How can standardised tests reduce bias?
Simply put, standardised tests can reduce bias by removing much of the variation inherent in assessments administered by humans. While some standardised tests may involve marking by people (others use computer marking), those markers will be trained and monitored, and the reliability of their judgements noted and, where needed, improved through training and moderation.
Good standardised tests have also undergone investigations which provide an understanding of how valid and reliable they are (assessment providers should be able to describe these aspects of their tests clearly). Those which are delivered by computers (and don’t use human marking) can reduce bias considerably, and increase the reliability and objectivity of the assessment process.
Applications and limitations: using standardised tests as bookends
While standardised tests may be more reliable and objective than some teacher assessments, they are limited both by what they can assess and the cost of using them. Developing good assessments is a very hard process (a single test can take months to develop), and even the best can only assess a narrow (though often important) range of attributes or characteristics (such as reading ability).
Having acknowledged that they do not provide an assessment panacea, a good mix of standardised assessments and well-administered teacher assessments is recommended for schools. The standardised assessments – often used at the beginning and end of academic years – can offer robust ‘bookends’ to help guide teaching strategies and differentiation in the first instance and, subsequently, to make statements about progress. In between these points, however, reliable, valid, purposeful and valuable teacher assessment is needed.