Objective testing of educational achievement.

Abstract

Educational achievement tests can be classified in many ways. Perhaps the most global classification is the distinction between standardized and teachermade tests. This distinction clarifies the degree to which content, format, administration, and scoring are standardized or left to the discretion of the teacher or testing agent (see Volume 3, Chapter 16, this handbook). Another distinction in educational tests is the format of the test items. More commonly used tests are classified as objectively or subjectively scored. The extent to which tests are objective or subjective is usually a function of the type of scoring required given the item or task type. These terms are components of a continuum rather than a dichotomous choice. Although some items might be considered strictly objectively scored because a correct response is known and responses may be machine scored (or scored with a key), other items may be subjectively scored because they require human scoring or a complex scoring guide such as a rubric, with which many correct (or partially correct) responses are possible. Even so, many forms of subjectively scored items, such as short-answer constructed-response (CR) items or even essays, may be scored with simple to complex computer scoring engines. Perhaps most performance assessments, with complex multidimensional aspects, are most unlikely to be objectively scored (see Chapter 20, this volume).When focused more specifically on the nature of the test items or tasks themselves, additional classification schemes are possible. The two categories used in this chapter are selected-response (SR) and CR items. For SR items, options are available from which to select responses, whereas for CR items, the respondent must construct or produce a response because none are available for selection. Similarly, items may be considered fixed-response items (SR) versus free-response or production items (CR). In practice, most educational test developers use the terms multiple choice (selected response) and constructed response because these formats are predominant in current educational tests. However, multiple choice as a label refers to a specific class of selection items, and so it is not all encompassing. Within the two classifications of SR items and CR items, there are many possible formats. In the class of CR items are formats that could be objectively scored, including the cloze, fill-in-the-blank, grid-in response, and short-answer formats and even some experiments or tasks that result in specific products. More subjectively scored production items include demonstrations, debates, essays, exhibitions, interviews, observations, and more complex performance tasks, portfolios, projects, and research papers (see Haladyna & Rodriguez, in press, for a comprehensive review of these types).The focus of this chapter is on objective testing of educational achievement and thus items that lend themselves to objective scoring (see Chapter 19, this volume, for additional perspectives). This category includes a wide range of SR i...