How Postgraduate Students Engage With Online Course Material and Activities

Abstract

This paper describes the final phase of a project to investigate the use of a virtual learning environment (VLE) with distance learners supporting each other through study groups. The specific topic was the development of academic literacy skills by postgraduate computing students. An initial survey and pilot study were used to plan the final study in which a set of online activities was provided for a complete cohort of students. Because student participation in these activities was limited, a series of telephone interviews was carried out with a sample of this cohort to investigate their attitudes to participation in optional activities. The VLE logs were also used to examine patterns of student activity. Analysis of this material has identified a number of factors that influence the way in which students work with the VLE: how they choose which activities to take part in and how they engage with course materials.

Keywords:

• Collaborative learning
• academic literacies
• virtual learning environments
• requirements engineering
• learning strategies

1. Introduction

There is increasing concern about overall standards of academic literacy amongst students, including those studying computing (Hetherington et al, 2007). At the same time, online technologies provide a range of opportunities for helping students to develop different types of writing and learning (Lea, 2000). In particular they support feedback from fellow students as well as from tutors. The project described in this paper was designed to look at ways of supporting students through online study groups in a virtual learning environment (VLE). Its specific objective was to help students on a postgraduate course in computing to develop appropriate academic literacies. A successful pilot study with a group of volunteers indicated that they had found the activities provided helpful in preparing for the assignments and examination. Their expressed preference was for three or four activities over the lifetime of the course, using discussion forums. However, a final study based on these findings failed to engage students in meaningful interaction. A programme of telephone interviews was carried to investigate the reasons for this and to identify which of a range of materials provided they had found engaging and why.

1.1 Background

The Centre for Open Learning of Mathematics, Science, Computing and Technology (COLMSCT) at the Open University (OU) currently supports a variety of projects including areas such as online communities and assessment. This study focuses on groups of OU students taking a postgraduate course, Software Requirements for Business Systems, which is presented twice a year, from November to April and from May to October. The course is taught through a mix of printed and online materials. Although students are allocated to tutorial groups for support and the marking of assignments, there are no face-to-face tutorials. There is limited contact with other students through an online conference. In addition the course makes use of some features of the Open University’s new VLE, including a wiki and a reflective journal (Thomas and Minocha, 2007).

From an educational viewpoint, there is a need for tutorial work to develop the academic literacies appropriate to postgraduate study. The learning outcomes state that, on completion of the course, students should: “understand the state-of-the-art in requirements engineering, leading to a deep and systematic understanding of the subject area”. They are expected to achieve this “through the use of books, reference papers (e.g. conference and journal publications) and other audio/visual media” (Open University, 2006). These outcomes are assessed both in three tutor-marked assignments and in the examination. In each case students are expected to write a critical evaluation of one or more papers in the field of requirements engineering. General guidance on how to do this is provided in the assignment booklets. However, there are at present no tutorial sessions to assist students in developing the appropriate skills.

The students on this program are part time, physically separated from the University and from each other. Most of them are in full time employment in the general area of software development. Although this provides good practical experience, it can be difficult for them to acquire the appropriate techniques of academic discourse. Unlike full time students they are not immersed in an academic environment. How therefore do they learn what is expected of them? Many academic tasks make little sense to students (Gibbs and Simpson, 2004). They need to engage actively with these tasks in order to test their understanding of what is required and develop their ability to apply the appropriate criteria (Boud, 2000). This understanding may be achieved through taking part in dialogues (Hetherington et al, 2007), and even through the observation of the dialogues of others (McKendree et al, 1998). One objective of the project therefore was to look for ways to make the learning outcomes more explicit, by creating dialogues around a set of relevant tasks.

1.2 Academic Literacies

Student writing and literacy in academic contexts may be understood through three models that overlap to some extent (Lea and Street, 2006). The study skills model focuses on the surface features of language, viewing writing as a transferable skill. The academic socialization model describes the need for students to acquire a set of ground rules for a literacy that is typical of a particular discipline. In contrast to both of these, the academic literacies model is concerned with how meaning is created in relation to individual, social and institutional identities. The social processes involved in acquiring literacy and using it effectively are viewed as more complex than those of the other two models. Although all three models are important, the development of academic literacy is particularly challenging for postgraduate distance learning students for whom there is little opportunity for discussion with members of the course team or other students. This situation is exacerbated because assessment is firmly based on their written assignments, which often take the conventional essay form. In some cases this will be their only means of communication with their tutors.

