H. Sepasi
Abstract
In order to find a reasonable solution to measurement problems in education and psychology, new test theories have recently been proposed by test specialists. The purpose of this paper is to survey the theoretical concepts and assumptions underlying the two very popular psychological and educational ...
Read More
In order to find a reasonable solution to measurement problems in education and psychology, new test theories have recently been proposed by test specialists. The purpose of this paper is to survey the theoretical concepts and assumptions underlying the two very popular psychological and educational test theories. The classical test theory (CTT) first proposed by Spearman, the English psychologist, in early twentieth century, also called today the old theory, is based originally on three basic concepts: observed score, true score and error score. Most of the studies conducted in the field of education and psychology have used CTT in order to interpret the reliability and validity of their research data. Due to the limitations that were found in old theory (CTT), a new test theory, the item response theory (IRT), has been proposed by modern test specialists. On the basis of the assumptions of IRT, many logical interpretations could be made in solving measurement problems in constructing educational and psychological tests. In this paper, attempts have also been made to compare the basic assumptions of these two theories. A further analysis is made to clarify the similarities and differences between item parameters used by these two test theories. In addition, advantages and limitations of each theory that test constructors believe should be considered in item writing are discussed. Finally, based on the findings of this paper, it is suggested that graduate students and faculty members who are interested in measurement and evaluation of student learning or in doing research in education or psychology, should he familiar with the basics of the old theory (1T) and the new one (IRT).
H. Sepasi
Abstract
Abstract Teachers have long been concerned with the contributions of’ different types of evaluation to their students achievement. Many believe that the use of multiple- choice items provide greater students achievement and promotes more positive attitudes toward learning. On the other hand, ...
Read More
Abstract Teachers have long been concerned with the contributions of’ different types of evaluation to their students achievement. Many believe that the use of multiple- choice items provide greater students achievement and promotes more positive attitudes toward learning. On the other hand, psychometricians like Ebe] and Frisbie, believe that there is no significant difference between multiple-choice and true-false test items. This study was designed to test the hypothesis that the use of true-false test is as effective as multple-ehoice test in mesauring students’ achievement- A total of 304 university students enrolled in the course, were randomly divided into two groups. One group participitated in a multiple-choice test whereas the other group received true-false items. Instructional strategies including objectives of the course, tests contents, and testing procedures were the same for both groups. Then, in order to find out whether the two groups were similar or different in term of their achievements, the correct response in the two tests were compared. The results showed no significant difference between the final scores of the two groups. This finding indicates that teachers should give attention to the contribution of both multiple- choice and true- false test items in measuring their students achievement.
H. Sepasi
Abstract
In this article attempt has been made to introduce the new test theory, i. e. item response theory, (IRT). The major reason for the development of this new theory was the criticism raised by new psychometrician against old one, i. e. The classical test theory (CTT). Traces of this theory can be found ...
Read More
In this article attempt has been made to introduce the new test theory, i. e. item response theory, (IRT). The major reason for the development of this new theory was the criticism raised by new psychometrician against old one, i. e. The classical test theory (CTT). Traces of this theory can be found in the work of Lowely. In recent years, other test specialists like Lord, and Brinbuam have played a major role in introducing the IRT. In this theory the total test score does not appear to play a very significant role in interperting examinee performance, rather, the test items are studied individually. The theory uses item characteristics curve parameters such as difficulty and discrimination indices to estimate the probability of examinee’s correct response to each test item. This theory is based mainly on two basic assumptions: local independence and unidimensionality of latent trait space. Based on these two assumptions, the probability of correct answers to every test item was first investigated by cummulative normal ogive curve and then by logistic curve for its simpilicity to work with. Lord introduced normal ogive curve, then it was Brinbaum who replaced the normal ogive by logistic curve to facilitate computation. He also designed one-parameter model, (difficulty parameter), two-parameter model (difficulty and discrimination), and three parameter model (difficulty, discrimination, and, guess), all of which can be used in constructing psychological and educational tests.