Gender and Other Factors in Student Online Activity

Stephen Bostock, Department of Computer Science, Keele University, U.K. and Wu Lizhi, Tianjin Foreign Studies University, Peoples' Republic of China

Summary

A large online course included 18 asynchronous discussion groups as part of assessed student activity. The number and content of student messages were analyzed with 19 variables describing student characteristics and attitudes. Older students wrote fewer messages. Factors positively related to writing messages were positive attitudes to online discussions and to information technology generally, and a high course grade. These factors meant that females wrote more messages. There were gender differences in half the characteristics and attitudes measured. Discussion groups had were created with different gender mixes. In mixed groups females wrote fewer messages than in female groups but males wrote more messages than in male groups.

Introduction

A course about the Internet has been taught annually since 1995/96 to 300-400 first year undergraduates at Keele University, U.K.. The course was designed on constructivist educational principles, and encourages independent student learning with discussion (Bostock 1997). It is taught largely online, using the Web as a flexible medium for providing course resources and activities. Part of the design was to provide student discussion groups, using asynchronous text messages in online virtual rooms, taking place over several weeks but with little tutor input being possible due to limited staff time. In 1998/99 there were 18 such online discussion groups of 20 students. To understand what makes some discussions more successful, the transcripts of these online text discussions have been analyzed, along with data about individual students from questionnaires including student age, gender, experience and attitudes.

Course design principles

Many authors have discussed 'constructivism' as a theory of learning and teaching. Grabinger and Dunlap (1993) reviewed the literature and summarized it as five principles, summarized below. Lebow (1993), Jonassen et al. (1993) and Simons (1993) came to similar conclusions.

1. Provide authentic assessment, with realistic tasks that need a deep understanding of the subject. Assessment is the major motivation for students. The learning outcomes to be assessed must be clear to everyone at the start, and the assessment must require students to demonstrate them and not, for example, merely require a description of how they could be demonstrated, or what the tutor said on the subject.

2. Encourage student responsibility for learning.  If students are to become autonomous, lifelong learners they need to practice having responsibility for their own learning, including having some control of content, pace and tasks, and space to use initiative.

3. Learning activities should be 'generative' in that they require students to use skills and knowledge in making a knowledge product. This might be text or software or a physical object, or the solution to a real problem.

4. Provide an authentic context for learning. Learning should be anchored to real world events, problems and issues, rather than in a thin ‘academic’ abstraction of the real world.

5. Encourage co-operative support. While the previous principles could all be embodied in an individual learning alone, it is well known that collaboration and debate deepen understanding. Debate clarifies and elaborates concepts and identifies misconceptions, by requiring learners to defend their own views and accommodate those of others. Collaborative working pools good ideas to synthesize better ones. The main feature of the course supporting co-operation was the online discussion groups.

The course

Keele University has an interdisciplinary degree structure where humanities and social science students have to do one natural science course. Most of them choose and IT course of two semesters and, in the second semester, most of these do this module about the Internet: a one-semester (12 week) course from February to May, with four weeks' break at Easter. They must pass it to get their degree. The 300-400 students are mostly 18 or 19 years old, with a few older students.

The aim of the Internet course is to give an understanding of the concepts, procedures and issues involved in using the Internet and online environments effectively for study and work. The learning outcomes are that students will be able to use the Internet to find, access and evaluate resources relevant to any particular subject; participate appropriately in online discussion environments; and discuss the impact of electronic networking on aspects of society such as business, education, gender, and the nature of work. It’s a mixture of technical skills, evaluation skills and an engagement with the wider impacts of the technology.

The scheduled elements of the module are, firstly, a two-hour practical session per student per week in a PC laboratory. The demonstrators (postgraduate students) present help students with technical skills individually. Alternatively, some students can do work from home or from the library. Secondly, a one-hour period per week in a lecture theatre was mostly used to show videos of real Internet applications and issues. Contact between individual students and the tutor (SJB) was before and after lectures, and occasionally in practicals, but mostly by email.

All course information, tasks and resources are on course web site on Keele’s intranet, except the video tapes, which were available in the library. No paper materials were provided and none were required of students for assessment. An online tutorial introduced the course content in place of three hours of lectures. Figure 1 shows the 1998/99 course home page, with links to a collection of notes on Internet topics, instructions for practical tasks, summaries of and questions about the lectures and videos, the assessment requirements, other course information, and a page of links to relevant resources on the World Wide Web.