In order for students to identify whether they have met the appropriate standards for the task in hand, they need to receive feedback from peers and practitioners. Indeed, the more complex the learning, the less likely it is that it can be accomplished in isolation from others (Boud, 2000). Nichol and Macfarlane-Dick (2006) propose a framework of seven principles for good feedback practice that supports self-regulation. These include helping to clarify what good performance is, for example, by providing suitable examples and increasing discussion and reflection about criteria. This is often difficult to achieve with distance learners, although online forums appear to provide a supportive environment for such discussions.

Gibbs and Simpson (2004) argue that the only way to gain facility with the discourse of a discipline is to undertake plenty of practice in using it, for example, through writing. It is important to develop the learning activities of reading around the topic and constructing arguments. Timing is also important: imperfect feedback from a fellow student that is provided almost immediately may have more impact than better feedback from a tutor several weeks later. Such peer dialogue enhances in students a sense of self-control over learning. It also provides information to teachers that can be used to help shape teaching. For example, a group of students can be asked to identify a question that they think is worth asking about a paper (Nichol and Macfarlane-Dick, 2006), a technique that could be adapted for use in a VLE.

Distinctions may be drawn between different types of academic content, between the different contexts in which writing takes place, and the different practices associated with these contexts (Lea, 2000). One context that is especially relevant here is the use of electronic bulletin boards, which provide a way in which dialogue may be reintroduced or expanded for students who might otherwise miss out on this aspect of higher education (McKendree et al, 1998). There are a number of ways in which tutors and course designers may assist students who are using online conferences. These range from understanding the academic content that is being explored, through understanding the nature of the contributions that students make to conferences, being clear about the kind of environment that the conference represents (e.g. seminar, lecture or tutorial), recognising the contrasts and differences between writing in the conference and writing for assessment, to recognising the institutional relationships of power and authority that exist between student and tutor (Lea, 2000).

1.3 Individual approaches to studying

The learning strategies adopted by students have been the subject of much research. Entwistle (2001) draws many of these theories together, incorporating the work of Marton on approaches to learning, that of Pask on learning styles and strategies, and that of Biggs on the dimensions of the study process. A key concept is the difference between deep and surface approaches to learning. In the former the intention of the student is to understand ideas for themselves, in the latter to cope with course requirements. Ramsden (1979) describes a third or strategic approach to learning, whose most notable characteristic is the ability to adapt positively to the context in which learning takes place. He also identifies three ways in which students may demonstrate a deep level approach: relating a task to personal experience, making active attempts to relate different parts of a task to each other, and imposing a structure on the whole task and thinking about its meaning.

It might be expected that differences between deep, surface and strategic approaches would affect the way in which students learn online. However, it is not clear to what extent these approaches reflect individual differences or preferred strategies, or to what extent they are influenced by the medium of instruction. Evidence that some students adapt the VLE to their own purposes is reported by Heaton-Shrestha et al (2007) who carried out interviews with 43 first year undergraduates who were using Blackboard and found that volume of use was lower for strategic learners than for deep learners. The former said that they consulted the VLE to answer specific queries, whilst the latter appeared to have integrated the use of the VLE into their study. The authors did not find evidence to show that the VLE encouraged either deep or surface learning, although it accommodated both. They concluded that learning style did not depend upon the medium of instruction.

Molesworth (2004) studied the online learning of 60 final year undergraduates who used First Class to support a variety of activities, including virtual seminars in which students were asked to respond to specific questions by posting to a conference. He found that they were more likely to read messages than to join in discussions. Indeed many students simply posted their responses and then never revisited the discussion. He concluded that the benefit of flexibility is actually the flexibility to neglect a mode of learning, particularly when there is a strong assessment orientation amongst students. This is consistent with the findings of Maltby and Mackie (2009) that, whilst a VLE provides a flexible environment for learning and access to a wide range of resources for well-motivated ‘model’ students, it can also encourage the less engaged ones to skim material or put off learning until a later date. The VLE was not used primarily as a means of communication in these studies. However, the students involved were based on campus and therefore had other opportunities for collaboration and discussion.

