WCETs 15 Annual Conference November 5, 2003 San Diego, CA TCM and miniBRIDGE: Using and Interpreting

1
(No Transcript)
2
WCETs 15 Annual ConferenceNovember 5, 2003 San
Diego, CATCM and mini-BRIDGE Using and
Interpreting TCM Cost Data

• Frank Jewett, Principal Investigator TCM/BRIDGE
  Project
• The Technology Costing Methodology Project is
  sponsored by the Western Cooperative for
  Educational Telecommunications (WCET) and funded
  by FIPSE. The TCM/BRIDGE component of the
  project is funded by a supplemental grant to WCET
  from the Andrew W. Mellon Foundation.

3

• The TCM Project has clarified many of the major
  issues related to costing the use of information
  technology for instructional delivery.
• The supplemental TCM/BRIDGE Project
  (mini-BRIDGE) provides a way to interpret and
  better understand the implications of the TCM
  cost data that are obtained.

4

• To illustrate, I will use results from one of the
  pilot campuses that was involved in the
  TCM/BRIDGE Project.
• The cost data and the report on all six campuses,
  Applications of the Mini-BRIDGE Model to TCM
  Cost Data, are available at www.wiche.edu/telecom
  /projects/tcm/bridge
• Also see F. Jewett and T. Henderson, The TCM
  Project, Collecting and Interpreting
  Instructional Cost Data, Planning for Higher
  Education, Sept-Nov 2003

5
TCM data for courses at WSU

• Mediated courses (10)
• Design Develop Interact and
  evaluate
• Average 4,137 4,985 4,891
  
• Total cost 14,013 average
  enrollment 17
• Average cost per student
  824
• __________________________________________________
  ________________________________________________
• Classroom courses
• The average cost per student in a regular 3 unit
  undergrad. (lecture/lab) course was 448, or
  about half the cost of the mediated courses.

6
What should we conclude?

• Are mediated courses more expensive than
  classroom courses?
• Yes, if we rely upon the simple comparison.

7
Take another approach (as discussed in the TCM
Handbook)

• We need a model or theory or hypothesis in order
  to interpret the data. For example, providing
  course instruction involves certain tasks
  including
• preparation of course materials,
• delivery/presentation of course materials,
• interaction with students,
• evaluation of students.

8
When a course is taught in a classroom...

• An individual faculty member is usually
  responsible for all of these activities, bundled
  under the heading of faculty workload
• There is no logical reason why the activities
  must be bundled
• Information technology provides an opportunity to
  unbundle the activities into course related
  (prepare and present) and enrollment related
  (interact and evaluate)

9
Based upon these categories, we can rearrange the
WSU cost data as follows
10

• The table contains the elements of a cost
  model, a relationship between course costs and
  course enrollment. (The model is called
  mini-BRIDGE.)
• For a (3 unit) classroom course we estimate costs
  at 448 per enrollee.
• Direct Cost 448 x Enrollment
• For an online version of the same course we
  estimate costs at 9,122 (course related costs)
  plus 228 per enrollee (enrollment related
  costs).
• Direct Cost 9,122 228 x Enrollment

11
We could do a table showing the costs associated
with different enrollments for each mode but it
is easier to use a graph of mini-BRIDGE to
illustrate...
12

• Based upon the initial TCM cost data I had from
  WSU, the mini-BRIDGE model showed a crossover
  point point, where the costs of the online and
  classroom instruction are equal at an enrollment
  of 57 students.
• For enrollments below 57, classroom instruction
  was less expensive, for enrollments above 57,
  online was less expensive. (go back to graph)

13
But wait! After reviewing the original cost
estimates in the mini-BRIDGE context...

• WSU determined that the original TCM cost data I
  had used were overstated.
• Initial development costs are not incurred every
  time the course is offered online (a weighted
  average of such costs over 4-5 offerings was
  5,013 instead of 9,122).
• Variable costs per student were lower for large
  enrollment courses which constitute about 30 of
  the online total (a weighted average including
  large enrollment courses was 271 instead of
  288).
• Revised Direct cost 5,013 271 x Enrollment

14
WSU Cost Estimates for Online Courses Adjusted
for Multiple Offerings and Large Enrollment
Courses - Mediated (b)
15
Warnings and Reservations

• The WSU cost parameters are averages of ten
  courses that were converted to online. These
  averages do not represent the aggregate costs
  associated with the ten courses unless
  enrollments are the same in all courses.
• The costs were estimated based upon activity
  surveys of faculty and staff. They include only
  a small component of overhead costs such as
  departmental research and committee assignments
  because most of the faculty were part-time.
• Applying an adjustment factor to account for
  these additional activities increased the
  crossover point about 75, from 57 to 100.

