Writing tests in a multilingual society

Like it or not, the United States is already a multilingual society. According to the Pew Research Center, in 1980 the U.S. Hispanic population was 14.8 million people and made up 6.5% of the total U.S. population. By 2014, that population had exploded to 55.3 million and made up 17.3% of the total U.S. population, and Spanish speakers are only one of the numerous populations for whom English is not a first language and who are growing at fantastic rates. The Los Angeles Unified School District estimates that 92 different languages are spoken in their schools.

This growth has led to an increased demand from U.S. schools and districts for tools and services that serve the specific needs of English-Language Learners (ELLs). In a recent webinar, EdWeek reported survey results that showed an estimated 35 percent of school districts are interested in purchasing software for ELLs.

Products and services that serve non-native English speakers well may be different from their English language counterpoints in profound ways. Language structures the way we see, think about, and interact with the world around us. It would be an injustice to assume that, for example, standardized test items that were developed by and for native English speakers will be valid and reliable indicators of ELL learning without close attention to the latter’s particular needs.

How can educators write and develop curriculum, instruction, and evaluation in English that nonetheless makes sense to non-native English speakers and gives them a fair change at success? Matthew Kushinka has spent over 12 years in the language services industry and is the owner of RedLine Language Services, which since 2011 has provided translation, editing, and related services in 20 languages on over 500 projects to approximately 100 clients. Kushinka was kind enough to address these issues when I discussed them with him over email recently.

Even when translation is not an issue, the construction of individual test items is a challenge. Kushinka recommends that items be short, clearly worded with only one possible interpretation, and culturally appropriate. “The following passage is a conversation between two classmates" is much more accessible for young people than something like "the following email exchange is between a supervisor and her direct report. The supervisor has just learned that her report has been belligerent with a client." The latter question is too long and contains terms like “direct report” that may may be difficult for young people to understand. Items for English Language Learners (ELLs) should avoid overuse of cognates and contain a degree of information redundancy for students whose English language comprehension is very low.

Kushinka also reminds us that we can convey complex ideas without complex language. Item writers should constantly ask themselves if they can simplify the word choice and structure of an item. Of course, sometimes using difficult terminology and/or complex processes is unavoidable. “If you’re writing a passage on the advent of the automobile, you might not be able to get around using the term internal combustion engine.” In such cases, Kushinka recommends explication, examples, illustrations/graphics, analogous statements, etc.

Bryan Shellytesting