Objective testing
An objective test is one in which each question is designed to have
an unequivocally correct answer or set of answers. In contrast a
subjective test, for which there is no uniquely correct answer, examines
the ability of the student to respond clearly and develop an argument.
The Multiple Choice Question (MCQ) is the most commonly used type of
objective question. While objective tests are used largely to test
knowledge and comprehension, with careful design they can also assess
higher learning outcomes such as the application of knowledge and
analytical skills.
Because they can be marked quickly and easily and can cover a wide
range of course content, objective tests are particularly appropriate
for informal self-assessment to give students ongoing feedback about
their progress. They can be used in this way throughout an undergraduate
course. For formal assessment, however, their contribution to final
marks is likely to decline as students progress through their degree
programmes and the range of abilities being examined becomes more
comprehensive and demanding.
The use of computers for objective testing
Approaches to computer-assisted assessment (CAA) can be divided into
two systems:
automated marking of paper forms, using an optical mark reader
(OMR) with hard copy question paper and an OMR-readable answer
form;
computerised marking in which questions are presented and responses
assessed entirely by computer software, with no paper
involved.
CAA software can provide immediate supportive feedback for each
question, tailored to the answer given, making this particularly
suitable for informal self-assessment. Superior graphical and
visualisation exercises can also be included with the use of multimedia
files such as graphics, sound and video clips. Most software offers a
facility for reporting results and analysing student responses. An
increasing number of systems are web based.
Benefits
It is objective (i.e. highly reliable in the assignment of
marks)
Tests can be marked and returned very speedily
Tests can incorporate a variety of media (images, video, audio)
Relevant feedback can be given automatically (during or after the
test)
Randomised selection can be made from large question banks
Flexible access gives students opportunities for
self-assessment
Built-in test management (collation, analysis, tabulation, report
generation) eases the administrative burden of
assessment
Concerns
Questions tend to address factual content and may therefore
encourage learning of surface detail rather than an appreciation of
underlying concepts
The nature of the process makes it difficult to retrieve mistakes
in compiling questions before the test is taken, placing a heavy duty
of responsibility on the question setter
Formal assessment by computer is subject to concerns over
security
Students may obtain the correct answer for the wrong reason or may
be encouraged to reinforce misconceptions about the subject while
answering questions incorrectly
Hardware and software resources and support must be available
Producing well designed, appropriate questions involves a
considerable investment of staff time
Objective question types
There are various forms of objective question:
multiple choice: choose the correct answer form a list of
alternatives
multiple response: select a number of correct answers from the
list
true/false
selection/association: match items from two related lists
assertion/reason: choose the correct reason for an assertion (a
special case MCQ)
While these question types can be delivered on paper and then
processed by OMR, the following two types specifically require a
computer for input:
text match/gap filling: enter a word, short phrase or number
visual identification/hotspot: move a marker is moved to identify a
particular ‘hotspot’ on an image
Designing questions
An MCQ question typically comprises three parts:
stem: the question component;
key: the correct answer;
distractors: incorrect answers provided as alternatives to the
key;
and optionally:
feedback: a mark and/or comment reflecting the student’s
performance. This may be presented immediately after the question, at
the end of the test or not at all. Feedback may also be provided on
the basis of overall test performance.
For MCQs students might be discouraged from guessing by the
imposition of a penalty for incorrect answers, although it is generally
considered that negative marking simply alters the baseline of results.
In any case, the rewards of ‘blind’ guessing diminish rapidly as the
number of questions increases. Other viewpoints hold that ‘intelligent’
guessing may be no bad thing, and that well designed MCQ questions will
lead a genuinely blind guesser astray.
It is self-evident that the questions should reflect the aims and
objectives of the course and be appropriate to the abilities of the
students for whom they are intended. A pre-test during which questions
can be tried out, edited and reformulated is considered essential.
Beyond that:
stem: ensure that it is concise and unambiguous, avoiding
negatives (‘Which of the following is NOT…?’) and grammatical clues to
the key
key: this should be the same length as the distractors and its
position in any list should be randomised using the software (unless
it is a numeric answer in which case answers should be in ascending
order, avoiding any systematic placement of the key)
distractors: these should be apparently realistic alternatives
that, as far as practicable, cover the full range of options and
easily identify misconceptions
feedback: this should be appropriate, helpful, encouraging,
varied and unpatronising
Implementing tests
Students find that objective tests require considerable concentration
and it is probably wise to limit the test content to no more than one
hour’s duration, although imposition of a time limit for self-assessment
purposes is usually counterproductive. The use of open tests to assess
coursework is subject to abuse through collaboration unless supervised,
but this is true of conventional activities executed out of class as
well. Encouraging this sort of student collaboration may even be
considered desirable. The provision of feedback under such conditions is
also a significant issue: withholding information can compromise other
educational objectives. Using CAA in examination conditions requires
careful preparation including back-up plans in case of hardware or
software problems. Step-by-step tutorials and check lists are available
for summative assessment planning.
Question banks
Writing good objective questions can be a lengthy and laborious
exercise. It is usually faster to edit existing questions. Some question
banks are available commercially, generally associated with text books.
It is worth finding out whether the textbooks you use for your courses
have associated questions. It is also possible to adapt questions found
in textbooks for objective testing. Otherwise, contact colleagues at
other institutions to see whether they are willing to work with you to
produce questions for similar courses. This may best be done via a
relevant learned society. Your CTI Centre can probably point you in the
right direction.
Resources
Computer-assisted assessment centre: http://caacentre.ac.uk/
Computer-assisted assessment mailing list: http://www.mailbase.ac.uk/lists/computer-assisted-assessment/
For information on OMRs see Data & Research Services: http://www.drs.co.uk/
For information on computerised systems see Question Mark Designer
and Perception: http://www.qmark.com/ or TRIADS: http://www.derby.ac.uk/assess/talk/quicdemo/html/
For Web-assisted assessment see CASTLE: http://www.le.ac.uk/castle; the
Web-assisted assessment mailbase list: http://www.mailbase.ac.uk/.
A tutorial on setting effective objective tests is available from the
CTI Centre for Land Use and Environmental Sciences: http://www.clues.abdn.ac.uk:8080/caa/caatut.html
A protocol for the implementation of summative computer-assisted
assessment examinations is at: http://www.clues.abdn.ac.uk:8080/caa/protocol.html
Original
Authors
Eamonn Twomey
Jacqui Nicol
Christina Smart
Version: 4.1.10.1