A scrutiny of the relationship between cognitive load and difficulty estimates of language test items

Abstract

Recently, researchers have expressed their growing concern over the scrutiny of language test items in light of cognitive load theory (CLT). While cognitive load has been central to language learning research, it has not enjoyed due attention in high-stakes language tests. The current study set out to delve into the relationship between difficulty estimates and cognitive load of language test items. To measure cognitive load, examinees’ perceived level of difficulty and response time were considered. In this regard, empirical data were collected from 60 MA students and graduates through a quantitative correlational design. The current study further employed the Rasch model to estimate difficulties of the vocabulary and grammar items of the Iranian university entrance examination (IUEE) for MA in English majors held in 2018 and 2019. The study’s findings revealed statistically significant correlations between difficulty estimates and perceived level of difficulty for vocabulary items. As for grammar items, no statistically significant correlations were detected between the variables. Whereas the results indicated strong positive correlations between response time and difficulty estimates regarding vocabulary items, no statistically significant correlations were observed between the variables concerning grammar items. All in all, perceived level of difficulty, response time, and difficulty estimates appeared to be sound indicators of cognitive load with respect to vocabulary test items, but not with regard to grammar test items. The implications of the findings will be discussed.