|
The Ethics of Evaluating Instructional Computing
Marvin J. Croy
- Introduction
- 1. Moor’s Analysis
- 2. Overall’s Analysis
- 3. Considerations of these Analyses
- 3.1 Equal Educational Opportunity
- 3.2 Optimum Instruction
- 3.3 Informed Consent and Student Obligation
- 4. Minimizing the Risks to Students
- 5. Summation
- End Notes
- References
1. Leslie Burkholder, “Computer Ethics on Campus,” in Terrell Ward Bynum, Walter Maner and John L. Fodor, eds., Computer Ethics Issues in Academic Computing, Research Center on Computing & Society, 1992. (This volume, pages 1 – 12 above.)
2. This is case number 5 in the Ethical Case Studies in Teaching Philosophy series edited by Philip Pecorino. It is published in full in Teaching Philosophy, Volume 9, 1986, p. 351.
3. James H. Moor, “Computer-Assisted Instruction and the Guinea Pig Dilemma,” Teaching Philosophy, Volume 9, 1986, p. 354.
4. Ibid., p. 353.
5. Ibid.
6. Ibid., p. 354.
7. Ibid.
8. Christine Overall, “Innovation and Injustice,” Teaching Philosophy, Volume 9, 1986, p. 356.
9. Ibid., p. 357.
10. Ibid., pp. 357 – 358.
11. Ibid., p. 355.
12. Moor, p. 354.
13. Overall touches on this issue when she asks whether students divided into experimental and control groups are being provided with the “same educational experience.”
14. Overall, p. 357.
15. There is a potential problem with comparing student performance over several semesters before and after the introduction of some innovation, however. Many additional changes may occur over extended periods and may thus confound any attempt to attribute observed improvements to the instructional innovation.
16. One alternative to the standard experimental control group study is provided by a “within subjects” design. Students in this arrangement serve as both experimental and control subjects. That is, each student is exposed to both the experimental and control treatment condition. In respect to evaluating a CAI program, for instance, half of the students might use the program for several weeks while the other half make use of standard methods. Then, the students using the CAI program would make use of standard techniques for the same period of time and vice-versa. Performance measures would be taken at several points during this process for comparative purposes, and several such alternations might be even scheduled. (One practical drawback is that scheduling more than one swapping of treatments may be disruptive from both a curricular and pedagogical perspective.) The advantage of this approach is that, except for order of conditions, students are given equal treatment. The exception just noted is indicative of some ethical difficulty. Subject matter and its level of difficulty vary throughout the course. The CAI program might be most effective when the subject matter is most difficult, and students using CAI during that period would have an advantage. Consequently, Moor’s resolution should be employed as necessary.
17. K. Strike and J. Soltis, The Ethics of Teaching, Teachers College Press, 1985, p. 63.
18. Glaskow, Sadowski, and Davis (1977) found that even in experiments involving deception and stress, student ratings of their unacceptability decreased when information was provided about the worth of these studies. In one of my own studies, after explanation of its purpose, students randomly assigned to experimental and control groups were given the opportunity of changing groups, but none chose to do so (Croy, 1991).
19. Two points are in order here. First, Moor himself does not claim that his resolution makes in-class experimentation a risk-free venture for students. Second, I believe that, despite the complexities just noted, Moor’s resolution makes a significant contribution to the fair treatment of students in this context. It is not, however, without an additional complication. This complication is related to the concept of aptitude-treatment interaction. ATI research aims at discovering whether different students possess different characteristics that may predispose them toward or against various instructional techniques. That is, a given student may possess an aptitude that would make computer-assisted instruction more effective for that individual than some other means of instruction, or vice versa. Students with different aptitudes may well populate both experimental and control groups of an evaluative study and consequently some students may be disadvantaged more than others by assignment to a group which underperforms its counterpart. In fact, some students may not be disadvantaged at all. So, not every student will deserve to be compensated and those who do may deserve to be compensated by degree rather than by a fixed amount. Nevertheless, without prior knowledge of this interaction, little or nothing can be done to correct for it. A study which aims at evaluating an instructional technique will probably precede analyses of aptitude-treatment interactions with that technique. Consequently, Moor’s resolution should be adopted as formulated unless evidence dictates otherwise.
20. See the final common rule expressed in the U.S. Government’s Federal Register, Vol. 56, No. 117, June 18, 1991.
21. Part of the cost, of course, will come in terms of subject self-selection. Self-selection, however, is already present to a larger degree and should be increased little by providing advanced notice concerning in-class research.
22. I would like to thank Jim Moor for his critical comments on an earlier version of this paper.
Home > Teaching Resources > Computer Ethics Issues in Academic Computing > The Ethics of Evaluating Instructional Computing
HOME | IN
THE NEWS | RESEARCH RESOURCES
TEACHING
RESOURCES | STUDENT
RESOURCES
LINKS | THE
GALLERY | STAFF
The Research Center on Computing & Society
at Southern Connecticut
State University
501 Crescent Street • New Haven, CT 06515
Director: (203) 392-6790 •
e-mail: webmaster@computerethics.org
© 2000 – 2007 – Research Center on Computing & Society