Reliability of assessment

Reliability and validity of assessments are important, but they have entirely different meanings in the standards model (pp. 188-90). A frequent criticism of qualitative assessment is that it is 'subjective' and 'unreliable'. This is the measurement model talking.

Reliability here is not a matter of statistical operations, but of being very clear about what we are doing, what learning outcomes we want, what is to be the evidence for those outcomes and why. In other words, reliable assessments are part and parcel of good teaching. We have been explicating the framework and the specific criteria for making informed and reliable judgments about students' learning from Chapter 5 onwards.

Can we rely on the assessment results – are they reliable?

In the measurement model, reliability means Stability, Dimensionality, and Conditions of testing. Here reliability is seen as a property of the test.

In the standards model reliability means something rather different: intra-judge reliability and inter-judge reliability.

Here, reliability is not a property of the test, but of the ability of teachers/judges to make consistent judgements.

Source: Biggs, J.B., & Tang, C. (2007).