Michael W. Maher
Professor of Management and Accounting
Barbara Sommer
Lecturer of Psychology
Gerald Russell
Senior Lecturer of Food Science and Technology
Curt Acredolo
Associate Adjunct Professor of Human and Community Development
Harry R. Matthews
Principal Investigator
Professor of Bio-Chemistry and Director of Media Works
June 7, 2002
Submitted for publication in a prospective book to be edited by Saul Fisher.
This paper provides cost-effectiveness information comparing traditional lecture and online offerings of a large undergraduate course at the University of California, Davis. The student performance data show that the students relying on the lecture class performed better than those relying on the online material, similar to the results in Brown and Liedholm (2002). Interviews with students who relied on the online materials revealed that they had concerns about missing cues about what was important (e.g., for examinations), and they found the mostly text presentation of online materials to lack the liveliness of class lectures. As this was one of the first courses in the Mellon project at UC Davis, the online materials were limited in their interactive features, which reduced their effectiveness vis-a-vis a popular instructor for the lecture course. Using time sheet information for the instructor’s time, cost data for TAs, costs of putting the course on the web, costs of providing software and hardware for the online course materials and costs of building space, we constructed cost estimates as if the instructor had offered a stand-alone online course and as if the instructor had offered a stand-alone lecture course. These cost estimates showed a 10 percent cost savings for the online course compared to the traditional lecture course. The result is lower costs and lower student performance for the online course.
UC Davis is a Research One university with 19,000 undergraduate students, and expects to add 5,000 undergraduate students over the next ten years. University administrators are seeking ways to educate additional students in a cost-effective manner, including online education. We selected Introduction to Food Science and Technology 2 (FST 2) as an appropriate candidate for exploring the potential of online course presentation. FST2 is a popular general education course that enrolled 430 students in the term that we studied. The class provides a description of how raw agricultural commodities are preserved and converted into edible foods. The content covers the regulation of food manufacture as well as the chemistry and microbiology of food quality and safety. The instructor, who consistently gets instructor ratings greater than 4.5 on a 5.0 scale, has a lively in-class presentation. He enhances his lectures with PowerPoint slides that are liberally peppered with images and cartoons relevant to the subject.
A relational database designed for flexible presentation of course materials provided the technical core of the course. The pages were built dynamically by Cold Fusion templates. The course was presented on the web.
To prepare the course for online presentation, we hired student assistants and programmers who worked closely with one of the authors. We used 1) videotapes of the lectures recorded in class the quarter prior to the online course, from which we digitized and transcribed the audio, and 2) the PowerPoint presentations that were used during the lectures of the traditional course offering. We redrew the PowerPoint slides as Flash animations. We designed a template for slides with a uniform layout, which was built dynamically on demand using Flash Generator to merge information from the database with the Flash animation. Using the videotapes as a guide, we broke up the sound and text into blocks that corresponded to slides or steps (such as bullet points) within the slides. We entered these components, together with the graphics, into a relational database. Cold Fusion templates were used to display the modules on the course web pages.
The online version of FST 2 was put into place early in Fall Quarter 2000. As the multimedia files were too large for ease in accessing the materials online, the browser-based lectures and graphics were copied onto a CD-ROM and sold for $2.75 in the student bookstore. The student had to be online to use the CD.
Due to campus policy, there was no way to restrict students to either a strict online or traditional mode of presentation of course materials. Students were free to attend class or to avail themselves of the online material. The instructor announced the availability of the online material, and use of it was assessed by self-report. Each of the examinations contained an item asking students to indicate usage of the online lectures: A = all lecture, B = mostly lecture, C = mostly web, D = all web. All students were obliged to take the examinations in the traditional classroom setting.
The instructor gave three 70-point examinations during the quarter. As technical problems interfered with access to the online course up to the first examination, we restricted our analyses to the second and third examinations, which we will refer to as exams #2 and #3. Preliminary analysis showed that there was a class level effect, with performance improving from frosh to senior standing. Thus, class level was included as a covariate in the analysis, with web usage as a predictor of examination score. We also included individual grade point average (GPA) as a covariate.
Table 1 shows that for Exam #2, students who relied on the lecture tended to do better [F(3,374) = 2.91, p =.034]. A contrast analysis (LSD test) revealed that the “All lecture” group performance was significantly better than that of each of the other three groups (p < .01). The differences among the other 3 groups were not statistically significant.
