ABSTRACT
This chapter discusses the problem of modeling items and tests and then assembling items into one or more forms of an achievement test when relatively few test takers are available. Achievement tests are tests developed for purposes such as certification or licensure judgments, measurement of knowledge acquisition from a course of study, or decisions about whether a person has adequate prerequisite knowledge to be admitted to an educational opportunity, advance in a career development track, or be considered for a particular job. Small-scale is arbitrarily defined as pretest data obtained from 0 to 200 representative test takers. It is assumed that relatively few test takers are available to participate in the development process, either because the testing program is new or because the target population is inherently small. A testing system includes the procedures used to develop the test forms, the infrastructure used to deliver the test forms and record test-taker behavior, and the decision-making policies and procedures that are implemented to evaluate and use the test results. The testing system creates, administers, maintains, and ultimately revises or retires test forms.
