By Dalat TESOL
Helping novice researchers understand and apply statistics in English language education research
📌 Introduction
“I’m okay with teaching methods and theory… but statistics? That’s where I panic.”
If you’re a graduate student in TESOL or applied linguistics, chances are you’ve had a similar feeling. The very mention of terms like t-tests, correlation, or standard deviation can trigger anxiety — especially if you come from a non-mathematical background.
But here’s the good news: you don’t need to be a statistician to use statistics effectively in your research. What you do need is a clear understanding of:
- What basic statistical tools do
- When to use them
- How to interpret and report your findings in simple academic language
In TESOL research, statistics are not about crunching numbers for their own sake — they’re tools to help us:
- Describe learner performance or beliefs
- Identify trends in large groups
- Compare teaching interventions
- Explore relationships between key variables (like self-efficacy and AI use)
This guide walks you through the core statistical concepts and tools commonly used in TESOL studies. Whether you’re analyzing test scores, survey data, or experimental results, you’ll learn how to apply and explain basic statistics clearly — with real examples along the way.
1. 🎯 What Is Statistics in TESOL Research?
Statistics refers to techniques for summarizing and analyzing data in order to make meaning from it. In TESOL, it helps us answer research questions like:
- Do students who use mobile apps learn vocabulary faster?
- Is there a relationship between teacher feedback and student confidence?
There are two main types:
🔹 Descriptive Statistics
These are used to describe what the data looks like.
✅ Example:
You collected IELTS writing scores from 50 students.
- The mean score (average) is 6.4
- The standard deviation is 0.7 → scores are close together
You now know that most students scored around 6.4, and scores don’t vary much.
🔹 Inferential Statistics
These help you make conclusions or predictions based on a sample.
✅ Example:
You survey 100 students about their AI tool use. Can you generalize to all EFL students in Vietnam?
→ You use a t-test or correlation to determine if patterns are statistically significant.
2. 📦 Types of Variables in TESOL (with Examples)
Understanding variable types helps you choose the right test. Here’s a simple breakdown:
| Type | Meaning | TESOL Example |
|---|---|---|
| Nominal | Categories without order | Gender (male/female), AI tool (ChatGPT, Grammarly) |
| Ordinal | Categories with order | Proficiency level (Beginner, Intermediate, Advanced) |
| Interval | Numbers with equal spacing, no true zero | Likert scale responses (e.g., 1–5 for motivation) |
| Ratio | Like interval but with a true zero | Vocabulary test scores (0–100), number of writing drafts submitted |
3. 🧪 Basic Statistical Tools and When to Use Them
Let’s go over five core tools, each with a TESOL-friendly example.
📌 a. Mean and Standard Deviation
Used to summarize scores or ratings.
✅ Example:
You give a self-efficacy survey (1 = strongly disagree, 5 = strongly agree).
Students’ average score = 3.9, SD = 0.4
→ They generally agree, and responses don’t vary too much.
📌 b. Correlation (Pearson’s r)
Used to see if two variables are related.
✅ Example:
You examine whether students who use ChatGPT more often feel more confident writing essays.
r = 0.52, p < .01 → Moderate positive correlation
→ As ChatGPT use goes up, confidence tends to go up too.
📌 c. Independent Samples t-test
Used to compare two groups.
✅ Example:
Do male and female students differ in writing self-efficacy scores?
- Group A (males): Mean = 3.8
- Group B (females): Mean = 4.1
t(58) = 2.12, p = .038
→ The difference is statistically significant at the .05 level.
📌 d. ANOVA (Analysis of Variance)
Used to compare more than two groups.
✅ Example:
You compare motivation levels in Year 1, Year 2, and Year 3 students.
ANOVA shows F(2, 87) = 4.89, p = .01 → Significant
Post-hoc test reveals Year 3 students are significantly less motivated than Year 1.
📌 e. Cronbach’s Alpha
Used to test the reliability of a questionnaire (are the items consistent?).
✅ Example:
You design a 10-item scale to measure anxiety.
Cronbach’s α = .81 → Good internal consistency
✅ Rule of thumb:
- α ≥ .70 → Acceptable
- α ≥ .80 → Good
- α ≥ .90 → Excellent
4. 🧾 Reporting Results in Academic Writing (APA Style)
Use clear structure and include all relevant statistics.
📌 t-test Example:
An independent-samples t-test revealed a significant difference in willingness to communicate between students with high vs. low self-efficacy, t(64) = 2.34, p = .022.
📌 Correlation Example:
There was a moderate positive correlation between AI tool use and writing confidence, r = .41, p < .01.
✅ Use italics for t, r, p, etc.
✅ Always explain the finding in words: “This suggests that…”
5. 🧠 Common Mistakes to Avoid (With Fixes)
| Mistake | Fix with Example |
|---|---|
| Confusing correlation with causation | Say: “Students who use AI more tend to report higher confidence” (not “AI caused confidence”) |
| Using the wrong test | Know your variables: To compare three groups, don’t use a t-test — use ANOVA |
| Ignoring assumptions | For t-tests, check normal distribution and equal variance |
| Forgetting reliability | Before using your questionnaire results, run Cronbach’s alpha |
🔎 Quick Reference Table: What Test Should I Use?
| Goal | Variables | Test |
|---|---|---|
| Compare 2 groups | Group (nominal) + score (interval/ratio) | t-test |
| Compare 3+ groups | Group (nominal) + score | ANOVA |
| Test relationship | Two interval variables | Pearson’s r |
| Measure internal consistency | Likert-type scale items | Cronbach’s alpha |
| Describe responses | One variable | Mean, SD, frequency |
✍️ Final Thoughts
You don’t need to master every statistical test — just understand the basics:
- What your research question is asking
- What type of data you have
- Which tool answers that question clearly
- How to report it transparently and simply
In TESOL research, clarity and appropriateness matter more than complexity. With practice and purpose, even “non-math people” can become confident in using statistics to support strong, evidence-based claims.
🌿 Dalat TESOL – Chia sẻ kiến thức giảng dạy, nghiên cứu khoa học và cơ hội xuất bản