How to review an IELTS practice test so the next one scores higher
Scoring a mock test tells you almost nothing. The error log that turns every practice test into a list of things to fix, with worked categories.
Study timelineGeneralUpdated 6 min read
The number at the bottom of a marked practice test is the least useful thing that test produces. "32 out of 40" tells you where you are; it tells you nothing about what to do on Tuesday. The useful output is the eight wrong answers, sorted by why they were wrong, because eight wrong answers almost never share one cause: three were rushed, one was a vocabulary gap, one was a trap you walked into with your eyes open, and three were blank because the clock beat you. Those are four different problems with four different fixes, and averaging them into a band score destroys exactly the information you need.
Tag every wrong answer with one of six causes
The taxonomy has to be small and fixed, or you will invent a category per item and the log becomes a diary. Six tags cover nearly everything:
- careless - you knew it, and the answer that reached the answer sheet was not the one in your head. Wrong line, wrong letter, misspelling, plural dropped.
- ran-out-of-time - blank, or answered in a panic in the last ninety seconds. Untested knowledge, not missing knowledge.
- did-not-know-the-word - the item hinged on a word or collocation you could not decode, in the passage or in the question.
- misread-the-question - you answered a question the paper did not ask. Word limit exceeded, "TWO letters" when you gave one, Task 2 prompt half-addressed.
- fell-for-the-trap - you read the right sentence and chose the answer the paper wanted you to choose. The classic FALSE that is actually NOT GIVEN — a three-step check tells them apart — and the Listening distractor that was corrected two seconds later.
- genuinely-too-hard - you understood the question, located the right text, gave it proper time, and still could not decide. If a third of your errors land here, you are using it as a place to hide.
One tag per item. Forcing the choice is the point: it makes you articulate the failure instead of feeling bad about it.
The log, worked
Here is what five errors from one Academic Reading paper look like once tagged. Keep it in a spreadsheet, one row per wrong answer, all papers in one sheet so the pattern accumulates.
| Q | Task type | I wrote | Answer | Tag | Fix |
|---|---|---|---|---|---|
| 7 | True/False/Not Given | FALSE | NOT GIVEN | fell-for-the-trap | Passage contradicted a neighbouring claim, not this one. Re-check: does the text address this exact proposition? |
| 14 | Matching headings | vi | iii | careless | Right heading, written against paragraph D instead of C. Number the paragraphs on the question page before starting. |
| 22 | Sentence completion | "highly permeable layers" | "permeable" | misread-the-question | Limit was ONE WORD. Circle the limit and write it at the top of the answer column. |
| 31 | Summary completion | blank | "sediment" | did-not-know-the-word | Did not know sediment / deposition. Added to geology word list with the sentence it appeared in. |
| 38 | Matching features | blank | C | ran-out-of-time | Spent 26 minutes on Passage 1. Hard cap of 18 minutes on Passage 1 next time. |
Note what the Fix column contains: an instruction, in the imperative, that changes behaviour in a measurable way next paper. "Be more careful" is not a fix. "Number the paragraphs before starting" is.
The repeat offender is your syllabus
Now the part almost nobody does. At the end of each week, count the tags across every paper you sat. Anything appearing three or more times gets its own dedicated session that week. Not more mock tests - a session aimed at that one cause:
- Sort your log by tag, then by task type, and count.
- Take the tag with the highest count. Read the Fix column for those rows: they will say roughly the same thing three times over.
- Write a single sentence describing the session. Four fell-for-the-trap errors in True/False/Not Given means ninety minutes of nothing but those items, no timer, reasoning aloud about why each distractor is a distractor.
- Do the session. Then sit one paper and see whether that tag's count drops.
- Archive rows you have fixed to a second sheet so the live log stays short and current.
Sequence matters, because the tags are not equally expensive. careless and misread-the-question errors are the cheapest marks on the paper: they need procedural discipline, not English. ran-out-of-time comes next: timing is a plan rather than an ability, and Reading gives 60 minutes for 40 questions with no transfer time, so a per-passage budget is arithmetic — a minute-by-minute plan for the three passages already exists. Only then is it worth grinding at did-not-know-the-word and fell-for-the-trap, the slow, real ones.
Writing and Speaking need a different log
For Reading and Listening the answer key settles it. For Writing and Speaking the criteria are the key. Writing Task 2 is assessed on Task Response, Coherence and Cohesion, Lexical Resource and Grammatical Range and Accuracy; Task 1 replaces Task Response with Task Achievement; Speaking is assessed on Fluency and Coherence, Lexical Resource, Grammatical Range and Accuracy and Pronunciation. Judge your response against each criterion separately, and instead of tags log one artefact per criterion: the sentence that lost the mark, and the rewrite that saves it. "Vocabulary weak" is useless. "Wrote big problem twice in Task 2; replace with acute shortage and serious drawback" is a fix you can still act on in a week.
The test itself is fixed, published and predictable. Your errors are the only variable in the room, and they are the only thing worth studying. Score the paper in ten minutes, then spend forty on the log.