Why did my IELTS writing score drop on the retake?

A writing band that comes back lower on the retake is almost never a sign that your English went backwards. Each attempt is marked as a fresh best-fit judgement on a different script, and the way the eight criterion scores are averaged and weighted means the printed number can fall half a band with only one criterion having moved.

That does not make the result meaningless. It means the useful question is not "why did I get worse" but "which criterion moved, and what did I change that moved it". What follows is the arithmetic that produces a lower band, the three changes that most often produce it, and what to check before booking a third test.

The same candidate, two scripts, two independent judgements.

Your band is a judgement on one script, not a running total

Nothing carries over between attempts. The examiner marking your second essay does not see your first script, your previous band, or how long you have been preparing. They read what is in front of them and, for each criterion, choose the descriptor level that is the best fit for what that script does.

"Best fit" is the load-bearing phrase. Your essay will not match every line of one band's wording, so the examiner picks the band it fits most closely, criterion by criterion. Each criterion is awarded a whole band — there are no half bands at criterion level — and the four are averaged to produce the band for that task.

So "6.5, then 6" is not one measurement repeated. It is two independent judgements, on two different prompts, of two texts written on two different days. Two things follow. Small differences between the scripts can produce different bands without either judgement being wrong. And your second result describes your second essay, not your progress between the two.

Half a band down is inside the arithmetic

The reported writing band is assembled from eight criterion scores. Each task is scored on its own four criteria and averaged, and the two task results are then combined with Task 2 counting double:

(Task 1 + Task 2 × 2) / 3

rounded to the nearest quarter band. Two consequences of that formula account for most of the drops people bring to a second test.

One criterion is worth about a sixth of a band. A single criterion is a quarter of its task's average, and Task 2 carries two-thirds of the total, so one criterion falling by one band in Task 2 — Lexical Resource going from 7 to 6, for instance — moves the reported figure by roughly 0.17. On its own that is invisible. Two criteria moving, or one criterion moving while your first attempt was already sitting on a rounding boundary, is enough to change the number printed on the report form.

Task 1 is half the weight, and it is where most people stop looking. Suppose the first attempt was Task 1 = 6, Task 2 = 7: (6 + 14) / 3 = 6.67, which rounds to 6.75. Second time round Task 2 holds at 7, but the chart type is unfamiliar and Task 1 comes out at 5: (5 + 14) / 3 = 6.33, rounding to 6.25. Task 2 did not move at all. The band did. The same weighting seen from the other side, where a strong Task 1 cannot rescue a weaker Task 2, is set out in I get 7 in Task 1 but 6.5 in Task 2.

The practical upshot is that a half-band drop is one or two criterion scores, not a collapse — and until you know which ones, nothing you do next is aimed at anything.

Cause 1 — the repair you made cost you a different criterion

This is the most common cause by some distance, and the one that feels least fair, because the drop follows directly from having done the homework.

The four criteria are not separate budgets. They are four readings of the same text, so a change made to lift one of them is also a change to the other three. Three versions of this show up again and again.

Told your vocabulary was the problem, you added less common words. Band 7 Lexical Resource asks for a resource "sufficient to allow some flexibility and precision", with an evident "awareness of style and collocation". Words inserted for effect rather than for meaning fail on that clause specifically, so the essay stays at the band 6 line — "a rather restricted range or a lack of precision in word choice" — while the register wobble they introduce now reads across into the examiner's view of the whole script.

Told your grammar was the problem, you wrote longer sentences. Band 7 Grammatical Range and Accuracy requires that "error-free sentences are frequent". Range and accuracy sit inside one criterion, so lengthening every sentence trades one against the other, and punctuation errors arrive with the subordinate clauses. The four structures that count as range without that trade are in complex sentences for band 7.

Told your essay did not flow, you added connectives. Band 6 Coherence and Cohesion names this failure directly: cohesive devices "may be faulty or mechanical due to misuse, overuse or omission". Adding signposting is the standard first response to a cohesion comment, and it is the response the descriptor penalises — the alternative is in linking words are keeping you at band 6.

If you changed one thing between attempts and the band moved down, that change is the first suspect, ahead of the test itself.

Lifting one criterion is also a change to the other three.

Cause 2 — the second prompt asked for something the first one did not

Task Response is scored against the prompt in front of you, and prompts are not interchangeable.

Band 6: "The main parts of the prompt are addressed (though some may be more fully covered than others)."

Band 7: "The main parts of the prompt are appropriately addressed. A clear and developed position is presented."

A prompt carrying two demands in one sentence — discuss both views and give your own opinion, or a "to what extent" attached to a second question — caps Task Response at 6 when you cover one part fully and the other thinly. That cap holds no matter how much your grammar and vocabulary improved in between, because it is a property of the prompt you drew and of how you divided your paragraphs, not of your English.

