Invited presentation,Australasian Curriculum,
Assessment&Certification Authorities(ACACA)2003
National Conference,Adelaide,July 31
VALIDITY AND RELIABILITY IN THE SENIOR SCHOOL CURRICULUM:
NEW TAKES ON OLD QUESTIONS
留学生论文枪手https://www.51lunwen.org/StudentPapers.html
Tim McNamara
The University of Melbourne
Validity theory has made important advances in the last decade,and has thepotential to illuminate old problems in senior secondary
Assessment in a new way.This presentation introduces the Evidence Centred Design model ofMislevy et al.(2003),which enables us to distinguish the claims we wish to make about candidates on the basis of assessments from the evidence we wouldneed to gather to support such claims.The practical design of assessment procedures follows from the nature of the evidence that we need to gather.The current status of this model is evaluated,together with its relevance to assessment at senior secondary level.Evidence Centred Design is in turn an operationalization of the validity model of Messick(1989),who proposed a unified model of validity in which reasoning and empirical evidence are used tosupport the inferences(claims)about candidates that are made on thebasis of
test scores.Messick’s insistence on the need to articulate and defend the valuesimplied in test constructs,and to take responsibility for the impact of thedecisions made on the basis of test scores,have proved the most novel andcontroversial aspects of his work.The paper concludes with comments on theadequacy of Messick’s model,particularly in dealing with the social andpolitical context in which senior secondary assessment is conducted.
INTRODUCTION
In this paper,I will attempt to present some recent work on the validity of educational assessments.This work makes available a set of conceptual tools that we can use to2talk about the purposes of our assessments,how we can make them more flexible without sacrificing quality,what their vulnerable points are,and how we mightaddress those,and how we can make assessment serve the cause of learning better.I will begin by presenting a simple model of the fundamental character of an assessment,and then relate it to work by the American validity theoristsBob Mislevyand the late Samuel Messick.I will give examples of the points I ammaking from
work in the Australian context,where possible at Senior Secondary level.I’llconclude with a recent idea for building a practical link between the criteria used in judging student work and opportunities for student learning.
WHAT IS A TEST?
Assessments involve three steps(Figure 1).
Criterion Construct Test
domain of real-
world
performance,
knowledge,
capacity
characterization
of essential
features of
performance;
theory of domain
?test design?
test
performances;
item responses
Target
(unobservable)?about
(via
theoretical
model)
?inferences
Observations/
Evidence3
Figure 1 Criterion,construct and test
First,we identify our test target–what we would like to know(if only we could)about test takers:what they know,what they can do,the performances they arecapable of in non-test settings.This target of our inferences is known as the criterion.However,the criterion cannot be‘known’in any direct sense;it can only be knownabout indirectly,‘through a glass,darkly’,in a testing version of Labov’s Observer’sParadox,or Heisenberg’s Uncertainty Principle.We have to approach it through
inference and surmise.
本论文由英语论文网提供整理,提供论文代写,英语论文代写,代写论文,代写英语论文,代写留学生论文,代写英文论文,留学生论文代写相关核心关键词搜索。