Tutored Video Instruction:
A Preliminary Qualitative Evaluation
of Computer Science at the
University of Washington
Brett E. Shelton
Lori Postner
Spring 2000
TABLE OF CONTENTS
General Description of Tutored-Video Instruction
Original Study (Gibbons, 1977)
University of Washington, 1999-present
U.W. Video Lecture
Technology and Presentation Environment
Users of U.W. TVI: Students
Users of TVI: Tutors
The Passive Instructor
An Attempt at Integration
The Work-around
Goals of the Redesign
A Pattern from Field Observations
Reasons for Differences
Gibbons TVI
Original Lecture
Tutor Expertise
Student Characteristics
Subject Matter
Technology & Presentation Environment
Results and Reasons for Success
Redesign of TVI
Tutor Expertise
Original lecture pedagogy
Subject Matter
Technology & Presentation Environment
Conclusions
References
Appendix A
General Description of Tutored-Video Instruction
The idea of video-based instruction is not new. Since the advent of recorded media, people have attempted to learn from pre-recorded material. Most recently, a large effort has been put forth by the educational community to make the most desirable lectures available to people across great distances, the idea being to take the same experience that exists in a classroom to those who cannot be physically present. When these lectures are shown to small groups of students, an attractive feature of this technology emerges: interaction between students and educators. These discussions allow students to share concerns, difficulties, and questions about the lecture material. Tutored-video instruction (TVI) has the potential of creating such an environment.
The TVI technique responds to the educational needs of the students by combining the positive features of lectures with those of small group discussions. The lectures provide for depth and continuity in the subject matter, while the tutorial discussions afford a means of making the lectures respond to individual needs and differences (Gibbons, 1977, p. 1140).
Original Study (Gibbons, 1977)
In the original study by Gibbons in 1977, a program was outlined in which TVI was to be implemented for off-campus students pursuing graduate studies at Stanford University. Gibbons put forth certain recommendations in order to facilitate the successful implementation of TVI.
Recommended TVI guidelines
1)the program should be specific
2)the program should have clearly defined goals
3)the program should utilize the most appropriate and desirable technologies for delivery
4)tutors should be motivated and trained
5)personal interaction among students is crucial
6)the educational effectiveness of the program should be monitored over time
University of Washington, 1999-present
The University of Washington (U.W.) implemented its own form of TVI in 1999 and it continues to be used. Senior faculty members within the Computer Science Department decided to try TVI primarily because the surrounding community colleges requested their students receive credit for their beginning computer science courses (CS 142 & 143) at the University of Washington. The University faculty wanted to ensure that the community college students would be properly prepared to transfer, enroll, and succeed in upper division courses. The apparent success of the students in Gibbons’ study encouraged the faculty at the University to make available the U.W. lectures and the accompanying slides of the lecture notes via the Internet. Observations of the TVI program at two of the surrounding community colleges have been made and are reflected throughout this paper.
U.W. Video Lecture
The lectures of an experienced University of Washington computer science faculty member are videotaped for distribution to the community colleges. However, it is unknown how well versed the faculty member is in lecturing to a taped audience or his/her opinion of TVI. The taped lectures are stored and can be re-shown in later courses.
Technology and Presentation Environment
The TVI classes are designed for a capacity of thirty students. The rooms used have similar facilities where all seats face the front of the room so that students can see the TVI content projected on a white screen or television. The video is comprised of two main parts: a small streaming video frame of the taped lecturer in the corner and a text-based slide that shows notes and sample code. The instructional content takes up the majority of the screen as shown in Figure 1. At times the content of the video is unclear due to poor recording quality. It may also appear tilted on the screen or out of focus. During the original taping the instructor can be out of camera range. The audio portion of the presentation was periodically unclear or delayed due to streaming video technical difficulties. The instructor commonly repeated in-class questions for clarity purposes, presumably to ensure that the viewers of the tape understood the query. While the instructors seemed comfortable with the taping procedure, the act of being video taped may have affected the proceedings of the lecture class.
Figure 1: TVI screen capture.
