Psychology of Assessment - Week 6
Test Construction
The 5 Stages of Test
Development
1. Test Conceptualization
2. Test Construction
3. Test Tryout
4. Item Analysis
4. Test Revision
Test Conceptualization
- The objective of the test (what will
it measure?)
- The need for the test (what
advantages will it have?)
- The test population (who will take
it?)
- The content (what will the test
cover?)
- The style of administration (how
will it be given)
- The item format (will it be forced
choice? Multiple choice?)
- The inclusion of alternate forms use
(is it necessary for this test?)
- The training requirements (what
professionals are allowed to give the test?)
Test Construction - Scaling
- Scaling or the process of assigning
numbers in measurement
- Thurstone was the first to push the issue,
stating that scales should be developed based on the performance
of multiple test takers...Allows you to say someone has more or
less of the characteristic you are measuring
- Summative scale - when the item
scores are totaled (e.g., a questionnaire composed of Likert
Scales)
- Comparative scaling - a subject must
select one of two responses, called paired
comparisons
- Categorical scaling - placing items
into different categories which are then scored
differently.
- Guttman scales - items range
sequentially from a weak to strong expression of an attitude or
trait
Test Construction - Item
Writing
- Dichotomous format - offers two
possible responses to each item (True-False is most
common)
- Polytomous format - More than one
choice for each item, credit given for only one. The others items
are called "distractors." (e.g., such as multiple
choice)
Test Tryout
- Pilot the test with a small number of
subjects to determine the correct population, and number/quality
of items
Item Analysis
Two Types:
- Item difficulty (or easiness) - The
proportion of test takers getting any specific item
correct
- Item discrimination - Measures how
individual items correlate with the entire test for a group of
test takers
Test Revision
- Involves cross validation or giving
the test to another group of test-takers to see if results are
similar to the original group (remember that the validity
coefficient of the test will be smaller due to
shrinkage)
Where to Find Tests in Print
- Tests in Print - gives title, for whom the
test is designed, copywriter, author, publisher, whether or not a
manual exists
- Mental Measurements Yearbook (Buros)-
Includes reviews of each test, and comments about reliability and
validity
- Tests: A Comprehensive Reference for
Assessment in Psychology, Education, and Business - Descriptions
of tests broken down into these subheadings
Back
to the Psych of Assessment Course Outline