The next examination served as a replication. The students who attended lectures performed better [F(3,375) = 4.08, p = .007] (Table 2). The group contrasts showed that the “All lecture” and “Mostly lecture” groups outperformed the “All web” and “Mostly web” groups (p < .01).
There was some variability among the students with regard to use of the web across the two examinations. Combining the 'all' and 'mostly' categories, 250 relied on the traditional lectures for both exams, and 68 relied on the web. Another 34 went from lecture to web, and 22 gave up on the web, returning to reliance on the traditional lectures.
Comparing the two clear usage groups -- students relying on the lectures for both exams versus those relying on the web for both exams -- the course pass rates (grade of C or above) were 99.8% and 95.6%, respectively. The overall mean for points accumulated for the traditional lecture group was 94.80 (SD = 9.59, N = 242); the mean for the web group was 91.47 (SD = 9.03, N = 65) [F(1,303) = 4.09, p = .044]. In order to assess the magnitude of the effect, we took the difference between the "lecture" (control) mean and the "web" (experimental) mean, and divided it by the standard deviation of the "lecture" group (Levin, Glass & Meister, 1987). The effect size was .35; a magnitude characterized as small to medium by Cohen (1988). In the context of this study, the effect is smaller than that of being a Freshman versus being a Sophomore with regard to the total points for the class (effect size = .52). It is larger than the magnitude of the difference between Sophomores and Juniors (.13), and between Juniors and Seniors (.06) (using the upper class standard deviation in the denominator for each contrast).
In addition, we compared the overall performance of the Fall 2000 class with that of the students enrolled in Fall 1999. With the exception of the online component, the course material was practically the same, and the instructor uses a consistent criterion for assigning grades. Using a 4-point system (A = 4), the class grade point average for Fall 2000 was 3.43 (SD = 6.85, N = 424). For Fall 1999 it was 3.44 (SD = 6.75, N = 418). The addition of the online materials did not affect overall class performance.
A focus group interview, with eight students who had taken the course with mostly or all web presentations, revealed that the students felt a need for better-written web material. They found it difficult to distinguish between the more important and less important information. They preferred to print out the online material for study rather than read it on their computer screens. In general they felt that the printed lectures were too much like a poorly-written textbook and that a regular textbook would have been a better alternative. They missed the motivation provided by a live instructor, the information cues of inflections and gesture, and the insights gained from the instructor's response to student questions. They expressed the attitude that their tuition entitled them to a live professor, and claimed that online classes were available at a much-reduced price at community colleges.
On the positive side, these students appreciated the convenience of the web-based materials, the fact that they could proceed at their own pace, and review at will. However, they also pointed out the numerous distractions of studying online at home, stating that classroom time was set aside to listen and focus on the course material.
Our aim is to estimate what it costs the university to offer this course online or in the traditional lecture format using activity-based costing. Activity-based costing (ABC), which is essentially the same as the “ingredients method” recommended by Levin and McEwan (2001, pp. 45-76), is the state of the art in business and public sector settings (Hilton, Maher and Selto, 2003, ch. 4). ABC identifies the activities that use instructional resources, assigns costs to those activities, and then assigns costs to “products” (i.e., courses) based on the products’ respective use of activities.
The 'cost' or sacrifice of a resource should be distinguished from the expenditure for the resource. For example, the cost of classroom space for the traditional course offering is the sacrifice of the resource (i.e., the building space), even if the instructor or department makes no expenditure of funds to acquire the building space.[1]
Our cost estimates are potentially 'differential costs.' These are costs that we expect to differ between the online and traditional course offerings.[2] We exclude non-differential costs such as departmental administration, faculty space, and costs that are part of the university infrastructure, such as security, library, student housing, and computer labs. We separate the startup costs for the online course offering from the costs of ongoing course delivery.
Table 3 presents our estimate of the costs of a typical offering of FST 2 in both online and traditional lecture formats to a class of 430 students, which was the enrollment during Fall Quarter 2000. We intend for these estimates to be valid for a stand-alone course using the traditional lecture format that enrolls 430 students and also for a stand-alone online class that enrolls 430 students. We measured costs of the following activities and made the assumptions described below in estimating the costs of the two formats. The numbers by the headings indicate the row numbers in Table 3.
