Alan Dix > HCI Education

quality matters

SIGCHI Bulletin HCI Education column November 2001
About this time last year I wrote a little about Cheltenham and Gloucester College of Higher Education (I'll call it C&G for short) where I have been external examiner for the whole master's level provision in the college. Each subject has its own external examiner, but I had an additional responsibility to overlook the entire postgraduate scheme. I say 'was' as this was my last year of a 4 year stint and am now preparing final reports etc.

As I mentioned previously, it is an odd position. The individual subject examiners have already dealt with more academic issues and I'm not qualified to judge the academic standards of MBAs, fine art exhibitions or theology MAs. So process issues and 'quality' come to the fore.

I must admit to being a sceptic about 'quality' systems in education. In my experience they fall into two main camps. First, typical in the older UK universities, is where the quality procedures are seen as an irrelevance and are completed in order to satisfy external agencies, but make no material difference to real teaching or assessment. The second, more typical in newer UK universities and perhaps even less satisfactory, is where the processes become self serving bureaucracies - forms are produced and procedures enforced irrespective of their effect on the actual student experience.

OK, perhaps I should have said 'cynical' rather than 'sceptical'.

The reason I mention all this is that in C&G, or at least the postgraduate part that I have been associated with, it seems that ... wait for it ... the quality procedures actually work to increase educational quality. And cynic that I am I am not alone, for as often as I tell UK academics this they cannot believe that it is true.

So what is it about these QA procedures that work when so many fail, and why is it that it seems so difficult to get it right?

The first success factor is attitude: both a strong and widespread belief in the importance of education (as opposed to running an educational business, having a job, doing a bit of teaching so you have time for research) and a sense of ownership of the process leading to 'buy-in'. We all know from participatory design and similar techniques that getting users to own a system, irrespective of its technical merits, is likely to lead to success. On the other hand, we have all, I'm sure, also seen systems imposed from 'above' with corresponding widespread subversion.

The second success factor, is willingness to change. Although procedures and rules are taken seriously, where difficult cases show some deficiency in the rules they are re-examined and where necessary reformulated. Real quality procedures are not handed down on tablets of stone, but worked out in praxis.

Those who know software engineering may think of the SEI capability maturity model. Software development teams/companies are graded at various levels. At the lowest level are ad hoc undocumented processes, with documented processes above this. However, the highest levels of 'maturity' are for those who have meta-processes to constantly re-examine and where appropriate modify their development processes, This meta-reflection is a very 'academic' thing, but in so many contexts is also so very practical. In a confusing and complex world we may not be able to get things right first time, but we can make them better.

So finally, why is it that quality procedures seem so hard to get right in a university context. One reason is that they are typically modelled on (successful) industrial QA. But, industrial QA is all about uniformity, not excellence. In a fork factory it is important that every fork is the same - output is assured to be uniform. In a university the output (students understanding and grades) are not expected to be uniform. Uniformity of opportunity yes, but not a guaranteed B+ grade for everyone!

Furthermore, we would never accept an industrial process with the diversity of input as we do in education. Imagine putting aluminium, stainless steel and wood into the same fork factory!

Funnily enough interfaces are similar, we have so many different users, different goals, different contexts - do we want systems to produce uniform outputs, or do we want systems that enable each to achieve their own purposes in their own ways?

So industrial QA may certainly be appropriate for educational administrative processes, admissions, building services, cleaning ... but for education itself, perhaps we need something a little different. But perhaps the same something different we need for quality interface design.