Pretesting is actually
a process where the test items (the questions) are actually tested before the
items are released into tests and also exams. Pretesting ensures the exams are
accurate and fair for all students. It also ensures the test content is
appropriate and that students from various background can equally score for it as
well. There are few parts of pretesting that would be explained in this blog.
Pilot testing
Before the main
trial is carried out, a less formal pretesting would be carried out that is
called “pilot testing”. This pilot testing varies from testing it on some colleagues
to students. The aim of this pilot testing would be to identify and rectify all
problems before the main trial. There are few methods of carrying out the pilot
testing. There are two types given here:
>Testing on colleagues/fellow
teacher
-at least 2 of
them are native speakers of English
They are to
check whether:
-The instructions
are clear
-The language of
the items are acceptable
-The answer key
is accurate
>Testing on a
group of students from the similar background and level. This will provide
information on:
-The time
students need for completing it
-The clarity of
instructions
-The kind of
language
-The accuracy and
comprehension of any answer keys
-The usability of
the marking scales
-Pilot testing
will reveal many flaws and thus will save time and effort when the main trials
are carried out.
The
main trials
Here we would be discussing on how to
conduct main trials.
The main trials are where the test items
are tested on students possibly replicating the real exam. The focus here would
be on multiple-choice items as they are difficult to construct. The test constructers
usually miss the ambiguities in the distractors. Thus these type of questions
would require more pretesting compared to other types of questions.
There are few aspects that should be
taken into consideration when it comes to the main trials:
v The
sample should be representative of the final candidature
The pretest must have a similar range of abilities
and background. The results of the pretest would be useless if the expected
test population is not similar. Henning 1987 recommends 1000 students for trial
multiple-choice test. Basically, the more students taking the pretest, the
better it would be.
v Trial
test students should take the test seriously
If the students do not take the test seriously then
the whole trialling process would be useless.
v Trial
test should be administered exactly the same way as the final exam will be
So that the administration
guideline together test item can be presented under same circumstances as in
the live exam.
No comments:
Post a Comment