Examining the relationship between computer familiarity and performance on computer-based language tasks

The planned introduction of a computer-based Test of English as a Foreign Language (TOEFL) test raises concerns that language proficiency will be confounded with computer proficiency, introducing construct-irrelevant variance to the measurement of examinees' English-language abilities. We administered a questionnaire focusing on examinees' computer familiarity to 90,000 TOEFL test takers. A group of 1,200 "low-computer-familiar" and "high-computer-familiar" examinees from 12 international sites worked through a computer tutorial and a set of 60 computer-based TOEFL test items. We found no meaningful relationship between level of computer familiarity and level of performance on the computerized language tasks after controlling for English language ability. We concluded that no evidence exists of an adverse relationship between computer familiarity and computer-based TOEFL test performance due to lack of prior computer experience.