IELTS Writing Error Taxonomy

In this guide

1. How to use this taxonomy
2. 1. Task Response / Task Achievement errors
3. 2. Coherence & Cohesion errors
4. 3. Lexical Resource errors
5. 4. Grammatical Range & Accuracy errors
6. Try this yourself
7. Frequently asked questions

Not all Writing mistakes are equal. A missing overview, a wrong collocation, and an article error hurt different criteria, and need different fixes. This taxonomy from IELTS AI Tutor by IELTSGRADER groups common issues the way a tutor reads your script: by criterion first, then by pattern.

Use it to interpret reports from the essay checker or Task 2 checker.

How to use this taxonomy

1. Find your lowest criterion on a recent report.
2. Scan only that section below.
3. Pick the 1–2 patterns that match your writing.
4. Drill those for a week using feedback to study plan.

Trying to fix the entire list at once is how candidates stay busy without improving.

---

1. Task Response / Task Achievement errors

TR1: Partial answer

What it looks like: Discusses only one side of a two-part question; ignores “advantages and disadvantages”; gives no opinion when required.

Fix: Underline every task instruction word before planning.

TR2: Undeveloped ideas

What it looks like: Topic sentence + one short claim + move on. Common at Band 6, see education sample.

Fix: Force explain + example in every body paragraph.

TR3: Position drift

What it looks like: Introduction agrees; conclusion becomes neutral or opposite.

Fix: Write your extent stance in one line and paste it above the page while writing.

TR4: Generic examples

What it looks like: “For example, in many countries…” with no mechanism.

Fix: Name a concrete scenario (school policy, workplace rule, household habit).

TA5: Missing Task 1 overview (Academic)

What it looks like: Data listed; no summary of main trends.

Fix: Two overview sentences before details. Model: bar chart Band 7.

TA6: Missed letter bullet (GT)

What it looks like: Beautiful tone; one prompt bullet ignored.

Fix: Tick bullets. Checklist: GT formal letter checklist.

TA7: Wrong letter tone

What it looks like: Slang in a complaint to a manager; stiff formality to a friend.

Fix: Identify recipient relationship first.

---

2. Coherence & Cohesion errors

CC1: Mixed-paragraph syndrome

What it looks like: One paragraph contains two unrelated main ideas.

Fix: One job per paragraph. Plan with labels.

CC2: Mechanical linker chains

What it looks like: Firstly / Moreover / Furthermore / In addition with no logical progress.

Fix: Ban empty linkers for a week; use because / however / as a result. Plan: 14-day CC plan.

CC3: Weak or missing topic sentences

What it looks like: Paragraph starts with a detail or example.

Fix: First sentence must predict the paragraph.

CC4: Poor referencing

What it looks like: Same noun repeated eight times; or unclear this with no antecedent.

Fix: Practise this approach / these measures / such a policy.

CC5: Conclusion introduces new arguments

What it looks like: Brand-new example appears only at the end.

Fix: Conclusions summarise and close, no new case studies.

CC6: Illogical order

What it looks like: Example before the claim; contrast before the idea being contrasted.

Fix: General → specific → significance.

---

3. Lexical Resource errors

LR1: Repetition loops

What it looks like: increase / people / government / environment recycled endlessly.

Fix: Small paraphrase banks. Guides: trends vocabulary, LR Band 6→7.

LR2: Thesaurus stuffing

What it looks like: Rare words used inaccurately (ameliorate the students).

Fix: Prefer precise common academic collocations.

LR3: Wrong collocation

What it looks like: make a meeting, strong rain, big damage.

Fix: Learn verb–noun pairs by topic, not isolated words.

LR4: Informal lexis in Academic essays

What it looks like: kids, stuff, a lot of guys think.

Fix: children, issues, many people argue.

LR5: Degree mismatch in Task 1

What it looks like: skyrocketed for a 2% rise.

Fix: Match adverb/adjective strength to the data.

LR6: Copying the rubric

What it looks like: Introduction repeats the question almost word for word.

Fix: Paraphrase prompt key terms in the first two sentences.

---

4. Grammatical Range & Accuracy errors

GRA1: Agreement errors

What it looks like: People is, The number of students are.

Fix: Error log; rewrite only agreement mistakes for three days.

GRA2: Article problems

What it looks like: Missing the with unique references; extra articles with plurals.

Fix: Weekly article clinic on your own sentences.

GRA3: Tense timeline breaks

What it looks like: Task 1 jumps randomly between past and present; narrative essays mix past forms without reason.

Fix: Decide timeline before writing; circle verbs in review.

