Chapter 7
This chapter discusses designing web-based tests and
exercises. A study was conducted which showed that
Internet-conducted tests showed no differences in
scores, redesign needed, or any bias due to gender,
economic disadvantage, or educational disability relative
to the same test on paper. The author notes the good
and bad reasons for using tests and specifies that
designers should know what kind of skill that will
measure student learning: knowledge, skills, or attitudes.
Designer/teachers should also know how test should
be graded (computer graded, teacher graded, peer-grading,
etc.), when feedback will be delivered (after each
question, after a delay for human evaluation, after
all of the questions have been asked, etc.), the amount
of time learners should have, specifying if learners
can retake the test, and what things should be done
if there are technical problems. Since there is a
lot technology at work (network, servers, browser,
plug-in, etc.), designers/teachers should always prepare
for these situations.
The author later discusses the different types of
questions in which can be applied in a web-based environment.
Designers/teachers should select the most appropriate
type of questions for their subject and students.
Following are a list of the different types:
- True/False Questions – these
types of questions should be used in a right or
wrong, work or not work, safe or unsafe types
of questions. To make these types of questions
more effectively, there should be more than one
question on a subject, phrase it in different
ways, discourage guessing, provide clear hints
or explanations for incorrect answers, phrase
the question to fit the answer, and provide alternative
forms.
- Multiple-Choice (MC) Questions
– these types of questions display a list of answers
for learners to choose from and are fairly easy
to construct and understand. MC Questions can
have one answer or multiple answers.
- Text-Input Questions – these
types of questions require the learner to type
in the answer to a question. Text-input questions
should be used to recall technical or business
terms, part numbers, abbreviations, commands and
statements in a programming language, and vocabulary
in a foreign language.
- Matching-List Questions –
these types of questions require learners to specify
which items in one list correspond to items in
another. Some ways to design matching-list questions
include writing list items clearly, keep the list
short so that they both fit in the same display,
let learners indicate matches simply, and eliminate
the “process-of-elimination” effect.
- Click-in-Picture Questions
– These types of questions ask the learner to
select an object or area in a picture by pointing
to it with the mouse and clicking the mouse button.
Click-in-picture questions should be used when
it is important for learners to know where something
is or what it looks like than what it is called.
- Drag-and-Drop Questions –
These types of questions require learners to move
icons or images to specific locations on the screen.
Drag-and-drop should be used when testing learner's
ability to (1) name things by dragging names onto
them, (2) classify things by dragging them into
boxes representing the categories, (3) rank things
by dragging them into position along a scale,
(4) assemble the pieces of a whole by dragging
them into correct positions relative to one another,
and (5) show relationships among a group of objects.
- Simulation Questions – these
types of questions uses a simulation to let learners
perform a highly interactive task. Simulations
should be used when testing the ability to perform
a procedure, when the procedure is complex, and
when you are qualifying people to perform a task
in the real world.
- Fill-in-the-Blanks Question
– These types of questions require learners to
supply missing words in a paragraph of text or
missing items in a table. Fill-in-the-blank questions
should be used to test incremental knowledge,
where context matter, to measure ability to apply
verbal knowledge in context, to ask complex questions,
and to provide scaffolding.
Before giving the test, an important thing to do is
explain the test thoroughly. Learners should know
if the test is graded, thing that it test covers,
if test is timed, the length of the tests, and so
on. Learners are usually very curious test, because
it can weigh heavily on the course. Questions should
be designed effectively. Some ways include word questions
and answers clearly, keep it simple for students,
ask just one question at a time, ask job related questions,
avoid obsolescence, be careful when asking absolute
questions, emphasize important words, keep all answers
about the same length, keep questions challenging,
and simplify selecting answers.
The way that test questions are sequenced is important.
Also, questions should have meaningful feedback. For
correct answers, feedback should be brief, but for
incorrect answers, feedback should be neural and not
embarrass the learners. Incorrect answers should be
provided with a thorough explanation such as why the
answer is wrong and where to get remediation.
Finally, the author closes by nothing ways to improve
testing and prevent cheating. Some ways to improve
testing include monitoring results (see which questions
are hard and easy and which questions learners may
skip), encourage learners to provide feedback, make
tests fair for all learners, test early and often
as possible, define a scale of grades, vary test and
practice forms. Some ways to prevent cheating include
finding out why learners cheat in the first place,
ways to detect cheating, finding ways that learners
cheat, finding ways to reduce cheating (trust approach,
fence approach, and threat approach), and validate
test-takers (identify background information and other
ways to ensure that the person taking the test is
actually taking the test).
