Assessment in 2008: looking forward

Gales are howling, trains in chaos, so it must be January and time to look ahead to what 2008 has in store…

The final release of QTI v2.1 should be out this spring, and it’ll be interesting to see what uptake is like.  This will be the most stable and mature version of the specification to date, supported by a long public draft stage and a number of implementations.  Angel Learning are a significant commercial early adopter, and other vendors are bound to be looking at their experiences and whether Angel’s embracing of the specification has an impact on their own customer demand for QTI 2.1. 

Other significant implementors of 2.1 are the JISC Capital Programme projects which will be concluding around March.  AQuRate offers an item authoring tool, Minibix provides support for a range of item banking functions while ASDEL is an assessment delivery engine which supports both standalone use and integration with a VLE.    These projects should deliver quality resources to the community which will provide a firm foundation for use of the specification.  There was a sneak preview of these projects at our last SIG meeting.

Talking of SIG meetings, dates for the next two meetings can now be confirmed. 

On 19 February there will be a joint meeting with the CETIS Educational Content SIG in Cambridge.  This meeting will cover a range of shared concerns such as new content related specifications such as Common Cartridge and Tools Interoperability, and innovative approaches to educational material and assessment.  Information about this meeting and online registration will be available very soon.  This will be preceded by a workshop hosted by the Capital Programme projects discussed above.

The focus shifts from assessment as content to assessment as process with another joint meeting on 1 May in Glasgow.  This meeting will be a joint meeting with the CETIS Portfolio and Enterprise SIGs and will offer an opportunity to explore some of the shared issues in these domains.  Again, information on the event will be available on the mailing lists, on this blog and on the website in due course.

Another event of note is the annual International Computer Assisted Assessment Conference on 8 and 9 July at Loughborough.  The call for papers is already out, with submissions due by 29 February.  As always, this should be a lively and important event in the CAA calendar.  Alt-C 2008, Rethinking the Digital Divide, will be held in Leeds on 9 – 11 September; again, the closing date for submissions is 29 February.  There’s also a regularly updated list of non-CETIS assessment related events on the wiki.

And what about the trends for eassessment in 2008?  The results of Sheila’s poll, with a strong emphasis on Web 2.0 technologies and possibilities, do seem to reflect to some extent the comments on the last meeting’s evaluation forms which suggested increasing interest in innovative technologies, signficant concern with transforming and enhancing the assessment experience and direct engagement with teaching and learning rather than the more abstract issues of standards and specifications for their own sake.  It will be interesting to see how the more ‘traditional’ XML-based QTI v2.1 fares in the light of the increasing popularity of mashups and web services in 2008.

Assessment SIG meeting, 26 September 2007

Academics and developers met in Glasgow recently to participate in the most recent Assessment SIG meeting. The very full agenda covered a range of topics, both technical and pedagogic, and presentations led to some lively discussions.

Myles Danson of JISC opened the day by presenting JISC’s views and priorities for eassessment, as well as pointing to some future work they will be undertaking in the domain.

Yongwu Miao of the Open University of the Netherlands discussed work undertaken by the TENCompetence Project, with a particular focus on the relationship between IMS QTI and IMS Learning Design and the work they have done in this area. Dick Bacon of the University of Surrey and the HEA discussed the relationship between different varieties or ‘dialects’ of QTI, exploring some of the implementation and interpretation issues that hinder or break interoperability between systems nominally implementing the same version of the specification. CAL Consultant Graham Smith pleased the audience with news that a new Java version of his QTI demonstrator will be available shortly with updated support for QTI 2.0 items, which should help in the identification and resolution of implementation problems.

Martin Hawksey of the University of Strathclyde presented the work of the Re-Engineering Assessment Practices project. With a focus on real world assessment experiences, including an impressive collection of case studies exploring the impact of transformation within assessment practices, the REAP project was of particular interest to participants. Also of great interest, and perhaps unsuprisingly sparking the greatest amount of debate, was the exploration of ‘Assessment 2.0′ presented by Bobby Elliott of the Scottish Qualifications Authority. Bobby looked at ways in which Web 2.0 technologies can be used to enhance and modernise assessment in ways which can engage and appeal to increasingly digitally literate learners.

