- Standardized tests are developed by psychometricians and content experts and follow a set of pre-determined guidelines for content, administration and scoring. Defining the purpose of the test is the first step in development and guides the other steps.
- There are two types of standardized tests. One type can be categorized by the type of test, such as achievement or aptitude. The other can be categorized by the way the scores from the test are interpreted, i.e. criterion- or norm-referenced.
- After the purpose is defined, the next step is to decide on the appropriate content Domain for the test. A test serves as a representative sample of the total content domain.
- If the test measures a Construct, the Construct must be operationally defined.
- Test specifications provide an outline of how the test should be constructed from the content domains and defines the test format, item format, scoring rules or Rubric, and method of score interpretation.
- Test specifications and blueprint serve as a guide to the people writing the specific test items.
- Following a blueprint, items are developed. After an initial pool of items is written, the items are reviewed for accuracy, clarity, appropriateness, and sensitivity. Those items that do not meet these standards are either revised or removed from the test.
- The items then go through a preliminary item tryout and progress to larger scale field tests with the target population.
- After the items have been tested, they undergo item analysis, which involves analyzing each item´s: Difficulty level, Discrimination index, and Differential Item Functioning (DIF).
- After analyzing the reliability and Validity of the test (discussed in separate lessons), the final step in test development is constructing administration guidelines outlining exactly how and to whom the test should be administered. This ensures standardized administration.
- Equating test scores allows teachers to make direct comparisons between different tests, different versions of the same test, or tests across different grade levels.
- Values used in the equating process are determined by giving the tests to a sample of test takers representative of the population and comparing their performances on the different tests.
- Scores can be adjusted a number of different ways. Sometimes a constant number is just added to one form´s score, other times the scores are manipulated so they have a specific desired mean and standard deviation.
- Equating can be used to examine changes in scores of the same group over time. This type of equating is known as Horizontal Equating.
- Equating can be used to examine changes in scores between groups at different grade levels. This is an example of vertical equating.
|
