A FLIGHT SIMULATOR FOR UNIVERSITY DEPARTMENT PLANNING
John F. Herrmann
California State University, Long Beach, CA
John L.
DeOlden
Hughes Electronics Corp., Westchester, CA
ABSTRACT
A simulation model for university
department planning is presented that allows
department chairs, faculty, or administrators to
“test fly" their decision choices for key
variables in the model, observe the impact of
these choices, and learn how to improve
system performance, given their respective
objectives. Simulator "pilots" have control
over average class size and faculty teaching
load, and they can observe the impact of their
decisions on a number of system variables
such as enrollment, percentage of classes
taught by part-time faculty, and the number of
full-time faculty.
Class size and faculty teaching load were
chosen as decision variables to illustrate use of
the model at California State University, Long
Beach because the CSU System budgeting
environment has recently been changed so that
sabbaticals and faculty research time must be
funded at the department level rather than the
university level. Policy constraints on the
hiring and firing of tenured, tenure-track, and
part-time faculty are considered. Student
enrollment is influenced by the university
reputation and student success or failure is
influenced by SAT scores. Class size is
related to student and faculty "burnout" and
faculty teaching load is related to faculty
morale, quality of teaching, and need for
part-time instructors. The academic year
budget for a department is determined by
recent enrollment trends.
The structure of the model used for
this flight simulator is general enough to apply
to any educational system.’ Figure 1 shows a
planning model for an academic department
that is comprised of four related subsystems or
“sectors” and two control blocks. The sectors
are Student Acquisition and Assignment,
Student Education Process, Faculty and Staff
Acquisition and Assignment, and Financial
and Physical Resource Acquisition and
Application. Each of these sectors is
influenced and controlled by the choice of
Policy Variables and the calculations implicit
in Class Requirements Determination. The
model explicitly considers interaction and
feedback among sectors rather than treating
each sector separately.
Policy & Class
Environment }—-—7#™|_ Requirements.
Variables Determination
NG y A
Financial Faculty
&Physicll [yy & Staff
Resource Acquisition
Acquisition & Assignment
& Application
Figure 1. Academic Planning Mode! Overview
The model also considers the impact of
several variables with systemic effects and
shows key relationships that are characterized
ON
by complex behavioral patterns. For example,
the quality of teaching is likely to change as
the knowledge and experience level of
instructors change when use of part-time
faculty increases. Likewise, faculty may
experience "teaching burnout" in conjunction
with larger class sizes and fewer research
opportunities. "Burnout" effects equally
apply to students. As class size increases, the
number of sections of core courses decrease,
reducing scheduling flexibility of students. In
addition, when class size grows, class
availability decreases, and the quality of
teaching declines, the performance and morale
of current students will decline, attrition will
increase, and recruitment of desirable new
students will be increasingly difficult.
The model assumes a given enrollment
environment that leads to a potential market of
students. These students are influenced in their
decision to actually seek enrollment by SAT
scores and the campus reputation. The
reputation of the campus is, in tum,
determined by reputation of the faculty, the
quality of students already attending (as
indicated by SAT scores), class sizes and
availability, and student fees. The reputation
of the faculty is a function of research
performed, quality of teaching provided, and
extent to which part-time faculty is used.
Potential students are those seeking to enroll,
but entry is constrained by capacity (as
determined by available faculty and financial
and physical resources) and the demand of
active students continuing their enrollment.
Active students continue until they graduate,
fail or drop out.
The control panel for the department flight
simulator has been designed to keep
complexity at a minimum while not limiting
the analytical capability of the model. When
too many factors are allowed to change
simultaneously, it becomes very difficult to
sort out cause or determine relative impact for
any given variable. On the other hand, if the
flight simulator does not allow users
substantial freedom in making decisions and
shaping the environment to their liking, then
interest is easily lost and the model may be
dismissed out of hand as unrealistic, too
limited, and therefore not useful for real
decision situations. At CSULB, the authors
have attempted to deal with this dilemma by
providing four pages on the control panel for
users to work with.
Figure 2 shows the first page of the control
panel. Model users can vary average class
size and faculty teaching load and observe
graphical output for enrollment, percentage of
teaching done by part-time faculty, full-time
faculty and other variables of interest. The
graphical display shown in Figure 2 is the
initial page of a “graph pad" that contains
output graphs specified by the user. The
output graphs can display multiple variables
for a given simulation run (e.g., enrollment,
student morale, and percent of sections taught
by part-time faculty), or comparative
information from different simulation runs for
a given variable (e.g., enrollment when class
size is changed from 20 to 30 to 40 and so on).
Numeric displays are provided for wait-listed
students, class capacity, enrollments, number
of sections taught, and number of full-time
and part time faculty.
The second page of the control panel is a
“high-level” map that shows how sectors of
the model are related (Figure 1, shown earlier,
is a simplified representation of this map).