Course assessment had two elements. The course grade is given for a final report describing the searching for, and evaluating, web resources on a specific topic agreed with the tutor by email. Students placed their report in their own web spaces and completed a web form to submit its URL. Secondly, evidence of completion of ten practical tasks had to be submitted by email or web forms. The tasks included completing the online tutorial, creating a personal home web page, comparing search engines, and using real-time 'chat'. One task (the subject of this paper) was to demonstrate appropriate contributions to an online discussion group, by demonstrating they had contributed a number of messages, and showing one message with significant content. The online discussion software stores all messages with date stamps, so students cannot easily cheat.

Figure 1 The 1998/99 course home page.

Course Evaluation

Web forms were used to collect information and evaluations from students throughout the course. Many aspects of the course were successful but the views of students were varied. Table 1 summarizes the answers to open questions volunteered by more than 10 students.

Table 1. Student evaluations

Many students said they liked learning to use the global Internet, creating a home page, real time chat across the network, and online discussions. In general the assessments worked well (the dislike of the online tutorial was due to the task involved). Most students liked the freedom to choose the subject of their report, and liked some of the practical tasks, especially creating home web pages. 

The use of discussion spaces was liked by some, disliked by others. Those disliking the discussion spaces were mostly because of technical problems they experienced in registering with the discussion system.  Collaboration was the hardest design principle to implement successfully; making online discussions  work well has been a problem area in the course (Bostock 1998, Clark 1996). Different group sizes have been tried; in 1998/99 groups of 20 were used. But there was still a great deal of variation in the amount of online activity between individuals and between groups. This paper attempts to explain that variation.

Online discussions

In 1998/99 private discussion spaces were created for groups of 20 students and the tutor within the BSCW web board (Basic Support for Cooperative Work, http://bscw.gmd.de/ see Figures 2 and 3). The response time was generally acceptable. The main practical difficulty was registering large numbers of students with the system, which could be slow and produce errors. Once registered, students experienced few technical difficulties.

The discussion spaces were used from week 5 to week 12 of the course, to discuss the issue of the Internet in education. Students were told how to find their group on the BSCW web site by email. There was a Welcome message from the tutor waiting for them and they responded to it by introducing themselves to the group. The only other tutor message was after several weeks, asking for the URL (web address) of a useful  resource.  Each week, each student was to ask a question, answer a question, or comment on an answer. A final practical task was to submit by email a summary of their contributions to discussion (message headers only) plus one good message by themselves.

Figure 2 A BSCW discussion as seen by the tutor, showing some message headers.

Figure 3 A BSCW message

Gender and IT  

Gender is an issue for IT (Grundy 1996). Fewer girls and women study or have jobs in engineering, in schools and homes boys can dominate computer use, often for playing games; and females are typically less confident about using technology and have less experience with it. (Rosen, Sears and Weil 1987, Bostock, Seifert and McArdle 1987, Perry and Grebber 1990, Brosnan and Davison 1994, Hatton 1995, Ford and Miller 1996, Blum 1999). However, females are generally more ‘chatty’ and collaborative and the Internet supports such human communication as well as person-computer interactions. So are females better at online discussion?  (Ory, Bullock and Burnasks 1997, King 2000)

To investigate both the simple effects of gender and the effects of interactions between genders, the gender mix of online groups was manipulated but students were not told. Of the 18 discussion groups, groups 1 to 7 were all female, groups 8 to 13 were 50% female, 50% male, and groups 14 to 18 were all male. The tutor determined the group membership. Members of a group were spread across the practical sessions so they did not work usually together face to face.

Analysis of discussion transcripts

The transcripts of the discussions were analyzed. There were 1386 messages in total but after disregarding the ‘Welcome’ and ‘URL' messages prompted directly by the tutor, and duplicates messages sent in error, 1015 messages were classified and used in the analysis.

Recent discussions of content analysis often refer to the classification of online messages proposed by Henri (1992), which described five aspects of messages: participative (the quantity of messages);  social (introductions, personal support); interactive (referring to other messages); cognitive (exhibiting knowledge and skills); and metacognitive (showing awareness and self control).

What was of interest here was the contribution to learning in the group so the messages were classified only for cognitive and metacognitive attributes, in a single scale, 1 to 5 (Table 1). Messages with more cognitive content were also typically longer.  

Table 1  Classification of messages
Type	Typical length (lines)	Characteristics
1	1-2	Either with no academic content or simple, short open-ended questions, as prompted by the task instructions.
2	2-3	Simple statements of fact or opinion, or a question to start a discussions.
3	3-6	Some academic content relevant to the topic being discussed. Acknowledged personal opinion, or argument based on fact.
4	5-9	Clarification and elaboration, supported argument and inference. Speculation.
5	9-16	Substantial argument based on quotations or claimed facts. Development of new questions from facts. Clarification of issues. Attempts to summarize and make balanced judgements.

