Test discrimination and test construction for cognitive diagnosisUS 8348674 B2摘要A method, system and computer-readable carrier for using a discrimination index to select test items from an item bank for a test are disclosed. At least one parameter may be identified for each of a plurality of test items in an item bank. A first test item may be selected from the item bank based on at least the parameter for the test item. Each unselected test item may be evaluated to determine whether one or more constraints would be satisfied if the test item were selected. A next test item may be selected from the unselected test items that satisfy the one or more first constraints based on at least the parameter for each test item. The Evaluation and test item selection processes may be repeated until one or more second constraints are satisfied.图片(4)说明CROSS-REFERENCE TO RELATED APPLICATIONS This patent application claims priority to, and incorporates by reference in its entirety, U.S. Provisional Patent Application No. 60/600,899, entitled “Test Discrimination and Test Construction for Cognitive Diagnosis” and filed Aug. 12, 2004BACKGROUND Standardized testing is prevalent in the United States today. Such testing is often used for higher education entrance examinations and achievement testing at the primary and secondary school levels. The prevalence of standardized testing in the United States has been further bolstered by the No Child Left Behind Act of 2001, which emphasizes nationwide test-based assessment to measure students abilities to ensure appropriate grade placement and quality of education. However, unlike measurements that are made in the physical world, such as length and weight, measuring students skills, knowledge and attributes that cannot be directly observed is challenging. Instead of measuring a particular skill, knowledge or attribute directly, the student must be measured based on a set of observable responses that are indicators of the skill, knowledge or attribute.For example, if an examiner wanted to measure extraversion, it is not obvious what tool, or questionnaire, would be effective. Even if the examiner had an appropriate questionnaire, changes in repeated measurements of an individuals extraversion could be due to changes in both the construct and the error of measurement. Classical Test Theory (CTT) and Item Response Theory (IRT) provide methods for developing instruments to measure constructs such as extraversion. In addition, CTT and IRT both provide methods for obtaining an examinees score, such as a score on a constructed extraversion scale.The typical focus of research in the field of assessment measurement and evaluation has been on methods of IRT. A goal of IRT is to optimally order examinees along a low dimensional plane (typically, a one-dimensional plane) based on the examinees responses and the characteristics of the test items. The ordering of examinees is done via a set of latent variables presupposed to measure ability. The item responses are generally considered to be conditionally independent of each other.The typical IRT application uses a test to estimate an examinees set of abilities (such as verbal ability or mathematical ability) on a continuous scale. An examinee receives a scaled score (a latent trait scaled to some easily understood metric) and/or a percentile rank. The final score (an ordering of examinees along a latent dimension) is used as the standardized measure of competency for an area-specific ability.Although achieving a partial ordering of examinees remains an important goal in some settings of educational measurement, the practicality of such methods is questionable in common testing applications. For each examinee, the process of acquiring the knowledge that each test purports to measure seems unlikely to occur via this same low dimensional approach of broadly defined general abilities. This is, at least in part, because such testing can only assess a students abilities generally, but cannot adequately determine whether a student has mastered a particular ability or not.Alternatively, estimation of an examinees “score” is not the focus in some cases. For example, a teacher may be interested in estimating students profiles. The profile for each student specifies a set of dichotomous skills, or attributes, that a student has or has not mastered. A profile of discrete attributes provides the teacher with information about the instructional needs of groups of students (unlike multidimensional IRT which provides a profile of scores). Cognitive Diagnosis Models (CDMs) can be used when the interest of a test is to estimate students profiles, or attribute mastery patterns, instead of providing a general estimate of ability.Many high stakes decisions, such as admission to a school, require that examinees be ordered along several one-dimensional scales. Dichotomous decisions (e.g., accepted or not) a