6.3 Increase the Granularity

The reliability of assessment suffers when the value of the marginal (unit) mark is small.

---ooOoo---

This bundle focuses student (and staff) attention on the extent of the difference in achievement that is reflected by a single mark - and changes the way assessors award marks so as to ensure the scale rises in appropriate steps.

The way it works is, if each category in the marking scheme (e.g. deign, implementation, evaluation) has an associated maximum mark, and within each category marks in the range 0-100 are awarded then it is very difficult to achieve numerical agreement between one assessor's 57% and another's 52%. However, if the granularity is increased, to reflect the marks "earned" by each component (i.e. if Design carries 40 per cent of the total project marks, mark it out of forty instead of 100). Agreement between assessors on what 15/40 means is more likely, and the award of extra marks on appeal from the student becomes more meaningful.

It works better if there is agreement among the staff that granularity is a problem and if agreement about what the weightings should be has been secured.

It doesn't work unless you have a number of categories in the mark scheme.

VariationStudents think that 1% on an assessment that counts for of 25% of 1/12th of their degree is worth arguing over. It isn't. Don't argue, but give it away freely.

---ooOoo---

So: make the marginal mark worth something