Teaching in a university in the United Arab Emirates presents a unique set of language challenges. The most obvious one is instructing in a language that is not the native language of the students. It is true that as ESL learners, Emirati students are required to have an acceptable level of English proficiency in speaking, writing, reading, and listening for university admission. It is also a fact that students have been exposed to English in their earlier schooling and in their communities. However, it becomes very obvious that this level of proficiency and exposure may be inadequate to tackle university-level studies, and many students struggle. Typically, instructors see these academic performance deficiencies primarily in test results. It is logical then for instructors to wonder whether simplifying the language used on tests could help students do better on these assessments. As faculty members at Zayed University, this was exactly what we asked.

Reading available research studies did not really help answer this question. Surprisingly, not much has been written about university students learning mathematics in a second language. There have been studies done with younger students. For example, grammar and mathematics vocabulary were found to be sources of difficulty for Irish secondary students who had been previously instructed in Gaelic and who were about to move into their tertiary level. Another example was an earlier study among Spanish-speaking fourth and tenth grade students in the United States, which showed that more items were answered correctly on a first-language mathematics test than on an equivalent English-language mathematics test.

To answer our question, we devised an experiment wherein we invited groups of students to answer two mathematics word problems. Different students, however, worked on different versions of the same word problems. A control group attempted to answer the questions in their original format – just as they appeared in the text book. Another group received versions of the same word problems couched in more complicated language, while a third group worked on versions where the language was simplified. A fourth group received an Arabic translation of the problems, and the final group answered the same word problems with supporting pictures (such as the one shown above). The idea was to compare the five different groups to see what, if any, were the differences in their scores.

To make the more complicated versions, we changed grammar and vocabulary, made sentences longer and even included elements that were not really necessary for the problem to be understood. Here’s an example:

Original version: Sara is paid commission according to the following scale:

Complicated version: Working at a large retail outlet with high-end clientele, Sara gets paid a monthly commission by her company:

As can be seen, the second, modified sentence is much longer, there are redundant words and the grammar is more complex.

For the simplified versions, we did the opposite. Taking the same example from the original, here’s the alternative that should be easier for students to understand:

The company pays Sara extra money when she sells more products.

Here the sentence is short; the passive (Sara is paid), which studies have shown is more difficult for students to understand, is replaced with an active structure (The company pays Sara); and the word commission, which might be unfamiliar to students, has been replaced with the phrase extra money when she sells more products.

After the students did the tasks, and after much pondering and analyzing on our part, we found that the results were not encouraging at all: the score of students in those different groups was generally the same (and rather low), regardless of which version they tried. This implies that regardless of language complexity, student performance was similar; it also suggests that simplifying the language of mathematics tests will not necessarily lead to better test scores.

What does this really mean then? Does it mean that language does not affect mathematics performance? Perhaps not. We suspect it means that as researchers, we were looking at the wrong thing and framing the problem in a narrow way, focusing only on the test. Maybe the test simply allows all (or most) students to display a lack of mathematical skill which has already been 'achieved' during a course that did not help them? Instead of looking at the results of tests, we plan in the future to look at how language is being used in the classroom, while students are learning the mathematics. For example, perhaps listening is more important than reading. This makes sense because lectures are listening intensive: if students don’t understand the instructor, confusion builds up.

A study such as this sometimes produces more questions than it does answers. Yet it does seem surprising that none of the modifications seemed to make any difference to the outcome. The other revealing, and worrying, finding was that almost none of the students successfully managed to answer either of the problems! Future studies may perhaps shed more light on these issues. For now, the details of our study can be accessed via this link: http://doi.org/10.18538/lthe.v14.n2.288