Methods of Assessment
23. What are norm-referenced tests? What are criterion-referenced tests?
Norm-referenced tests
The purpose of a norm-referenced test is to compare the performance of an individual with the general standard of performance demonstrated by the group to which the individual belongs, or with another appropriate reference group. An IQ test is a familiar example.
A norm-referenced test therefore compares the performance of an individual with that of other people.
The primary purpose of norm-referenced testing is to discriminate between candidates by identifying differences in their levels of performance.
Item difficulty is determined by finding the percentage of people in a trial group who answer an item correctly.
Items which are too easy or too difficult are normally rejected in a norm-referenced test because they contribute relatively little to the overall discrimination of the test.
To establish the discrimination value of an item, the performance of individuals who answer the item correctly is compared with their performance on the test as a whole. Ideally, candidates who score highly on the test as a whole should be more likely to answer the item correctly, while candidates who score poorly overall should be more likely to answer it incorrectly.
Criterion-referenced tests
In a criterion-referenced test, the emphasis is not on how an individual performs in relation to his or her peers. Instead, the concern is whether the individual knows or can do something specific that he or she is expected to know or be able to do.
A criterion-referenced test describes an individual's performance with reference to externally predetermined and specified objectives. The criterion is therefore an externally defined standard of performance.
Examples include attainment scales used by Cambridge Oral Examiners and B. J. Carroll's operational specifications, which define sets of performance criteria. For example, an objective might be that a learner possesses the level of language skills necessary to function adequately as a tourist.
In criterion-referenced testing, each objective is considered separately. The tester is not primarily interested in the candidate's overall score but in whether particular objectives have been achieved.
The basic assumption is that there are two groups of individuals: those who can carry out the required operation and those who cannot. The test is therefore intended to distinguish between these two groups.
However, this apparent dichotomy needs qualification. There is a third group within the population: people who are currently learning to carry out the operation. Among these learners there will be a continuous rather than a simple yes/no distribution of ability.
The items in a criterion-referenced test should therefore reflect the specific criteria being tested. Members of this third group may pass some items while failing others, depending on which particular skills or objectives they have mastered.