A test's items adequately represent the skills in its target domain. This supports
A Test-retest reliability
B Interrater reliability
C Content validity
D A percentile rank
Answer: C. Content validity. Content validity evidence concerns whether test content represents the intended domain.
A test score is compared with a relevant external outcome or criterion
A Internal consistency
B Standardization
C A raw score
D Criterion-related validity
Answer: D. Criterion-related validity. Criterion-related validity evidence links scores with a relevant criterion.
A test's score patterns fit the theory of the ability it measures
A Construct validity
B Test-retest reliability
C Face validity only
D Norming
Answer: A. Construct validity. Construct validity evidence supports the interpretation that scores reflect the intended construct.
A test predicts later course success, an example of
A Concurrent validity
B Predictive validity
C Interrater reliability
D Content sampling
Answer: B. Predictive validity. Predictive validity concerns score relationships with outcomes measured later.
A test is compared with a current measure of the same outcome
A Predictive validity
B Internal consistency
C Concurrent validity
D Test security
Answer: C. Concurrent validity. Concurrent validity evidence relates test scores to a criterion measured at roughly the same time.
A scale relates strongly to similar constructs and weakly to distinct ones
A Test-retest and interrater
B Norms and raw scores
C Screening and diagnosis
D Convergent and discriminant evidence
Answer: D. Convergent and discriminant evidence. Convergent evidence supports expected links with related measures; discriminant evidence supports separation from distinct constructs.
Consistent test procedures do not by themselves establish
A Fairness or validity
B Standardization
C Uniform instructions
D Common scoring
Answer: A. Fairness or validity. Standardization supports consistency but does not by itself show that interpretations are valid or fair.
A z-score of plus one is about
A One percentile below mean
B One standard deviation above mean
C One raw point above mean
D Two standard deviations below
Answer: B. One standard deviation above mean. A z-score expresses a score in standard deviation units relative to the mean.
An 80th percentile rank is different from
A A comparison group
B A norm score
C 80 percent correct
D A score position
Answer: C. 80 percent correct. Percentile rank is relative to a reference group and is not the percentage of items answered correctly.
The likely range around an observed score can be estimated using
A Content validity
B A projective method
C An achievement level
D Standard error of measurement
Answer: D. Standard error of measurement. The standard error of measurement estimates uncertainty around an observed score.
A reliable test may still fail to measure its intended construct, so reliability
A Does not guarantee validity
B Always proves fairness
C Means scores are biased
D Replaces norms
Answer: A. Does not guarantee validity. A measure can be consistent without supporting the intended score interpretation.
A stable score over time may reflect either consistency or
A A norm group
B Actual change in the person
C A content domain
D A criterion score
Answer: B. Actual change in the person. Test-retest differences can reflect true change as well as measurement error.
A score pattern that functions similarly across groups supports
A Social desirability
B Content sampling
C Measurement invariance
D A percentile rank
Answer: C. Measurement invariance. Measurement invariance concerns whether a measure represents the construct comparably across groups.
A test item that disadvantages a group for reasons unrelated to the target skill raises concern about
A Criterion mastery
B Test-retest consistency
C Construct coverage
D Item bias
Answer: D. Item bias. Item bias can occur when an item introduces irrelevant barriers for some groups.
A representative norm sample helps make score comparisons
A More appropriate
B Automatically diagnostic
C Culture-free
D Error-free
Answer: A. More appropriate. Norms are most useful when the reference sample is appropriate for the person and intended use.
A screening scale that exceeds a threshold usually signals
A A confirmed diagnosis
B Need for further assessment
C A treatment outcome
D A fixed personality
Answer: B. Need for further assessment. A positive screening result often indicates that further assessment may be warranted.
A broad assessment may combine tests with interviews and
A One cutoff alone
B A single label
C Collateral information
D A percentile only
Answer: C. Collateral information. Collateral information from relevant sources can contribute to a fuller assessment.
A computer adaptive test depends on an item bank with
A Unscored practice items only
B Identical item order
C No scoring rules
D Calibrated difficulty levels
Answer: D. Calibrated difficulty levels. Computer adaptive testing selects items based on response history and calibrated item characteristics.
A norm sample that no longer reflects the population may make comparisons
A Less appropriate
B More valid automatically
C Independent of age
D Free of error
Answer: A. Less appropriate. Norms may need updating when the reference population or intended use changes.
A very high internal consistency score does not prove that a measure
A Has related items
B Assesses the intended construct
C Produces consistent responses
D Can be scored
Answer: B. Assesses the intended construct. Internal consistency is one kind of reliability evidence, not proof of validity.
Two scorers use clear rubrics and practice on sample responses to improve
A Predictive validity
B Norm representativeness
C Interrater agreement
D Construct coverage
Answer: C. Interrater agreement. Training and clear rubrics can improve consistency among raters.
The same raw score can mean different things without
A A test title
B A response sheet
C A time limit
D A reference framework
Answer: D. A reference framework. Scores need an appropriate norm group or criterion to support meaningful interpretation.
A translated measure should be checked for conceptual equivalence and
A Local validity evidence
B Only spelling accuracy
C Popularity
D Identical raw scores
Answer: A. Local validity evidence. Translation should preserve meaning, and score interpretations need evidence in the target language and group.
Using one test score as the whole person risks
A Standardization
B Overinterpretation
C Interrater scoring
D Norm referencing
Answer: B. Overinterpretation. A score is one source of information and should not be treated as a complete description of a person.
A test score that supports one use may not support
A Any comparison
B A raw score
C Every other use
D A standard procedure
Answer: C. Every other use. Validity evidence is tied to interpretations and uses rather than being universal for every purpose.