The day also featured several demonstrations of tools under development. Niall Barr of NB Software demonstrated his current work, an assessment tool which utilises the IMS QTI, Content Packaging and Common Cartridge specifications, while Steve Bennett of the University of Hertfordshire demonstrated MCQFM, a JISC-funded tool which provides a simple text-based format for converting and editing items between formats. Two more JISC projects closed the day. AQuRate, presented by Alicia Campos and David Livingstone of Kingstone University, is an elegant item authoring tool while ASDEL, presented by Jon Hare of the University of Southampton, is an assessment delivery tool which builds on the R2Q2 project to provide a fuller test tool. A third project, Minibix (University of Cambridge) on item banking, is working closely with AQuRate and ASDEL.

Links to presentations (via slideshare), project websites and other information can all be found on our wiki: http://wiki.cetis.org.uk/JISC_CETIS_Assessment_SIG_meeting%2C_26_September_2007.

Assessment for Learner Responsibility

On Monday, I attended a Learning Enhancement Network event here at Strathclyde on Assessment for Learner Responsibility.  Strathclyde is in the process of revising its assessment policy, and this event brought together staff from across the university, together with a number of student representatives and some colleagues from other universities. 

The University of Edinburgh has recently completed a similar process, and Nigel Seaton from Edinburgh’s College of Science and Engineering presented the outcomes of this process.  It was particularly interesting to see that all assessment in the College is now formative, in that students receive feedback for all the work they do – including formal examinations.  I really like this: as both a student and a tutor I always found it hugely frustrating to not be able to get or give feedback from the most important assessments beyond a bald grade or classification.  This is particularly important for students who experience signficantly worse performance in these assessments than in earlier coursework, and who are often bewildered, demoralised and demotivated by the lack of information provided.  The Data Protection and Freedom of Information Acts make the disclosure of examiners’ comments a legal obligation, but it’s nice to see this spun positively and used as a real learning opportunity rather than just warning markers not to make rude comments on exam scripts.

The College will also be introducing an eportfolio system, not for PDP but for use as a subject-specific learning and reflection aid for students, another very appealing idea. 

Jim Baxter from Strathclyde’s Department of Psychology gave a very entertaining presentation on collaborative WebCT-based activities introduced to the first year course in collaboration with the REAP project.  This involved a lot of collaborative work, something which had already been raised by Nigel and which was returned to in the afternoon’s breakout discussions.  I’ve never been a fan of group assessment, peer assessment, and other activities which force people into social models regardless of whether that is what is right for them, and I was pleased to see that I wasn’t the only one who felt some unease about compelling students to take part in such activities.  There was genuine agreement that there’s a need to respect all different learning styles and that there may be a tension between fashionable approaches and what’s actually best for an individual.  Staff reported that the majority of students themselves said that they dislike group work, although it’s regarded far more positively in post-graduation surveys – which is rather interesting itself, perhaps extroverts are more likely to complete surveys..?

One issue which concerned me is the observation that a student can be prevented from sitting the final examination for a course if they haven’t participated in group activities.  The justification for this is that these activities are detailed in the course materials so they knew what they were signing up for – but surely students should be studying a course because they’re interested in the subject and not because they can cope with the teaching style, and really shouldn’t be prevented from studying their chosen subject because they are don’t have a particular learning style.  The principle that students should have a choice in the methods and timing of assessment is therefore very welcome.

David Nicol presented the eleven principles of good assessment practive which were the initial outcomes from the university’s working group examining the assessment policy and from David’s long interest in this area.  These covered engagement to stimulate learning through clarificiation of what constitutes good practice, encouraging ‘time and effort’ on educationally purposeful tasks, high quality feedback, opportunties to close the feedback look, encourage positive motivational beliefs and self-esteem and encourage dialogue about learning between all stakeholders.  Other principles focused on empowerment, sharing responsibility for learning with students by facilitating self-assessment and reflection, giving learners choice in the nature, methods, criteria and timing of assessment, involving students in policy and practice decisions and supporting the development of learner communities and social integration. 

A particular strength of the day was the involvement of Strathclyde students and the opportunity the day gave for dialogue between staff and learners.  The greatest concerns which emerged from the afternoon breakout discussions were time, feedback and over-examining.