Other studies have shown that students vary greatly in the amount of time that they spend accessing material online. Nachmias and Segev (2003) analysed web logs for 177 online courses and observed wide differences between individual students. Nearly 40% of students did not view content items at all. Of those who did view content, many viewed only a few items, although some viewed many items. Morris et al (2005) found that students who completed a course successfully spent almost 40 hours viewing discussion and content pages during a semester, compared to only 15 hours viewing by non-successful completers. However, creating and replying to discussion posts did not prove to be significant for success.

Lack of active participation by some students should not necessarily be seen as negative. Amongst the unintended effects of remote interactions are the possible benefits of vicarious learning. In other words, lurking can be valuable for some students. They can see their peers and tutors modelling the process of interpretation and can compare their own understanding with that of others (Lea, 2000). Other studies confirm that students find it useful to read other people’s comments in a discussion forum, even if they themselves do not contribute (Hetherington et al, 2007). However, this presupposes that there is a sufficient volume of contributions of a higher enough quality to make reading worthwhile.

2. Project Structure

For each of its courses, the Open University provides an online conference using First Class. This is open to all students enrolled on a presentation and is moderated by one of the tutors. In addition the University has recently adopted a VLE based on Moodle, an open source course management system that supports a variety of online tools, each of which provides a different way of communicating online (Cole, 2005). These include Chat, which provides synchronous communication, similar to MSN, amongst students and between students and tutor; Forum, which provides asynchronous discussion, amongst students and between students and tutor; Wiki, which provides a collaborative website that all students and tutors can edit, retaining a history of all the changes; Blog, which provides a website generated by a student or tutor to which other students and/or the tutor can add comments; and Journal, which provides a space for individual students to write down thoughts and reflections about a topic that are visible only to the student and the tutor.

The VLE was used as the setting for the project, which was designed in three phases. In the first phase, which followed the presentation of the course that ended in April 2007, students from one tutorial group completed a questionnaire about the five different tools described above. They were also asked to identify up to three requirements (functional or non-functional) that might be important for the success of online activities.

The results of the survey in Phase 1 were used to plan the second phase of the project, a pilot study with a group of volunteers drawn from the body of students for the presentation that started in May 2007. These volunteers undertook to take part in one or more online activities during the presentation and to provide feedback on these activities at the end of the course. They were offered the opportunity to carry out a set of short formative activities centred on developing the academic analysis and writing skills required for postgraduate study. In addition, by gaining experience with some of the tools available in the VLE, they would help to identify how best these tools could be used to support future students of this course (and other OU courses). A separate VLE space was provided for these students. Access to this space was limited to the researcher and the volunteers, who were told that all data collected from contributions to the activities and from feedback would be held anonymously and used for research purposes only. The volunteers were provided with a set of five short activities designed to help with developing appropriate skills in writing at Masters level. Two of these activities used wikis and the other three discussion forums. The selected activities were designed to encourage the development of appropriate strategies for approaching tasks such as evaluating a paper and writing a critical summary.

The final phase of the project was a full study, open to all students on the May 2008 presentation. The structure was based on feedback from the students involved in the first two phases of the study, who indicated that they would be willing to take part in three optional activities over the lifetime of a course, using discussion forums. Each activity was based on one of the papers provided by the course team and identified as essential reading for the course. A link to each activity was provided from the appropriate week within the online study calendar and a reminder was provided following the link to the paper within the resources page. The activities were designed to encourage students to relate concepts introduced in the papers to their own experience. For example, one paper described several rules of thumb (heuristics) that can be used to identify functional requirements. Students were asked to select at least one rule of thumb that could be used to identify a functional requirement for one of their work projects that might otherwise have been missed and post example(s) to the forum.

The original intention was to follow the approach adopted in Phase 2 by gathering data from student participation in the online discussions, supporting this with log data. Interviews with students were to be carried out at the end of the course to evaluate the approach. However, participation in the activities was so limited that it was decided to place more emphasis on the interview programme and to broaden its scope. It was considered that interviews would complement the data from the initial survey and the VLE logs by providing a deeper insight into students’ attitudes and motivation than could be obtained from a further survey. The interview strategy was based on purposive sampling. There were 47 students on the course, of whom 38 were available for survey (31 male and 7 female). Log data was used to identify three groups of students who were of interest: those who made active contributions to the First Class conference and the optional discussion forums, those who generally looked at the conference and forums but rarely or never contributed, and those who neither contributed nor looked at the conference and forums. Two representatives of each of these groups were identified and invited to take part in the study. They were told that their responses would be stored anonymously and used to improve the way the VLE is used to support students. Two of these students did not reply and were replaced by other suitable candidates. The six respondents in the interview programme are coded C-1 to C-6. The final distribution is shown in Figure 1.