16
What can we conclude? (1)

• TCM is the only costing approach in this country
  that relies upon a common, written methodology
  (contained in the TCM Handbook).
• Simple comparisons of specific cost values for
  classroom and mediated courses are interesting
  but not very helpful for policy purposes.
• TCM has a cost model embedded in it that helps us
  to interpret the specific cost results observed
  over a range of enrollments.

17
What can we conclude? (2)

• The cost structure of mediated courses is
  basically different than that of classroom
  courses.
• Mediated courses tend to have (a) higher fixed
  costs and (b) lower incremental (per student)
  costs than classroom courses. (This result was
  observed in 5 of the 6 mini-BRIDGE cases and 6
  of the 8 benefit cost cases I did in 1996-98.)
• The lower incremental (enrollment related) costs
  are a trade-off for the higher fixed (course
  related) costs of mediated instruction.

18
What can we conclude? (3)

• Because of the fixed costs, at relatively low
  enrollments classroom instruction is usually less
  expensive (to the left of the crossover point).
• But, as enrollment increases, mediated costs
  increase at a slower rate than classroom costs,
  at some point the mediated version of the course
  costs less (to the right of the crossover
  point).

19
What can we conclude? (4)

• It is the fixed costs associated with mediated
  courses that give rise to scale economies,
  average cost per student declines as enrollment
  increases.
• To fully realize the savings potential of these
  scale economies, it is necessary to substantially
  reorganize the way instruction is produced and
  delivered.
• This has major implications for the way
  institutions are structured and managed.

20
For Further Information

• Frank Jewett, WCET, Principal Investigator
  TCM/BRIDGE Project. ltFrnkJewett_at_aol.comgt
  714-990-9506
• also see
• Dennis Jones, TCM Handbook (Version 1) available
  at ltwww.wiche.edu/telecom/projects/tcmgt
• For an analytic discussion of the crossover
  model used here see F. Jewett, A Framework for
  the Comparative Analysis of Classroom Instruction
  vis-a-vis Distributed Instruction, Ch. 5 in M.
  Finkelstein, et al, Dollars, Distance, and Online
  Education the New Economics of College Teaching
  and Learning, ACE/Oryx, Phoenix, 2000.

21
TCM 3

• A Proposal for Connecting and Updating the
  Costing Software

22
We have three independent components of costing
software

• The TCM Tabulator supports the TCM Handbook it
  allows a user to input original cost data
  elements and obtain course cost estimates.
• Mini-BRIDGE is a conceptual model and an EXCEL
  spreadsheet that uses TCM cost data to graph and
  compare the cost functions for a course delivered
  using various instructional modes.
• BRIDGE is a computer cost simulation model that
  compares the total cost of growing a campus using
  classroom delivery vs. various mixes of mediated
  delivery, e.g., 50 classroom, 20 television,
  30 online.

23

• I developed the BRIDGE campus cost simulation
  model in 1996-98 (pre-TCM) as part of a U.S.
  Dept. of Ed. grant.
• It was developed when TV was a major mode of
  mediated delivery, and before the terms online
  learning and e-Learning had come into general
  use.
• BRIDGE is based upon three generic types of
  course delivery classroom, asynchronous network,
  and broadcast television.
• A substantial part of the BRIDGE software is
  designed to allow a user to specify cost
  parameters for a generic broadcast (TV) course
  and an asynchronous network course.

24
A question that arose early (1999) in the TCM
Project

• Can we connect the TCM course cost data to the
  BRIDGE campus cost simulation model?

25

• This connection was the objective of the
  TCM-BRIDGE Project.
• The result of the project is mini-BRIDGE which
  provides the course cost parameters from TCM that
  can be used in BRIDGE.
• Mini-BRIDGE is the conceptual connection between
  the TCM course cost data and the BRIDGE campus
  cost simulation model.