Users of U.W. TVI: Students
In order to understand the use of TVI in a community college setting it is necessary to see how the students in such institutions vary from those who attend four-year universities like the University of Washington. Students enrolled in community colleges are not as academically able on average as those enrolled at four-year institutions. Zigerell (1970) states "… the midpoint of the curve describing the abilities of the two-year college population, when laid over the curve for the four-year, or 'regular' college, will fall within the lower one-third of the latter curve" (p. 705). He reports that community colleges do not attract the brightest students wanting to pursue a college degree. In fact, many students who choose to attend a community college do so as their last choice. Dougherty (1992) cites the following characteristics of community college students.
On the average, community college students are poorer, more often nonwhite, less academically apt, less ambitious, and less likely to attend full-time than are four-year college students. (p. 190)
The audience that makes up the TVI student body has important implications on the effectiveness of the model. What kind of student responds well to TVI instruction? How large a factor does motivation have on the resulting instruction? Are certain students better suited for TVI, and if so, which students are they?
The students in the TVI computer science classrooms were observed to be relatively young, most between the ages of 19 and 25 with some non-traditional students over the age of 25. There was a larger number of males in the classes overall but the male/female ratio varied from class to class. In one case there was only 1 female in a class of 16 students while another class had 5 females and 5 males. English was not the primary language for many of the students, as is traditionally the case in Seattle area community colleges. The classes ranged in size from 8 to 20 students.
Users of TVI: Tutors
Although the students may be the most obvious users of TVI, the tutors bear much of the responsibility for its success or failure as they facilitate discussion around the technology. The tutors’ levels of experience vary. For the TVI classes, three tutors were observed; each having different levels of expertise and representing diverse teaching styles. One tutor was new to computer science instruction and was educated in mathematics (a common domain from which to draw instructors). The other two instructors had previous computer science teaching experience at their respective community colleges institutions. None of the tutors had any experience in teaching with TVI, nor were they provided with training in how to optimally implement TVI in their courses. The tutors observed fall into three categories as described below.
The Passive Instructor
One tutor preferred to let the taped lectures run for long stretches of time while offering suggestions and occasionally stopping to provide clarification. He initiated a technique in which he would open a text-box over the streaming video instruction and type sample computer code that further explained or illustrated a main point of the lectured instruction. The in-class time was largely split between long stretches of watching video and lengthy class instruction. This instructor rarely used the whiteboard and did not initiate many questions himself but rather re-iterated those questions posed by the University professor. During the non-video portions he would lead the class in a homework discussion or general topical issues. It did not appear that the tutor could predict the areas where students may have trouble, a possible explanation for why he did not initiate questions or interactions.
An Attempt at Integration
A different tutor spent most of the class trying to meld the streaming video into his normal “lecture” or traditional instruction style. He used the video as more of a “teaching aid” than as the primary instructional source. He often spoke over the video (and audio) restating the concepts he felt were most important, or discussing related side issues he felt needed clarification. At times he skipped-over portions of the video that he felt he had adequately explained.
The Work-around
Yet a third tutor chose to use the video instruction only as a guide for the course. She followed the structure provided by TVI but remained the primary presenter of information. In many cases she stopped the video before the U.W. professor had discussed a topic and would teach it herself. As she did this it appeared as though she was gauging the students reactions to and comprehension of the material. After she thought most students understood her explanation she would start the video and so that the students could hear the U.W. professor's explanation. She also fast-forwarded through portions of the tape that she felt she had sufficiently covered. Her main motivation for using TVI was to provide her students with the opportunity to get credit at the University. Thus she had some difficulty integrating the video into her own traditional lecture style. She felt as if she had to “work-around” the intrusion of the video lecture and would not opt to use it if given a choice.
The tutors displayed certain similar characteristics in their interactions using TVI in their classrooms. They all seemed to fully understand the material although some did not appear to be completely comfortable with presenting it. The instructors initiated nearly all of the stoppages of the video, often when the taped professor asked an in-class question but also to substantiate certain points. They all worked with a similar classroom set-up, similar number of students, and spent roughly the same total amount of time watching video and discussing topics. See Appendix A for details.