Users can “drill down" into the detail level for
each sector by "clicking" on a symbol in the
sector's title block. Once a user has accessed
the detailed model for a sector, it is possible to
“drill down" still further to examine
definitions of variables and functional and
graphical relationships assumed by the model.
212
bie
Page 1 of 4
University Department Budget
Planning Flight Simulator
to Page 30f 4 >>
Glass Size Choson =
This Simulator has three levels of detail :
1) Paget of the Control Panel provides an introduction
model and its elements.
2) The next level of detail is seen by ‘clicking’ on any
Sectors shown on the Overview map of the model, Page2.
3) The element level is accessed by “clicking* on a
specific element to see that element's definition or
felationship to other variables.
15 eee 45
ZV
(Restore) Sectn per Ft Fac Chosen = [3.3]
20 enfss0 <7
to the
Wal Usted Siod
of the
By changing selected elements, users can see how
various management choices will impact the
department, the faculty, and the students.
4: Class Ent 2
2200.00"
‘50.00
Page 1 is setup so that the impacts of changes to
Class Size [slider labeled “Class Size Chosen*)
and
Sections Taught per FT Faculty
[slider labeled "Sectn per FT Fac Chosen*]
can be seen on the Graphs on this page
The Graphs are :
© Class Enrollments
0 Number of Fulltime Faculty
© Class Enrollments + Number of Fulltime
Faculty + Percent of Section Taught by
Parttime Faculty
30.00
Secins by PT Facully
9: Fulltime Faculty
1500.00
90.00} =
15.00 =
bod
20:
23
712.00 24.00
@__Ervtant & Eovtoment: Page 2
80.00 96.00 132.00
Months 07:04 PM 4N19/96
Page 2 shows the sectors of the model.
The detail mode! for any of the displayed
sectors can be viewed by "clicking" on
the triangle symbol In the sector title
block
to Page 2 of 4
Vv
Vv
Page 3 displays the Faculty and Student Morale
Indicators. The graphically defined functions
underlying these indicators can be changed at
this level
{ie., you do not have to go to the model detail fevel)
Figure 2. The Control Panel (Page 1)
Page 4 contains additional sliders and
Graph Pads for user selected variables.
Ifor example, changes to the potential student
market such as growth or deciinel
The third page of the control panel
contains two graphical displays that determine
student morale and two different graphical
displays that determine faculty morale. Also
included is a "graph pad" for tracking student
and faculty morale either in combination for a
given simulation run or individually for
different runs. If a user wishes to change any
of the functional relationships shown in the
graphic displays, the user simply "clicks" on a
graphical display, and then "clicks" on the
dialogue box version of the graph which
appears. A graphical function of any shape
can then be specified by "clicking" on the
graphical grid of the dialogue box.
The fourth page of the control panel is
intended to provide flexibility to expand the
scope of the user's investigation. Input sliders
and output displays can be added in response
to a given user's desires. For example, a user
might wish to add a slider for setting an
assumed growth rate for the market of
potential students.
As users work with the simulator to
investigate possible consequences of actions
they would like to see taken (e.g., a reduction
in teaching load or smaller classes), and as
they become more familiar with the model and
its assumptions, they may call into question
values of parameters and/or the existence or
strength of relationships posited by the
authors. A key advantage of the model is that
the software allows users to make selected
changes. For example, the authors have
posited a particular inverted, U-shaped curve
that shows how class size is assumed to
impact student morale. If users do not believe
the authors have accurately portrayed the
impact of changing class size on student
morale, they need only specify their own
preferences.
Likewise, as users become still more
familiar with the model, it is likely that they
may wish to broaden the range of variables
over which they have control. Page 4 of the
control panel is designed for this purpose. For
example, changes in class size and teaching
load impact enrollment, so it is logical for
users to wish to examine other factors which
will affect enrollment as well. Obvious
candidates would be SAT score requirements
and tuition fees. It is a simple matter to add
input sliders to the model for these or other
variables. This was not done on page | of the
control panel in order to keep complexity at a
minimum and to promote better understanding
of the interactive effects of only two decision
variables.
The simulator presented here illustrates
that a simple intuitive approach or even a
spreadsheet analysis of the decision to fund
sabbaticals and research time by increasing
class sizes is inadequate. This is because
neither approach explicitly takes into account
complex interactive effects. This study also
helps to illustrate the importance of taking into
account the potentially wide influence of
selected decision variables. For example, as
class size changes are seen to affect
enrollment and pedagogy, enrollment changes
may impact admission standards, and changes
in pedagogy may affect educational objectives
or institutional strategy. Lastly, although the
focus of this study was a framework for
analysis of department budgeting issues, it is
clear that the model lends itself well to
problems of enrollment management. By
shifting concern from class size and faculty
teaching load to admission standards, relative
pricing, and growth of potential student
population, the model can easily be adapted to
address a broad range of other academic
planning issues.
' The model was constructed using ithink
software, Version 3.0.5, High Performance
Systems, Inc. 1994, Hanover, NH.
Xo
CC BY-NC-SA 4.0