1. Instructor time
Perhaps the most problematic task in computing educational costs is to obtain accurate measures of instructor time. We required that the instructor be willing to complete daily time logs as part of the buy-in for this study. Because this course was offered at the beginning of the Mellon project, we started sending a daily e-mail prompt to fill out an online time log starting July 1, 2000, three months before the October 1 start date of the course.[3] (For subsequent courses in the project, we started the e-mail prompts much earlier.)
We requested the instructor to fill out the daily time log in 0.5 hour increments in six major categories plus several subcategories. We present the timesheet portion of this log in Figure 1. In addition to the time sheet in Figure 1, the e-mail prompt included instructions for recording time in the log.[4]
Figure 1 shows how the timesheet divided into ‘direct time’ and ‘indirect time,’ with the first eight subcategories under direct time. Direct time referred to time easily associated with a particular course offering in a particular term. Indirect time, in contrast, referred to time spent on teaching that benefited teaching the course but could not be easily associated with a particular course offering in a particular term. An example of indirect time was the time spent attending a seminar on teaching technology.
As Figure 1 shows, the timesheet’s columns divided into ‘online,’ traditional’ and ‘joint’ time. Joint time referred to time spent on the course that the instructor could not be easily identify as pertaining to the online offering only or to the traditional offering only. An example of joint time would be the time spent preparing an examination that would be given to both sets of students.
The timesheet provided data on faculty activities for each of the course offerings. To convert time data to costs, we used a cost driver rate of $61 per hour, which equals the instructor's salary and benefits for the year divided by 1,5 00 hours (nine-month academic year).
In Table 3, the instructor time in the columns labeled "Traditional, Cost Driver Volume" and "Online, Cost Driver Volume" is the sum of the joint time plus the separable time for the online and traditional versions, respectively. 'Joint time' is time that the instructor could not separate into either online or traditional categories (e.g., preparing examinations). Specifically, the 112.0 hours shown in the traditional column are the sum of the hours that the instructor recorded as joint (57.5 hours), plus the hours that the instructor recorded as separately identifiable with the traditional course (54.5 hours). In the same way, the 107.0 hours shown in the online column are the sum of the joint time (57.5 hours) plus the hours that the instructor recorded as separately identifiable with the online course (49.5 hours).
The instructor time in Table 3 should be interpreted as the time required to offer the traditional course on a stand-alone basis (i.e., without offering the parallel online course). Similarly, the time required for the online course offering assumes that the online version is a stand-alone course.
We summarize the subcategories of instructor time that were presented in Figure 1 into two categories in Table 3. The first category, labeled 1.1, refers to tasks that generally take place prior to course delivery. Because the instructor probably would be more efficient in preparing the course content of the online version in subsequent years, it is possible that the amounts reported in both the online and traditional columns in row 1.1 overstate the amount of time that would be required in subsequent years for this course because the instructor did a major overhaul of the course for this study. The instructor was unable to separate the time that would be incurred each year from the time that was unique this year. The numbers reported in row 1.2 refer to the time spent delivering the course, which are likely to be the same whether this was the first or later offerings of the online course.
These figures are based on the hiring of three Teaching Assistants at 50 percent time, paid $5,853 each, which includes benefits (based on University payroll data). The instructor used the Teaching Assistants mostly to help write examinations and to grade examinations. A small proportion of Teaching Assistant time was spent dealing with students. Considering that the Teaching Assistants were used mostly to prepare and grade exams, we assume that both the online and traditional course offerings would each use three Teaching Assistants for a class of 430 students.
Table 4 presents estimates of the cost of launching the online course that were derived from time sheet information and combined with salary and benefit data. The staff time used to put FST 2 on the web is based on two categories of time. The first category (startup hours) is startup time that would benefit future offerings of the course until it was completely redesigned. We assumed that the basic design of the course would remain intact for five years, so these "startup" costs would be amortized over five years. The second category (annual course revision) is time required for yearly routine revision of the course. We relied on the instructor and the staff who put the course online to inform us how much staff time was used for startup and how much was used for the routine annual revision of the course. These costs are shown in Table 4.
3.1 Amortization of startup costs
Startup time is composed of time for content development, programming and supervision/administration of the project. We computed startup costs using actual time sheets, along with salaries and benefits for the staff that worked on putting FST 2 on the web. The total startup cost was $20,307, which we amortized over 10 course offerings. We amortized the startup costs over 10 course offerings because the course is offered twice per year, and we estimated a useful life of five years for the startup work, not counting annual course revisions that are discussed in section 3.2 below (10 course offerings = 2 courses per year for 5 years).
We calculated the amortization using the capital recovery factor that takes into account both the annual cost amortization and the interest that could have been earned on the money spent in a lump sum at the start of the project instead of spread over the life of the project (Levin and McEwan, 2001, pp. 65-70). For our computations, we used the estimated five-year life of the startup costs and an interest rate of 6.5 percent, which is the rate used by the university to compute the space usage costs described later in this paper. The computation appears at the bottom of Table 4, and results in assigning a cost of $2,437 to the online course offering as shown in row 3.1 of table 3.
3.2 Annual course revision costs
The instructor indicated that he made some revision of the course each year. These annual revisions did not entail a complete teardown and rebuild of the course, but were those required to keep the course up-to-date. Those annual revisions would require programming and content development. The total costs (of time) for annual course revisions was $2,718, as shown in Table 4. As there are two offerings per year, we used $1,359 = $2,718/2 as our estimate of the course revision costs for this offering of FST 2, as appears in row 3.2 of table 3.
4. Computer hardware and software
This project required obtaining computer hardware; namely, a server and two computers for programming and software. The computer hardware cost $5,800; the software cost $1,600.
4.1 Hardware
The server provides a capacity of 10 courses per quarter and has a useful life of three years. Therefore, we first computed an annual depreciation amount using the capital recovery factor in Levin and McEwan (2001, p. 69) for a 6.5 percent interest rate and three-year asset life, then assigned 1/30 of that annual amount to FST 2 course because the annual server capacity is 30 courses (10 courses per quarter times three quarters per year). This computation resulted in a cost of $73 per course, which appears in row 4.1 of table 3.
4.2 Software
The software would be used for 18 months. Using the same approach as above to compute the annual amortization, we assigned 1/30 of the annual amount to FST 2. This computation resulted in a cost of $41 per course, which appears in row 4.2 of table 3.
4.3 Maintenance
The maintenance cost of the server is $6,600 per year. Therefore, we charge each course for 1/30 of the costs of maintenance (1/30 = one course in one quarter divided by the product of 10 courses per quarter times 3 quarters per year). The computation results in $220 per course, which appears in row 4.3 in table 3.
5. Space
For space costs, we used amounts provided by the University's Department of Budget and Planning, which were derived from plans for a new lecture hall to be built on campus. These costs represent the opportunity cost of students’ using space in lecture halls. Taking into account the projected cost of the new lecture hall, an annual debt service rate of 6.5%, a 30-year life, and an assumed allowance for utilities and maintenance, the university figured the total annual cost of the lecture hall to be $430,000. Using standard university assumptions about the number of hours that lecture halls are used, the personnel from the Department of Planning and Budget converted this annual cost of $430,000 into a cost per student of $17.86, for a course that uses the lecture hall three hours per week for the 10-week quarter. For a class of 430 students, the total space cost was $7,680.
The space costs do not include the cost of land nor the cost of campus infrastructure, such as roads, pathways, libraries, grounds-keeping, and security. We expect that increasing the number of the traditional lecture courses will have a greater impact on these costs than will increasing online courses. We take the perspective of the university in our cost analysis, so the costs to students (e.g., enrollment fees, using computers, time spent attending lectures or studying online, purchasing textbooks and other course-related materials) is outside of the scope of this study and remains a potential area of inquiry for future research.
As Figure 1 shows, we included a row in the time log for "Working with project staff" which is the time that the instructor spent with us in meetings and in general Mellon project administration. We excluded that time from our estimates of costs of offering FST 2 because that time was uniquely required for this project, and not specifically for course offerings.
Also with regard to instructor time, we excluded indirect time from our estimates. "Indirect time" is time spent on teaching in general, but not for the particular course being studied (see bottom portion of Figure 1). We requested that information on the time log in order to capture all teaching-related activity. In the case of FST 2, the online indirect time was 28.5 hours, 38 hours for traditional, and 43 hours of joint time.
Table 3 presents the cost of offering the online course as well as the cost of using the traditional lecture approach to offer the course. Table 5 shows the differential cost of online versus traditional instruction. Replacing the traditional course with an online version would result in a savings of $4,348. These savings were obtained at the cost of a modest, but reliable, decline in student performance with an effect size of .35.
To calculate cost effectiveness ratios, we use the total costs from Table 3, a base number of 430 students and the pass rates of 99.8% for the traditional course ($45,177/429 students), and 95.6% for the online course ($40,829/411 students). The resulting cost per student who passes the course is $105 for the traditional course and $99 for the online version.
The data from Table 5 show that the major source of the savings from the online course is in the savings of space costs. In this study, there was little differential cost for instructor time and no differential cost for teaching assistant time. Most of the online preparation time involved getting the material ready for the staff to put online.
The added expense for the online version, that is not present for the traditional class, is primarily in staff costs for administration, content development, and programming. The server costs, which can be spread over multiple courses, are relatively low.
Conclusions and discussion
Cost behavior with increased enrollments
In thinking about accommodating increasing course enrollments, the online course scales more smoothly with increasing student numbers. There is a large up-front cost of acquiring resources necessary to offer the online course (e.g., course preparation; server costs) which is incurred however small the number of students. The incremental cost per student (largely TA time) above the up-front cost is approximately constant as student numbers increase. Thus, there is a large cost to mount the course but the cost of the course then rises slowly and smoothly as student numbers rise.
On the other hand, the upfront cost of acquiring the resources for the traditional course (e.g., course preparation, lecture hall space) is generally lower than for the online course and initially rises slowly with student numbers (e.g., TA time). However, when the lecture hall becomes full, additional students generate a large step increase in costs due to moving to a larger lecture hall, or creating an additional section of the course with additional instructor salary. Thus, the traditional course scales in a step-wise fashion with student numbers. Administrators desiring efficient use of capacity will tend to "cap" the size of the course just before one of the steps. These steps can be very large; for example, adding a second section might come close to doubling the expenditure of resources for the course. In this sense, the online version is more scalable than the traditional lecture version.
Student performance
The fact that the online students performed less well than the students who attended the traditional lecture was an unanticipated outcome. This result might be because students who took the online version lacked the discipline of lecture attendance and the motivation provided by this excellent instructor. Our effect estimate is based on just one example, however. Given the large number of uncontrolled variables, we need to repeat the experiment a number of times to obtain a more valid estimate. As we were unable to randomly assign students to the two versions of the class, we do not know the degree to which the difference is due to the quality of the materials or to self-selection factors. Although a random assignment of students to conditions would be a desirable test, in actual circumstances, students are likely to be self-selected with regard to their enrollment in online courses. It is probably safe to assume that the absence of a more immediate contact with the instructor is a factor that needs to be taken into account and compensated for in some way when preparing course materials.
In addition to the costs shown above, there were costs of a more subjective nature. On the technical side, the process of moving from a boutique project to an enterprise application involved shifting to an Oracle database housed in the campus data center, and led to a significant loss of control of the application by the project programmer. There were configuration problems with the Oracle database. The conversion required a much larger commitment of Oracle database administrator time than was originally expected. Similarly, the Cold Fusion templates were moved to a UNIX computer. Apart from the conversion of file names to a case-sensitive format and a greater sensitivity to thread locking, moving the Cold Fusion part was relatively trouble-free.
Fewer students than hoped availed themselves of the online materials. There were two reasons: 1) despite pilot testing, producing the materials and providing them in a browser format required more time and training of programmers than expected. As a result the presentation was not as polished as it might have been; and 2) the server was not reliable, and students became frustrated and irritated at the unpredictable lack of access.
A major challenge in the realm of online pedagogy is to construct something with at least comparable quality and appeal as the lecture format. Taking advantage of the potential of the Internet requires a structure and prototyping that differs from that currently in place. Innovators are up against decades of established practice and procedure. Faculty are accustomed to presenting their material in a lecture format, and many students come to college with a cherished vision of a professor at the podium providing them with the wisdom of the ages.
In many colleges, particularly at the research university level, there are few incentives for change. As this case study demonstrates, putting a course online requires considerable time on the part of the instructor -- time away from a research pursuit. The degree to which the institution is willing to credit the time spent for course development and preparation will be a major influence on whether or not faculty avail themselves and their students of the potential of online education.
Figure 1. Faculty Time Log
|
Direct Time: |
Online |
Traditional |
Joint |
||
|
1. |
Working with project staff. |
|
|
|
|
|
2. |
Planning the course, developing materials, developing lecture content. |
|
|
|
|
|
3. |
Preparing the course for online delivery. Reviewing materials for online delivery. |
|
|
|
|
|
4. |
Delivering the course in a particular term. Preparing for lectures, delivering lectures, dealing with problems with the online and traditional delivery of course content to students. |
|
|
|
|
|
5. |
Interacting with students outside of class. |
|
|
|
|
|
6. |
Evaluating student performance. Preparing and grading examinations, grading papers and projects. Assigning grades. |
|
|
|
|
|
7. |
Training and supervising TAs and other assistants. |
|
|
|
|
|
8. |
Other (please specify). |
|
|
|
|
|
Indirect Time: |
|
|
|
||
|
1. |
Attending seminars, reading, collecting materials useful for teaching. |
|
|
|
|
|
2. |
Interacting with students that cannot be identified with a particular course offering. |
|
|
|
|
|
3. |
Planning and developing materials for future offerings of the course. |
|
|
|
|
|
4. |
Other (please specify). |
|
|
|
|
|
Expenditures (Please describe briefly and put $ amounts in the spaces to the right). |
|
|
|
||
Table 1. Mean (SD) scores for Exam #2, including class level and GPA
|
Group |
Exam score |
Mean Class level* |
Mean GPA** |
N |
|
All lecture |
50.73 (8.60) |
2.23 (1.11) |
2.92 (0.54) |
205 |
|
Mostly lecture |
48.38 (7.93) |
2.32 (1.19) |
2.78 (0.53) |
84 |
|
Mostly web |
47.68 (7.46) |
2.36 (1.18) |
2.79 (0.46) |
53 |
|
All web |
47.29 (7.33) |
2.84 (1.20) |
2.67 (0.49) |
38 |
*1 = FR, 2 = SO, = JR, 4 = SR
**4 = A, 3 = B, 2 = C, 1 = D, 0 = F
Table 2. Mean (SD) scores for exam #3, including class level and GPA
|
Group |
Exam score |
Mean Class level* |
Mean GPA** |
N |
|
All lecture |
59.65 (6.51) |
2.33 (1.12) |
2.97 (0.51) |
181 |
|
Mostly lecture |
59.46 (7.59) |
2.22 (1.10) |
2.83 (0.55) |
96 |
|
Mostly web |
55.00 (9.48) |
2.38 (1.14) |
2.64 (0.54) |
52 |
|
All web |
56.71 (9.36) |
2.65 (1.25) |
2.77 (0.48) |
52 |
*1 = FR, 2 = SO, = JR, 4 = SR
**4 = A, 3 = B, 2 = C, 1 = D, 0 = F
Table 3. Cost estimates for the traditional
course and its online alternative
|
|
|
|
Traditional |
Traditional |
Online |
Online |
|
|
Activity |
Cost Driver |
Cost Driver Volume |
Cost
|
Cost Driver Volume |
Cost |
|
1. |
Instructor time: |
|
|
|
|
|
|
1.1 |
Course planning and development |
Faculty hours |
112.0 hours |
$6,944* |
107.0 hours |
$ 6,634* |
|
1.2 |
Delivering the course |
Faculty hours |
213.0 hours |
$12,993* |
205.0 hours |
$12,505* |
|
2. |
Cost of teaching assistants |
Teaching assistants |
3 teaching assistants |
$17,560** |
3 teaching assistants |
$17,560** |
|
3. |
Staff time to put the course onto the web: |
|
|
|
|
|
|
3.1 |
Startup costs (see Table 4) |
Staff hours |
0 |
$0 |
172 hours |
$2,437 |
|
3.2 |
Annual course revision (see Table 4) |
Staff hours |
0 |
$0 |
121 hours |
$1,359 |
|
4. |
Server costs:
|
|
|
|
|
|
|
4.1 |
Hardware |
Number of courses using server |
0 |
$0 |
1 course |
$73 |
|
4.2 |
Software |
Number of courses using server |
0 |
$0 |
1 course |
$41 |
|
4.3 |
Maintenance |
Number of courses using server |
0 |
$0 |
1 course |
$220 |
|
5. |
Space usage costs |
Number of students |
430 students |
$7,680***
|
0 |
$0 |
|
|
Total |
|
|
$45,177 |
|
$40,829 |
* Cost driver rate = $61 per instructor hour based on the instructor’s actual salary and benefits and using a nine-month (1,500 hour) academic work year.