Figure 1 Original classification of students in the sample for interview

At the end of the presentation in October 2008 semi-structured interviews were carried out by telephone with these six students to obtain their views on the specific activities based on the discussion forums and, more widely, on the use of other optional course material. This included the use of the First Class conference, academic papers provided as additional reading, recordings of seminars on DVD, and online quizzes. The interviews were recorded, transcribed and analysed using grounded theory methodology. In addition the log data from the VLE was used to provide statistical data about the date/time of access and general type of activity undertaken by these students.

3. Findings from the final study

Analysis of the interview transcripts was carried out to provide an insight into the choices that students make in deciding which elements of a course to engage with and which to neglect. Although lack of time plays a part, student choices are also affected by the relevance of the material to their current work, how well the activity fits in with their preferred methods of working, and whether or not their contributions are assessed.

3.1 Limitations of time

The original focus of this phase of the project was to be on the three optional activities in the discussion forums. Of the 38 students on the survey list, 22 looked at the discussion forum at least once, but only three posted a message. Student C-2 was one of those who did so and provided a positive response:

“Good, yeah I just wish I had more time to take part, to participate and to improve the quality of my work on the course.”

Student C-3 also posted a message but did not find the forums very useful:

“I could find all the information I needed from the website, the First Class conferences.”

The third active student, C-6, posted to the second forum in response to prompting from her tutor:

“… no-one had put any comments so I added one comment no idea if anybody responded to it, I think I must be the only one… I posted one message, and I think I looked back a couple of times to see if anybody had responded, and nobody had. I found it quite a fast paced course, so moving onto the next thing.”

As Entwistle (2001) remarks, even when the intention of the student is to gain a deep understanding, time constraints often prevent them completing the process that leads to full understanding. A number of other factors may have contributed to this very limited response. Firstly, the activities were labelled as optional in order not to increase the workload of students who were already finding it difficult to keep up with the course. They were thus an obvious candidate for neglect:

“I’ve used the first one but I didn’t have time to really input any depth into them… think I looked at it twice. I looked at the first time and then we were prompted to do your journal and stuff like that, and I looked through the different forums that were going on.”

[Student C-4]

The structure of the VLE may also have played a part in limiting student activity. The default display of the online Study Calendar is to show three weeks at a time. Thus activities that are embedded in the Calendar, as these were, are visible for only a short time:

“I didn’t realise until quite a lot later about that discussion forum, so by the time I’d gone to look at it, it was past the point in the study calendar.”

[Student C-5]

Finally the comparatively low number of students on this presentation of the course meant that it was difficult to generate sufficient online activity, both for the optional discussion forums and in the First Class conference. Of the 38 students on the survey list, 20 posted nothing at all to this conference. Only 19 read the introductory message from the moderator. Although certain messages from the course team were also posted on the VLE, it is clear that some students never look at the First Class conference for the course that they are studying. The usual reason given is that it is too time consuming. However:

“There is another reason I suppose. Had it been something that I was really, really, really interested in, for example, I don’t know Homer or something like that, then I would have made the time to do it for personal interest, but because it was work related…“

[Student C-1: she did not post any messages and did not use First Class on her two previous courses either]

In contrast Student C-3 posted and read messages regularly. He also used First Class conferences on his previous courses and found them:

“Very useful for two reasons. First it makes you feel less lonely so to speak. With this facilitation you can think you’re always on your own or you don’t think there are other people, students going through the same process as you, and second thing, well many students have the same questions, the same doubts I had, so it was very useful to browse the conference and see what the answers to these doubts were”.

However, the low level of activity was a factor in discouraging students from making more use of this conference:

“To be honest I didn’t find it that useful because it wasn’t very active”

[Student C-5].

While lack of time makes it difficult for many students to participate, lack of participation then discourages those who do have time from contributing to conferences and forums. This is consistent with the findings of Molesworth (2004) who reports that some of the students in his study were frustrated by the lack of contributions from other students and disappointed that their own contributions had not been read.

