TOEIC Link Writing — Parallelism and Coordinate Structure Balance Under Timed Response: Making Every List and Comparison Read as Deliberate Prose
Parallelism is the principle that items joined in a series, a comparison, or a paired construction should share the same grammatical form. "The training improved accuracy, speed, and confidence" is parallel — three nouns in a row. "The training improved accuracy, speed, and made workers more confident" is not — two nouns and then a clause, and the reader feels the sentence stumble even if they cannot name the fault. In TOEIC Link writing, faulty parallelism rarely produces a single obviously wrong word; it produces a sentence that reads as slightly out of control, and that impression quietly erodes the organization and range scores.
The reason parallelism is worth deliberate practice is that it is almost never a knowledge problem. Writers who produce faulty parallelism under timed conditions can, when shown the sentence afterward, fix it in seconds. They break it in the first place because the second and third items in a list are generated one at a time, and by the time the writer reaches the third item they have lost track of the grammatical shape the first item established. The repair is not learning what parallelism is; it is building the reflex to check that every item in a coordinate structure answers to the same template.
For related work on controlling sentence structure and reading as deliberate prose, see the writing cleft and fronting for emphasis and end-focus control guide, the writing end-weight and given–new information ordering for readability guide, and the writing transition signal precision and conjunctive adverb selection guide.
Why parallelism breaks under time pressure
A coordinate structure is any construction where two or more elements are joined as grammatical equals: a list joined by "and" or "or," a comparison joined by "than" or "as," or a paired construction joined by "not only … but also," "both … and," or "either … or." English expects the joined elements to be the same kind of thing — all nouns, all verb phrases, all clauses, all prepositional phrases.
Under timed conditions the writer does not plan the whole structure before starting it. They write the first item, reach the coordinator, and then generate the next item as a fresh thought — often in a different grammatical shape because the idea arrived in that shape. "The report was clear, concise, and it covered every point" starts with two adjectives and ends with a clause, because "it covered every point" is how the third idea occurred to the writer. Nothing about the writer's grammar knowledge failed. The planning horizon was simply shorter than the sentence, and the coordinator did not carry the template forward. The discipline below reinserts that template.
The three positions where parallelism matters most
Parallelism can break anywhere, but three positions account for most faults in timed essays, and knowing them tells you where to look.
1. The three-item list
The most common failure. The writer establishes a form with the first two items and drifts on the third, because the third is generated last and under the most time pressure. "We need to hire staff, buy equipment, and training is required" mixes two verb phrases with a clause. The fix is to make the third item answer the same question as the first: hire staff, buy equipment, provide training. Test by reading each item directly after the verb or preposition that governs the list — "we need to … hire staff / buy equipment / training is required" — and the odd one out becomes audible.
2. The comparison
Comparisons joined by "than" or "as … as" must compare the same kind of thing. "The new system is faster than the old process" compares a system to a process — close enough to pass unnoticed, but "the new system is faster than the old one" or "faster than the old system" is cleaner. The sharper failure is comparing unlike categories: "Her results were better than her colleagues" compares results to colleagues, when the writer meant "better than her colleagues'" — better than her colleagues' results. Under time pressure the possessive marker or the second noun drops, and the comparison quietly compares the wrong pair.
3. The correlative pair
Constructions like "not only … but also," "both … and," and "either … or" demand that the same grammatical form follow each half of the pair. "She not only reviewed the data but also the conclusions" is faulty because "reviewed the data" is a verb phrase and "the conclusions" is a bare noun. The fix places the pair around matching elements: "She reviewed not only the data but also the conclusions" — two nouns — or "She not only reviewed the data but also questioned the conclusions" — two verb phrases. The rule is simple: whatever kind of thing follows the first half of the correlative must follow the second half too.
The repair reflex: template and test
The whole discipline reduces to two moves the writer performs at the coordinator.
Set the template at the first item. When you write the first element of a list, comparison, or pair, register its grammatical shape — noun, verb phrase, adjective, clause. That shape is now the template every following item must match.
Test each following item against the template. As you reach each subsequent item, read it directly after the word that governs the structure — the verb, the preposition, the "than," the first half of the correlative — and confirm it is the same kind of thing. If it is not, either reshape the item to fit the template or reshape the template to fit the item, whichever is cleaner. The one thing you cannot do is leave them mismatched.
This test costs about a second per coordinate structure, and it is the single highest-return grammar check in timed writing because faulty parallelism is both common and invisible to the writer who produced it.
A four-week protocol for parallel structure at speed
Parallelism is a reflex, and reflexes come from repetition, not from re-reading the rule.
Week 1 — Recognition. Take ten of your own past essays and mark every coordinate structure — every list, comparison, and correlative pair. For each, name the template the first item sets and check whether the following items match. You are training your eye to find coordinate structures before you can train it to balance them.
Week 2 — The three-item list. Write twenty sentences that each contain a three-item list, deliberately generating the third item last and then testing it against the template. Rewrite any that drift. The goal is to make the third-item check automatic, because that is where the fault lives.
Week 3 — Comparisons and correlatives. Write ten comparison sentences and ten correlative-pair sentences, checking that each compares or pairs the same kind of thing. Pay special attention to the dropped possessive in comparisons and the misplaced "not only" in correlatives.
Week 4 — Timed integration. Write full timed essays and, in the last two minutes of review, scan only for coordinate structures — nothing else. Confirm each is balanced. By now the template-and-test move should fire during composition, and the review pass should find almost nothing.
What balanced coordinate structure signals to a rater
A rater does not consciously award points for parallelism. What they register is that your lists, comparisons, and paired constructions read as controlled — that when you set up a series, every item answers the same question, and the sentence lands cleanly. That control is exactly what the organization and range criteria reward: it is the difference between prose that sounds assembled and prose that sounds composed. Parallelism is one of the cheapest ways to buy that impression, because the ideas are already yours — you are only making sure the grammar that carries them lines up.
The writer who masters this stops producing the small stumbles that make an otherwise strong essay read as slightly amateur, and starts producing the clean, balanced sentences that tell a rater the writer is in full control of the structure — which, by the end of four weeks of deliberate practice, they are.