** Cost driver rate = $5,833 per teaching assistant per quarter using actual salary and benefits for a 20-hour work-week for the quarter.
*** Cost driver rate = $17.86
per student per quarter for this course according to the University Office of
Planning and Budget.
Table 4. Costs of putting the course on the
web
|
Staff |
Startup hours |
Startup costs |
Annual course revision hours |
Annual course revision costs |
|
Administrative and supervision |
491 |
$9,190 |
69 hours |
$1,289 |
|
Content development |
686 |
$5,718 |
172 hours |
$1,429 |
|
Programming |
540 |
$5,399 |
0 |
$0 |
|
Total |
1,717 |
$20,307 |
241 hours |
$2,718 |
Computation of cost assigned to FST 2 for this study
Course’s share of startup costs:
Assume a 5 year life and interest rate of 6.5 percent per year.
Capital recovery factor (Levin and McEwan, 2001, p. 69):
Annual amortization = Principal [r(1 + r)n]/[(1 + r)n – 1]
= $20,307 [0.065(1 + 0.065)5]/[(1 + 0.065)5 – 1]
= $4,874
Assuming two courses per year, the cost per course = $4,874/2 = $2,437 (used in Table 3)
Share of annual course
revision costs = $2,718/2 = $1,359 per course (used in Table 3)
Table 5. Savings by substituting an online
course offering for a traditional course offering
|
|
Activity |
Cost Driver |
Difference in Cost Driver Volume |
Cost Savings with Online Offering |
|
1. |
Instructor tasks |
|
|
|
|
1.1 |
Course planning and development. |
Faculty hours |
5.0 |
$310 |
|
1.2 |
Delivering the course. |
Faculty hours |
8.0 |
$488 |
|
2. |
Teaching assistants |
Teaching assistants |
0 |
$0 |
|
3. |
Getting course onto the web |
Staff hours |
|
$(3,796) |
|
4. |
Server: hardware, software, and maintenance |
Courses (capacity utilization) |
1 |
$(334) |
|
5. |
Space usage. |
Number of students enrolled in the course |
430 |
$7,680 |
|
|
Savings from substituting an online course offering for a traditional course offering |
|
|
$4,348 |
Brown, B. W. and C. E. Liedholm (2002). “Can Web Courses Replace the Classroom in Principles
of Microeconomics?” American Economic Review, 92 (2), forthcoming.
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. (2nd ed.). Hillsdale, NJ: Erlbaum Associates.
Hilton, R. W., Maher, M. W. and Selto, F H. (2003). Cost Management: Strategies for Business Decisions (2nd ed.).Burr Ridge, IL: McGraw-Hill/Irwin.
Levin, H. M., Glass, G. V., & Meister, G. R. (1987). “Cost-effectiveness of computer-assisted instruction.” Evaluation Review, 11(1), 50-72.
Levin, H. M. and McEwan, P. J. (2001). Cost-Effectiveness Analysis (2nd ed.). Thousand Oaks, CA: Sage Publications.
Maher, M. W. and Marais, M. L. (1998). “A Field Study on the Limitations of Activity-Based Costing When Resources Are Provided on a Joint and Indivisible Basis.” Journal of Accounting Research, 36 (1), 129-142.
Mathews, H. R., Maher, M. W., and Sommer, B. A. (2001). “Costing Online Courses—A Contemporaneous Time Log.” Presented at Syllabus2001, Santa Clara, CA., July 2001.
[1] See Levin and McEwan (2001, pp. 64-66), Hilton, Maher and Selto (2003, ch. 4, 5) and Maher and Marais (1998) for further discussion of the distinction between the cost of using resources and the costs to acquire resources.
[2] We use the term ‘differential’ to encompass the concept of incremental costs (cost increases) and decremental costs (cost decreases). To us, differential analysis is the same as marginal or incremental analysis (Levin and McEwan, 2001, pp. 101-102).
[3] 40 percent of the instructor’s total time developing and offering the course was spent before July 1, 2000 and had to be collected retrospectively. Time collected retrospectively was mostly for course revision and comprised 68 percent of the total time spent only on the traditional offering, 58 percent of the total time spent only on the online offering, and 14 percent of the time spent that was ‘joint’ time not clearly separable into online or traditional offerings.
[4] Detailed information about the time log appears in Mathews, Maher and Sommer (2001), available at http://cloudybay.ucdavis.edu/mellon/publications.html