measure can be applied to measure

the risk in these embedded options.

Dr. Wirch also has research interests in multivariate risk distributions and the application of

copulas to obtain marginal distributions. Insurance asset portfolios are often allocated among

funds which may be controlled by different fund managers. Determining the risk in each fund

is often as important as determining the risk of the total portfolio. A similar decomposition

can be done using multiple risk factors such as the interest rate and exchange rate.

Currently, Dr. Wirch is investigating the generalization of static capital requirements to

dynamic capital requirements using a characterization of dynamic risk measures developed by

Tan Wang

(1996).

.

r

16

---

Appendix B

The Statistical Consulting Service

The Statistical Consulting Service (SCS) is a consulting unit within the Statistics Group

at Simon Fraser University formed in 1980 to promote more effective links between academic

statisticians and applied researchers. The SCS provides expert statistical advice to individu-

als involved in data-based research projects, promotes links between the statistical education

program at SFU and individuals involved in data-based research, and improves the .quality of

both data-based research and the teaching of graduate students in statistics.

The SCS has access to a variety of modern computing facilities and software. It has a full-

time Director, who is a qualified statistical consultant. All statistical faculty in the Department

of Mathematics & Statistics at SFU contribute to the SCS; their combined experience includes a

full spectrum of statistical application areas as well as a broad coverage of statistical techniques.

SCS staff have contacts to other university resources that can be arranged on a contractual

basis.

Typical SCS projects include design of experiments or surveys, sample size determination,

suggestion of appropriate software for statistical analysis, provision of advice on graphical data

summaries, determination of appropriate inferential techniques, evaluation of study protocols in

grant applications, quality assurance and reliability techniques and the offering of short courses.

Itemized below are examples of completed projects.

. Modeling the distribution of pulp fibre length.

• . Designed experiments in the pulp and paper industry.

. Estimation of variability in the Consumer Price Index.

• Study of ventilation in humans during exercise.

• Effects of mild anesthesia on the thermoregulatory response during cold exposure in hu-

mans.

• Career profile of women scientists in B.C.

• Parental decision-making for student recruitment into early French immersion.

• Sediment discharge regimes of B.C. rivers.

• Comparison of diets for salmon nutrition.

• Mark-recapture method of estimating size of a criminal population from police records.

• Analysis of DNA damage in bladder cells.

.

17

4.

---

Appendix C

Students who completed the M.Sc. between 93-1 and 99-3

Year Semester

Name

Supervisor

Present Occupation

1993

1

Waweru

Lockhart

Statistician, Government of Kenya

1

Mwangi

Eaves

Statistician, Government of Kenya

2

Deng

Dean

Ph.D. Computer Science

2

Zhong

Weldon

Ph.D. Engineering

1994

1

Grunwald

Eaves

U.S. College Statistics Instructor

1

Yu

Weldon

Lost to follow-up

1

Zhan

Dean

Ph.D. Statistics Waterloo

2

Kristiansen

Dean

Statistician, ICBC

3

Hu

Lockhart

Lost to follow-up

1995

1

Cheng

Schwarz

Ph.D., Computer Science

1

Carpenter

Dean

Self-Owned Business

1

Ma

Eaves

CIBC, Toronto

2

Rae

Dean

Biostatistician, St. Paul's Hospital

1996

1

Rajwani

Schwarz

Statistician, ICBC

2

Sun

Schwarz

Sessional Lecturer

3

Sutherland

Schwarz

Statistician, Health Canada

1997

1

Brownell

Routledge

Statistician, US Marketing Firm

2

Lui

Swartz

Lecturer, Coquitlam College, Sessional

2

Kalbfleisch

Sitter

Statistician, Capital One Market Research, VA

1998

1

Taylor

Schwarz

Statistician, UBC

2

Han

Routledge

Anderson Consulting

2

Ghahramani

Dean

Statistician, A.T. &T

2

Thompson

Dean

Biostatistician, U.S. National Cancer Institute

1999

2

Paltiel

Schwarz

UBC Statistics Consulting Laboratory

2

Zhang

Weldon

Ph.D., Computer Science

3

Macleod

Schwarz

Anderson Consulting

Students who completed the Ph.D. between 93-1 and 99-2

Year

Semester

Name

Supervisor

Present Occupation

1994

1

Spinelli

Stephens

Faculty, UBC

1

Coronel-Brizio

Stephens

Faculty, Mexico

1995

- ?

2

Banneheka

Routledge

Faculty, Sri Lanka

2

Tsao

Routledge

Faculty, UVic

1996

2

Perera

Lockhart

Faculty, Sri Lanka

1997

2

Butler

Lockhart

Postdoctoral Fellow

3

Baishaw

Dean

Biostatistician, Synectics Health Corporation, UBC

1999

1

Bingham

Sitter

Faculty, U Michigan

2

MacNab

Dean

Faculty, UBC

2

Wu

Sitter

Faculty, UWaterloo

18 ?

1J3

.

---

Appendix D

0

?

Graduate Students' Coop Employment

Year Semester

Student

Employer

1995

3

Jason Sutherland

Health Canada, Ottawa

1996 1

Peter Lui

BC Ministry of Forests, Victoria

1

Karim Rajwani

Fed. Dept. of Fisheries & Oceans, Naiaimo

1

Jason Sutherland

Health Canada, Ottawa

2

Stephanie Brownell

Health Canada, Ottawa

2

Melody Ghahramani

St. Paul's Hospital, Vancouver

2

Charles Paltiel

Statistics Canada, Ottawa

3

Myra Andrews

Statistics Canada, Ottawa

3

Stephanie Brownell

Health Canada, Ottawa

3

Hillary Han

Imperial Oil, Calgary

3

Peter Lui

BC Ministry of Forests, Victoria

1997

1

Hillary Han

Imperial Oil, Calgary

1

Heidi Kalbfleisch

BC Cancer Agency, Vancouver

2

Melody Ghahramani

St. Paul's Hospital, Vancouver

2

Darby Thompson

Health Canada, Ottawa

3

Heidi Kalbfleisch

Statistics Canada, Ottawa

3

Darby Thompson

Health Canada, Ottawa

1998

2

Ellen Chan

Statistics Canada, Ottawa

2

Cohn MacLeod

Statistics Canada, Ottawa

2

Charles Paltiel

Statistics Canada, Ottawa

2

Greg Pond

Fed. Dept. of Fisheries & Oceans, Nanaimo

2

Ruihua Yin

Health Canada, Ottawa

3

Greg Pond

BC Ministry of Forests, Victoria

1999 1

May Lee

BC Ministry of Forests, Victoria

2

Milena Simic

Health Canada, Ottawa

2

Wanlin Wei

BC Ministry of Forests, Victoria

3

Milena Simic

Health Canada, Ottawa

3

Alex Zhao

Technical University of BC, Surrey

.

19 ?

Lpf.

---

Appendix E

Letters from Doctoral Graduates

.

RTI

---

OIPRTMFNT Or STATISii

S

144 SOUTh I S1ATL SIRLLI

4062 I RU /I

HIM DIN(;

?

ANN AkI3 JR. MIt

I MAN

.18109.1

USA

I)L4AI! 1 Mi. N IAI (M I 1(1:

I?

4 . 1j 71,4 ! 19

rA)

1

/44

1

/(,$.4(7(,

It) vlioiii it

IIiV

t(3Ifl(j

\ly

IIah1Io

is Drrok !3iiighani and I am an Assistant Professor in div U pnI'lIruiIt of Statis-

tics at t In

.

t)iiivtrsitv of Michigan, and I was a graduate stw.k'iit in SFUS Department of

\lathtniaf irs and Statistics froill 1995-1999. 1 am writing with regards to time proposal for

I

lit' crvnt.it,ii

(If

the Department of Statistics at. Sinioii Fraser Universit

y

. Iii particular, I am

writing to offer

mv

Support fur this venture and

Lu

congratulate he University For

1P(:OgTii'i-

lug St at it

i(S

as it

t1icidttie that is

IIHICItW

and differt q it

froni Mat luu4latIes.

En itiv opifituhi. there is a I)('ICt'l)t

1011 }liIIOIIg

statistics factilt

?

at. large, and

1*11101mg

the ?

st iich'uts that t Iivv m lvit'. that gIatItiaLe

ediicfflioii in

statistics is better in a statistics tie-

l

)1

1111I11t

rat III'I

t

I1U)

U

j()iL11

mathematics and statistics department.

?

V1iih'

till' i(lelL

HIM

grinluatc' t'diic;itititi ill statistks is actually better um it statistics department is arguable, the

pon-cptioll that it is b0 ter

is vvRkIit.

B

y

formi

forminga separate' department, til(' Statistics

I'm-1111

1

N.

will 1)4' able to atti'tct ttlol'(' of the host graduate students.

The now St at isu its i)opartiuoiit at Siunomi Fraser will also help attract new high

Ciil!l)I'O

facult y

iiivinlors. As s',iti

p one who Was roventl

y

going through the niivrvww 1noce.ss, I (-.;tit

smw that (ho facult

y

positions in statistics departments were inure attractive to inc than

tilosO

Ill

jtntit dt

•

part Iui('iit

S

l)t'raus(' I would be judged (e.g.. the

teiIllV('

pi'ttt's) b y intLivil IHaIS

III

statistics

4

1111Y.

III

H.

•

p4011L

tt&qnitt

III&'IIL

tllL'Fe

eXiStS

the

possibility

that you will lit

I'OViCV('(l

i,v iiueuimIors

ol I

hit' ht('LtItV who are ess&'iituillv

II! U

tiifft'i'otit

(liSCii)lii1('

azal ma

y

not be able

I()

iltT1li11 O)V

assess vtiur iun3,art Utl the field. My froling

IS

that this

Opilui(

III

is

Vi(l('ly

hold

11%'

Vt tiluig t('S('*rrlI('rS ill statist

its.

Last ly. I should say it few words alunit t he Statistics group at SFU. As it gra(llULtt' st.tuiieiit.

I SpvIlt

a

lot

of

t iluit' with the Statistics faculty in omm capacity

01'

another.

rhOil-

dedication

to

their si intents and to their work Was always impressive. the

1111.1St

iut,til>lc'

mtSI)4Ct,

of rite

group. to itit'.

IS

till' integrity of its uneuimbers.

hi short. fortimiug a new I)epart umicut of Statistics at Siiitoii Fraser will.

ill my

(il}I1I1C)iI,

alLov

1!

good pro g

raill l)O(Oliit' ('V('1I bet-1-cr. I feel that

this

willallow tlit' Lmveisity to attract

t't('lt lilt Ii('

high talil ttr graduate Students andfru1t.v ittcntlx'rs

ill

(lit' 1110U

of statistics.

Siitueit'lv.

j)

l)eurk l3ittghauu

.

.

14.

Od

?

vDv:go oo-Lr-u('

---

.

From gchencoxbox.math.uregina.ca

Fri Jan 14 07:33:14 2000

Date: Fri, 14 Jan 2000 09:33:12 -0600 (CST)

From: Gemai Chen <gchen@math.uregina.ca

>

Subject: Re: Separate Department of Statistics

To: lockhart@cs.sfu.ca

X-Status: $$$$

X-UID:

Dear Richard:

I was very glad to hear that you and your colleagues at SFU

are working on the creation of a Statistics Department.

As a former student, as a person who has had experience in

both studying and working in a Statistics department as well

as in a Math and Stats department, I feel certain that what

you are doing is good for the statistics discipline to grow,

good for the mathematicians to stay more focused on their job,

and undoubtedly good for the future of SFU.

Statistics is NOT a sub-discipline of Mathematics. Statisticians

use unique principles to guide their work and utilize whatever

.

?

tools available (including math tools) to get the job done.

Recent changes in science and society have expected all statisticians

to relate their work to fruitful applications; the number of

theorem-proof statisticians, who looked most like mathematicians,

is shrinking dramatically. A Statistics Department will provide

the most natural environment for the above to happen, and students

will find it much easier to choose a discipline which has led many

to challenging, enjoyable and financially rewarding jobs.

Best wishes!

Gemai

Dr. Gemai Chen ?

Associate Professor

Department of Mathematics and Statistics, University of Regina,

Regina, Saskatchewan S4S 0A2 Canada

Phone: (306) 585-4342 Fax: (306) 585-4020 E-mail:gchen@math.uregina.ca

L.

---

___ ?

B.C. RESEARCH INSTITUTE FOR CHILDREN'S & WOMEN'S HEALTH

Centre for Health Evaluation Research

4480 Oak Street, Room E4 I 4A

Vancouver, BC Canada V6H 3V4

Telephone (604) 875-3130

Fax (604) 875.3124

6 February 2000

Dr. Ying MacNab

Professor Charmaine Dean

Director of Statistics

Department of Mathematics and Statistics

Simon Fraser University

Burnaby, BC, Canada, V5A 1S6

Dear Professor Dean,

Thank you for informing me that the statisticians at SFU have put forward a

proposal for the establishment of a Department of Statistics at Simon Fraser

University. I am writing to you to express my support for this important

initiative. I also wish to congratulate the University for taking such a signif-

icant step forward. By building a Department of Statistics, the teaching, the

development, and the application of statistical science can be fully explored.

It was about sixty years ago in 1940 when Professor Harold Hotelling pub-

lished his historical essay

The Teaching

of Statistics,

and later in 1949, his

essay on

The place

of

Statistics in University.

Later, in 1988, both essays were

reprinted by the IMS in Statistical Science (Vol 3, No. 1) with in-depth dis-

cussion by distinguished discussants, many of them prominent statisticians.

Now statistics has been widely recognized as a distinct and fundamental dis-

cipline, a science of rapid development and wide application. The number

of institutions with Department of Statistics has increased and such a move

has proven to be successful and rewarding. Those universities where a De-

partment of Statistics/Biostatistics is firmly established are able to attract

students of the highest caliber. Studyinging in a Department of Statistics is

particularly important for those who wish to pursue higher degrees in statis-

tics.

- The members of the Statistics Group work actively in the field of statistics

and have put together a sound program of statistical teaching. I believe that

the establishment of a Department of Statistics will bring good prospects for

the University and provide an environment where statisticians can be more

0.

0

RESEARCH INSTITUTE PARTNERS

Children's & Women's Health Centre of British Columbia (C&W), The University of British Columbia

---

focused on their teaching and research work and the future departmental chair

can work with fellow statisticians to refine and extend the current programs.

I wish you

all the best and success in this endeavour.

Sincerely yours,

Ying MacNab

S

S

L

?

Mg

---

Changbao Wu, Assistant Professor

Dept of Stats and Actuarial Sci., University of Waterloo

Waterloo, ON N2L 3G1 CANADA ?

Office: MC 6148

Tel: (519) 888-4567 Ext. 5537 ?

Fax: (519) 746-1875

?

Email: cbwu@icarus.math.uwaterloo.ca

My name is Changbao Wu. I finished my Ph.D. studies in

Statistics at Simon Fraser last August and took an

Assistant Professorship in the Department of Statistics

and Actuarial Science at University of Waterloo

immediately after that.

Last Spring, during my job interviews, I had a chance to

visit some statistics departments in Canadian universities

(UBC, Western Ontario, Waterloo, Manitoba and Carleton) and

some others in the United States as well. It was only after

those detailed investigations I realized that how solid the

statistics program at SFU is. The undergraduate and

graduate programs are solid, the faculty members are strong

and active.

However, the fact that the statistics program is within the

combined Math and Stat department has a tremendous

negative impact on this group. Students who consider

universities for graduate studies usually believe that the

program should be small and weak. As a matter of fact, I

almost abandoned the idea of pursuing a Ph.D. at SFU.

It puts the program at an

unfavorable position for getting best students. It also has

an unseenable impact on its graduates when they start

looking for jobs. During my four years at SFU, I felt that

there were too many struggles in battling for computing

power, conflicts for other resources and spaces, delays in

decision-making process. The combined department will hurt

the Math program less than it does for other programs,

unless the department is to be renamed as

the Department of Statistics and Mathematical Science.

As a Ph.D. graduate of SFU who benefitted very much from

the statistics program, I strongly support the move

toward a separate department of statistics. A top

statistics department among Canadian universities will be

there.

.

is

61.

---

Appendix F: Space-Sharing with Mathematics

A.

Space typically used solely for the Statistics and Actuarial Programs.

1.

15 regular offices in the main Math-Stat corridor: 10 for regular faculty; 1 for the Statistics

Lab Co-ordinator; 1 for the Director of the Statistical Consulting Service; and 3 for

Emeritus professors, sessionals, adjuncts, and visitors.

2.

1 small office, SCB 9651 shared by 3 statistics Ph.D. students.

3.

The Statistics Workshop

B. Space currently shared between Mathematics, Statistics and Actuarial Science.

1.

5 regular offices in the main Math-Stat corridor: 1 for the Departmental Assistant; 1 for

the Graduate Secretary; 1 for the Chair's Secretary; 1 for the Co-op Coordinator; 1 for

the computer technician.

2.

1 larger office for the Chair.

3.

1 storage room.

4.

1 large departmental office, currently occupied by a receptionist and 2 secretaries, but

has room for another secretary, with attached photocopy/fax/storage room.

5.

1 Departmental Library.

6.

1 small room with a sink, fridge and microwave; note that the department does not have

a common room - this room can just about hold four people standing closely together.

7.

1 large shared graduate student office, K 9501, in which currently 25 students share 20

desks; of these 17 are statistics graduate students. This room also serves as a graduate

computer room housing 5 SUN workstations.

8.

1 resource room, K 9509 for seminars, departmental meetings and colloquia.

9.

The WCAT Laboratory, K 9514.

10.

1 graduate student computing room opposite to the WCAT Laboratory, K 9512.

11.

1 printer/computer room, TLX 10515.

12.

1 meeting room, TLX 10501.

The renovations required for the reallocation procedure require a modest amount of resources

to separate the space between the two new units and at the same time gain a small amount

of more useful space. This would be combined with a space- and resource-sharing proposal

between the two units and a plan by the university to make available additional space for both

• units as soon as such is available. Our proposal is for the new Department of Statistics and

Actuarial Science to occupy the east end of the main corridor of the current department. This

would include:

21

---

I. Space for the Statistics and Actuarial Science Program.

1. 15 regular offices allocated as above under item 1 of part A.

2. 1 regular office for a Chair.

3.

1 regular office for a Departmental Assistant.

4.

The space currently occupied by the Departmental Library renovated to be the new

Statistics and Actuarial Science Departmental office. In doing this renovation a small

amount of space could be absorbed into one of the adjacent offices to create a slightly

larger office for the Chair.

5. 1 regular office renovated to house the statistics library, and to act also as a coffee/meeting

room.

6. The Statistics Workshop.

7.

The small office, SCB 9651, to be shared by finishing Ph.D. students.

8.

The graduate student shared office, K 9501. A co-operative effort between the two new

units will ensure that office-space is made available to all graduate students.

H. Space shared with Mathematics.

1. 2 regular offices in the main Math-Stat corridor for personnel who will work for both new

units: 1 for the Co-op Coordinator; 1 for the computing technician.

2. the resource room, K 9509 for seminars, departmental meetings and colloquia.

3.

The WCAT Laboratory, K 9514.

4. the graduate student computing room opposite to the WCAT Laboratory, K 9512.

5.

the printer/computer room, TLX 10515.

6.

the meeting room, TLX 10501.

The above proposal should be justified in terms of its fairness to the Department of Math-

ematics. In part I, the 15 regular offices in item 1 are already typically used by statistics and

actuarial science. The offices for the Chair and Departmental Assistant would be offset by a

decrease of one secretarial office in the current department. This office and space for a math

-

ematics departmental library could be recovered by renovations to the current departmental

office. The adjacent coffee room and storage room could also be renovated to create more usable

space for mathematics, perhaps a common room. This would make sense since presumably the

new Department of Mathematics would have only 4 instead of 6 support staff and the present

office would be far too large for the smaller unit. If done in an intelligent manner, space could

be gained since at present the current departmental office has space for one more secretary

than it currently holds. The separate storage room next to the small coffee room might be

incorporated as the new smaller department would not need as much storage.

The current shared space which will not be shared by the two units includes the current

library, main office, Chair's office, Departmental Assistant's office, graduate secretary's office,

22 ?

53.

---

Chair's secretary's office, storage room and small coffee room. The proposal here is equivalent to

viewing that, of this space, statistics and actuarial science retains the current library, and three

offices, one for a Chair, one for a Departmental Assistant and one other as a library/meeting

room, while mathematics retains the rest. In terms of square footage, the new department then

obtains one-third of this current shared space.

