Exam Structure

The exam development process is presided over by the test development committee. Exams are assembled to conform to the test specifications determined by the committee as a result of their review of the curriculum survey. These specifications dictate the percentages of items representing various aspects of the subject.

For example, for the Introductory Sociology exam, the test specifications require the questions to reflect the following topics in the percentages indicated:

20% Institutions

  • Economic
  • Educational
  • Family
  • Medical
  • Political
  • Religious

15% Social patterns

  • Community
  • Demography
  • Human ecology
  • Rural and urban patterns

20% Social processes

  • Collective behavior and social movements
  • Culture
  • Deviance and social control
  • Groups and organizations
  • Social change
  • Social interaction
  • Socialization

30% Social stratification (process and structure)

  • Aging
  • Power and social inequality
  • Professions and occupations
  • Race and ethnic relations
  • Sex and gender roles
  • Social class
  • Social mobility

15% The sociological perspective

  • History of sociology
  • Methods
  • Sociological theory

Each exam must also include items that represent a range of difficulty levels. This range allows psychometricians and other testing specialists to ensure that the test discriminates well among the different levels of content mastery among test-takers.

Testlet Structure of CLEP Exams

When the CLEP program moved to computer-based administration of exams (CBT), one objective was to increase test security by varying the content of the exams to a greater extent. CLEP exams are assembled using the testlet method. A testlet is a collection of questions organized around a content area or methodology within the larger scope of the exam.

For example, the Biology exam is composed of five testlet sections: A, B, C, D and E, each of which covers a portion of the content specifications (Molecular and Cellular Biology, Evolution/Diversity, Organisms, Genetics/Molecular Genetics, and Ecology/Social Biology) for the test as a whole. For each content area, the test development committee creates several testlets that are parallel in terms of content and statistical properties.

For testlet section A, which focuses on Molecular and Cellular Biology, the committee creates several testlets — A1, A2, A3, and so forth — each containing different questions, but all related to this area of biology. The same is done for testlet B, which focuses on Evolution/Diversity, and for the other content areas covered by C, D, and E. During an exam administration, the computer software selects at random any A-B-C-D-E sequence of testlets, without impacting the content, level of difficulty, or statistical validity. Each test-taker sees one A testlet, one B testlet, and so forth, but whether they see A1 or A2, B1 or B2 is determined only when the test is administered.

To clarify: suppose there were three testlets for each of three content areas (A1, A2, A3, B1, B2, B3, C1, C2, and C3). While in this scenario there are only nine testlets, there can be 27 distinct variants of the exam, where any pair of the 27 exams has at least one testlet not in common, as shown below:


Because the individual exams are assembled from parallel testlets, the resulting exams are equivalent. In addition, test security is enhanced. For example, if 27 candidates test simultaneously at the same testing center, each of them will likely see a different form of the same exam. As more items are added to the question pool, the number of individual testlets and possible unique forms increase. If a test has three versions of each of four testlets, 81 unique forms can be generated; if a test has three versions each of five testlets, 243 unique forms are possible.

Advantages of the testlet approach:

  • Multiple exams can be assembled and administered while providing a reasonable degree of control over content and statistical specifications.
  • The test administration software rotates testlets so that each variation of the exam is seen by equal numbers of test-takers across all test centers. This ensures a consistent statistical analysis.
  • Because the assembly of testlets follows strict statistical and test content requirements, comparability of testlets with one another is predictable, controlled and reliable.