The assessment problem
Assessment drives education - or to be more precise, summative assessment drives education.
Basic requirements of summative assessment
Summative assessment needs to meet a number of basic requirements (the essential criteria) in order to be viable/cost effective. It must:
- enable comparisons to be made between learners (eg Person X did/is better than Person Y)
- be scalable (ie be able to be applied to huge numbers of learners at a reasonable cost)
- provide concise outputs (ie provide 'results' in a format that makes it easy and quick to make comparisons across learners)
- be credible (people must believe that the comparisons that the assessment enables are meaningful/useful - this does not necessarily equate with assessment validity)
Ideally summative assessment should also meet the following desirable criteria:
- assess things that are relevant
- include a formative assessment element
The limitations with current assessment
At present we have a range of forms of assessment that meet the first batch of essential criteria quite well, but fail to meet the second batch of desirable criteria adequately.
Exams, for example, clearly meet the comparison, scalable, concise criteria and are seen as being credible (despite evidence that they do not actually provide the 'objective measure' that folk would like to claim). However, exams generally fail dismally in relation to the desirable criteria. Exams tend to focus on content rather than process. Exams also tend to adopt a behaviourist or constructivist view of learning (which focuses on what an individual can do in isolation) rather than a socio-cultural view of learning (which focuses on what individuals can do in a social context - which includes in collaboration with others).
Essays are quite similar to exams, though they are seen as being less credible (there is a greater recognition of their lack of reliability/validity) and often are less scaleable (particularly in contexts where feedback is expected to be provided).
Other forms of assessment tend to do better against the desirable criteria, but fail to meet the essential ones. For example, portfolios may perform well in terms of assessing things that are relevant but are not scaleable (if you don't believe me then try marking 200 portfolios in one batch - life is too short!).
One of the consequences of all of this are that much learning is not valued, because it is not assessed (in a way that meets the criteria set out above). Other consequences include:
- predefinition of the curriculum - cos that helps with making 'credible comparisons'
- a limited focus on those things that current forms of assessment can handle
- <Add more here?>
The problem that we need to crack is how to develop forms of assessment that meet all of the essential and all of the desirable criteria.
Utilising peer assessment may be the way forward. Wikied represents one ideas about how this might work.