GRA4: Complex sentence collapse

What it looks like: Long sentences with although… but, broken relative clauses.

Fix: Two clean complex patterns you can control; then expand.

GRA5: Punctuation that blocks meaning

What it looks like: Run-ons; comma splices in key opinion sentences.

Fix: Full stop discipline in conclusions and topic sentences.

GRA6: Range without control

What it looks like: Forced passives and conditionals that confuse the reader.

Fix: Accuracy first; add range only where meaning stays clear. See Band 6 vs 7.

---

Cross-criterion “combo” mistakes

Some issues hit more than one band descriptor:

Pattern	Criteria affected
Memorised template off-topic	TR + CC + LR
Listing every Task 1 number	TA + CC + LR
Angry informal complaint letter	TA + LR
Endless introduction, thin bodies	TR + CC

Your AI tutor may flag the symptom in one criterion; your study plan should still address the root.

Error taxonomy → weekly plan

If most flags are…	Spend the week on…
TR/TA	Question analysis + development drills
CC	Paragraph jobs + topic sentences
LR	Collocation banks + paraphrase
GRA	Personal error log + controlled complex sentences

After patterns repeat across essays, learning editions can group them into a focused study path.

Worked example: one essay, four families

Imagine a Task 2 script with these tutor flags:

• Body 2 is only three sentences and restates the claim (TR2)
• Three paragraphs start with Moreover (CC2)
• Important appears nine times (LR1)
• People is appears twice (GRA1)

A weak response is “study everything.” A tutor response is:

1. This week: TR2 only, expand every body with explanation + example
2. Next week: CC2, ban empty linkers; rebuild topic sentences
3. Keep a mini log for GRA1 so it does not return
4. Build a 15-word paraphrase bank later for LR1

That sequencing is how taxonomy becomes a plan instead of a panic list. Full method: feedback to study plan and mock to 14-day sprint.

Severity: meaning-blocking vs cosmetic

Not every flag deserves equal time.

Meaning-blocking (fix first):

• Missing letter bullet / missing overview
• Position that contradicts itself
• Grammar errors that confuse who did what
• Paragraphs so mixed the reader cannot follow

Cosmetic (fix after the above):

• One repeated linker
• A slightly informal word in an otherwise clear academic sentence
• A rare article slip that does not hide meaning

Band 7 scripts can still contain small cosmetic issues. Band 6 scripts often mix blocking problems with cosmetic ones, and students polish cosmetics first because they are easier to notice.

Task 1 vs Task 2: do not mix taxonomies blindly

A “thin example” note on Task 2 is a TR issue. On Academic Task 1, inventing an example about why sales rose is usually a TA problem (speculation). On GT letters, “thin” often means a bullet answered without detail, still TA.

Keep module context clear with Academic vs GT writing and TR vs TA decoded.

Building a personal error codebook

After three checked tasks, create a one-page codebook:

Code	My example sentence	Corrected version	Criterion
GRA1	People is worried…	People are worried…	GRA
CC1	(paste mixed paragraph)	(split version)	CC
LR3	make a research	conduct research	LR

Review the codebook for five minutes before every timed essay. This is more effective than rereading generic grammar books. Collocation support: lexical resource Band 6 to 7.

How IELTS AI Tutor uses this lens

IELTS AI Tutor does not only say “grammar wrong.” It aims to show which family of problems is capping your band so you can practise like a tutor would assign homework, not like a spellcheck. Dual-checked feedback (dual AI grading explained) helps keep those family labels stable enough to plan around.

For timed discovery of your taxonomy profile, take a mock writing test. If the same codes return after a sprint, move into a learning edition rather than collecting new random tips. Plateau context: stuck at Band 6.5.

Try this yourself

Prompt: Highlight errors in your last essay by criterion family, then rewrite only the top two patterns.

1. Write under realistic time limits.
2. Check your essay free (or the Task 1/2 checker if relevant).
3. Rewrite only the lowest criterion once, then re-check.

Next steps

Stop treating every mistake as equal. Classify, prioritise, drill.

Run an essay through the essay checker, map flags to this taxonomy, then build a week from feedback to study plan. Compare pricing and signup when you want repeated checks to track whether each error family shrinks.

Frequently asked questions

What mistakes most often cap Band 6.5?

Under-developed Task Response, mechanical coherence, collocation errors, and repeated grammar slips.

Should I fix grammar first?

Only if GRA is clearly the lowest. Otherwise fix the criterion that actually caps you.

How do I use an error taxonomy?

Label mistakes by criterion, count repeats, and drill the highest-frequency family.