LANGUAGEED
Issue 17 Mar / Apr 2009

Keeping Score: How a Corpus Can Help Improve Teaching

Which words do Math teachers use most? Do Science teachers have a different style of teaching? What kind of talk goes on in the typical Singapore classroom? Researchers have been able to find answers to these sorts of questions by using a corpus of classroom data.

Article highlights
  • What is a corpus?
  • What does a corpus of classroom language tell us?
  • How can you create your own corpus?

A corpus is not just simply a collection of texts. Different corpora may contain different types of content, depending on their purpose. A corpus can help us analyse how language is used by different people in different contexts.

The SCoRE So Far

Researchers at the National Institute of Education have been building up a Singapore Corpus of Research in Education, or SCoRE in short, since 2005. It is a collection of real classroom dialogue from over 450 lessons in over 120 Primary 5 and Secondary 3 classrooms.

SCoRE comprises, literally, scores and scores of transcribed data from English Language, Mathematics, Science and Social Studies classrooms.

It is now a rich source of empirical data for education researchers, linguists, teacher trainers and curriculum designers who want to understand the interactions of students and teachers in Singapore classrooms.

What Does SCoRE Tell Us?

A corpus can show us how language is really used. An analysis of the data provides rich evidence of the pedagogy and language patterns that are typical of the classroom environment in Singapore.

For example, the SCoRE researchers have made discoveries about the following:

Frequently used words

  • Math teachers frequently used words for numbers, concepts (like “triangle”) and operators (like “minus” and “divide”). They also used the word “must” far more than other teachers!
  • Science teachers used a lot of subject-specific words (e.g., “cell”, “photosynthesis”).
  • Social Studies teachers tended to use words that relate to key social issues (e.g., “water”, “Singapore”, “Malaysia”).
  • However, English Language teachers had a more diffuse core set of keywords. And, interestingly, there were more content words (e.g., “dinosaur”) than subject-specific words (e.g., “adjective”, “clause”, “verb”).

Common phrases

By looking at “lexical bundles” or fixed phrases, we know that teachers have characteristic ways of getting things done in the classroom through the use of language:

  • Knowledge building, which often takes the form of a question (e.g., “How do you make…?”)
  • Classroom management, which is used to manage interpersonal relationships (e.g., “I want you to…”)
  • Discourse facilitation (e.g., “Okay, now, I’m going…”)

Discourse patterns

Most lessons have a basic “flow”. The most common is a three-part pattern known as the IRF: teacher initiation (I), student response (R), and teacher follow-up (F). A very high proportion of Singapore classroom talk was found to follow the IRF pattern, regardless of subject, level or stream, but they differ in the types of initiations, responses and follow-ups commonly used.

By combining these analyses, and others, we can explore the characteristic patterns of teaching.

Create Your Own Corpus

A corpus can make a powerful teaching resource. With current technology, it isn’t too difficult to create one of your own. Here’s a simple recipe for creating your own corpus:

  1. Collect a large sample of your students’ work
    To make this step as easy as possible, get your students to submit their essays to you electronically (e.g., as an MS Word document). For best results, save the documents as plain text (.txt) files with an indexical file name (e.g., “student_text_001”, “student_text_002”, etc.).
  2. Process the text
    You can use a concordance program to produce a key-word-in-context (KWIC) concordance for a specific word. Free concordance software is easily available on the Web (e.g., Simple Concordance Program, ConcApp Concordancer, KWIC Concordance for Windows).
  3. Design curriculum tasks
    Now you have data with which you can design authentic tasks and assessments. For example, you can create a simple multiple context cloze task (fill in the blanks) using your concordance search results. The use of the students’ own work for such an exercise will also help to raise students’ consciousness of the ways they are using grammar in their own texts.

References
Biber, D., Conrad, S., & Reppen, R. (1998). Corpus linguistics: Investigating language structure and use. Cambridge, UK: Cambridge University Press.

Biber, D., Johnasson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spoken and written English.Harlow, UK: Pearson Education Ltd.

This is an example of a multiple-context cloze task based on data generated by a concordance program. Find the missing word that fits ALL of the blanks below.

“The law has just been passed, making it illegal to give _____ or sell alcohol to anyone under the age of 18.”

“Restaurants have become so desperate that they no longer turn _____ customers who are not properly dressed.”

“Awareness has led to a shift in local consumer sentiment _____ from processed, imported foods to locally-produced, organically grown fruit and vegetables.”

“Though this latest condominium development is just five minutes _____ from Tiong Bahru MRT station, figures have shown that sales are low.”

“When she discovered yoga in 2002, Catherine knew right _____ that it would transform her both spiritually and physically.”

Resources for Research
Doyle, P. (2002). Grammar learning and thinking. In S. C. Chang & Y. M. Cheah (Eds.), Teachers’ handbook on teaching thinking skills across disciplines (pp. 127-133). Singapore: Prentice Hall, Pearson Education Asia.

Doyle, P. (2007). Teaching grammar with a learner corpus. In T. Ruanni, F. Tupas, Y. Yi, & C. Nur (Eds.), Changing perspectives on pedagogical grammar (pp. 77-86). Singapore: Singapore Association for Applied Linguistics.

Doyle, P. (2009). Language development in Singapore classrooms: A corpus-based description of the ‘school variety’. In R. Silver, C. Goh, & L. Alsagoff (Eds.), Language learning in new English contexts: Studies of acquisition and development (pp. 91-111). London: Continuum.

Never Miss A Story