Next Generation TOEFL®: New test, new prep

Page No.: 
Bruce Rogers

It's finally coming.

It will be here in September of 2005. And it's going to create some sizeable waves in English language classrooms all over the world. It's taken awhile, and there have been a few delays.

It's also gone by a number of names. At first the project was called TOEFL 2000. (Cynics started calling it TOEFL 2000ish.) It was once known as the IBT (Internet-based TOEFL). Now, perhaps taking a cue from the famous television series, Star Trek: The Next Generation, it is called the Next Generation TOEFL (the NGT).

The goal of the NGT is to test "academic competence": that is, the ability to understand and communicate effectively in the university environment. The new test is based on years of research and development. ETS did its homework and analyzed a wide variety of authentic college textbooks and academic discourse. ETS researchers showed their patience by attending lectures that they didn't even need to attend in order to get a grade. They visited classrooms and cafeterias all over North America to learn how professors talk and how students talk to professors and to each other.

One of the most pointed criticisms of the current computer-based exam (the CBT) is that it tests skills unrelated to real academic tasks performed by international students at North American universities. For example, finding the error in an out-of-context sentence in the current Structure section, critics pointed out, is irrelevant to anything one might be called on to do at Ohio State, Stanford, or Old Miss. Now, one might argue that writing an essay on the NGT about whether it is better to live in a dorm or in off-campus housing is not a very common academic task either. Still, there is no doubt that the new form of the test poses far more authentic tasks than the current one.

For many years, the TOEFL was divided into three parts. Sections 1 and 3 tested the "passive" language skills: listening and reading comprehension. Section 2 measured test-takers' knowledge of English grammar and usage. Later, the TOEFL began to test the productive skill of writing. The Test of Writing English (TWE) was added, first as an optional section and then as a mandatory part of the test. In the beginning, the writing grade was reported separately. However, on the current CBT, the weight of the essay writing section increased. It now counts for half the score in Section 2.

On the NGT, the grammar section will be history. Grammar will be tested only indirectly, mainly through the writing and speaking tasks. This change has some people cheering, others jeering. Students, at least those with strong backgrounds in structure, find that solving these little grammar puzzles is fun and as addictive as eating popcorn. Many TOEFL Prep teachers like this section because it is the most coachable. With a few weeks of focused practice, students often see a measurable gain in their structure score, and a good score on structure pulls up their overall score. In addition, grammar teachers like this section because it can be used to motivate, or even intimidate, students who are struggling with a difficult structure. ("You have to study past participles because they'll be on the TOEFL.") TOEFL Prep material writers (including myself) like Section 2 because it is, for most of us, the easiest section to write for. On the CBT, Section 2 counts for only 1/6 of the total score, but in most TOEFL Prep books, more ink is dedicated to the structure section than to any other section.

While the Structure section will soon vanish, there will be a Speaking section on the Next Generation test. For test-takers whose comfort zone does not extend beyond multiple-choice questions, the speaking tasks will be the most daunting challenge of the test. However, for many in the ESL community, a speaking score will provide the missing piece of the assessment puzzle. Few in our profession have not met or at least heard of students who have a score of 250 or higher on the computer-based exam but who struggle to make themselves understood.

The new test goes beyond testing skills in isolation. It also contains "integrated" tasks that require test-takers to use a combination of skills.

Design of the Next Generation TOEFL

The NGT will be an amalgam of familiar and new.

Overall, it will take about four hours (about the same as the current test). It will be delivered on computers via the Internet but will not be computer adaptive. In other words, a right answer or a wrong answer will not affect the level of difficulty of the next item. The test will no longer be given on demand; there will be prescribed dates and times for taking the test, as there were in the past for the paper-based test.

The test will be divided into four sections.

Reading: This section is only a little different from the current reading section. It consists of three readings and questions about them. Although there are fewer readings than on the CBT, the readings on the Next Generation test are considerably longer. The readings are based on material from textbooks. There are familiar detail, main idea, inference, and vocabulary questions. Also included are the "black-box" questions introduced in the computer-based version that ask students to add a sentence to an existing passage. A new item-type requires test-takers to drag answer choices into an incomplete outline. One nice feature of the new test is the glossing of difficult content vocabulary. By clicking on certain words highlighted in the text, test-takers get an explanation via pop-up windows. Another welcome feature is a review section that allows test-takers to see which items they did not answer and enables them to return to these items quickly.