The authors of messages were identified as Keele usernames so they could be matched with the data from questionnaires, which always asked for the username.  

Student characteristics

Students completed 3 questionnaires, on weeks 2, 4 and 12, as web forms that sent data to files for later analysis (Figure 4). Of the data collected, 19 variables were used to describe students. Some variables totaled answers from several questions, for example a measure of confidence with computers totaled answers to 7 questions.

Student characteristics and experience was described by, for example, gender, age, experience with IT, online experience, report % mark, and whether they knew the others in their online group. Student attitudes were described, for example, by attitude to using IT and to submitting work electronically, to BSCW discussions, to confidence about the role of computers, and to approaches to studying the course (deep/surface approaches to learning, Richardson 1990),.

Figure 4 A web form collecting student data

Results

The number of messages per student is positively skewed (figure 5). This is typical of online discussions - many students send a few messages, and fewer send more messages. Unfortunately this means that the data is not normally distributed so it had to transformed for statistical tests or analyzed with methods that accommodated this type of data.

Figure 5  Messages per student

When we look at the quality of messages, the pattern is similar for all gender groups (figure 6). The pattern for 'good' messages (types 3 to 5) was the same as for total messages (1 to 5). So later analysis used only the total number of messages.

Figure 6 Message quality  

Gender differences

Of 19 student variables, 10 variables were statistically significantly different between males and females (Table 2). Sample sizes were 321 or less depending on the questionnaire responses. 

More females than males thought they had less experience of IT applications, were less confident of using them, were less positive about using computers, thought they would use them less in their courses, thought that they would find the course difficult, and preferred paper to online information. But females had higher average report marks, more preferred online discussion to face-to-face discussion, and they rated BSCW as a discussion environment more highly. They were equally confident as males were in the use of BSCW software, even though they had lower confidence with other IT applications.

Table 2 Gender differences

Figure 7  Mean number of messages per student in different types of groups

All-female groups have statistically more messages per student than male groups (p<0.05). Mixed groups were intermediate. In mixed groups the messages per female (FX) drops while the messages per male (MX) increases. The online presence of females encourage messages from males; the presence of males deters messages from females. They probably guess the gender of the message writers from the screen names the students chose within BSCW, but the students do not meet - the effect is through online interactions.

Multiple regression

To attempt to explain why some students write more messages than others, multiple regression of the 19 variables on message number per student was performed. As the variables are of different data types (continuous, ordinal, category) ‘regression with optimal scaling’ (CATREG in SPSS 10). By successively eliminating least significant variables, four variables (numbering as table 2) provided the most significant multiple regression (p< 0.001) on total number of messages, in order of importance:

(8) a preference for discussions online rather than face-to-face discussions

(19) the  percentage grade for the final course report 

(4) age  (a negative factor)

(10) a positive attitude to using IT

The multiple regression equation is significant statistically (figure 8), but it only explains 22% of the variation in message number per student (Pratt’s importance just shows how important each variable is in the 22%). So it does not tell us very much if we wanted to increase the discussion activity number in future; there are many other factors having a small influence. 

Variables 8, 19 and 10 are positively related to message number.  Variable 4, age, is negatively related – older students write fewer messages. Variables 8 and 19 are most important (Table 3) and have higher scores for women (Table 2). This explains why women, on average, write more messages. Age (variable 4) is negatively related to message number. There is a non-significant difference between the median ages of the genders in this group which may add to the gender difference in message number but the effect, if any, would be small. Variable 10 is the least important and is working against the gender difference in message number: women have slightly less positive attitudes to computers and so this would tend to reduce the message number of women compared to men, but it is a smaller effect than factors 8 and 19. 

Table 3 Multiple regression results - The best four variables

When we plot message number against age, we see that students older than 21 write few messages in discussions (Figure 9). This may be due to their lack of experience and confidence with IT - when they were at school they did not use computers.

Figure 9 Age effect on message number - the dots on the graph for 18-21 years cover many individual students

The other three significant variables are not surprising.  If a student said they liked discussions online, they did more of it. Students with high report marks contributed more to discussions. If they had less anxiety about the use of IT they wrote more (anxiety was measured as the total of scores for 7 questions about the use of IT).

