Equating

Basic Concept of Equating

The goal of equating is to find an equating relationship that transforms the scores on Form X (i.e., a new form) to the scale of Form Y (i.e., an old or base form). It is assumed here that some process has been used to establish a raw-to-scale score transformation for the base form Y.

Typically, the first step in equating is to determine a transformation function based on raw scores (e.g., number-correct scores) for both forms. Then, the equated raw scores on Form X are converted to scale scores on Form Y. These scale scores are the scores reported to examinees. When this process is performed successfully, the reported scale scores have the same meaning regardless of which form was administered. For example, a scale score of 500 on Form X indicates the same level of performance as a scale score of 500 on Form Y.

Equating Designs

For equating to be successful, differences attributable to forms must be separated from differences in the examinee groups taking the forms. Accomplishing this requires using a data collection design in which the sets of data for Form X and Form Y have some link between them—either common items or common (or similar) persons. The three most commonly used designs are the random groups design, the single group design, and the common-item nonequivalent groups (CINEG) design, sometimes called the nonequivalent anchor test design.