3.2 Relevance to work

An important factor in considering student engagement with this course was their motivation for enrolling on it. The postgraduate computing programme is designed for students working in industry. The reasons that all six students in the telephone survey gave for choosing this course were mainly work-related. Their objectives might reflect their past experience:

“Because at that time I do a lot of Requirements work anyway, so I was keen to progress with that, and learn a bit more about what I was doing, or what I should have been doing better or not doing at all”

[Student C-5]

Their current practice:

“It was relevant for the project I’m working on at work… the production of a control system for a nuclear submarine.”

[Student C-4]

Or their future plans:

“I thought that [Requirements Analysis] was an area that I’d be interested in moving into, and I thought that it would be a good start.”

[Student C-1]

This emphasis on work coloured the responses of students to the material provided for them. For example, they were asked to say which, of all the papers provided by the course team, apart from the ones that were used for assessment, made the most impact on them. Their responses gave value to the relevance of the material to past, present and future issues at work. For example, Student C-1 chose:

“The one using creativity [to] elicit requirements, I suppose that one, and that I’d say it had the most impact on me… For two reasons, well the first was I do a lot of that kind of work, not obviously from requirements engineering, but I do a lot of facilitating workshops which I really enjoy. I thought this would be something really positive, especially as I was thinking that I would like to move into that field.”

On this evidence, linking online activities to work might seem a way to make them more relevant to these students. As described above, this was the approach adopted in this phase of the project, without much success. The discussions were based on three papers selected from the list identified by the course team as essential reading. Five of the six students in the survey remembered reading the first paper, which appeared early in the course. Three remembered reading the second paper, but only one remembered reading the third paper, regarded as essential for the final unit of the course. In some cases the respondents expressed doubts about the nature or even the existence of the papers, one even wondering whether the interviewer was putting in titles that weren’t on the website.

Most of the students started by trying to read all the papers provided. Their recall of material presented earlier in the course was much better than that presented towards the end, which may reflect pressure of work as the final assignment and examination approach. None of the papers selected for discussion was chosen by the respondents as having made an impact on them. The papers that they preferred were those that describe specific techniques for which they could see an application in their work environment, one of the factors that may demonstrate a deep approach to learning (Ramsden, 1979). However, these techniques tend to differ from one student to the next, making it difficult to select a paper that will engage a wide range of students.

3.3 Methods of working

Moodle maintains user activity logs that can be sorted by user, date/time and activity. These provided a way of monitoring the activities of the students over the presentation. The log entries could be divided into active contributions by students, which included adding, deleting and updating discussions and posts, or editing the wiki, as opposed to the passive viewing of discussions and resources. The number and type of online actions varied widely between students, a factor identified in the pilot study (Jeffcoate, 2008). Figure 2 shows the total number of actions logged for each of the six students in the interview programme. These are broken down into wiki actions, which were part of the assessment for the first two assignments, the quizzes and the discussion forums, which were optional, viewing resources and downloading files, and some miscellaneous housekeeping actions.

It is interesting to compare these results with the initial classification of these students given in Figure 1. Although both C-1 and C-4 were originally classed as non-contributors, their patterns of activity are very different. C-1 did not access either the First Class conference or the discussion forums and her contributions to the wiki activity, which was assessed, were very limited. In contrast, C-4 appears to have spent a lot of time on the wiki, although further analysis shows that the majority of these actions were passive — reading contributions made by others. C-1 appears to have devoted most of her online time to the quizzes, which were popular: all but one of these students tried them on a regular basis, and generally valued them highly. They were used in a variety of ways:

“I didn’t take the quizzes while I was studying. I kept them for part of the revision before the exam, they were very useful.”

[Student C-3]

“I tried to do them like before and after I’d done the unit, and I think I used them a bit in the revision as well.”

[Student C-5]

Figure 2 Student activities logged by type

In addition to the variation in the number of actions, the timing of those actions also differs. Some students prefer longer online sessions:

“Actually to be quite honest there were only a few times when you actually had to go to the computer to get through your… course booklet, and I think there should have been more, and the other thing is to have them in chunks, because its an effort isn’t it to go and switch your computer on to do an activity that will last possibly 10 minutes. It would be much better to have an hour’s worth of computer activities as opposed to just those little fleeting visits on the computer, so probably an hour’s worth at a time.”

