Psychological Assessment & Measurement Techniques Flashcards

Which type of validity evidence examines whether a test's internal structure matches its theoretical constructs?

Answer: Internal structure validity

Internal structure validity evidence (part of construct validity) examines whether item relationships and factor structure align with the theoretical model.

A selection test with a 0.40 criterion-related validity coefficient is used with a base rate of 50% and a selection ratio of 20%. Using the Taylor-Russell tables, what effect does lowering the selection ratio have on the success ratio?

Answer: It increases the success ratio

Lower selection ratios (hiring fewer, more selective) increase the success ratio because you are selecting only top scorers from a larger pool.

The Angoff method, Bookmark method, and Contrasting Groups method are all approaches to:

Answer: Setting passing scores (cut scores)

The Angoff, Bookmark, and Contrasting Groups methods are standard-setting procedures used to establish defensible cut scores on employment or credentialing tests.

Which phenomenon occurs when candidates' scores on a selection instrument improve substantially upon retesting, threatening the validity of the instrument?

Answer: Coaching and practice effects

Coaching and practice effects occur when retesting leads to score gains unrelated to the underlying construct, threatening test validity for repeat test-takers.

In Item Response Theory (IRT), the 'b' parameter represents:

Answer: Item difficulty (location)

In IRT, the b parameter is the item difficulty or location parameter, indicating the trait level at which the probability of a correct response is 0.50 (in 1PL/2PL models).

A company uses a cognitive ability test that shows adverse impact against a protected group. Under the Uniform Guidelines, what is the FIRST recommended action?

Answer: Demonstrate the test's job-relatedness and business necessity

Under the Uniform Guidelines, when adverse impact exists the employer must demonstrate job-relatedness and validity evidence to justify continued use of the selection procedure.

Which reliability coefficient is most appropriate when a test measures a single, unidimensional construct and all items share equal variance?

Answer: KR-21

KR-21 is appropriate when items measure a single construct and all items are assumed to have equal difficulty (equal variance), making it simpler to calculate than KR-20.