ABSTRACT

In the wake of Covid-19, language testing finds itself rather suddenly in an era where computer- delivered and at-home test administration is commonplace rather than a peculiarity (Isbell & Kremmel, 2020). It is worth noting that at-home language testing pre-dates Covid-19, and in that sense Covid is best seen as accelerating a trend. More importantly, it would seem there’s no
going back now: the widespread adoption, commercial success, and test-taker benefits of at-home tests may have fundamentally altered the language testing landscape.

Matters of construct validity for high-stakes, computer-delivered language tests have been extensively investigated, with basic concerns like typing skill and computer familiarity accounted for and issues related to computerized delivery and scoring of test tasks increasingly well understood. Similarly, with respect to at-home administration, large-scale comparisons of score distributions across test-centre and at-home language test administrations have already been hashed out by psychometricians (e.g., Zumbo, 2021).

 

We are less focused on and less prepared to grapple with issues of fairness and justice in at-home language testing. While we have good frameworks for thinking about fairness and justice in language testing (e.g., Kunnan, 2018), we must now contend with variations and possibilities in high-stakes testing that have only recently become ubiquitous enough to warrant broad concern. In this talk, I highlight what I see as key, (re)emerging concerns for fairness and justice of high- stakes language tests in the remote age, including construct and decision comparability, proctoring and privacy, public health, internet and communications infrastructure, and geographic/temporal/financial access to testing services.