1. Function - Formative versus summative
   • What are the actual functions of these assessments?
   • Assessment is the only formative if it has an impact on pupils learning
2. Quality - Validity versus reliability
   • How reliable are the tests given to students?
   • What about the students on the boundary of levels and grades? Do we need to use classical measures of reliability.
   • Research has shown certain percentages of pupils have been awarded the wrong grades
   • How valid are these actual tests? Is a maths test with a high content of reading actually required valid?
   • A tests validity should include reliability
   • Are schools achieving good results good schools?
3. Format - Multiple-choice versus constructed response
   • Are multiple choice tests fair? Are they biased against a certain gender?
   • Do these tests have misconceptions construed by pupils?<
   • What do our tests leave behind apart from scores?
   • We can develop good quality tests if we spend more time and money, develop more systematic ways
4. Scope - Continuous versus one-off
   • Are some assessments too narrow? What are the effects?
   • Do they encourage teachers to teach to the test?
   • Does the learning become narrow, shallow and transient?
5. Authority - Teacher - produced versus expert-produced
   • Who sets the assessment?
   • Are there problems of biased inferences?
   • Is teacher assessment valid?
6. Locus - School-based versus externally marked
   • Do teachers need to become more involved in all aspects assessment - marking tests?
   • We need to learn from other countries, but would these systems work in this country?

The Challenge

To design an assessment system that is:

• distributed
• synoptic
• extensive
• manageable
• trusted

Click on the link for a copy of Dylan Wiliam's keynote presentation