Comparing the Psychometric Characteristics of Speeded and Standard C-Tests

Document Type: Original Article

Authors

English Department, Islamic Azad University, Mashhad Branch, Mashhad, Iran.

Abstract

The C-test as a measure of first and second language proficiency suffers from ceiling effect. That is, the C-test is very easy for native speakers and for advanced second language learners and many examinees obtain full scores. Therefore, the test cannot distinguish among advanced learners or native speakers. Speeded C-test is suggested as a remedy for this problem. Previous research has showed that if C-test is administrated under time constraint it can better differentiate among advanced learners and native speakers. In this study we aimed to compare the psychometric characteristics of speeded and standard C-tests and investigate their relationship with reading comprehension. One-hundred Iranian undergraduate English students were selected to take part in this research. A standard (power) C-test, a speeded C-test, and a reading comprehension test were given to the students. Findings revealed that both speeded C-test and standard C-test could measure reading comprehension ability but standard C-test is a better predictor of reading comprehension ability and enjoys a higher reliability. Exploratory factor analysis of the data resulted in a one-factor solution. All C-test passages (speeded and power) and reading comprehension passages loaded heavily on a single factor. Nevertheless, power C-test passages had higher loadings compared to speeded passages.

Keywords