Reliability Test-re-test, equivalent forms, internal consistency. Test-re-test, equivalent forms, internal consistency. Consistency across markers, occasions, test items, test types, marking conventions, grading procedures, contexts. Consistency Inter-raterreliability Dependability Degree of confidence Transparency
Threats to reliability from markers Threats to reliability from markers Errors in Errors inmarking Inter-rater Inter-raterreliability Inconsistency in Inconsistency inmarkers Variation at Variation at grade boundaries Halo effect Halo effect
Copyright Keith Morrison, 2004 THREATS TO RELIABILITY FROM STUDENTS AND TEACHERS Motivation and interest in the task Relationship between assessor and assessee Test conditions/environment Hawthorne effect Distractions Teacher expectations Time of day/week Students are not always clear on what they think is being asked
Copyright Keith Morrison, 2004 THREATS TO RELIABILITY FROM STUDENTS AND TEACHERS Students may see two similar questions as dissimilar Teachers teach to the test Teachers and students practise test-like materials Students can perform the skill in the test but not in real life, and vice versa Cultural and ethnic background affect meaningfulness of the test Unreliability of marking practices
Copyright Keith Morrison, 2004 THREATS TO RELIABILITY FROM ASSESSMENT ITEMS Task may be multi-dimensional Validity of items Language of the assessment Readability levels Size and complexity of numbers Number and type of operations Form and presentation of questions Early error in a sequence may block later stages of the sequence Questions using mechanical toys may favour boys over girls
Copyright Keith Morrison, 2004 THREATS TO RELIABILITY FROM ASSESSMENT ITEMS Questions requiring use of dolls or kitchen work may favour girls over boys Essays favour boys if they concern impersonal topics and girls if they favour personal topics Boys perform better than girls on multiple choice items Girls perform better than boys in written work Continuous assessment favours girls Questions are culture-bound The length of the test affects performance
Copyright Keith Morrison, 2004 CONSTRUCT Predictive validity Content validity Face validity Consequential validity VALIDITY
Copyright Keith Morrison, 2004 KEY FEATURES OF ASSESSMENT (1) Assessment to empower pupils as learners. Classroom assessment impacts significantly on the pupils’ sense of self, expectations, motivation and confidence. Assessment should provide guidance to both teachers and pupils about what needs to be learnt next. Assessment should embody an approach to teaching and learning in which the development of long-term dispositions is more important than short-term performance.
Copyright Keith Morrison, 2004 KEY FEATURES OF ASSESSMENT (2) The purposes are to be diagnostic and formative, providing feedback and being educative. Teaching should be adjusted in light of assessment evidence. Assessment should promote, not damage, student motivation and self-esteem. Assessment should be constructively critical and provide rich, positive feedback and feedforward. The assessments should be criterion-referenced and the criteria should be public. The assessments should lead to diagnostic teaching. Assessment should promote student self-evaluation.
Copyright Keith Morrison, 2004 KEY FEATURES OF ASSESSMENT (3) The assessments should be built on evidence rather than on intuition. Assessment data should be derived from everyday classroom activities. Assessment opportunities should be sought in everyday classroom activities. Semi-structured approaches to gathering data are recommended, generating words rather than numbers (measures). Assessments should be linked to the student teacher’s and the student’s action planning and target setting.
Copyright Keith Morrison, 2004 KEY FEATURES OF ASSESSMENT (4) Involve the students in the assessment process. Communicate the assessment criteria to students. Demonstrate validity and reliability. Demonstrate fitness for purpose in deciding the method(s) of gathering assessment data and setting assessment tasks. Select assessment methods that accord strongly with everyday teaching and learning processes.