(in testing) a measure of the ease of a test item. It is the proportion of the test
takers who answered the item correctly, and is determined by the following
formula:
ItemFacility(IF) =where R=number of correct answers
N=the number of test takers
The higher the ratio of Rto N, the easier the item.item pooln
see item bank
item response theoryn
also IRT
a modern measurement theory, as opposed to classical test theory, based
on the probability of a test taker with a certain underlying ability getting
a particular item right or wrong. The difference among the three main
IRT models is in the number of parameters estimated in each model. The
one-parameter model, also called the Rasch model, estimates only item
difficulty (b-parameter); the two-parameter modeltakes into account item
difficulty and item discrimination (a-parameter); and the three-parameter
modelestimates a guessing factor (c-parameter) in addition to item dis-
crimination and item difficulty parameters. The more parameters an IRT
model has, the more complex it becomes and the larger sample size it
requires. IRT is used to detect test bias (e.g., diffrential item function-
ing) and develop a computer-adaptive test, among other applications.
see also classical test theory
item specificationsn
a set of item-writing guidelines consisting of the following elements: (a) a
brief general description of the skills to be measured by the item, (b) a
description of the material that test takers will encounter and respond to in
the item, (i.e., a prompt), (c) a description of what test takers are expected
to do in response to the prompt (e.g., select an answer from four options in
a multiple-choice itemformat) and how their responses will be evalu-
ated (e.g., a set of rating criteriafor essays), and (d) an example item,
written according to specifications.
see also test item, test specifications.
ITIn
an abbreviation for the Institute of Translation andInterpreting.
item poolR
N