Listening: This section consists of two longer conversations in campus settings and four classroom lectures. The language of the lectures and conversations is more "natural" than it was in previous versions of the test. There are false starts, back-ups, and lots of ums, uhs, you knows, and now let's sees. Personally, I appreciate ETS's efforts to make speech on the test sound more natural, but frankly, I find that hearing these kinds of fillers recorded somehow magnifies them. What sounds natural in real life sounds unnatural, at least in the samples ETS has so far released. And presumably, a large part of a typical native speaker's lexicon—the "colorful" language, references to taboo topics, and endless likes—will have to be edited out. After hearing portions of the conversations and lectures, test takers answer questions about the spoken material. Most are familiar multiple-choice items about details and main ideas. One new item type asks test-takers to put Yes/No marks next to a list of concepts to indicate whether the concept is or is not part of a logically related group. Another significant change from the current test is that test takers are encouraged—not forbidden—to take notes.

Writing: There are two tasks in this section: an "independent" writing task and an integrated task. The independent task asks test-takers to draw on their own experience and knowledge to respond to a rather generic and non-controversial prompt. ("Some people say that the most important feature of a university is its library. Do you agree with this opinion?") This is very similar to the current Essay section on the CBT. The integrated task involves first reading a short passage, then hearing a lecture on a related topic, and finally writing about what was read and heard. Part of the task is to explain the relationship between the reading and the listening.

Speaking: As in writing, speaking tasks can be classified as independent or integrated. There are two independent tasks. Test takers see prompts similar to the one in the writing section. They then have 15 seconds to prepare a response and 45 seconds to respond. The integrated tasks are also similar to those in the writing section. Test-takers listen to a lecture, and sometimes read a related passage. They then have 30 seconds to prepare a response to a question based on the lecture and reading and 60 seconds in which to deliver that response.

Washback from the Next Generation TOEFL

Assessment experts often refer to the "washback effect." This is the way in which classroom teaching is changed by the nature of a test. A quick stroll through the ELT section of a bookstore—especially in East Asia—shows that the washback from the current forms of the TOEFL (and the TOEIC) has reached flood proportions. There are probably more books dedicated to test preparation than to any other aspect of language learning. Given the effect that these high-stakes tests have on people's lives, this is not surprising.

How will changes in the test change the washback?

For one thing, test preparation courses and materials will probably change their focus from micro to macro aspects of language. Gone are the short conversations from Listening and the single, uncontextualized sentences from Structure. (The short vocabulary sentences vanished from Reading some years ago.) Test prep will no longer focus on memorizing individual vocabulary words and idioms and mastering unrelated grammar points. It will have to focus on understanding and producing larger chunks of language.

For another thing, the line between TOEFL Prep and general language learning will blur. A TOEFL prep classes can teach general English and a general class will help students get ready for the test. Conversation classes and classes that teach note-taking skills will suddenly provide a fast track to better TOEFL scores. No longer will students be able to say, "I didn't come to English class today because I had to study for the TOEFL." Classes that teach TOEFL tactics in a language other than English will become largely counter-productive. As the emphasis shifts from learning grammar rules to building fluency, classes will center more on the students and less on the teacher.

Certainly there will still be classes where students take practice tests that duplicate the new test. However, there will also be room for activities that are communicative, interesting, and fun. There is no doubt that this is what ETS intended by changing the test. And while it may take awhile for students, teachers, and material writers to get used to it, I believe that the changes in this test will have a tremendously positive effect on the way people learn English.

Bruce Rogers is the author of The Complete Guide to the TOEFL: CBT Edition (Heinle & Heinle) and The Complete Guide to TOEIC (Thomson Asia). He taught at the Economics Institute at the University of Colorado for twenty years, as well as in Indonesia, Vietnam, Korea, and the Czech Republic. He is the immediate past president of Colorado TESOL. He is currently at work on a new edition of The Complete Guide for the Next Generation TOEFL.