Deadline coordination is perhaps a particular issue at Strathclyde which has a very broad first year curriculum: in the Faculty of Law, Arts and Social Sciences, for example, students have to study five potentially quite disparate subjects, all of which tend to set the same deadlines for work.  There were occasional examples of the principle of ‘giving learners choice in the timing of assessment tasks’, for example the lecturer who negotiates deadlines with his classes.  Our group spent some time lusting after a hypothetical ‘assessment booking system’ which course coordinators could use to pick ‘slots’ for assessment deadlines – perhaps something the Enterprise SIG could look at :-)

Feedback was another major issue for the students.  They appreciate detailed constructive feedback, particularly where it explains marks in relation to published assessment criteria and offers suggestions for improving weak areas as well as highlighting strengths.  There were a few examples of bad practice, with ‘feedback’ which consisted of a smiley face and a mark being a particular low, but not many examples of truly detailed feedback.  The timeliness of feedback was also a concern: as one student observed, ‘how can I learn from my feedback if I don’t get it until after I’ve submitted my next assignment?’

Students and staff were both concerned that students are being over-examined.  Virtually every piece of work a student submits contributes towards the final mark for the course, meaning that students don’t have a space to fail.  The students in our group actually wanted more formative assessment, practice excercises particularly when undertaking a new type of task they hadn’t encountered before, and the opportunity to experiment and learn before being summatively assessed.

Other issues that emerged were the desire to offer staff incentives and reward innovative and engaging projects, training for staff in how to write content that actually will engage learners, and the real desire to share innovation throughout the university.

Massively Multi Learner

The HEA Information and Computer Science subject centre recently ran a workshop, ‘Massively Multi Learner’, on learning in multi user virtual environments which I was fortunate enough to be able to attend.

Perhaps inevitably, the presentations on the day were heavily skewed towards Second Life, a fact that I was glad to see the organisers themselves acknowledged as not necessarily ideal.  Unfortunately, Carl Potts, who had been scheduled to speak on learning within guilds in World of Warcraft, was unable to attend, but Laz Allen of TPLD (standing in for Helen Routledge) provided a non-SL and more game-orientated perspective on emerging technologies.  Of particular interest was the emphasis in this presentation on the assessment of game-based learning and of gaming activities, through reflection and debriefing, and through the logging and interpretation of ingame activities with reference to an identified set of skills.  Unlike commercial off-the-shelf games (COTS) and other resources such as SL, games specifically designed for learning can offer a more effective balance of learning objectives, subject matter content and gameplay, with assessment – often itself highly innovative – integrated from the outset.

The rest of the presentations all referenced SL to a greater or lesser extent.  I hugely enjoyed Aleks Krotoski‘s work on social networking in virtual worlds, in particular her identification of 75 avatars (“they know who they are”) who form “the feted [fetid?] inner core of Second Life”.  Unlike either single-player or MMO games, MUVEs such as SL are inherently socially orientated rather than goal-orientated; ‘success’ doesn’t come necessarily from accumulation of in-game objects or from PvP or PvE pwnage but from occupying key, extremely powerful positions within social networks.  As an infrequent and rather ‘resistive’ SLer, I feel strongly that the lack of scaffolding within SL, in contrast to the carefully balanced quest structure in games such as WoW which directs players through the game world and encourages casual grouping, makes social relationships within SL disproportionately important.

Other presentations explored some of the many purposes to which SL is being put.  Dave Taylor of the National Physical Laboratory discussed some very exciting international collaboration which has been taking place in the Space Island cluster, while Peter Twining demonstrated the Schome island pilot on the teen grid which is trialing SL as a learning space for a group of ‘gifted and talented’ learners.  Jeremy Kemp discussed Sloodle, an integration of SL and Moodle which uses mashups to connect the two systems.  The integration of SL and Moodle also offers the potential for resolving accessibility issues around SL by offering meaningful real time alternatives to inworld communications.

The final three speakers had all integrated SL closely into their teaching practice.  Mike Hobbs of Anglia Rushkin University described scripting tasks undertaken by second year Computing Science students to create learning resources used to explain computing concepts to first year students, while Annabeth Robinson (well known in SL as AngryBeth for her creative and practical objects) described the options her Design for Digital Media students had for woriking in SL and particularly for using it as a tool for machinima.  Mike Reddy provided an entertaining end to the day, looking at various ways in which Second Life can be integrated into a range of courses.