26
Need to update/upgrade BRIDGE

• Since 1998 we have gained more experience with
  instructional delivery modes that are not
  explicitly in BRIDGE including
• regular classroom instruction augmented with IT
  (IT is an add-on)
• classroom instruction partially displaced with
  online instruction (the hybrid course)
• In addition, there is a need to deal explicitly
  with a wider variety of online and television
  courses rather than just one of each

27
Need to update/upgrade BRIDGE (cont.)

• to add revenues to the campus cost analysis
• finally, we need to find a home for mini-BRIDGE
  the logical choice is to incorporate it in TCM
  Tabulator

28
Integrating the pieces...a software interface

• We need to develop a software interface that
  allows TCM cost parameters to be transferred to
  BRIDGE directly thus eliminating the need to
  design mediated courses in the BRIDGE model
  itself
• This software interface could be designed as a
  data repository where TCM cost data and cost
  parameters could be maintained and retrieved for
  use in BRIDGE or for other analytic purposes
• Such a repository could have a central component
  where, at campus option, cost data could be
  maintained to support comparative studies,
  regional analyses, etc.

29
TCM 3 Prospectus Review Draft Software
Components and Data Flow (page 7)
Denotes Data Output and Input
 
  BRIDGE II  
LOCAL REPOSITORY
DATA EXCERPTS Campus Option
Local Cost Simulation Analysis
WCET SCREENING
 
CENTRAL REPOSITORY
Methodology Updates for Tabulator
Data for BRIDGE II
Custom Reports Statistics Regional Analysis State
Analysis National Analysis
30
If there is interest, we can look at BRIDGE after
the break.
31
Older Stuff re TCM Project
32
The costs of mediated instruction (1)

• What is the problem?
• Higher Eds costing structures and conventions
  were developed in 1960s-1970s when almost all
  instruction was delivered in classrooms by
  individual faculty.
• Since the mode of instructional delivery was
  essentially homogeneous it was easy to aggregate
  across courses and calculate cost per FTE by
  disciplines (Social Science, Humanities...) and
  levels of instruction (lower division...).

33
What is the problem ? (2)

• These traditional cost structures help us to know
  that Engineering instruction is more expensive
  than English, or that upper division costs more
  than lower division.
• But they are of little use if one asks whether
  classroom is more or less expensive than mediated
  instruction.
• We needed a new perspective on instructional
  costs.

34
The costs of mediated instruction (3)

• Studies from the UK Open University -
• Tony Bates, Technology, Open Learning, and
  Distance Education, 1995.
• Sir John Daniel, Mega Universities and Knowledge
  Media, 1996.
• Greville Rumble, The Costs and Economics of Open
  and Distance Learning, 1997, includes reports on
  cost studies in the UK back through the 1970s.

35
The costs of mediated instruction (4)

• Projects -
• Center for Academic Transformation (RPI/Pew)
  The KISS Approach to Costing, Learning
  Marketspace, Oct. 22, 2003.
• Flashlight Project - 1994 to present.
• The Costs of Networked Learning (United Kingdom,
  but not OU) - 1997 to present.
• Case Studies in Evaluating the Benefits and Costs
  of Mediated Instruction and Distributed Learning
  1996 to 1998.

36
The TCM Project

• The Technology Methodology Costing Project (TCM),
  1998-2003, is a federally funded effort to
  develop a generally agreed upon costing
  methodology for mediated instruction.
• Sponsored by WCET, funded by FIPSE, led by Dennis
  Jones of NCHEMS.
• Worked with 20 pilot campuses to develop an
  authoritative methodology for costing course
  instruction of all types.

37
The TCM Project (cont.)

• Phase I of the project is complete Technology
  Costing Methodology Handbook (1.0), 2001.
  available at ltwww.wiche.edu.telecom/projects/tc
  m/index/htmgt
• Phase II is in final stage. Also funded by
  FIPSE, it includes extending the methodology to
  more complex situations, a second version of the
  TCM Handbook, and provides for dissemination and
  training

38
The TCM adopted a cost perspective that had
evolved from the benefit cost case studies

• Calculate the costs of individual courses.
• Use the data to compare the costs of alternative
  forms of delivery (e.g., classroom vs. online).
• Focus upon the direct costs of activities
  required to provide courses to students.
• In general, ignore indirect costs because they
  will appear on both sides of the classroom versus
  mediated cost comparison.