Task 1 behaves the same way, on Task Achievement rather than Task Response; the difference between the two is set out in Task Achievement vs Task Response. A line graph with one clear trend, a process diagram and a map do not ask for the same answer, and they do not reward the same practised sentences. If your preparation was all charts and you drew a map, Task 1 can fall a band on the day for reasons that have nothing to do with the weeks you spent working.

This is also the cause that most often explains a drop that felt like an improvement. A harder prompt handled competently can score below an easier prompt handled well, and the experience of writing it is of doing better work.

Cause 3 — prepared material reads differently to an examiner

The natural response to a disappointing band is to prepare harder: a template introduction, a stock set of paragraph openers, a list of phrases to place. Some of that genuinely helps with timing. Two parts of it work against the score.

The first is fit. A memorised opening is still marked on how well it addresses this prompt. An introduction that restates the question and announces that the essay will discuss both sides has, in descriptor terms, not yet presented a position — a Task Response loss created by the fix that was meant to save time.

The second is the seam. Prepared sentences usually sit a band above the writing around them, and the change in register is visible at reading speed. It does not read as strength. It reads as two texts spliced together, and the examiner is marking the whole script for control and style.

There is a plainer version of this as well. If preparation went into Task 2 and Task 1 got the minutes that were left, or if the second essay came in short, the loss is mechanical rather than linguistic: what really happens under 250 words sets out what an underlength script actually costs.

What to check before you book a third test

Work from the scripts, not from the two numbers.

Find out which criterion moved. Two overall bands cannot tell you anything on their own. Reconstruct the second essay while it is still fresh — most people can rebuild the structure and much of the wording within a day or two — and mark it against each criterion separately, then do the same with the first. A drop that turns out to be one criterion in Task 1 calls for a different month of work from a drop spread across Task 2.

Change one thing at a time. If the last attempt changed vocabulary, sentence length and paragraph structure at once, no result can tell you which of them moved. Take the lowest criterion, work on that alone, and check that the others hold steady in the next practice essay.

Score Task 1 on its own terms. Given the weighting, an unfamiliar Task 1 type is the cheapest item on this list to fix and the easiest to leave undiagnosed, because attention goes to the essay that felt harder.

Write the next one under exam conditions before you judge it. An essay written untimed at home is not evidence about a timed test, and the difference between timed and untimed is often the whole of the difference between two attempts.

If you believe the second mark was simply wrong, a remark is available and is a decision worth taking on evidence rather than on the impression that the essay was better. Either way that means getting the script assessed criterion by criterion first. The full band 5 to 8 wording is in IELTS Writing band descriptors explained, and the four criterion patterns that hold most candidates at the same result twice are in why am I stuck at band 6.5.

FAQ

Does a lower IELTS writing score on a retake mean my English got worse?

Almost never. Each attempt is a separate best-fit judgement on a different script written to a different prompt, and the examiner marking the second one does not see the first. A half-band fall is one or two criterion scores moving inside the weighted average, which is a much smaller change than the reported number suggests.

How can my band drop when the second essay felt better?

Two common reasons. The second prompt may have asked something the first did not, so a competent answer to a harder question scores below a strong answer to an easier one. Or the improvement you made — richer vocabulary, longer sentences, more linking words — cost you accuracy or precision in a criterion you were not watching.

Which task is more likely to have caused the drop?

Task 2 carries twice the weight, so it moves the total faster, but Task 1 is where drops go unnoticed. With Task 2 held at 7, a Task 1 falling from 6 to 5 takes the weighted result from 6.67 to 6.33 on its own, which is enough to change the band that gets printed.

How many criteria have to move for the band to fall half a band?

Usually one or two. A single criterion in Task 2 is worth about 0.17 of the reported figure, so one criterion is enough when your previous result was sitting near a rounding boundary, and two are enough in most other cases.

Should I ask for a remark if my score went down?

A remark is worth considering, but it is a decision to take on evidence rather than on the feeling that the essay was better. Have the script marked criterion by criterion first, so you know whether there is a specific criterion you would expect to move.

Are the band estimates on this site official?

No. They are estimates made against the public band descriptors. Only a certified examiner produces an official band.

Related reading

• Why am I stuck at band 6.5 in IELTS writing — the four criterion patterns behind a repeated result
• I get 7 in Task 1 but 6.5 in Task 2 — the same weighting seen from the other side
• No grammar mistakes but still band 6.5 — why a clean essay can still sit below band 7
• IELTS Writing band descriptors explained — the full band 5 to 8 wording for all four criteria

All descriptor wording quoted here is from the public Writing band descriptors updated in May 2023, published by IELTS and the British Council.