PSYCHOLOGICAL ASSESSMENT I
A. STEPS IN TEST CONSTRUCTION
- Test construction has become a highly technical matter, and the test procedure varies for different types.
I. Specification of the purpose of the test and the target test population.
- This begins the whole process.
- These test parameters can be described in terms of:
a. The specific area in which assessment is to be done
b. The characteristics of the attribute or some theoretical conception of the psychological nature of the
trait that the test aims to measure
c. The criterion that the test scores are to be related with for its validation.
II. Item writing
- This is done by experts in the various areas to be tested.
- Items are composed to cover a wide range of difficulty, from quite easy to very difficult items, for the intended
population.
- Many more items should be created than will be used in the final test.
III. Item Editing
- This stage is done by several persons, to increase the likelihood of spotting formal defects in the items.
- Each item is checked for clarity of wording, appropriateness of vocabulary level for the intended population,
stylistic equivalence of multiple choice distracters and the correct answer, and the “face validity” of the items.
o FACE VALIDITY
- Property of an item that gives it the appearance of measuring what the test as a whole is supposed
to measure.
- It may or may not be related to the actual validity of the item of the test, but it can affect the
“reasonable” “fairness” and acceptability of the test in the eyes of those who are taking it.
- In item editing, questionable, weak, or defective items are either revamped or discarded.
- Unless the remaining pool of acceptable items still contains a much larger number than will be needed for the
final test, more items are composed and subjected to editing until the required number is obtained.
IV. Item Try Out
- It consists of administering the entire pool of items, presented in the format of an actual test, to a large,
representative sample from the population for which the test is intended.
- The sole purpose of this try out is to obtain enough data for the next stage in the process from which the test is
intended.
V. Item Analysis
- This stage presents the most technical aspect of the whole process, involving a number of psychometric and
statistical methods.
- The essential information provided by an item analysis are the following:
a. The difficulty level (percentage passing) of an item
b. The discriminability of each item that is how clearly it differentiates (in terms of percentage passing
the item) the highest from the lowest scorers on the test as a whole, or how highly each item correlates
with the total score in the test.
c. Analysis of incorrect responses to determine, for example, which multiple - choice alternatives are so
rarely chosen as to be virtually non - functional.
VI. Standardization
- The distribution of raw scores (number correct) in a large representative sample of the target population is
converted to some meaningful, interpretable scale such as percentiles, IQs, or other forms of standardized
scores.
A. STEPS IN TEST CONSTRUCTION
- Test construction has become a highly technical matter, and the test procedure varies for different types.
I. Specification of the purpose of the test and the target test population.
- This begins the whole process.
- These test parameters can be described in terms of:
a. The specific area in which assessment is to be done
b. The characteristics of the attribute or some theoretical conception of the psychological nature of the
trait that the test aims to measure
c. The criterion that the test scores are to be related with for its validation.
II. Item writing
- This is done by experts in the various areas to be tested.
- Items are composed to cover a wide range of difficulty, from quite easy to very difficult items, for the intended
population.
- Many more items should be created than will be used in the final test.
III. Item Editing
- This stage is done by several persons, to increase the likelihood of spotting formal defects in the items.
- Each item is checked for clarity of wording, appropriateness of vocabulary level for the intended population,
stylistic equivalence of multiple choice distracters and the correct answer, and the “face validity” of the items.
o FACE VALIDITY
- Property of an item that gives it the appearance of measuring what the test as a whole is supposed
to measure.
- It may or may not be related to the actual validity of the item of the test, but it can affect the
“reasonable” “fairness” and acceptability of the test in the eyes of those who are taking it.
- In item editing, questionable, weak, or defective items are either revamped or discarded.
- Unless the remaining pool of acceptable items still contains a much larger number than will be needed for the
final test, more items are composed and subjected to editing until the required number is obtained.
IV. Item Try Out
- It consists of administering the entire pool of items, presented in the format of an actual test, to a large,
representative sample from the population for which the test is intended.
- The sole purpose of this try out is to obtain enough data for the next stage in the process from which the test is
intended.
V. Item Analysis
- This stage presents the most technical aspect of the whole process, involving a number of psychometric and
statistical methods.
- The essential information provided by an item analysis are the following:
a. The difficulty level (percentage passing) of an item
b. The discriminability of each item that is how clearly it differentiates (in terms of percentage passing
the item) the highest from the lowest scorers on the test as a whole, or how highly each item correlates
with the total score in the test.
c. Analysis of incorrect responses to determine, for example, which multiple - choice alternatives are so
rarely chosen as to be virtually non - functional.
VI. Standardization
- The distribution of raw scores (number correct) in a large representative sample of the target population is
converted to some meaningful, interpretable scale such as percentiles, IQs, or other forms of standardized
scores.
