Most candidates who get stuck at band 6.5 are not making dramatic errors — they are making consistent, small ones that examiners are trained to notice immediately. Understanding what band 7 writing actually looks like, at the descriptor level, is the fastest way to stop guessing and start improving deliberately.
#What the band 7 descriptors actually say
The official IELTS band descriptors assess Writing Task 2 across four equally weighted criteria: Task Achievement, Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy. Each criterion is marked independently, and your final writing band is an average of the four.
Here is what band 7 specifically requires across each criterion, quoted directly from the public descriptors:
Task Achievement (band 7): "addresses all parts of the task… presents a clear position throughout the response… presents, extends and supports main ideas, but there may be a tendency to over-generalise."
Coherence and Cohesion (band 7): "logically organises information and ideas… uses a range of cohesive devices appropriately although there may be some under- or over-use."
Lexical Resource (band 7): "uses a sufficient range of vocabulary to allow some flexibility and precision… uses less common lexical items with some awareness of style and collocation… may produce occasional errors in word choice, spelling and/or word formation."
Grammatical Range and Accuracy (band 7): "uses a variety of complex structures… produces frequent error-free sentences… has good control of grammar and punctuation but may make a few errors."
Notice what is not in those descriptors: perfection. Band 7 explicitly permits occasional errors, some over-generalisation, and minor cohesion lapses. The jump from 6.5 to 7 is not about eliminating all mistakes — it is about demonstrating consistent control and range, even when small slips occur.
#The single biggest gap between band 6.5 and band 7
Candidates at band 6.5 routinely produce responses that are grammatically acceptable and topically relevant but structurally thin. Their arguments exist — they just do not develop. This is the over-generalisation trap the Task Achievement descriptor references.
#What over-generalisation looks like
Consider this paragraph responding to a prompt about whether governments should fund the arts:
Before (band 6.5 level): "The arts are important for society. They help people express themselves and make communities better. Therefore, governments should give money to support them."
This is not wrong. It is just empty. Three sentences have been used to say one thing. There is no extension, no concrete reasoning, no concession.
#What development looks like at band 7
After (band 7 level): "State funding for the arts serves a function that private markets tend to overlook: the preservation of culturally specific forms that lack commercial viability. A regional folk theatre, for instance, may draw limited audiences and generate no profit, yet it encodes historical and linguistic identity that, once lost, cannot be recovered. Without public subsidy, such institutions close — not because they lack value, but because their value is non-monetary. This is precisely the kind of market failure that justifies government intervention."
The second version contains a claim, a specific illustration, a causal explanation, and an explicit logical link back to the claim. That structure — claim, example, explanation, significance — is what examiners mean when they describe responses that "extend and support" ideas.
#Lexical resource: what "less common vocabulary" actually means
The Lexical Resource descriptor at band 7 rewards vocabulary that demonstrates precision and collocation awareness. It does not reward word-swapping for its own sake. Replacing "important" with "pivotal" every time is not sophistication — examiners recognise it as mechanical substitution, and it can actually lower your score if the collocations are wrong.
#Collocation errors that cap candidates at band 6
Candidates routinely struggle with the difference between knowing a word and knowing how it behaves. Here are the kinds of errors that signal limited Lexical Resource:
- Wrong preposition: "contribute for society" instead of "contribute to society"
- Register mismatch: mixing formal argument ("it is imperative that") with colloquial phrasing ("loads of people") in the same paragraph
- Overcomplicated synonyms: "remuneration" when "pay" or "income" is more natural in context, producing awkward-sounding prose
- Repeated lexical stems: "economic economy economically" within a single paragraph — the descriptor penalises failure to paraphrase underlying concepts
Band 7 Lexical Resource is demonstrated by using mid-range academic vocabulary accurately and flexibly, not by deploying the longest word available. Phrases like "exacerbate existing inequalities," "mitigate environmental damage," or "pose significant challenges for policymakers" are band 7 because they are precise, correctly collocated, and used in context — not because they are obscure.
#Task 1 academic band 7: the overview problem
In Academic Task 1, the overview is the single element that most separates band 6 from band 7. Many candidates describe data thoroughly but never step back to state what the data means overall. The Task Achievement descriptor at band 7 requires a clear overview of main trends, differences, or stages — not just accurate reporting of individual figures.
#What a weak overview looks like
Weak: "The bar chart shows the percentage of people using the internet in five countries between 2000 and 2020."
