Predictors of Performance in the Virtual Classroom
Identifying and Helping At-Risk Cyber-Students
The ability of instructors to identify at-risk cyber-students quickly is critical because the usual cues associated with student anxiety, inattentiveness or apathy are not present in the virtual classroom. For instance, cues such as frowning, fidgeting and day-dreaming, which are often readily apparent in the conventional classroom, are not observable by Web-based instructors. Due to the lack of these traditional cues, cyber-instructors must develop other strategies for identifying at-risk students in the virtual classroom. In addition, conventional solutions, such as office hours and graduate teaching assistants, for assisting low-performing students are not typically available in Web-based classes. Therefore, cyber-instructors must be creative in devising strategies for helping their at-risk students. We describe several strategies whereby cyber-instructors can take advantage of the technologically rich learning environment of the Internet in helping their students. Effective use of these strategies can also help reduce attrition rates in Web-based courses.
We have taught three different Web-based psychology classes. In the past five years we have taught more than 30 online sections and, in any given semester, we typically teach concurrent conventional and online sections of a class. Students freely choose to register for the class format they desire. We use the same course syllabus, textbook, homework assignments and examinations to facilitate comparisons across these two learning formats. The equivalence of course materials across class formats is purposeful. Our approach to online instruction has always been one in which pedagogy rather than technology guides the design of our Web-based classes. Therefore, we have attempted to find ways to use Internet technology to re-create important pedagogical aspects of our conventional classes in the Web-based format.
Many of our course materials are available on the course home page. This means that prospective students can obtain information about the course content, such as the syllabus, grading policy and calendar, as well as our approach to online instruction before deciding to register for the course. Further, the home page has links to Web sites that describe characteristics of successful cyber-students. These links allow prospective students to evaluate their own learner characteristics and technological proficiencies regarding Web-based courses. For instance, the distance learning link found at www.petersons.com offers students a short online survey that evaluates their readiness for an online course. Our home page also encourages prospective students to contact us before registration if they have any concerns or questions about the content and technological demands inherent to the online course.
In our online class, delivery of course materials and instructor-student interactions occur via both asynchronous and synchronous modes of communication. Asynchro-nous modes include e-mail, fax, forum (discussion) postings and downloadable information from the Web site. Synchronous communication occurs primarily in our 90-minute online lectures using a chat room, which are scheduled on a weekly basis. In this article we will describe several predictors of cyber-student performance. If instructors are vigilant, these predictors can serve as early warning indicators for student failure as well as success in the virtual classroom.
Demographic and Educational Predictors
It may come as a surprise, but basic demographic characteristics such as gender and age are not reliable predictors of cyber-student performance (Wang and Newlin 2000). While there may be a perception that male teen-agers might have a technological advantage, research d'es not show systematic differences in performance as a function of gender and age for the college population. Indeed, we are not aware of any research demonstrating that there are reliable demographic predictors of performance among college students who choose to take online courses.
However, the educational backgrounds of cyber-students can serve as early warning indicators for failure or success in the virtual classroom. For instance, Osborn (2001) has shown that the number of previous distance learning courses taken by students reliably discriminates between students who drop out compared to those who remain in either Web-based or videoconferencing courses. This is attributed to students' prior experiences with distance learning courses, which increase their familiarity with the technological demands of the virtual classroom. Just as importantly, these students have developed confidence in their ability to take advantage of the learning opportunities available to them in these types of learning environments. Osborn also found that students who remain in distance learning courses take a greater number of college courses and have higher college GPAs compared to those students who drop out of these courses.
Recently, there have been investigations as to whether learner characteristics are correlated with performance in the virtual classroom. Some of the research has investigated characteristics that may be conceptualized as 'global' traits or dimensions, such as learning style, sensory preference, hemisphericity/brain dominance and locus of control (Ehrman 1990). Global traits refer to characteristics that are relatively enduring and stable across time and various learning environments (i.e., cross-situational). However, as far as predicting cyber-student success, research on these global characteristics has yielded very little in the way of definitive results.
The safest conclusion that can be reached is that only one global trait, locus of control, is moderately correlated with performance in the virtual classroom. Specifically, students with an internal locus of control (e.g., 'The success I have is largely a matter of my own doing') are more likely to succeed in an online class than students with an external locus of control (e.g., 'The success I have is largely a matter of chance'). This is because learners with an internal locus of control leave little to chance or fate, but attempt to manage their activities in a thoughtful manner.
