discussion method(in testing) a type of construct validitythat is based on the extent to
which two or more tests that are claimed to measure different underlying
constructsare in fact doing so. For example, to establish discriminant
validity of two tests that are claimed to measure the different constructs
(e.g. L2 listening and L2 vocabulary), both tests are administered to the
same group of test takers using the same method (e.g. multiple-choice
itemsfor both tests) and the test scores are correlated. If a weak or no
correlation is obtained, this is an indication that they are indeed measuring
different constructs.
see construct validity, convergent validity, multi-trait multi-
method methoddiscrimination^1 n
see stimulus-response theory
discrimination^2
also discrimination power
(in testing) the degree to which a test or an item in a test distinguishes
among stronger and weaker test takers. For example, if test takers are
known to have different degrees of ability but all score around 85% on a
test, the test fails to discriminate. A measure of the discrimination of a test
is known as a discrimination index.
see also item discrimination
discrimination index n
see discrimination
discrimination power n
another term for discrimination
discursive practices n
a term used in critical discourse analysisto refer to the processes of
production, distribution and interpretation that surround a text and
which must be taken into account in text analysis. These practices are
themselves viewed as embedded in wider social practices of power and
authority.
discussion method n
an approach to teaching which consists of a goal-focused group con-
versation involving either groups of students or the whole class, and which
usually involves interaction about subject matter between a teacher and