[Student C-1: She said that she logged on once a fortnight, for between 15 and 45 minutes at a time]

Others are limited in the time they have available:

“I found [the seminars on DVD] quite long. I think they’re about an hour long, and that’s quite a chunk of time for me to find.”

[Student C-6: she did the quizzes at work in the lunch hour]

Clearly it is difficult to provide a structure for online collaboration that will accommodate such different patterns of working. Experience with a collaborative activity using the wiki indicates that this is a frequent point of contention between students.

3.4 Optional or assessed?

When the time available for study is limited, the behaviour of undergraduate students tends to be guided by assessment (Kirkwood, 2008). It might be expected that these postgraduate students would also give priority to activities that were to be assessed in some way, as the wiki activities were in this course. In the interviews they were asked how they would feel about the other activities, currently optional, if they were assessed. They were generally unconcerned about the prospect of discussions, either in a forum or in a First Class conference, being assessed:

“Well if it was to be assessed, I’d contribute to it… In [this course] it wasn’t assessed so I wasn’t bothered… So if it was assessed, I’d get on a bit quicker.”

[Student C-6]

Or even in favour of the idea:

“… it would have been better if we’d been forced to write a bit more so I would be supportive of that.”

[Student C-5]

However, their attitude towards using the quizzes for assessment was generally more negative. Three of the six were opposed to the concept on the grounds that this would reduce the benefits that they currently obtained from doing them:

“I found the quizzes actually quite a useful thing to do repeatedly, so I thought they were quite a useful tool like that, and I thought if we were ever getting assessed on our answers then they’d be a bit more nerve racking, and it would completely change the way you looked at them.”

[Student C-5]

It thus appears that, although assessing activities that students do not undertake willingly may be beneficial, assessing those in which they readily participate is likely to be self-defeating.

4. Conclusions

Harasim (2000) observed that two basic models of online courses had emerged: one based on collaborative learning and interaction, and the other based on publishing information such as course materials and lecture notes online. Whilst the latter model was adopted in the early days as existing lecture-oriented courses were converted to electronic form, she argued that adoption of the collaborative model in online education would provide a paradigm shift based on the principles of collaboration in learning, access over time and space, and constructivism in knowledge work. During the past decade, the widespread adoption of VLEs for campus-based education as well as for distance learning has certainly provided an opportunity to encourage collaboration between students and interaction between students and tutors. A set of basic requirements for online activities was established in the first two phases of the current study (Jeffcoate, 2008). They demonstrate a set of concerns for the student as an individual, for example, support for developing creativity, and as part of group, such as the ability to identify the contributions of others in that group. Another set of issues relate to the online system, in particular its availability and the importance of usability. Finally there is a set of requirements related to the task: these include the need for a clear description of the task, a clear structure and clear criteria for completion. This last group is consistent with the findings of Devlin (2007) that it is better if students understand how to go about the task they are engaged in, especially if this understanding is shared. Furthermore, students involved in all three phases of this study expressed a willingness to take part in several online activities over the course of a presentation. However, in practice very few students did so.

An analysis of the responses to the interview questions indicates that there are other important factors that may not be identified explicitly by students. These relate first of all to their personal interests, both at work and at leisure. Experience with this group of students suggested that engagement is increased if the materials selected for an online course (e.g. papers, videos) describe specific techniques they can apply in their working lives. The availability of sufficient time and the way in which it is used in individual patterns of study also play a significant role. Some students prefer to log on for short periods each day, some want to log on for longer periods every week. Willingness to take part in online communities is affected by the volume of contributions from other students, the perceived quality of those contributions and whether discussion can be sustained beyond the minimum need for assessment. Course teams seeking to encourage online collaboration should provide a structure that will accommodate different modes of working, redesign web pages to reduce pressure on students, and encourage them to take part in activities at a time that suits them. The use of assessment should be selective. Whilst assessment encourages students to take part in activities, assessing an activity that they do willingly on their own may constrain the way that they use it. Achieving a balance between these different factors is likely to prove a challenge for those designing and running online courses. Without it, there is the danger that VLEs merely provide an increasingly wide range of materials and activities that students can selectively neglect.

5.

Acknowledgements

The author would like to thank COLMSCT team for their support during this project and the anonymous reviewers who provided helpful comments during the preparation of this paper.