In contrast to the research on global traits, which has searched for cross-situational consistency, investigations on 'situation-specific' (i.e., course-specific) learner characteristics have yielded much stronger predictors of cyber-student performance. Situation-specific characteristics refer to those behaviors and beliefs that are not generalized, but are associated with a particular activity or environment. One such concept is self-efficacy, which refers to an individual's belief that they can perform a specific behavior to attain a desired goal (Bandura 1997). It is situation specific because one can have high self-efficacy about accomplishing one task, but low self-efficacy about accomplishing another.
When applied to Web-based learning, research has shown that two types of self-efficacy reliably predict cyber-student performance (Wang and Newlin 2002). The first is self-efficacy for understanding course content and the second is self-efficacy for meeting the technological demands of an online course. One of the Web-based courses we teach is called Research Methods in Psychology, which includes a great deal of statistical content. At the beginning of each semester we ask students to assess their self-efficacy for statistics tasks. We find that students' self-efficacy is highly predictive of their final grade in the class. We also ask for students' self-efficacy for meeting the technological demands of the online class. We find that this measure was also predictive of students' final grades in class. Other researchers have reported similar findings. For example, students who have strong confidence in their computer skills and less computer anxiety were more likely to remain in a distance learning class than students with lower levels of computer confidence (Osborn 2001).
We have also collected a measure of students' motivations for taking an online class (Wang and Newlin 2002). Our measure was situation- specific in that we asked students about their motivation for taking our particular online course, rather than for college courses in general. This measure was collected at the beginning of the online section of Research Methods in Psychology. We found that students who had taken a Web course before and preferred this type of learning environment were much more likely to succeed than those who chose the Web course because it was the only section open when they registered. In fact, students who preferred Web-based learning environments averaged one final letter grade higher than those who enrolled solely because the course was available.
In light of the above findings, we believe that cyber-instructors should attempt to gather several early warning indicators of student performance (see 'Profiling At-Risk Cyber-Students' at left). However, these indicators are of value only to the extent that they assess student self-efficacy and motivation regarding a specific online course. These indicators will not have much predictive value for cyber-student performance if they are of a global, cross-situational nature.
Online Course Activity
A resource that is typically available on Internet servers is the ability to count and record Web site activity. Cyber-instructors will find that this can be a valuable asset when monitoring the course-related activity of their students. For instance, WebCT courseware has a student-tracking function that allows cyber-instructors to monitor the frequency and time of each student's visit to various pages on the course Web site.
Tracking a student's online course activity is important because it can reveal several early warning indicators of student performance in the virtual classroom. For instance, research has shown that the total number of home page visits during the first week of a course is predictive of final grades in the 16th week of the course (Wang and Newlin 2000). This research also found the total number of forum postings read and written by cyber-students during the first week is predictive of their final grade. Because this type of tracking information is automated and readily accessible by instructors, we strongly recommend its use as an early warning indicator of cyber-student performance.
Cyber-instructors who rely on electronic chats as a means of synchronous communication have an additional tool for identifying at-risk students. Specifically, cyber-instructors can monitor the number and type of student comments posted to online chats. For instance, when we performed a discourse analysis of online chats during the third week of our class, we found that the total number of student statements and the frequency of student responses to our queries were both predictive of final grades in the class (Wang, Newlin and Tucker 2001). Thus, we recommend that cyber-instructors monitor the frequency and type of student remarks that are made in an online chat room.
Helping At-Risk StudentsWe take a proactive approach to assist at-risk cyber-students. This means that we first allow prospective students to select whether or not to enroll in our online course. To facilitatethis process, our course home page describes the characteristics of successful cyber-students and provides links whereby prospective students can assess their readiness for the virtual classroom. In this manner, students who enroll in our online class are fully informed as to the technological demands of an online class.
As the semester begins, we also suggest that cyber-instructors engage in some quick and easy information gathering about their students. For example, our first quiz is a short survey assessing students' self-efficacies and motivations for taking our online class. The quiz is posted on a course Web page with students responding by e-mail. At the end of the first week of the course, we evaluate these survey responses in conjunction with student-tracking data.
