Should I solve insertion items before or after grammar gap-fills?

**Grammar gap-fills first.** They take **15-20 seconds each and bank reliable points**, while **insertion items take 75 seconds and carry higher logical cost**. **Skip insertions on first pass, finish all grammar items, then return to insertions** — this maximizes total Part 6 score. Spending all your time on insertions costs you the easy gap-fill points.

What if no candidate has a reference word?

Move to **connectors / adverbials**. Candidates with **However / Therefore / In addition / For example** at the start are checked against the previous sentence's relation (**contrast / cause / addition / illustration**). **If neither references nor connectors apply**, the hardest case, decide on **connection to the following sentence** — does the next sentence's opening reference word or connector pick up your candidate?

What is the priority when running short of time?

**Eliminate by references first**, then judge the remaining 2-3 candidates by **connector logic**. **References + connectors alone deliver 60-70% accuracy**. Memorizing the **five recurring patterns** (problem→solution, claim→example, sequencing, cause→effect, general→specific) lets you **predict the logical flow from the previous sentence**, saving time.

How do I reduce careless mistakes on insertion items?

Build a **3-item checklist** into the routine: **(1) reference resolution against previous sentence, (2) connector logic match, (3) natural connection to the following sentence**. **Saying the checklist out loud during practice** halves careless errors. **Under exam pressure** the checklist tends to collapse — **internalize it during the 4-week program** so it runs automatically on test day.

TOEIC LinkPublished May 2, 2026

TOEIC Link Reading Part 6 Sentence Insertion — A 3-Axis Logic Check (Connectors, References, Discourse Markers) That Lifts Accuracy from 50% to 80%

Q: Why is my insertion accuracy stuck around 60%?

Three likely causes: **(1) reading only the previous sentence and skipping the following one**, **(2) picking on topic relevance alone** (all four are topic-relevant), **(3) picking on grammatical naturalness alone** (all four are grammatical). The **3-axis mechanical check** prevents all three. If you stall at 60%, the most common gap is **failing to read the sentence after the gap** — fix that first.

TOEIC Link Part 6 (Text Completion) sentence-insertion items have the lowest pass rate among Reading sections. Unlike straightforward grammar gap-fills, they require inferring logical relationships across surrounding sentences and choosing one of four candidates. Without targeted training, even strong test-takers stall around 50%. This guide presents a 3-axis check (connectors, references, discourse markers) to mechanize the judgment, and a 4-week program that lifts accuracy from 50% to 80%+.

Structure and difficulty of Part 6 insertion

Part 6 carries 4 documents × 4 items = 16 questions, with 1-2 sentence-insertion items per document. Insertion requires context inference, not just grammar — the standard solve time is 35-50 seconds per item.

Why difficulty is high. All four candidates are grammatically valid and topically related. The only discriminator is logical necessity (connection to the previous sentence, advancement to the following sentence) — a pure logic test, not an English-skill test.

Common wrong-answer patterns. (1) Picking on topic relevance alone (all four are relevant). (2) Picking on grammatical naturalness alone (all four are grammatical). (3) Reading only the previous sentence (missing the link to the next sentence). The 3-axis check prevents all three.

Document-type frequency. Email (50%), notice / internal memo (25%), advertisement / article (15%), letter (10%). Email is most common so practice email-specific logic flow (request → details → next action) first.

Part 6 = 4 docs × 4 items
1-2 insertion items per doc
35-50 seconds per item
All four candidates are grammatical
Logic, not vocabulary, decides
Email format = 50% of documents

The 3-axis check framework

Axis 1 — connector / adverbial alignment. If the candidate begins with However / Therefore / In addition / For example, verify that the logical relation to the previous sentence (contrast / cause / addition / illustration) matches. However → previous sentence must contradict, Therefore → previous sentence must be the cause — fully mechanical.

Axis 2 — reference resolution. If the candidate contains this / that / these / those / it / they, verify that the referent (a specific noun phrase) exists in the previous sentence. No referent → eliminate the candidate. "this product" without a product mentioned earlier is wrong.

Axis 3 — discourse markers. First / Second / Finally / In conclusion / On the other hand signal structural position in the document. First in the previous sentence → expect Second in the candidate. In conclusion → near the end of a business document. Use the document's structural position as a constraint.

Axis priority. Apply in order: references → connectors → discourse markers. References can be checked mechanically and eliminate ~70% of choices. The remaining 30% need connectors / discourse markers.

Worked example (2026 Official Volume 3 Q140): - Previous: "Our new product line launched last month with strong customer reviews." - Candidates: (A) "However, sales were below target." (B) "These positive responses encouraged us to expand." (C) "First, we plan to add three new features." (D) "In conclusion, the strategy worked." - Axis 1 (connector): (A) However → previous positive, candidate negative — contrast valid / (D) In conclusion → wrong if not document-end - Axis 2 (reference): (B) These positive responses → "strong customer reviews" is an unambiguous referent ✓ - Axis 3 (discourse): (C) First → demands Second to follow; wrong if document-end - Answer: (B) — strongest reference alignment, natural logical advance

Axis 1: connector / adverbial logic relation
Axis 2: reference must exist in previous sentence
Axis 3: discourse marker = structural position
Priority: references > connectors > markers
~70% eliminated by references
Connectors / markers handle the remaining 30%

Step-by-step elimination procedure

Step 1 — 30-second document overview. Pin the opening, closing, paragraph structure to identify topic and logical flow. Email → request; ad → product type; notice → announcement — write a one-line summary.

