INTRODUCTION TO ASSESSMENT IN
EDUCATION
QUALITIES OF A GOOD TRADITIONAL TEST
Facilitator:
Shellon Samuels-White
QUALITIES OF A GOOD TEST
The effectiveness of a test is evaluated based on the following:
• Objectivity Balance Difficulty
• Fairness Relevance Discrimination
• Absence of Bias Efficiency Reliability
• Validity
EFFICIENCY
• Efficiency matches the number of items a student can be expected
to answer in the allotted time, against the number necessary to
adequately test the objectives of the test.
• Thus if a test is too short , it may not test the objectives adequately,
whilst if it is too long, it might become irrelevant, that it might be
testing endurance rather than the mastery of objectives.
EFFICIENCY
The number of items to use is modified by a number of practical
constraints:
• Age of students tested
• Time available for testing
• Type of test items
• Type of interpretation to be made (Cognitive rigour)
EFFICIENCY - EXAMPLE
• Comparison of two test scores
SAMMY’S PAPER ROY’S PAPER
2+3=5 2+3=6
5 + 4 = 10 5+4=9
6+ 3 = 8 6+ 3 = 9
4 + 8 = 14 4 + 8 = 12
9 + 6 = 16 9 + 6 = 15
DIFFICULTY
• Test should not be so easy that everyone gets 100%, nor so
difficult that no one can pass.
• Items should be at a level where at least half of the students get it
correct, though not the same students for every item.
DISCRIMINATION
• A test must discriminate between low performers
and high performers.
• If the same proportion of high and low achievers
make the correct or incorrect response to an item, it
has no discrimination.
• Poor discrimination should not be used as the only
reason to exclude an item from a test, as it may be
related to the lesson objective.
RELEVANCE
• When we assess, test items must match the objectives
they are designed to measure, in order to be judged
relevant.
• Tests should measure what has been taught and
nothing else.
RELEVANT- EXAMPLE
Objective:
Students will be able to add even numbers.
Test item:
• Add: 4 + 26 (Relevant)
• Add 4 + 26 and write your answer in (a) Words (b)
Standard notation (Irrelevant)
FAIRNESS
• The test should be constructed and administered
in such a way that all persons taking the test have
an equal opportunity to demonstrate their
achievement.
• All students taking the test should have been
exposed to the same content and should take the
test under the same conditions.
FAIRNESS - EXAMPLE
• IAE Groups A & B are both given the same end of term
exam. (Fairness taking place)
• Group A is not allowed to use calculators because they are
the Faster Group, but Group B is allowed because they are
little slower than Group A. (Not Fair)
OBJECTIVE
• A test is objective if all scorers agree on the right/wrong responses.
That is if it is marked by different teachers, the score will be the same.
The marking process should not be affected by the teacher’s
personality.
Tests must be Objective in:
(a) Test items ( items are perceived in the same way – free from
ambiguity.)
(b) Test scoring ( Scoring in the same way)
OBJECTIVE - EXAMPLE
• 15 + 4 = ? (objective test item)
• Describe a Sunset. ( subjective – scoring may be
influenced by the scorer’s judgement/personal
preference).
ABSENCE OF BIAS
• Tests should be free from bias.
• Assessment bias refers to qualities that offend or
unfairly penalize a group of examinees because of
gender, ethnicity, socioeconomic status ,religion or
other group-defining characteristics. (Poham, 2003)
ABSENCE OF BIAS - EXAMPLE
• Write an expository paragraph on “ Playing outdoors during
the snow”. (Biased - The topic might be outside of the
students scope of experience, since it doesn’t snow in
Jamaica and some of them might not have travelled outside of
Jamaica).
BALANCE
• The distribution of items in a good test should match the
emphases of the teaching course.
• Thus, if most of the time was spent on a particular topic
and less time on another, then the test should contain
more questions on the area that more time was spent
covering, and less on the area that less time was spent
covering.
BALANCE - EXAMPLE
• Mrs. Rowe spent 4 sessions teaching division of fractions and 2
sessions teaching equivalent fractions. Hence, on her Monthly
test she had 8 questions testing on dividing fractions but only 4
questions testing equivalent fractions.
(Balanced)
RELIABILITY
(Consistency, Dependency,Trust)
• Reliability involves the consistency with which a test yield the same
result in measuring whatever it measures. This involves the lack of
measurement error - the less error the better.
• Reliability is measured through testing and retesting.
• Reliability is a prerequisite to validity.
RELIABILITY
(Consistency, Dependency,Trust)
Questions to consider:
• If you mark the student’s paper today and re-mark tomorrow,
does the student get the same score?
• If an essay is marked independently by two teachers, do the
students get the same/similar score?
• Does a student who usually does well/poorly on tests score
well/poorly on your new test.
RELIABILITY- EXAMPLE
• We administer a test on Group A and got an average of 55. After 3
days we administered the same test to Group A and got the same
average. (Reliable – average is consistent.)
• Mary hates Spanish and normally does poorly on her Spanish tests.
She scored 30% on April monthly test. However, in May she scored
90% on her test.
(Test is not reliable – not consistent in the marks)
VALIDITY
Validity is the extent to which the test measures what it is intended to
measure.
Content validity is established by constructing test items that are
congruent with the objectives. Valid assessments must be reliable.
Face validity describes the degree to which an assessment
measures what it appears to measure.
VALIDITY
Select the type of assessment that best measures the intended
outcomes directly. E.g.. use supply response type items if
supplying the answer is an important element of the task e.g..
writing/calculating
Let’s Practice!
Critique
- Evaluate (a theory or practice) in a detailed and analytical way.
Checklist
• Description of the instrument (target audience, purpose, duration etc.)
• Arrangement of test – sections, directions etc.
• Domains of learning
• Alignment to curriculum
• Item types – against guidelines
• Suggestions for improvement – substantiated by evidence
SOURCE
Richardson, M. (1998) Classroom Testing and Evaluation.
Chalkboard Press
