ABSTRACT
A performance test is an examination designed to elicit performance on an actual or simulated real-life task. In contrast to observation of naturally occurring behavior “in vivo” (McGaghie, Butter, & Kaye, Chapter 8, this volume), the task is contrived for the purpose of the examination, and explicitly invites the examinee to demonstrate the behavior to be assessed. Thus a performance test is an “in vitro” assessment, at Miller’s “shows how” level (Miller, 1990); see Figure 9.1. Since the examinees know they are being assessed, their performance likely represents their personal best or maximum performance, rather than a typical performance. Examples of performance tests include a road test to obtain a driver’s license, an undersea diving test, and the Unites States Medical Licensing Exam (USMLE) Step 2 Clinical Skills Assessment. In this chapter we’ll review some of the purposes, advantages and limitations of performance tests, and provide practical guidelines for the use of standardized patients (SPs), a simulation modality commonly used in health professions education. Focusing on the use of SPs for assessment rather than instruction, we’ll discuss scoring options, multiple-station Objective Structured Clinical Exams (OSCEs), standard setting, and threats to validity in the context of SP exams; the same principles apply to performance tests using other modalities. There are several other types of simulations currently in use, such as bench models, virtual (computer-based) models, and mannequins. Many of the assessment issues addressed here also apply to these forms of simulation, which are discussed in greater detail in Chapter 10.