Specifically, we determine which students have low hit rates to the course home page and are relatively inactive in writing or reading forum postings. Then, we identify which of these students reported relatively low self-efficacy for the course and took our online section solely because of availability. Next, we make contact with these at-risk students via telephone and e-mail to discuss with them the behaviors linked to success in the virtual classroom. At this point, we also ascertain the reasons for their inactivity in class (e.g., technical or personal) and help them resolve those issues. We try to be as encouraging as possible, while reminding students of the technological and content demands of our online course. Students are typically appreciative of our personal contact, and many are immediately able to begin exhibiting the behaviors that will help ensure their success in the online course.
We have found that students who were members of a cyber-study group had higher final grades in our class than those who preferred to study alone (Wang and Newlin 2000). Therefore, during the semester we facilitate the formation of cyber-learning communities. First, we require that all lab reports be submitted as group projects. Second, many of our Web pages encourage students to 'be good citizens of cyberspace and help each other out.' Third, we instruct our students to use online chat rooms and forum postings to 'meet' other people and form study/lab groups.
There are two reasons why we believe that these cyber-study groups are beneficial for at-risk students. First, the peer-to-peer interactions needed for collaboration promote a collective sense of responsibility that is not ordinarily found in the virtual classroom. Second, cyber-students who have low self-efficacy or an external locus of control receive feedback and encouragement from their study partners. Consequently, this form of peer-to-peer interaction is an additional incentive for these students to perform well in the virtual classroom.
Finally, we believe that a heightened social presence by the instructor is beneficial for the at-risk cyber-student. Social presence refers to the degree of salience that another individual will enter into a meaningful dialogue (Short, Williams and Christie 1976). Accordingly, social presence is enhanced when another individual is perceived as real and genuine. This is critical for the virtual classroom because learner satisfaction is higher when computer-mediated communication is associated with high levels of social presence (Gunarwardena and Zittle 1997).
So how can cyber-instructors increase their social presence? By ensuring that there is immediacy and intimacy in the way they communicate with their students. This is best accomplished by the synchronous communication that occurs in regularly scheduled online chats and virtual office hours, and not simply by e-mail correspondence. At-risk cyber-students who have an external locus of control and low self-efficacy for the class will benefit the most from this sort of interaction.
Profiling At-Risk Cyber-Students
There are several indicators that, taken collectively, are reliable predictors of poor performance in the virtual classroom. Accordingly, we do not suggest relying on a single indicator as a means of identifying the at-risk cyber-student. Instead, we recommend that cyber-instructors compile several indicators to form a profile of the student who is potentially at-risk. In our view, any student who matches four or more of the characteristics on the following list has the potential for low performance in a virtual classroom:
- Does the student have an external locus of control?
- Does the student have low self-efficacy regarding their computer skills?
- Does the student have low self-efficacy regarding the course content?
- Does the student lack previous experience with online courses?
- Did the student enroll solely because of course availability?
- Does the student have a low login rate for the course home page?
- Is the student reading and writing few messages on the class forum?
- Is the student quiet or nonresponsive in the online chat room?
Bandura, A. 1997. Self-Efficacy: The Exercise of Control. New York: W. H. Freeman.
Ehrman, M. 1990. 'Psychological Factors and Distance Education.' American Journal of Distance Education (4) 10-24.
Gunarwardena, C. N. and F. J. Zittle. 1997. 'Social Presence as a Predictor of Satisfaction Within a Computer-Mediated Conferencing Environment.' The American Journal of Distance Education (11) 8-26.
Osborn, V. 2001. 'Identifying At-Risk Students in Videoconferencing and Web-Based Distance Education.' The American Journal of Distance Education (15) 41-54.
Short, J., E. Williams and B. Christie. 1976. The Social Psychology of Telecommunications. London: John Wiley & Sons.
Wang, A. Y. and M. H. Newlin. 2000. 'Characteristics of Students Who Enroll and Succeed in Web-Based Psychology Classes.' Journal of Educational Psychology (92) 137-43.
Wang, A. Y. and M. H. Newlin. 2002 (in press). 'Predictors of Web-Student Performance: The Role of Self-Efficacy and Reasons for Taking an Online Class.' Computers in Human Behavior.
Wang, A. Y., M. H. Newlin and T. L. Tucker. 2001. 'A Discourse Analysis of Online Classroom Chats: Predictors of Cyber-Student Performance.' Teaching of Psychology (28) 221-25.
This article originally appeared in the 05/01/2002 issue of THE Journal.