Step 2 — read the two surrounding sentences in detail. One sentence before, one sentence after the gap. Note subject-verb-object of the previous sentence and opening reference / connector of the following sentence.

Step 3 — scan candidates for references first. Check candidates with this / that / these / those / it. Eliminate candidates whose referent is not in the previous sentence. Typically removes 1-2 choices.

Step 4 — check connectors on remaining candidates. Candidates with However / Therefore / Furthermore / In addition are tested against the logical relation to the previous sentence. However after a positive sentence is wrong.

Step 5 — final check: connection to the next sentence. Read the remaining 1-2 candidates with the next sentence appended and verify natural logical flow. If the next sentence has a reference word, the candidate must be its source.

Time budget. Step 1 (30s) → Step 2 (15s) → Step 3 (10s) → Step 4 (10s) → Step 5 (10s) = 75 seconds total per insertion item. Slightly over the per-item average but compensated by 15-20 seconds for grammar gap-fills.

Step 1: 30s document overview
Step 2: 15s precise read of surrounding sentences
Step 3: 10s reference elimination
Step 4: 10s connector check
Step 5: 10s next-sentence connection
Total: 75 seconds per item

Five recurring logic patterns

Pattern 1 — problem → solution. Candidates with However / But / Unfortunately raise a problem against the previous sentence; the next sentence describes the fix. Most common in business email (request flows).

Pattern 2 — claim → example. For example / For instance / Specifically illustrates the previous claim. Common in ads / articles (product feature explanations).

Pattern 3 — sequencing (First → Second → Finally). First / Second / Third / Finally mark structural position. Common in notices / internal memos (procedure descriptions).

Pattern 4 — cause → effect. Therefore / As a result / Consequently carries the previous cause forward to its effect. Common across all business documents.

Pattern 5 — general → specific. In particular / Especially / Notably specializes a previous general statement. Common in ads / articles.

Sample drill (Pattern 1): - Previous: "Our delivery system has been reliable for the past five years." - Insert: ___ - Next: "We are implementing a new tracking system to address this issue." - Candidates: (A) "Customer satisfaction has improved." (B) "However, recent customer complaints have increased significantly." (C) "These improvements will continue." (D) "We have hired additional staff." - Answer: (B) — the next sentence references "this issue" which has no antecedent in the previous sentence; the candidate must introduce the issue.

P1: problem → solution
P2: claim → example
P3: sequencing First → Second → Finally
P4: cause → effect
P5: general → specific
Pattern recognition accelerates solving

Training program — four weeks

Week 1 — pattern recognition. Solve 8 Part 6 insertion items from Official Volume 1, writing 5 lines per choice on why each was eliminated. Record baseline accuracy. Speed is not the focus — the goal is logical hygiene.

Week 2 — implement the 3-axis check. 8 items from Official Volume 2 with the explicit order references → connectors → markers. Target 90 seconds per item; build the mechanical habit.

Week 3 — speed. 8 items from Official Volume 3 + one mock set. Target 75 seconds per item. Use elimination; final check is connection to the next sentence only.

Week 4 — mocks + weakness review. Two full mock sets under exam conditions. Plot accuracy from Week 1 to Week 4, classify wrong items by pattern, and deep-dive the patterns where you fail most.

Expected curve. 50% (Week 1) → 65% (Week 2) → 75% (Week 3) → 80%+ (Week 4) is the typical improvement trajectory. Whole-Reading impact is +20-30 points.

Week 1: recognition + baseline
Week 2: 3-axis check at 90 s/item
Week 3: speed at 75 s/item
Week 4: mocks + pattern-based review
50% → 80%+ accuracy curve
+20-30 Reading-section points

4-week program targets

Week	Target accuracy	Per-item time	Material	Focus
Week 1	50-60%	120 s	Official Volume 1	Recognition + elimination notes
Week 2	60-70%	90 s	Official Volume 2	3-axis implementation
Week 3	70-80%	75 s	Official Volume 3 + mock	Speed via elimination
Week 4	80%+	70 s	2 mock sets	Pattern-based weakness review

* Accuracy is measured on insertion items only. Whole-Part-6 accuracy = insertion-accuracy × 0.4 + grammar-gap-fill-accuracy × 0.6.

Six rules for sentence-insertion

References must resolve to the previous sentence
Connectors must match the previous-sentence relation
Discourse markers signal structural position
Eliminate methodically; finish in 75 seconds
4-week training: 50% → 80%+
Memorize the five recurring patterns

Try EnglishBlitz Part 6 drills Read the full Reading strategy

Frequently Asked Questions

TOEIC® and TOEIC Link™ are registered trademarks of ETS. EnglishBlitz is not affiliated with, endorsed by, or associated with ETS. The techniques in this article reflect trend analysis of official problem books and do not guarantee correct answers on individual items. Part 6 question characteristics vary across years and forms — verify against the most recent official volumes.