Gender in itself is not a significant factor in explaining number of messages. The gender differences described above are there because males and females are different in various characteristics, including factors a, b and c in this analysis.

Conclusions

The number of messages written in online discussions was related to many factors, primarily to liking online discussions more than face-to-face ones, gaining better marks overall, and being less anxious about IT. Such students tended to be females despite females having generally more negative attitudes to IT generally, they had more positive attitudes to using BSCW and in fact contributed more to online discussions on average. Gender as such does not contribute to message number per individual; gender differences in writing messages are explicable in terms of other student attributes.

This explains the average differences between single gender groups. However, there is also an effect of gender mix on discussion groups: females in the online presence of males wrote somewhat fewer messages than in females-only groups, and males wrote more messages when there were females writing messages. Something about the number or content of messages from one gender affected the number of messages written by the other.

What are the practical consequences for course design? Some students liked working online more than others, but students gaining high marks in other work for the course did more work online (and offline). We should make a variety of learning situations available. To encourage online discussions by all we should use mixed gender groups, and give older students special help to compensate their lack of experience with IT. And we should try to alleviate anxiety about IT generally, as it reduces message writing.

The course web is at http://www.keele.ac.uk/depts/cs/Stephen_Bostock/Internet/subinprg.htm

References

Blum, K.D. 1999 Gender differences in asynchronous learning in higher education: learning styles, participation barriers and communication partners. J. Asynchronous Learning Networks 3 (1)

Bostock, S.J.1998, Constructivism in Mass Higher Education: a Case Study British Journal of Educational Technology 29 (3)  225-240

Bostock, S.J.1997, Designing Web-Based Instruction for Active Learning, chapter 26 in Web Based Instruction, ed. Badrul Khan, February1997, published by Educational Technology Publications, Englewood Cliffs, New Jersey ISBN 0-87778-297-0

Bostock, S.J.Seifert  R.V.& McArdle J. 1987. The effects of learning environment and gender on the attainment of computer literacy. Studies in the education of adults 19, 37-45.

Brosnan, M.J.  and Davison, M.J. 1994 Computerphobia - is it a particularly female phenomenon? The Psychologist February 74-78

Clark J. 1996 Collaboration tools in online learning environments, Asynchronous Learning Networks 4 (1) http://www.aln.org/alnweb/magazine/

Ford, N and Miller, D 1996 Gender differences in Internet perceptions and use. Paper presented to the Elvira Conference 1996, http://www.shef.ac.uk/~is/home_old/gender.htm

Grabinger R.S. and Dunlap, J.C. Rich environments for active learning ALT-J 3(2) 5-34

Grundy, F. 1996 Women and Computers, Intellect, Exeter.

Hatton, D. 1995 Women and the "L": a study of the relationship between communication apprehension, gender and bulleting boards, A paper presented at the Annual Meeting of the Speech Communication Association (81^st, San Antonio, TX, November 18-21 1995) ERIC Clearinghouse number CS509270

Herring, S.C. Gender differences in CMC: findings and implications The CPSR Newsletter 18 (1) http://www.cpsr.org/publications/newsletters/issues/2000/Winter2000/herring.htm

Jonassen, D., Mayes, T. and McLeese, R. 1993 A manifesto for a constructivist approach to uses of technology in higher education, in Duffy T.M. , Lowyck, J. and Jonassen D.H. (eds) Designing environments for constructive learning Berlin: Springer-Verlag. 231-247

King, L.J. 2000 Gender issues in online communities The CPSR Newsletter 18 (1) http://www.cpsr.org/publications/newsletters/issues/2000/Winter2000/king.htm

Lebow, D. 1993 Constructivist values for instructional design: five principles toward a new mindset. ETR&D  41 (3) 4-16.

Ory, J.C., Bullock, C. and Burnasks, K 1997 Gender similarity in the use of and attitudes about ALN in a university setting J. Asynchronous Learning Networks 1(1)

Perry, R. and Greber, L. 1990 Women and computers: an introduction Signs Autumn

Richardson, J T E 1990 Reliability and replicability of the approaches to study questionnaire, Studies in Higher Education 15 (2) 155-168

Rosen, L.D., Sears, D.C. and Weil, M.M. 1987 Computerphobia Behavioural research methods, instruments and computers 19 (2) 167-179

Simons, P.R. 1993 Constructive learning: the role of the learner, in Duffy T.M. , Lowyck, J. and Jonassen D.H. (eds) Designing environments for constructive learning Berlin: Springer-Verlag. 291-313

s.j.bostock@cs.keele.ac.uk  May 2001