That is a description of the chart — it is not an overview. It tells the examiner what they can already see in the title.
#What a band 7 overview looks like
Strong: "Overall, internet usage rose substantially across all five countries over the period, with the most dramatic growth occurring in developing nations that began from a much lower baseline. By 2020, the gap between the highest- and lowest-usage countries had narrowed considerably compared to 2000."
This overview identifies the dominant trend, notes a meaningful sub-pattern, and signals change over time — all without referencing a single specific figure. Specific data belongs in the body paragraphs. The overview belongs at the end of the introduction or at the start of the response body, and it should read as a conclusion drawn from the whole chart, not a title restatement.
#Grammatical range and accuracy: the "variety of complex structures" requirement
The band 7 grammar descriptor requires "a variety of complex structures" and "frequent error-free sentences." This does not mean every sentence must be complex — it means your response should demonstrate that you can use complex structures and do so correctly most of the time.
#Structures that signal grammatical range
Examiners look for evidence that you can move beyond simple subject-verb-object constructions. The following structural types, used accurately, contribute to a band 7 Grammatical Range and Accuracy score:
- Relative clauses: "This is a policy that disproportionately affects low-income households, who have fewer alternatives available to them."
- Conditional structures: "Were governments to withdraw arts funding entirely, many cultural institutions would close within a decade."
- Passive constructions used for focus: "The decision was met with significant opposition from local communities, who had not been consulted during the planning stage."
- Nominalisations: Converting "governments decide" into "government decision-making" to vary sentence structure and achieve academic register.
The critical caveat: a complex structure used incorrectly counts against you more than a simple structure used correctly. If conditional forms are unreliable in your writing, use them selectively and only when you are confident in the construction. Grammatical range means demonstrating range — not forcing every possible structure into a single essay.
#Coherence and cohesion: the over-linking trap
Coherence and cohesion is the criterion most candidates try to fix by adding more linking words. This is almost always the wrong approach. The band 7 descriptor explicitly mentions "some under- or over-use" of cohesive devices as a characteristic of this band — meaning even band 7 responses are not perfectly calibrated. The real issue at band 6 is usually not insufficient linking but mechanical linking.
#Before and after: mechanical versus organic cohesion
Before (mechanical): "Firstly, pollution is a serious problem. Furthermore, it affects human health. In addition, it damages the environment. Moreover, it is expensive to treat. Therefore, action must be taken."
Every sentence opens with a connector. The result reads like a list, not an argument. The connectors do not show logical relationships — they just signal sequence.
After (organic): "Pollution carries compounding costs: the immediate damage to human health drives up healthcare expenditure, while the longer-term degradation of ecosystems reduces agricultural productivity and increases food insecurity. These interrelated effects make a purely reactive policy response inadequate."
The second version uses fewer connectors but achieves stronger cohesion through sentence construction — the colon introduces elaboration, the "while" clause contrasts timeframes, and "these interrelated effects" creates a referential link backwards. That is what examiners mean by coherence: ideas that connect because of how they are structured, not because of the signposting vocabulary applied to them.
#How to move from 6.5 to 7: a practical approach
The candidates who close the gap from 6.5 to 7 share a common characteristic: they practise with feedback against the actual descriptors, not against a general sense of "writing better." Vague improvement goals produce vague results. Targeted diagnostic work produces targeted gains.
Practically, this means the following:
- Write full timed responses — Task 1 in roughly twenty minutes, Task 2 in roughly forty minutes, following the standard IELTS Writing format. Untimed practice builds different habits.
- Score yourself on each descriptor independently before seeking external feedback. Identifying your weakest criterion is more useful than chasing an overall band score.
- Focus revision on one criterion per week. If your Lexical Resource is holding you at band 6, a week of targeted collocation work will have more impact than scattered general writing practice.
- Rewrite, not just review. After receiving feedback on a response, rewrite the weakest paragraph from scratch applying the corrections. Reading feedback is passive; rewriting is active learning.
The typical improvement trajectory is real but not instant. Candidates who practise with genuine descriptor-level feedback see meaningful gains over a structured preparation period — but the work has to be deliberate, not just prolific.
If you want to know which of the four criteria is actually limiting your band score right now, BandNine's AI scoring gives you criterion-level breakdowns on every practice essay you submit, matched against the official descriptors. Take a diagnostic test to find your baseline, then focus your preparation where it actually needs to go.