Teaching Load Summary
C. F. Niederriter - April 1997
For the past several years, the Dean's office and the Faculty Senate have struggled with defining teaching loads at Gustavus for the purposes of determining overload situations and staffing allocations. Several ad hoc committees have looked into the matter, without resolution. At least since the 1992-1993 school year, the Registrar has tabulated Faculty Load data by department, including number of sections (or contact hours) taught, full time load (as determined by the department), enrollment, and enrollment per faculty. All of the data is confused by the various ways that we calculate and/or discuss teaching load; courses, contact hours, and combinations of the two. This has been (and currently is) the source of much frustration for the committees discussing the issue. Some have suggested that things would be better if only we used semester hours instead of courses as a measure, but this ignores the inherent differences in the courses we teach and the evaluation tools that we use.
The author has taken data provided by the registrar's office for both the Fall Semester of 1996 and the Spring Semester of 1997 and tabulated a number of parameters associated with teaching loads. The data provided included First Term Seminars (for the Fall) and Curriculum II (both semesters), independent studies, etc. However, for the purposes of calculation, courses that have arranged times were not counted, including such things as music lessons and independent study research. This was not done as a result of any prejudice, but due to the fact that the number of contact hours and the combined scores could not be measured for arranged courses. This obviously effects some departmental data more than others, particularly Music, some science departments, and other areas with a significant number of arranged courses.
Figure 1: Average number of courses taught per faculty by department for Fall 1996.
If one looks at the average number of courses taught in a semester by department, as shown in Figures 1 and 2, it is apparent that some departments are lower than the college average (2.8 for Fall '96 and 2.7 for Spring '97) and some are higher. Departments that teach a significant number of laboratory sections are generally lower because labs often carry little or no credit, although they take a substantial amount of faculty time. Music appears low on these plots only because the author could not properly account for music lessons. Some departments clearly are teaching more courses than average, as well. However, it should be noted that all but five departments are within one standard deviation of the college average (mean), three low and two high.
Figure 2: Average number of courses taught per faculty by department for Spring 1997.
For a number of years contact hours have been used by some departments, mainly in the sciences and primarily for simplicity in comparing to national standards. As one might expect, the departments that appeared low in number of courses, are higher on a graph of contact hours by department, as shown in figures 3 and 4. The inverse is also true to some extent, departments that are higher on the courses graphs are lower on the contact hour graphs. Again, it should be noted that all but 4 or 5 departments are within one standard deviation of the mean, 2 on the high side and 2 or 3 (Spring) on the low side. The disclaimer about music lessons applies here as well.
Figure 3: Average number of contact hours per faculty member for Fall Semester.
Figure 4: Average number of contact hours per faculty member for Spring Semester.
Another important factor to take into consideration when discussing teaching load is the number of students in the courses taught. Plots of number of students per FTE are shown in figures 5 and 6. As we might expect, there are some departments who teach an above average number of students and some who teach fewer than the average. But only 5 departments are outside one standard deviation from the mean.
Figure 5: Average number of students per faculty for the Fall Semester.
From the data presented so far, one should conclude that each of these measures of teaching load, number of courses, contact hours, or number of students, is inadequate. Clearly faculty in different departments teach in different ways, some meeting students more often, some less often. Some of the differences can be attributed to numbers of students, recommendations of outside agencies, etc. Whatever the reasons for the differences, they do exist as should a reasonable and equitable measuring technique. The ideal approach would be for all faculty to record all of the time spent on each of their courses so their teaching load could be tabulated. Since this hasn't been done, and if it were begun now would not yield useful information for at least a semester, it is necessary to attempt to model the situation. The best model would somehow take into account all of the different ways that we teach and all of the constraints involved (like number of students, etc).
Figure 6: Average number of students per faculty for the Spring Semester.
The model that I am suggesting is an attempt to take into account several of the important factors in teaching load, class time, preparation, and grading. Class time and preparation time are both associated with the number of contact hours but are not directly related to the number of students in the course. Grading load, however, is strongly linked to the number of students and the type of course. For these reasons, the combined score for a faculty member is calculated by combining contact hours and number of students:
Combined Score = # of Students * X + # of Contact Hours * Y
The multipliers, X and Y are somewhat arbitrary, and I have settled on the following values because I believe that they most closely approximate reality. X can be 0, ½, or 1 depending upon the type of class that is being taught. A normal class would use the value ½ (times the number of students) to approximate the grading load, while a writing course would use a value of 1. Seminars, or other courses where there is little or no grading involved, would use the multiplier of 0 for X. To tabulate preparation and class time, I suggest using values of Y like 1, 2, and 3. A value of 2 would be used for normal courses, assuming that it takes just as long to prepare for a class as to teach it. If a faculty member has multiple sections of the same course, one of these would be assigned a value of 2 for Y and the other(s) would be assigned a value of 1. The last value, 3, might be used for a course that the faculty member has never taught before and must spend more time preparing for (I haven't made use of it yet).
Figure 7: Average Combined Scores for Fall Semester.
The resulting Combined Scores are much more tightly grouped around the college average as can be seen in figures 7 and 8. There are still 5 or six departments outside one standard deviation from the mean, but the standard deviation is smaller in this case. Also, one of the departments that is low is Music, for which I have already stated that I do not have an appropriate way to count lessons. Based on the smaller spread, I would suggest Combined Score is a better way to discuss teaching load across campus. The multipliers could, perhaps, be fine tuned a bit, but the basic concept of including all of the major components of teaching in some way is better than using only one component. It should also be clear from this perspective that teaching loads are fairly even across campus.
Figure 8: Average Combined Scores for Spring Semester.
There are limitations to this study, notably that it focuses on departmental averages of teaching load. But one should be able to conclude from the information presented here that teaching load is fairly evenly distributed across departments, contrary to perceptions that might exist. There may be individuals who are working significantly more or less than the averages presented here, but not whole departments or divisions.