arnold on Oct 05, 2008
With Domain Sampling, tests are constructed by randomly selecting a specified number of measures from a homogeneous, infinitely large pool. A sample of items is reliable to the extent that the score it produces correlates highly with these true scores (Nunnally and Bernstein, 1994, p. 217).
This is easier to understand if you look at it in terms of a situation by a test maker.
One situation: Norm Referenced: Psychological constructs are of importance because of the degree to which individuals differ on certain attributes. If you want to develop a test that differentiates you have to conceptualize the behavior components of the construct, produce the items, and select the items on which the expected degree of variation occurs (eliminating items with little variation).
In another situation: Criterion referenced: You need to measure an examinee’s performance in terms of absolute level of proficiency (achievement test). The developer begins with a set of instructional objectives and then proceeds to define a domain of performance to which inferences will be made. This is called the item domain or the domain sampling approach to test construction. An item domain is a well-defined population of items from which one or more test forms may be constructed by selection of a sample of items from this population.
Instead of creating all possible items, you can 1) produce a set of item-domain specifications that are so structured that item which are written would be interchangeable, 2) item specifications, and item forms or item algorithms.
Crocker and Algina, 1986
Nunnally and Bernstein, 1994