The above proposal does not solve the current lack of space. If such an approach is adopted,

both departments will still be underspaced. For example, even ignoring the expected growth,

the new Department of Statistics and Actuarial Science should really have more space for

conducting experiments in the Workshop, office space for graduate students and visitors, a

meeting room separate from the coffee room, a larger room for the Statistical Consulting

Office for client meetings, another office for visitors, an additional computer/printer room and

a small departmental library. These do not represent costs of separating the two units, but

rather represent current additional space requirements which exist within the Department of

Mathematics and Statistics. Similar space requirements could be listed for the Department of

Mathematics. We suggest that space for the two units be allocated from the space which will

be vacated when departments and faculty are shifted to the new applied sciences building.

Renovations to construct a Main Office for Statistics and Actuarial Science

We propose that the General Office for the new department be located in the present

Department of Mathematics and Statistics Library, Room K10542/10544, with offices for the

Department Chair and Departmental Assistant located on either side of this office. There would

be doors connecting these adjacent offices to the general office.

The General Office would house one full-time and one part-time support staff persons. To

. optimize the use of the available space, the existing wall between the room and the corridor

would be changed into a counter with a sliding grill. Blinds would be added to the windows,

and a small storage room for confidential materials would be created in the southeast corner of

the office.

We are recommending that all furniture be movable for two reasons. First, we view this as

an interim solution for accommodating the new department. Second, we need flexibility while

we learn how to utilize the space most effectively.

We wish to point out several undesirable features of this proposed arrangement. The entire

south side of this floor can be oppressively hot in the summertime, and K10542/10544 is

probably the hottest room of all. In addition, the large conduits seriously impinge on the

amount of usable space in this room. Furthermore, the hall outside the proposed general office

is narrow and cramped. Finally, the room contains doors that at present open onto an unusable

roof area which contains numerous exhaust vents. We urge the university to give high priority

to proposals and requests (some of them long-standing) to

1.

improve the airflow on the south side of the entire floor,

2.

remove the large conduits from K10542/10544,

3.

create an attractive, healthy roof-top area, similar to ones in the Multipurpose Complex,

and

0

?

4. find a more appropriate, long-term home for the new Department of Statistics.

23 ?

54.

---

Approximate immediate costs (as discussed with Erik Grafstrom, Manager, Minor Projects,

Facilities Management) are as follows:

Item

Approx. Costs

Counter and Grill

$15,000

Communicating Doors

$9,000

Shelving

$10,000

Mailsiots

$10,000

Blinds

$15,000

Furniture

$10,000

Confid. Storage

$8,000

Computers, Printer, Fax

$8000

Contingencies

$10,000

Total

$95,000

.

24

---

Appendix G

Budget

Salary and Operating Expenses

• Chair's stipend.

• Salaries for existing faculty, instructor—statistics laboratory, consulting director. These

are currently in place.

• Salaries for a Departmental Assistant, one full-time and one half-time secretary.

• The remainder of the position for David Eaves, a retired Statistics faculty member. We

propose that this be reallocated to the new department for use in establishing a permanent

Director of the SCS, in keeping with the plans of the ALRP of Mathematics and Statistics.

• Sessional budget. This will include teaching relief for the Chair and one course per

year for one other faculty member, as well as sessionals usually required for mounting

our programs. The sessional budget will be generated through the normal criteria for

allocation of fall-out and overhead-return funding.

• Teaching Assistant budget. We propose that the teaching assistant budget for statistics

and actuarial science be taken initially from the budget for the whole department. This

will be discussed with the Dean and the Mathematics Chair.

• Non-salary Operating Budget. Supplies, maintenance. We propose that (i) $20K allo-

. cated to the SCS and currently in the department's operating budget be moved to the

operating budget for the new department, (ii) one-third of the remainder of the operating

budget, outside of funds used specifically for the SCS and the CECM, be allocated to the

initial operating budget of statistics and actuarial science.

Capital Expenses

• Renovations to the departmental library/Chair's office, former joint general office, new

coffee/meeting room.

• Upgrade to undergraduate computing facility. Annual upgrades will be funded through

the operating budget.

• Computer and office furniture for Departmental Assistant. Note that office furniture for

the secretaries would be transfered from current available furniture for the department's

secretaries.

• Photocopier/facsimile/printer for the main office.

• Refrigerator, microwave, kettle, tables and chairs for meeting room.

.

54,.

25