Session 9: Learner Corpus Mini-Research

Masaki EGUCHI, Ph.D.

Housekeeping

🎯 Learning Objectives

As a wrap-up of the vocabulary and multiword sessions, you will be able to:

• Conduct mini learner corpus research.
  • Construct several learner corpus research questions with structured support.
  • Choose a learner corpus that suit your research needs
  • Justify choices of lexical richness measures to investigate a research questions
  • Compute lexical richness indices and analyze the results to answer research questions
  • Present the results and interpretation of the findings in a written prose

Mini-research outline

1. Construct research questions
2. Understand and choose the learner corpus
3. Construct hypothesis
4. Select index
5. Compute the index
6. Conduct analysis
7. Interpret and write-up the results

Step 1: Construct research questions

Researchers typically set RQs about the relationships between lexical characteristncs and variables that defines subsection of the corpus (e.g., grade, genre, or proficieincy score).

Step 2: Understand and choose the corpus

In this assignment, please choose one of the following corpora:

1. Growth in Grammar (GiG) corpus (Durrant, 2023)
2. ICNALE corpus (Edited Essay OR GRA)
3. If you have a corpus you want to analyze, let Masaki know!

Understanding the corpus contents

• You can look at followings:
  • meta-data (the additional information about the corpus files)
  • several text files

GiG corpus

• For detailed information, see Table 2.1 in Durrant (2023)

GiG metadata

ICNALE GRA

ICNALE GRA documents 120 essays evaluated by 80 people. - These 80 people are from different backgrounds. - You can basically use average essay ratings.

ICNALE paper

ICNALE metadata

Step 1 & 2: Formluating research questions

What would you like to know?

• How learners develop their language ability across time?

• What defines “more proficient” language use?

• How situational variables of writing/speaking impact the language production?

• How does production of X change across time/proficiency?

• Anything else?

• Do not forget to specify constructs!

Step 2: Choosing a right corpus

Based on the research question, which corpus should you choose?

1. Growth in Grammar (GiG) corpus (Durrant, 2023)
2. ICNALE corpus (Edited Essay OR GRA)

• Understand what it offers and doesn’t.
• Pick one corpus to move forward.

Step 3: Construct hypothesis

• Once you’ve articulated broad research questions and chose a corpus, speculate on possible results.
• Use what you’ve learned so far about lexical richness.
  • Lexical diversity
  • Lexical sophistication
  • Phraseological complexity
• Which ones are relevant to your questions?

Step 4: Select index

Let’s keep going!

Specifically, which index may capture the change in the vocabulary use in your context?

• List posssible indices to capture the differences in vocabulary use.
• Why do you think those are good choices?

Step 5: Compute the index

Okay now it’s time to do some analysis!

• Which ones are you going to use:
  • TAALED
  • TAALES
  • Lexical profiler
  • Custom list in Simple Text Analyzer?

Step 6: Conduct analysis

Once you obtained the indices, it’s time to understand the pattern by plotting.

• Use simple analyzer to plot data from:
  • GiG corpus
  • ICNALE
• The simple text analyzer is made for that purpose.

Step 7: Interpret and write-up the results

Now you understand what is happening in your corpus, you can write that up.

Mini-research outline

1. Construct research questions
2. Understand and choose the learner corpus
3. Construct hypothesis
4. Select index
5. Compute the index
6. Conduct analysis
7. Interpret and write-up the results

Let’s work!

Let us know if you have any questions.

Reflection

• You can now:
  • Provide reasons for your choice of lexical richness measures.
  • Conduct preliminary analysis to understand how text differ from one aother in relation to learner’s proficiency, learner groups, etc.