Integrated Assessment – IMS Webinar

On Monday night I attended IMS’s webinar on ‘Integrated assessment products and stategies: gauging student achievement and institutional performance’.  This was the first IMS webinar I’d attended, and I found it a useful session.  Over 80 people participated on Horizon Wimba for the session.

Rob Abel, CEO of IMS, introduced the session by describing integrated assessment as assessment which is designed into and throughout the learning experience.  He discussed the outcomes of a recent survey on satisfaction with elearning tools which showed that tools for quizzing and assessment had the highest satisfaction ratings amongst users; of the 88 products surveyed, Respondus came top in terms of user satisfaction.  This is a consequence both of the usefulness and maturity of this category as well as the availability and quality of tools available.

Rob also suggested that ‘standards are a platform for distributed innovation’, which is a nice phrase, although one of the criticisms often made of QTI is that it isn’t innovative.  It’s hard to see, however, how true innovation could be standardised.

Neil Allison (BlackBoard Director of Product Marketing), Sarah Bradford (eCollege Vice President of Product Management) and Dave Smetters (Respondus President) all spoke briefly about how their tools could be used for integrated assessment. 

Neil illustrated how BlackBoardcan ‘make assessment easier and more systematic’ by integration with other elements of the VLE such as enterprise surveys, portfolios and repositories.  One comment I found particularly interesting was an outcome from a December 2005 Blackboard Survey of Priorities in Higher Education Assessment, which found that portfolios are used in 86% of public institutions but only 43% of private, while interviews and focus groups are used by 78% of private institutions and only 48% of public.  I’ve tried to find this online without success; it’s referenced in slide 20 of the presentation.

Sarah noted that eCollegeusers are using Respondus and Questionmark’s secure browser to assess their learners.  Her talk focused on the eCollege outcome repository or database, which is linked to their content manager, stressing the importance of a good tagging system.  The eCollege Learning Outcome Manager addresses some of the problems for usage data management for quality assurance, an important issue given the current interest in item banking.

Dave’s talk was most wide-ranging, looking not only at the highly popular Respondus assessment authoring and management tool but at some of the wider issues around integrated eassessment.  He referenced research which found that only between 13 – 20% of courses with an online presence have one or more online assessment as part of that course – yet market research consistently shows that online assessment capabilities are one of the most appealing elements in drawing users to esystems.  As he said, once the system is in place, ‘reality kicks in': online assessment takes work, effort and time, raises difficulties in converting or creating content, and raises fears of the potential for cheating.  He argued that only a very small number of students have the desire to cheat, yet the impact can affect an entire class.  Students themselves like a secure assessment environment that minimises the possibilities for cheating.  Locked browsers are a big issue for Respondus at the moment; security of online assessments is also addressed by BS7988 which is currently being adopted by ISO.

Colin Smythe, IMS’s Chief Specification Strategist, provided a brief survey of the standards context for integrated assessment.  He noted that all specifications have some relevance for assessment, citing Tools Interoperability, Common Cartridge, ePortfolio, Content Packaging, LIP, Enterprise and Accessibility.  He also posted a useful timeline (slide 69) which shows that QTI v2.1 is scheduled for final release in the second quarter of 2007, to be synchronised with the latest version of Content Packaging.

He also said that Common Cartridge provides ‘for the first time content integration with assessment'; how much this will be adopted remains to be seen but IMS are marketing it quite forcefully.

There was time for a short question and answer session at the end.  I asked about the commitment of the vendors to QTI 2.1 and the use of QTI 2.1 in Common Cartridge.  The Common Cartridge specification uses an earlier version of QTI partly because there were some migration issues with 2.0 which have been resolved through transforms in 2.1, and also because IMS ‘didn’t want to force the marketplace to adopt a new specification’.  As Rob says, interoperability requires the ‘commitment of the marketplace’, and it would be useful to know what commitment these vendors have to the newer version. 

The session concluded with a reminder about the Learning Impact 2007 conference being held in Vancouver on 16 – 19 April 2007, which should be of interest to many.