Stuck at the Same Score? How to Find Recurring Problems in Your Practice Records (Part 2 of 3): Five Steps to Find the Weak Criterion, and a Log Template

IELTS · Practice Methods中文版　Updated 2026/11/09

In this post: A worked example with my 50 marked essays: five steps to find the criterion that keeps pulling your score down, plus a practice log template you can copy.

3. Worked example: finding the weak criterion in 50 essays (five steps)

All the numbers below are my own calculations from the scores of those 50 essays. Some criterion scores were given as a range (for example “7–8”); in the calculations I always used the midpoint, 7.5.

Step 1: Average Task 1 and Task 2 separately

This is the step people skip most often. The first criterion is different for Task 1 and Task 2 (TA vs TR), and the two tasks test different skills. If you average them together, the problem disappears into the mix.

This table alone makes it clear: my Task 1 TA averaged only 6.39, the lowest number in the whole table. But my Task 1 CC averaged 7.44, which means my paragraphing and linking were fine. The problem was whether the content was correct and complete.

Task 2 looks like it has all four criteria at around 7, as if there were no obvious weakness. If I had stopped here, I would have concluded, wrongly, that “Task 2 is fine”. That is why the next step is needed.

Step 2: Count which criterion is lowest in each essay

Averages get pulled up by high-scoring essays, so I also calculated a second measure: in each essay, which of the four criteria was lowest (including joint lowest), and how many times that happened.

The Task 1 answer is confirmed again: in 23 of the 27 essays, TA was the lowest or joint lowest score. That is not an occasional slip any more; it is a fixed pattern.

Task 2 is more spread out: LR, GRA and CC are all often the lowest. When that happens, you need to break it down further and look at how things change over time.

Step 3: Split by time and see which criterion is moving

I split the 50 essays into “the first two weeks” (20 May – 1 June, 16 essays) and “the last two weeks” (17 June – 2 July, 21 essays), and treated 2–16 June as a transition period.

This was the most useful table in my whole record. My Task 2 CC went from an average of about 6.4 in the first two weeks to about 7.9 in the last two; from the essay on 19 June onwards, all eight of my remaining Task 2 essays got CC 8. By comparison, LR and GRA barely moved.

What does that mean? In the early weeks my Task 2 TR was often 8, so my arguments themselves were fine. What was really pulling me down was paragraph organisation. And all five of my Task 2 scores of 7.5 (19 June, 28 June, 29 June, 1 July and 2 July) came in the last two weeks, exactly the period when CC settled at 8.

If you lay out every Task 2 CC score by date, the change is even easier to see:

The same view for Task 1 TA:

Task 1 TA has no such neat upward line: a TA of 5 was still turning up on 24 June. That told me the TA problem would not fix itself just by writing more. It needed a different approach (see step 4).

Step 4: Pull out the lowest-scoring essays and look for what they share

Six essays got TA 5. I listed them:

First finding: the TA 5s are spread across four chart types (bar charts, process diagrams, maps and line graphs). No single chart type was especially weak, so “practise more maps” would not solve the problem.

Second finding: in these six essays CC was mostly 7 or 8, and LR and GRA were mostly 7. In other words, my English itself was not the issue; I was losing marks on whether I had covered what needed to be covered, and covered it correctly. Checked against the public band descriptors, the key points for TA 5 are: details presented mechanically, no clear overview, key features not adequately covered, and data that may be inaccurate.

I went back through these essays one by one against the original questions and wrote up their problems in the Task 1 sample posts. For example, the bar chart post asks why TA was still 5 when the overview was correct, and the process diagram post asks what exactly was missing from a TA 5 answer. Reading several essays with the same score side by side is far more effective than reviewing them one at a time, because what you see is the thing that keeps repeating.

Step 5: Read the score sequence for each question type

Finally, I split the essays by question type to see which types were already “good enough” and which were still going up and down:

This table helps you decide where to spend your time. Agree/Disagree was already stable at 7, so if your target is 7, you don’t need to keep writing lots of them. The types with big swings (Two-question, process diagrams) are where next week’s practice should go.

4. Designing your log: a template you can copy

Below are the columns I recommend for a practice log. The first seven are the ones my original score log already had; the rest are the columns I think are most worth adding after doing the analysis above. The first row holds real scores from my 50 essays (Essay 5). The second row is an example row to show how to fill in “Main issue” and “Fix only one thing next time”; it is not from my original record.

If you use Google Sheets or Excel, you can paste this line straight into the first row as your headers:

How to fill in “Main issue”: use fixed tags, not long sentences

This column very easily ends up full of sentences like “overview not great” or “grammar needs work”, which you can’t count. I suggest using fixed tag codes so you can tally them later. Here is a suggested tag list; adjust it to your own mistakes: