What Caps Band 6.5 vs 7 on Task 2

Short answer: On graded Task 2 practice essays in our corpus, mid-band scripts are most often capped by Task Response or Lexical Resource, not by “not enough rare words.” Under-250 drafts show clearly weaker Task Response. Length alone does not create Band 7 — development does.

This data note from IELTS AI Tutor by IELTSGRADER summarises anonymized practice grades (not official IELTS). Related teaching: Band 6 vs Band 7 Task 2 and the Task 2 word counter. Use it to choose what to rewrite — then verify on your own draft.

In this guide

1. Corpus and limits
2. Which criterion is weakest at mid-band
3. Word count vs Task Response
4. How to use the tables without cargo-culting
5. What this means for a 6.5 → 7 plan
6. Try this yourself
7. Frequently asked questions

Corpus and limits

Fact	Detail
Skill	Writing Task 2 practice submissions only
Sample	170 graded Task 2 attempts with criterion bands
Wider writing pool	393 graded writing attempts (Task 1 + Task 2)
What we store for this page	Aggregates only — no essays, emails, or names
What a band means here	Practice estimate aligned to public descriptors

Limits (read these):

• Practice scores are not official IELTS results.
• Mid-band buckets (especially 6.5–6.99) are smaller than the full Task 2 set — treat percentages as directional, not destiny.
• Users who practise with us are not a random sample of all IELTS candidates.
• We do not claim a university or IDP/British Council endorsement.

Methodology context: how we grade and how IELTS writing is scored.

Which criterion is weakest at mid-band

For each Task 2 report we identify the lowest of the four writing criteria (ties assign the first matching criterion in TR → CC → LR → GRA order). Among attempts with overall practice bands in common mid-band ranges:

Overall practice band	Most common weakest criterion	Count in bucket
6.0–6.49	Task Response (TR)	9 of 22
6.0–6.49	Grammar (GRA)	7 of 22
6.0–6.49	Lexical Resource (LR)	6 of 22
6.5–6.99	Lexical Resource (LR)	5 of 11
6.5–6.99	Task Response (TR)	4 of 11
7.0–7.49	Task Response (TR)	5 of 11

How to read this without overclaiming:

• At 6.0–6.49, Task Response is the single most common floor — incomplete coverage and thin development show up first.
• At 6.5–6.99, Lexical Resource and Task Response both appear often; the sample is small (11), so do not treat “always LR” as a law.
• Even in the 7.0–7.49 practice band, Task Response still appears as the weakest criterion for several scripts — Band 7 is not “vocabulary solved everything.”

This matches classroom experience: candidates often rewrite vocabulary while the prompt is only half-answered. See also stuck at Band 6.5.

Word count vs Task Response

Same Task 2 corpus, bucketed by word count:

Word-count bucket	n	Average practice TR	Average overall
Under 250	47	4.14	3.49
250–290	47	5.57	4.66
291–350	52	5.84	5.27
Over 350	24	6.02	5.81

Honest takeaways:

1. Under 250 is expensive. Average Task Response sits clearly lower than the in-range buckets.
2. 250–290 is a practical floor, not a Band 7 guarantee. Average overall in that bucket is still mid/low in this practice set — development quality matters.
3. Over 350 is not free Band points. Average TR edges up, but long scripts often trade time for Grammar and Coherence errors on exam day.

Use the live Task 2 word counter on a timed draft, then grade the same text.

How to use the tables without cargo-culting

Do not conclude “I must fix Lexical Resource because the 6.5 bucket listed LR first.” Your next timed essay might be capped by Task Response. The aggregates explain common floors in this practice set. Your report explains your floor.

A safe workflow:

1. Take two timed Task 2s in one week.
2. Note the weakest criterion on each.
3. If the same criterion is lowest twice, make that your two-week focus.
4. Only then compare yourself to the mid-band table for motivation — not diagnosis.

Task Response failures that look like “vocabulary problems”

Candidates often say “I need more advanced words” when the script actually:

• Answers only half of a two-part question
• States a position without an example
• Repeats the same idea in new synonyms
• Finishes under 250 words

Those are Task Response issues. Synonym hunting will not fix them. The under-250 bucket’s average TR (4.14 in this set) is the blunt version of that lesson.

Lexical Resource as a real cap (when it is)

LR is a fair focus when:

• Meaning is clear and the task is fully answered
• Wording is repetitive or imprecise on key topic nouns
• Forced rare words create tone problems

Then replace two imprecise phrases per body paragraph — do not rebuild the whole essay. Pair with lexical resource Band 6 to 7 if you need drills.

Grammar as a silent overall drag

Even when GRA is not the single lowest mark, uncontrolled complex sentences pull the average down. If your report shows frequent meaning-blocking errors, simplify before expanding range. Mid-band 6.0–6.49 in this corpus still shows GRA as a common weakest mark (7 of 22).

Why we publish counts, not marketing percentages

Headlines that say “87% of students are capped by X” usually hide tiny samples. Here the 6.5–6.99 slice is only 11 attempts. We show raw counts so you can discount noisy cells. When the corpus grows, we can refresh this page’s updatedAt and tables.

Until then: treat this as a directional study for practice planning, then verify on grade my essay.

What this means for a 6.5 → 7 plan

1. Find your weakest criterion on a timed essay — do not guess from ChatGPT overall praise (ChatGPT vs criteria).
2. If TR is lowest: answer every part, add one concrete example per body, hit 260–290 words.
3. If LR is lowest: replace forced rare words with precise topic collocations; keep meaning clear.
4. If GRA is lowest: simplify one complex sentence per paragraph before adding range.
5. Re-check once after a single rewrite — see Band 6 vs 7 patterns.

A weekly rhythm that matches the data: timed Task 2 → criterion report → one rewrite → optional mock writing test. Keep notes on which criterion was lowest so you can spot a pattern across two weeks instead of reacting to one lucky overall.

Try this yourself

Prompt: Some people believe university education should be free for everyone. To what extent do you agree or disagree?

1. Write for 40 minutes. Target 260–290 words.
2. Submit to grade my essay.
3. Note which criterion is lowest. Compare it to the mid-band patterns above — but trust your report first.
4. Rewrite only the weakest paragraph. Re-check once.
5. If you were under 250, fix length before debating vocabulary.

Frequently asked questions

Is this official IELTS data?

No. These are anonymized practice grades from IELTSGRADER. Official results come only from your test centre.

Why are mid-band sample sizes small?

Fewer practice attempts land cleanly in 6.5–7.0 overall. We report counts so you can weigh confidence.

Does writing 350+ words raise Task Response?

On average in this set, TR is slightly higher above 350 — but time pressure and error rate often rise. Prefer developed 260–290 over bulk.

What if my weakest criterion is Coherence?

Coherence appears less often as the single lowest mark in these mid-band slices, but it still matters. Fix paragraph jobs before adding linkers.

Can ChatGPT tell me what caps my band?

It can guess. Same-essay criterion scoring is more actionable — ChatGPT vs IELTS criteria.

Will a university see these practice bands?

No. See will a university care about AI IELTS practice?.

Next steps

Stop collecting overall numbers. Find the criterion that caps you, rewrite that one, and verify under time.

Grade your Task 2 essay free · Task 2 word counter · Band 6 vs 7 guide