Goals of the Redesign
The TVI system utilizes two differing pedagogical philosophies. The first style is presented by the university professor video taped as he lectures to a presumably large audience. The second is the integration of the technology into a small classroom where students view the video and discuss the topics covered in it. In order to evaluate these pedagogical choices it is helpful to look at the literature on computer science education.
Tucker (1996b) points out that the "large lecture class is still the dominant model for teaching [computer science], amidst a preponderance of evidence that students do their best learning in individual and small group settings" (p. 2). He states that such traditional teaching methods employed by computer science educators are "strongly biased toward a male, individual, isolated work ethic. Lectures, texts and lab assignments favor the learning styles that are dominant within this narrow group" (p. 2). It is important to note that the classes observed did not fit this demographic as previously described.
"The management of large classes is a serious problem in many programs. Lectures alone, without student interaction in small groups and laboratories, are an inadequate learning mechanism for CS&E education" (Tucker, 1996a, p. 839). So why are lectures still the most common style for computer science classes? There may be a different reason altogether for using such approaches in four-year institutions. Often introductory computer programming courses are used to 'weed out' students who will not succeed as computer science majors. Therefore the method used to teach these courses has not concentrated on how to teach all students, but rather on how to limit the number of students who continue in the discipline. In these large lecture classes the professor introduces the material in a logical format, writes notes on the board often consisting of code examples and diagrams, and assigns programming tasks for the students to complete that allows them to practice the concepts transmitted to them via the lecture. Students enrolled in these classes frantically take notes and then struggle to code solutions to the programming assignments. These activities are solitary, students do not work together (or at least are not supposed to) and succeed or fail based upon their ability to make sense out of most of the material on their own.
In smaller classes the mode of instruction is typically similar to that described above in that the professor provides the content in a lecture style setting, but often the professor assigns in class coding and code tracing assignments. These two activities have different goals. Writing code segments helps students learn problem-solving skills by determining the necessary steps the computer performs into order to achieve the desired outcome. Tracing code helps students understand the internal workings of the computer as the code is executed and stores values in the various memory locations.
Some computer science teachers have realized that the traditional lecture is not the best strategy for teaching students and have tried a variety of other approaches. Kolesar et al. (1995) developed a course to be taught as an introduction to the computer programming sequence that uses spreadsheet applications to teach the basic concepts of computer science without teaching the complexities of programming language syntax. It is primarily a hands-on or laboratory course where students use the spreadsheet software to learn various programming concepts. Linn & Clancy (1992) used a case study approach to teaching programming where templates were used to help the student understand a specific problem, the approach taken by an expert to solve the problem and the expert's solution. They also provided test questions to assess the students' understanding. Numerous other approaches have been designed and run the gamut from whole language pedagogy to object-oriented programming and intelligence tutoring systems/computer-based environments (Cupper, 1996; Rosson, 1996; Soloway, 1993; Tureman, 1994).
A Pattern from Field Observations
One prominent observation is that the tutors invariably initiated stops in the video instruction. Once the video stopped, the tutors asked questions either posed by the taped instructor or of their own accord. Any subsequent conversation or discussion between student-and-student or student-and-instructor then remained on-topic or related to a corresponding assignment. There appeared to be hesitancy by the students to stop the “flow” of the videotape and to take notes while viewing the tape. Was this because the lecturer's explanation of the material was straightforward? Or was it difficult for the students to write and watch the lecture simultaneously? Were students who did not have a good command of the English language self-conscious about speaking up? Perhaps in larger classes timid individuals do not want to admit to having questions on difficult material, and will therefore be hesitant to stop the tape.
When interviewed, all tutors mentioned that they were skeptical of the success of the TVI format. The tutors were not convinced such a format worked for the computer science material in their classroom. One tutor thought training in the TVI model would be beneficial and was optimistic about its future. Another tutor felt that many factors needed to be changed in order for TVI to be successful.
Reasons for Differences
Many differences can be identified between the original study and the model brought forth by the University of Washington. By outlining the recommendations made by Gibbons, one can identify numerous differences. Some examples include the characteristics of the student body, the motivation of both the tutors and the students, the subject content of the course, the delivery method, the tutors’ expertise, and the grading procedure.